Applies To:

Show Versions Show Versions

Supplemental Document: BIG-IP 13.1.1.3 Fixes and Known Issues

Original Publication Date: 12/12/2018

BIG-IP Release Information

Version: 13.1.1.3
Build: 1.0

NOTE: This release includes fixes for the Spectre Variant 1 and Meltdown vulnerabilities (CVE-2017-5753, CVE-2017-5754).
In some configurations, installing software containing these fixes might impact performance. You can disable these fixes to recover performance. Please see K91229003 for additional Spectre and Meltdown information.

Cumulative fixes from BIG-IP v13.1.1.2 that are included in this release
Cumulative fixes from BIG-IP v13.1.1.1 that are included in this release
Cumulative fixes from BIG-IP v13.1.1 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.8 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.7 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.6 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.5 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.4 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.3 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.2 that are included in this release
Cumulative fixes from BIG-IP v13.1.0.1 that are included in this release
Known Issues in BIG-IP v13.1.x

Functional Change Fixes

ID Number Severity Solution Article(s) Description
745783-3 3-Major   Anti-fraud: remote logging of login attempts


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
684370-1 3-Major   APM now supports VMware Workspace ONE integration with VIDM as ID Provider
683741-1 3-Major   APM now supports VMware Workspace ONE integration with vIDM as ID Provider
635509-1 3-Major   APM does not support Vmware'e Blast UDP



Cumulative fixes from BIG-IP v13.1.1.2 that are included in this release


Vulnerability Fixes

ID Number CVE Solution Article(s) Description
726089-2 CVE-2018-15312 K44462254 Modifications to AVR metrics page
725815-1 CVE-2018-15320 K72442354 vlangroup usage may cause a excessive resource consumption
724339-1 CVE-2018-15314 K04524282 Unexpected TMUI output in AFM
724335-1 CVE-2018-15313 K21042153 Unexpected TMUI output in AFM
722091-3 CVE-2018-15319 K64208870 TMM may crash while processing HTTP traffic
717888 CVE-2018-15323 K26583415 TMM may leak memory when a virtual server uses the MQTT profile.
717742-5 CVE-2018-2798, CVE-2018-2811, CVE-2018-2795, CVE-2018-2790, CVE-2018-2783, CVE-2018-2825, CVE-2018-2826, CVE-2018-2796, CVE-2018-2799, CVE-2018-2815, CVE-2018-2800, CVE-2018-2814, CVE-2018-2797, CVE-2018-2794 K44923228 Oracle Java SE vulnerability CVE-2018-2783
704184-6 CVE-2018-5529 K52171282 APM MAC Client create files with owner only read write permissions
701253-5 CVE-2018-15318 K16248201 TMM core when using MPTCP
741858-1 CVE-2018-15324 K52206731 TMM may crash while processing Portal Access requests
734822-3 CVE-2018-15325 K77313277 TMSH improvements
726409-4 CVE-2017-8890
CVE-2017-9075
CVE-2017-9076
CVE-2017-9077
K61429540 Kernel Vulnerabilities: CVE-2017-8890 CVE-2017-9075 CVE-2017-9076 CVE-2017-9077
725801-4 CVE-2017-7889 K80440915 CVE-2017-7889: Kernel Vulnerability
725635-2 CVE-2018-3665 K21344224 CVE-2018-3665: Intel Lazy FPU Vulnerability
724680-4 CVE-2018-0732 K21665601 OpenSSL Vulnerability: CVE-2018-0732
719554-2 CVE-2018-8897 K17403481 Linux Kernel Vulnerability: CVE-2018-8897
710705-2 CVE-2018-7320, CVE-2018-7321, CVE-2018-7322, CVE-2018-7423, CVE-2018-7324, CVE-2018-7325, CVE-2018-7326, CVE-2018-7327, CVE-2018-7428, CVE-2018-7329, CVE-2018-7330, CVE-2018-7331, CVE-2018-7332, CVE-2018-7333, CVE-2018-7334, CVE-2018-7335, CVE-2018-7336, CVE-2018-7337, CVE-2018-7417, CVE-2018-7418, CVE-2018-7419, CVE-2018-7420, CVE-2018-7421 K34035645 Multiple Wireshark vulnerabilities
705799-2 CVE-2018-15325 K77313277 TMSH improvements
699453-4 CVE-2018-15327 K20222812 Web UI does not follow current best coding practices
712876-2 CVE-2017-8824 K15526101 CVE-2017-8824: Kernel Vulnerability


Functional Change Fixes

ID Number Severity Solution Article(s) Description
734527-1 3-Major   BGP 'capability graceful-restart' for peer-group not properly advertised when configured
715750-2 3-Major   The BIG-IP system may exhibit undesirable behaviors when receiving a FIN midstream an SSL connection.


TMOS Fixes

ID Number Severity Solution Article(s) Description
701826 0-Unspecified   qkview upload to ihealth fails or unable to untar qkview file
693611-3 1-Blocking K76313256 IKEv2 ike-peer might crash on stats object during peer modification update
743810-1 2-Critical   AWS: Disk resizing in m5/c5 instances fails silently.
743082-1 2-Critical   Upgrades to 13.1.1.2, 14.0.0, 14.0.0.1, or 14.0.0.2 from pre-12.1.3 can cause configurations to fail with GTM Pool Members
739507 2-Critical   How to recover from a failed state due to FIPS integrity check
739505 2-Critical   Automatic ISO digital signature checking not required when FIPS license active
739285-1 2-Critical   GUI partially missing when VCMP is provisioned
725696-1 2-Critical   A timer loop might occur when OCSP Stapling is enabled resulting in tmm getting aborted
723722-2 2-Critical   MCPD crashes if several thousand files are created between config syncs.
721924-2 2-Critical   bgpd may crash processing extended ASNs
721350-2 2-Critical   The size of the icrd_child process is steadily growing
717785-1 2-Critical   Interface-cos shows no egress stats for CoS configurations
716391-2 2-Critical K76031538 High priority for MySQL on 2 core vCMP may lead to control plane process starvation
711683-2 2-Critical   bcm56xxd crash with empty trunk in QinQ VLAN
707003-3 2-Critical   Unexpected syntax error in TMSH AVR
706423-1 2-Critical   tmm may restart if an IKEv2 child SA expires during an async encryption or decryption
703669-2 2-Critical   Eventd restarts on NULL pointer access
703045-1 2-Critical   If using TMSH commands with deprecated attributes in iAPP, the upgrade will fail.
700386-2 2-Critical   mcpd may dump core on startup
693996-5 2-Critical K42285625 MCPD sync errors and restart after multiple modifications to file object in chassis
692158-1 2-Critical   iCall and CLI script memory leak when saving configuration
691589-4 2-Critical   When using LDAP client auth, tamd may become stuck
690819-1 2-Critical   Using an iRule module after a 'session lookup' may result in crash
689437-1 2-Critical K49554067 icrd_child cores due to infinite recursion caused by incorrect group name handling
689002-3 2-Critical   Stackoverflow when JSON is deeply nested
658410-2 2-Critical   icrd_child generates a core when calling PUT on ltm/data-group/internal/
652877-5 2-Critical   Reactivating the license on a VIPRION system may cause MCPD process restart on all secondary blades
638091-6 2-Critical   Config sync after changing named pool members can cause mcpd on secondary blades to restart
739126 3-Major   Multiple VE installations may have different sized volumes
733585-3 3-Major   Merged can use %100 of CPU if all stats snapshot files are in the future
727467-1 3-Major   Some iSeries appliances can experience traffic disruption when the HA peer is upgraded from 12.1.3 and earlier to 13.1.0 or later.
722682-2 3-Major   Fix of ID 615222 results in upgrade issue for GTM pool member whose name contains a colon; config failed to load
721740-2 3-Major   CPU stats are not correctly recorded when snapshot files have timestamps in the future
720713-2 3-Major   TMM traffic to/from a i10600/i10800 device in vCMP host mode may fail
720461-2 3-Major   qkview prompts for password on chassis
718525-1 3-Major   PostGreSQL errors seen on startup after removing the mcpd binary database and rebooting
714974-2 3-Major   Platform-migrate of UCS containing QinQ fails on VE
714903-2 3-Major   Errors in chmand
714654-2 3-Major   Creating static route for advertised dynamic route might result in dropping the tmrouted connection to TMM
713813-2 3-Major   Node monitor instances not showing up in GUI
712102-2 3-Major K11430165 customizing or changing the HTTP Profile's IPv6 field hides the field or the row
710232-2 3-Major   platform-migrate fails when LACP trunks are in use
709444-2 3-Major   "NTP not configured on device" warning seen when NTP symmetric key authentication is configured
709192-1 3-Major   GUI/iControl SOAP - KeyManagement export_all_to_archive_file failing immediately after HTTPD restart
707990-2 3-Major   Unexpected TMUI output in SSL Certificate Instance page
707740-4 3-Major   Failure deleting GTM Monitors when used on mulitple Virtual Servers with the same ip:port combination
707509-1 3-Major   Initial vCMP guest creations can fail if certain hotfixes are used
707391-2 3-Major   BGP may keep announcing routes after disabling route health injection
706804-1 3-Major   SNMP trap destination configuration of network option is missing "default" keyword
706354-2 3-Major   OPT-0045 optic unable to link
706169-3 3-Major   tmsh memory leak
705456-1 3-Major   VCMP Guests unable to install block-device-image ISOs when http->https redirection is enabled
704755-1 3-Major   EUD_M package could not be installed on 800 platforms
704512-1 3-Major   Automated upload of qkview to iHealth can time out resulting in error
704336-1 3-Major   Updating 3rd party device cert not copied correctly to trusted certificate store
702227-3 3-Major   Memory leak in TMSH load sys config
700757-1 3-Major   vcmpd may crash when it is exiting
700576-1 3-Major   GUI - Server SSL Profile shows irrelevant options when "Server Certificate" is set to "Ignore"
700426 3-Major K58033284 Switching partitions while viewing objects in GUI can result in empty list
700250-3 3-Major K59327012 qkviews for secondary blade appear to be corrupt
698875-1 3-Major   Qkview Security Hardening
698084-3 3-Major K03776801 IPsec log messages in /var/log/ltm missing module ID to reach bigiq logs
696731-3 3-Major K94062594 The standard LinkUp/LinkDown traps may not be issued in all cases when an interface is administratively enabled/disabled
693578-2 3-Major   switch/color_policer tmctl table is not available in 12.1.0 - 13.1.0
692189-1 3-Major   errdefsd fails to generate a core file on request.
692179-1 3-Major   Potential high memory usage from errdefsd.
691609-1 3-Major   1NIC: Configuration error: Clearing IP on interface mgmt failed - Cannot assign requested address
690890-1 3-Major   Running sod manually can cause issues/failover
689375-1 3-Major K01512833 Unable configure 'Generic Alert' setting on SSL client/server profile through TMUI when 'Proxy SSL' is enabled
688406-1 3-Major K14513346 HA-Group Score showing 0
687905-2 3-Major K72040312 OneConnect profile causes CMP redirected connections on the HA standby
687534-1 3-Major   If a Pool contains ".." in the name, it is impossible to add a Member to this pool using the GUI Local Traffic > Pools : Member List page
684391-3 3-Major   Existing IPsec tunnels reload. tmipsecd creates a core file.
684218-1 3-Major   vADC 'live-install' Downgrade from v13.1.0 is not possible
681782-6 3-Major   Unicast IP address can be configured in a failover multicast configuration
679347-2 3-Major K44117473 ECP does not work for PFS in IKEv2 child SAs
678488-1 3-Major K59332320 BGP default-originate not announced to peers if several are peering over different VLANs
677485-1 3-Major   Discovery of DSC clustered BIG-IP systems fails due to secure value decryption error
671712-2 3-Major   The values returned for the ltmUserStatProfileStat table are incorrect.
670528-4 3-Major K20251354 Warnings during vCMP host upgrade.
651413-4 3-Major K34042229 tmsh list ltm node does not return an error when node does not exist
642923-6 3-Major K01951295 MCP misses its heartbeat (and is killed by sod) if there are a large number of file objects on the system
617643-2 3-Major   iControl.ForceSessions enabled results in GUI error on certain pages
551925-4 3-Major   Misdirected UDP traffic with hardware acceleration
464650-6 3-Major   Failure of mcpd with invalid authentication context.
727297-3 4-Minor   GUI TACACS+ remote server list should accept hostname
725612-1 4-Minor   syslog-ng remote destination needs unique name that changes on address change.
719770-2 4-Minor   tmctl -H -V and -l options without values crashed
714749-2 4-Minor   cURL Vulnerability: CVE-2018-1000120
713947-1 4-Minor   stpd repeatedly logs "hal sendMessage failed"
713932-1 4-Minor   Commands are replicated to PostgreSQL even when not in use.
707631-2 4-Minor   The 'SYN Challenge Handling' setting of a TCP profile can revert to defaults when updating the profile using the GUI
707267 4-Minor   REST Framework HTTP header limit size increased to 8 KB
691491-5 4-Minor K13841403 2000/4000, 10000, i2000/i4000, i5000/i7000/i10000, i15000, B4000 platforms may return incorrect SNMP sysIfxStatHighSpeed values for 10G/40G/100G interfaces
685582-7 4-Minor   Incorrect output of b64 unit key hash by command f5mku -f
683029-1 4-Minor   Sync of virtual address and self IP traffic groups only happens in one direction
679135-2 4-Minor   IKEv1 and IKEv2 cannot share common local address in tunnels
678388-1 4-Minor K00050055 IKEv1 racoon daemon is not restarted when killed multiple times


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
722594-2 1-Blocking K91300169 TCP flow may not work as expected if double tagging is used
737445-2 2-Critical   Use of TCP Verified Accept can disable server-side flow control
727044-2 2-Critical   TMM may crash while processing compressed data
726239-4 2-Critical   interruption of traffic handling as sod daemon restarts TMM
725545-1 2-Critical   Ephemeral listener might not be set up correctly
724906-1 2-Critical   sasp_gwm monitor leaks memory over time
724868-1 2-Critical K11662998 dynconfd memory usage increases over time
724213-1 2-Critical   Modified ssl_profile monitor param not synced correctly
722893-1 2-Critical K30764018 The TMM - host interface may stall when the kernel memory is fragmented
722387-3 2-Critical   TMM may crash when processing APM DTLS traffic
716900-2 2-Critical   TMM core when using MPTCP
716213-1 2-Critical   BIG-IP system issues TCP-Reset with error 'SSID error (Out of Bounds)' in traffic
713612-1 2-Critical   tmm might restart if the HTTP passthrough on pipeline option is used
710221-2 2-Critical K67352313 Ephemeral node stays Unavailable on 'standby' BIG-IP system after a forced-offline FQDN node is re-enabled
673664-1 2-Critical   TMM crashes when sys db Crypto.HwAcceleration is disabled.
635191-2 2-Critical   Under rare circumstances TMM may crash
727222-1 3-Major   206 Partial Content responses from ramcache have malformed Content-Range header
723300-2 3-Major   TMM may crash when tracing iRules containing nameless listeners on internal virtual servers
722677-4 3-Major   High-Speed Bridge may lock up
722363-2 3-Major   Client fails to connect to server when using PVA offload at Established
721261-1 3-Major   v12.x Policy rule names containing slashes are not migrated properly
720293-3 3-Major   HTTP2 IPv4 to IPv6 fails
719600-2 3-Major   TCP::collect iRule with L7 policy present may result in connection reset
717346-2 3-Major K13040347 [WebSocket ] tmsh show /ltm profile WebSocket current and max numbers far larger than total
715883 3-Major   tmm crash due to invalid cookie attribute
715785-2 3-Major   Incorrect encryption error for monitors during sync or upgrade
715756-2 3-Major   Clusterd may not trigger primary election when primary blade has critical filesystems mounted read-only
715467-2 3-Major   Ephemeral pool member and node are not updated on 'A' record changes if pool member port is ANY
714384-3 3-Major   DHCP traffic may not be forwarded when BWC is configured
707951-2 3-Major   Stalled mirrored flows on HA next-active when OneConnect is used.
704764-3 3-Major   SASP monitor marks members down with non-default route domains
703580-1 3-Major   TLS1.1 handshake failure on v12.1.3 vCMP guest with earlier BIG-IP version on vCMP host.
703266-2 3-Major   Potential MCP memory leak in LTM policy compile code
702450-1 3-Major   The validation error message generated by deleting certain object types referenced by a policy action is incorrect
701690-1 3-Major K53819652 Fragmented ICMP forwarded with incorrect icmp checksum
700696-1 3-Major   SSID does not cache fragmented Client Certificates correctly via iRule
699273-1 3-Major   TMM Core During FTP Monitor Use
695925-1 3-Major   tmm crash when showing connections for a CMP disabled virtual server
691785-1 3-Major   The bcm570x driver can cause TMM to core when transmitting packets larger than 6144 bytes
691224-3 3-Major K59327001 Fragmented SSL Client Hello message do not get properly reassembled when SSL Persistence is enabled
690778-1 3-Major K53531153 Memory can leak if the STREAM::replace command is called more than once in the STREAM_MATCHED event in an iRule
688629-1 3-Major K52334096 Deleting data-group in use by iRule does not trigger validation error
685110-1 3-Major K05430133 With a non-LTM license (ASM, APM, etc.), ephemeral nodes will not be created for FQDN nodes/pool members.
681757-3 3-Major K32521651 Upgraded volume may fail to load if a Local Traffic Policy uses the forward parameter 'member'
681673-4 3-Major   tmsh modify FDB command permits multicast MAC addresses, which produces unexpected results
679613-1 3-Major K23531420 i2000/i4000 Platforms Improperly Handle VLANs Created with a Value of '1'
672312-3 3-Major   IP ToS may not be forwarded to serverside with syncookie activated
602708-4 3-Major   Traffic may not passthrough CoS by default
716922-2 4-Minor   Reduction in PUSH flags when Nagle Enabled
712637-2 4-Minor   Host header persistence not implemented
700433-1 4-Minor K10870739 Memory leak when attaching an LTM policy to a virtual server
697988-3 4-Minor   During config sync, if many client-ssl profiles are on a virtual server the CPU may briefly spike to 100%
693966-1 4-Minor   TCP sndpack not reset along with other tcp profile stats
688557-1 4-Minor K50462482 Tmsh help for ltm sasp monitor incorrectly lists default mode as 'pull'
495242-4 4-Minor   mcpd log messages: Failed to unpublish LOIPC object


Global Traffic Manager (DNS) Fixes

ID Number Severity Solution Article(s) Description
718885-3 2-Critical K25348242 Under certain conditions, monitor probes may not be sent at the configured interval
723792-2 3-Major   GTM regex handling of some escape characters renders it invalid
719644-2 3-Major   If auto-discovery is enabled, GTM configuration may be affected after an upgrade from 11.5.x to later versions


Application Security Manager Fixes

ID Number Severity Solution Article(s) Description
737500-2 2-Critical   Apply Policy and Upgrade time degradation when there are previous enforced rules
726090-1 2-Critical   No Bot Defense Logs when ASM Policy is used without Proactive Bot Defense
724414-2 2-Critical   ASM UMU memory leak when WebSocket url has json profile with parse parameters enabled
724032-1 2-Critical   Searching Request Log for value containing backslash does not return expected result
721741-3 2-Critical   BD and BD_Agent out-of-sync for IP Address Exception, false positive/negative
704143-1 2-Critical   BD memory leak
701856-1 2-Critical   Memory leak in ASM-config Event Dispatcher upon continuous Policy Builder restart
740719-2 3-Major   ASM CSP header parser does not honor unsafe-inline attribute within script-src directive


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
737867-1 3-Major   Scheduled reports are being incorrectly displayed in different partitions


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
739716-2 1-Blocking   APM Subroutine loops without finishing
740777-1 2-Critical   Secondary blades mcp daemon restart when subroutine properties are configured
739947-1 2-Critical   TMM may crash while processing APM traffic
739674-1 2-Critical   TMM might core in SWG scenario with per-request policy.
722013 2-Critical   MCPD restarts on all secondary blades post config-sync involving APM customization group
713820-1 2-Critical   Pass in IP to urldb categorization engine
739939-1 3-Major   Ping Access Agent Module leaks memory in TMM.
739190 3-Major   Policies could be exported with not patched /Common partition
738582-1 3-Major   Ping Access Agent Module leaks memory in TMM.
738397-1 3-Major   SAML -IdP-initiated case that has a per-request policy with a logon page in a subroutine fails.
737355-1 3-Major   HTTP Strict-Transport-Security (HSTS) headers not being added to all APM generated files
737064-2 3-Major   ACCESS::session iRule commands may not work in serverside events
726895 3-Major   VPE cannot modify subroutine settings
726616-1 3-Major   TMM crashes when a session is terminated
726592-1 3-Major   Invalid logsetting config messaging between our daemons could send apmd, apm_websso, localdbmgr into infinite loop
725867-2 3-Major   ADFS proxy does not fetch configuration for non-floating virtual servers
725412-1 3-Major   APM does not follow current best practices for HTTP headers
724571-1 3-Major   Importing access profile takes a long time
722969-2 3-Major   Access Policy import with 'reuse' enabled instead rewrites shared objects
722423-1 3-Major   Analytics agent always resets when Category Lookup is of type custom only
720757-1 3-Major   Without proper licenses Category Lookup always fails with license error in Allow Ending
713655-2 3-Major   RouteDomainSelectionAgent might fail under heavy control plane traffic/activities
711427-2 3-Major   Edge Browser does not launch F5 VPN App
710884-1 3-Major   Portal Access might omit some valid cookies when rewriting HTTP request.
701800-2 3-Major   SSO-enabled native RDP resources cannot be launched from APM Webtop with Mac RDP client 10.x
701056-1 3-Major   User is not able to reset their Active Directory password
698984-1 3-Major   Auto-disable TMM.HTTP.TCL.Validation when APM is provisioned
696669-1 3-Major   Users cannot change or reset RSA PIN
696544-1 3-Major   APM end users can not change/reset password when auth agents are included in per-req policy
671323-1 3-Major   Reset PIN Fail if Token input field is not 'password' field
734595-2 4-Minor   sp-connector is not being deleted together with profile
721375-1 4-Minor   Export then import of config with RSA server in it might fail


WebAccelerator Fixes

ID Number Severity Solution Article(s) Description
706642-2 2-Critical   wamd may leak memory during configuration changes and cluster events


Service Provider Fixes

ID Number Severity Solution Article(s) Description
709383-2 3-Major   DIAMETER::persist reset non-functional
706750-1 3-Major   Changing CGNAT SIP ALG profile log settings while handling traffic may cause tmm crash.
691048-1 3-Major K34553736 Support DIAMETER Experimental-Result AVP response
688942-5 3-Major   ICAP: Chunk parser performs poorly with very large chunk


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
724532-2 2-Critical   SIG SEGV during IP intelligence category match in TMM
720045-1 2-Critical   IP fragmented UDP DNS request and response packets dropped as DNS Malformed
710755-1 2-Critical K30572159 Crash when cached route information becomes stale and the system accesses the information from it.
698333-1 2-Critical K43392052 TMM core upon matching a past (saved) dynamic signature of a specific family (network or dns) after the dynamic signature is disabled for that family on the parent context (but still enabled for other families)
694849-1 2-Critical   TMM crash when packet sampling is turned for DNS BDOS signatures.
672514-1 2-Critical   Local Traffic/Virtual Server/Security page crashed
630137-2 2-Critical   Dynamic Signatures feature can fill up /config partition impacting system stability
726154-2 3-Major   TMM restart when there are virtual server and route-domains with the same names and attached firewall or NAT policies
704528-2 3-Major   tmm may run out of memory during IP shunning
704369-2 3-Major   TMM on BIG-IP restarts if a dos profile is attached to a virtual with sip-routing enabled
696201-1 3-Major   Anomaly threshold floor calculation (for bins with very low learnt threshold) during signature generation
686376-2 3-Major   Scheduled blob and current blob no longer work after restarting BIG-IP or PCCD daemon
707054-1 4-Minor   SYN Cookie MSS for Fast L4 Profiles is limited to 256-9162
699454-4 4-Minor   Web UI does not follow current best coding practices
699452-4 4-Minor   Web UI does not follow current best coding practices


Policy Enforcement Manager Fixes

ID Number Severity Solution Article(s) Description
726647-3 3-Major   PEM content insertion in a compressed response may truncate some data
721704-1 3-Major   UDP flows are not deleted after subscriber deletion
709670-2 3-Major K44067891 iRule triggered from RADIUS occasionally fails to create subscribers.


Carrier-Grade NAT Fixes

ID Number Severity Solution Article(s) Description
721570-1 1-Blocking K20285019 TMM core when trying to log an unknown subscriber
734446-2 2-Critical   TMM crash after changing LSN pool mode from PBA to NAPT
688246-1 2-Critical   An invalid mode in the LSN::persistence command causes TMM crash
708830-2 3-Major   Inbound or hairpin connections may get stuck consuming memory.


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
738669-2 3-Major   Login validation may fail for a large request with early server response
737368-1 3-Major   Fingerprint cookie large value may result in tmm core.


Anomaly Detection Services Fixes

ID Number Severity Solution Article(s) Description
739277 2-Critical   TMM core when the virtual server is deleted during POST data and BADOS standard/aggressive mode
720585-1 3-Major   Signatures generated by Behavioral DOS algorithm can create false-positive signatures
689540-1 3-Major   The same DOS attack generates new signatures even if there are signatures generated during previous attacks.


Traffic Classification Engine Fixes

ID Number Severity Solution Article(s) Description
726303-1 3-Major   Unlock 10 million custom db entry limit


iApp Technology Fixes

ID Number Severity Solution Article(s) Description
726872-2 3-Major   IApp LX directory disappears after upgrade or restoring from ucs



Cumulative fixes from BIG-IP v13.1.1.1 that are included in this release


Functional Change Fixes

None



Cumulative fixes from BIG-IP v13.1.1 that are included in this release


Functional Change Fixes

ID Number Severity Solution Article(s) Description
693359-1 1-Blocking   AWS M5 and C5 instance families are supported


TMOS Fixes

ID Number Severity Solution Article(s) Description
721364 1-Blocking   BIG-IP per-application VE BYOL license does not support three wildcard virtual servers
716469 1-Blocking   OpenSSL 1.0.1l fails with 512 bit DSA keys
697615-1 1-Blocking K65013424 Neurond may restart indefinitely after boot, with neurond_i2c_config message
675921-2 1-Blocking   Creating 5th vCMP 'ssl-mode dedicated' guest results in an error, but is running
723130-1 2-Critical   Invalid-certificate warning displayed when deploying BIG-IP VE OVA file
700086-1 2-Critical   AWS C5/M5 Instances do not support BIG-IP VE
696732-3 2-Critical K54431534 tmm may crash in a compression provider
721985 3-Major   PAYG License remains inactive as dossier verification fails.
721512 3-Major   Config tool fails to configure management-ip when default routes exist for both IPv4 and IPv6.
721342 3-Major   No support for Per-App VE LTM and Per-App VE WAF images for various cloud environments.
720961-1 3-Major   Upgrading in Intelligence Community AWS environment may fail
720756-1 3-Major   SNMP platform name is unknown on i11400-DS/i11600-DS/i11800-DS
720651-2 3-Major   Running Guest Changed to Provisioned Never Stops
720104-1 3-Major   BIG-IP i2800/i2600 650W AC PSU has marketing name missing under 'show sys hardware'
719396-1 3-Major K34339214 DHCP Client on BIG-IP sends host-name 'localhost' in DISCOVER packet after first boot.
717832 3-Major   Remove unneeded files from UCS backup directories
714303-1 3-Major   X520 virtual functions do not support MAC masquerading
712266-1 3-Major   Decompression of large buffer might fail with comp_code=11 with Nitrox 3 hardware
697616-2 3-Major   Report Device error: crypto codec qat-crypto0-0 queue is stuck on vCMP guests
680086 3-Major   md5sum check on BMC firmware fails
673996-2 3-Major   Unable to set 'media-fixed' on management port on BIG-IP i15000 platforms
680388-1 4-Minor   f5optics should not show function name in non-debug log messages
653759-1 4-Minor   Chassis Variant number is not specified in C2200/C2400 chassis after a firmware update
720391-2 5-Cosmetic   BIG-IP i7800 Dual SSD reports wrong marketing name under 'show sys hardware'


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
737550 2-Critical   State Mirroring between BIG-IP 13.0.x and 13.1.x systems may cause TMM core on standby system during upgrade
701538-2 2-Critical   SSL handshake fails for OCSP, TLS false start, and SSL hardware acceleration configured
720460-1 3-Major   Compression traffic still goes to hardware accelerator when compression.strategy is set to softwareonly
694778-1 3-Major   Certain Intel Crypto HW fails to decrypt data if the given output buffer size differs from RSA private key size
686631-2 3-Major   Deselect a compression provider at the end of a job and reselect a provider for a new job
679494-1 3-Major   Change the default compression strategy to speed
495443-9 3-Major K16621 ECDH negotiation failures logged as critical errors.
679496-2 4-Minor   Add 'comp_req' to the output of 'tmctl compress'


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
717909 2-Critical   tmm can abort on sPVA flush if the HSB flush does not succeed
701637 2-Critical   Crash in bcm56xxd during TMM failover
644822 2-Critical K19245372 FastL4 virtual server with enabled loose-init option works differently with/without AFM provisioned
702738-1 3-Major K32181540 Tmm might crash activating new blob when changing firewall rules
698182 3-Major   Upgrading from 13.1.1 to newer release might cause config to not be copied over
697516 3-Major   Upgrading using a ucs or scf file does not autogenerate uuids when current config has the uuid-default-autogenerate flag enabled



Cumulative fixes from BIG-IP v13.1.0.8 that are included in this release


Vulnerability Fixes

ID Number CVE Solution Article(s) Description
716992-2 CVE-2018-5539 K75432956 The ASM bd process may crash
715923-1 CVE-2018-15317 K43625118 When processing TLS traffic TMM may reset connections
710244-3 CVE-2018-5536 K27391542 Memory Leak of access policy execution objects
710140-1 CVE-2018-5527 K20134942 TMM may consume excessive resources when processing SSL Intercept traffic
709688-3 CVE-2017-3144
CVE-2018-5732
CVE-2018-5733
K08306700 dhcp Security Advisory - CVE-2017-3144, CVE-2018-5732, CVE-2018-5733
695072-2 CVE-2016-8399
CVE-2017-1000111
CVE-2017-1000112
CVE-2017-11176
CVE-2017-14106
CVE-2017-7184
CVE-2017-7541
CVE-2017-7542
CVE-2017-7558
K23030550 CVE-2016-8399 CVE-2017-1000111 CVE-2017-1000112 CVE-2017-11176 CVE-2017-14106 CVE-2017-7184 CVE-2017-7541 CVE-2017-7542 CVE-2017-7558
693744-4 CVE-2018-5531 K64721111 CVE-2018-5531: vCMP vulnerability
651741-2 CVE-2017-5970, K60104355 CVE-2017-5970 kernel.el7: ipv4: Invalid IP options could cause skb->dst drop
717900-2 CVE-2018-5528 K27044729 TMM crash while processing APM data
710148-2 CVE-2017-1000111
CVE-2017-1000112
K60250153 CVE-2017-1000111 & CVE-2017-1000112
709256-2 CVE-2017-9074
CVE-2017-7542
K61223103 CVE-2017-9074: Local Linux Kernel Vulnerability
705476-2 CVE-2018-15322 K28003839 Appliance Mode does not follow design best practices
698813-2 CVE-2018-5538 K45435121 When processing DNSX transfers ZoneRunner does not enforce best practices
688625-5 CVE-2017-11628 K75543432 PHP Vulnerability CVE-2017-11628
662850-6 CVE-2015-2716 K50459349 Expat XML library vulnerability CVE-2015-2716
714879-3 CVE-2018-15326 K34652116 APM CRLDP Auth passes all certs


Functional Change Fixes

ID Number Severity Solution Article(s) Description
685020-3 3-Major   Enhancement to SessionDB provides timeout


TMOS Fixes

ID Number Severity Solution Article(s) Description
708956-1 1-Blocking K51206433 During system boots up, error message displays: 'Dataplane INOPERABLE - only 1 HSBes found on this platform'
719597 2-Critical   HA between VIPRION chassis with B2250 blades might not work when running v12.1.1 and v13.1.0
715820-1 2-Critical K61422392 vCMP in HA configuration with VIPRION chassis might cause unstable data plane
712401-1 2-Critical   Enhanced administrator lock/unlock for Common Criteria compliance
676203-3 2-Critical   Inter-blade mpi connection fails, does not recover, and eventually all memory consumed.
665362-2 2-Critical   MCPD might crash if the AOM restarts
581851-6 2-Critical K16234725 mcpd process on secondary blades unexpectedly restarts when the system processes multiple tmsh commands
711249-1 3-Major   NAS-IP-Address added to RADIUS packet unexpectedly
710976-1 3-Major   Network Map might take a long time to load
710827-2 3-Major   TMUI dashboard daemon stability issue
708484-2 3-Major   Network Map might take a long time to load
707445-3 3-Major K47025244 Nitrox 3 compression hangs/unable to recover
705818-1 3-Major   GUI Network Map Policy with forward Rule to Pool, Pool does not show up
704804-1 3-Major   The NAS-IP-Address in RADIUS remote authentication is unexpectedly set to the loopback address
704733-1 3-Major   NAS-IP-Address is sent with the bytes in reverse order
704247-2 3-Major   BIG-IP software may fail to install if multiple copies of the same image are present and have been deleted
701249-1 3-Major   RADIUS authentication requests erroneously specify NAS-IP-Address of 127.0.0.1
700895-1 3-Major K34944451 GUI Network Map objects in subfolders are not being shown
696260-1 3-Major K53103420 GUI Network Map as Start Screen presents database error
694696-5 3-Major   On multiblade Viprion, creating a new traffic-group causes the device to go Offline
694547-2 3-Major K74203532 TMSH save sys config creates unneeded generate_config processes.
689730-3 3-Major   Software installations from v13.1.0 might fail
687658 3-Major   Monitor operations in transaction will cause it to stay unchecked
686906-2 3-Major   Fragmented IPv6 packets not handled correctly on Virtual Edition
674455-5 3-Major   Serial console baud rate setting lost after running 'tmidiag -r' in Maintenance OS
678254-1 4-Minor   Error logged when restarting Tomcat


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
721571-1 2-Critical   State Mirroring between BIG-IP 12.1.3.* and 13.* or 14.* systems may cause TMM core on standby system during upgrade
718071-1 2-Critical   HTTP2 with ASM policy not passing traffic
715747 2-Critical   TMM may restart when running traffic through custom SSLO deployments.
709828-2 2-Critical   fasthttp can crash with Large Receive Offload enabled
707244-3 2-Critical   iRule command clientside and serverside may crash tmm
707207-1 2-Critical   iRuleLx returning undefined value may cause TMM restart
700597-1 2-Critical   Local Traffic Policy on HTTP/2 virtual server no longer matches
700056-1 2-Critical   MCPD process may lock up and restart when applying Local Traffic Policy to virtual server
690756-1 2-Critical   APM depends on undocumented internal behavior of HTTP iRule commands after a retry is initiated
571651-4 2-Critical K66544028 Reset Nitrox3 crypto accelerator queue if it becomes stuck.
713951-5 3-Major   tmm core files produced by nitrox_diag may be missing data
713934-2 3-Major   Using iRule 'DNS::question name' to shorten the name in DNS_REQUEST may result malformed TC response
712819-2 3-Major   'HTTP::hsts preload' iRule command cannot be used
712475-3 3-Major K56479945 DNS zones without servers will prevent DNS Express reading zone data
712437-3 3-Major K20355559 Records containing hyphens (-) will prevent child zone from loading correctly
711281-5 3-Major   nitrox_diag may run out of space on /shared
710996-2 3-Major   VIPRION outgoing management IPv6 traffic from primary blade uses the cluster member IP
709133-2 3-Major   When the BigDB variable tmm.ssl.loggingcreatedcerts is set, a double-free error might occur
709132-1 3-Major   When the BigDB variable tmm.ssl.loggingcreatedcerts is set a buffer overflow can occur
707961-2 3-Major K50013510 Unable to add policy to virtual server; error = Failed to compile the combined policies
707109-1 3-Major   Memory leak when using C3D
704381-5 3-Major   SSL/TLS handshake failures and terminations are logged at too low a level
702151-1 3-Major   HTTP/2 can garble large headers
700889-3 3-Major K07330445 Software syncookies without TCP TS improperly include TCP options that are not encoded
700061-4 3-Major   Restarting service MCPD or rebooting BIG-IP device adds 'other' file read permissions to key file
699598-2 3-Major   HTTP/2 requests with large body may result in RST_STREAM with FRAME_SIZE_ERROR
696755 3-Major   HTTP/2 may truncate a response body when served from cache
693308-1 3-Major   SSL Session Persistence hangs upon receipt of fragmented Client Certificate Chain
689089-1 3-Major   VIPRION cluster IP reverted to 'default' (192.168.1.246) following unexpected reboot
688744-1 3-Major K11793920 LTM Policy does not correctly handle multiple datagroups
686890-1 3-Major   X509_EXTENSION memory blocks leak when C3D forges the certificate.
682944-1 3-Major   key-id missing for installed netHSM key for standby BIG-IP system in HA setup
682283-2 3-Major   Malformed HTTP/2 request with invalid Content-Length value is served against RFC
678872-3 3-Major   Inconsistent behavior for virtual-address and selfip on the same ip-address
673399-3 3-Major   HTTP request dropped after a 401 exchange when a Websockets profile is attached to virtual server.
653201-2 3-Major   Update the default CA certificate bundle file to the latest version and remove expiring certificates from it
713533-2 4-Minor   list self-ip with queries does not work
708249-2 4-Minor   nitrox_diag utility generates QKView files with 5 MB maximum file size limit
692095-1 4-Minor K65311501 bigd logs monitor status unknown for FQDN Node/Pool Member
678801-4 4-Minor   WS::enabled returned empty string
677958-4 4-Minor   WS::frame prepend and WS::frame append do not insert string in the right place.


Performance Fixes

ID Number Severity Solution Article(s) Description
698992-1 3-Major   Performance degraded


Global Traffic Manager (DNS) Fixes

ID Number Severity Solution Article(s) Description
713066-1 2-Critical K10620131 Connection failure during DNS lookup to disabled nameserver can crash TMM
707310-2 2-Critical   DNSSEC Signed Zone Transfers of Large Zones Can Be Incomplete (missing NSEC3s and RRSIGs)
721895 3-Major   Add functionality to configure the minimum TLS version advertised and accepted by big3d (iQuery)
715448-2 3-Major   Providing LB::status with a GTM Pool name in a variable caused validation issues
710032-1 3-Major   'No Access' error when viewing GSLB Server's virtual server that has a name indicating a partition that does not exist on that BIG-IP system.
706128-2 3-Major   DNSSEC Signed Zone Transfers Can Leak Memory
703545-1 3-Major   DNS::return iRule "loop" checking disabled


Application Security Manager Fixes

ID Number Severity Solution Article(s) Description
718152 2-Critical K14591455 ASM GUI request log does not load on cluster
716788-2 2-Critical   TMM may crash while response modifications are being performed within DoSL7 filter
713390-1 2-Critical   ASM Signature Update cannot be performed on hourly billing cloud instance
685230-3 2-Critical   memory leak on a specific server scenario
606983-2 2-Critical   ASM errors during policy import
719459-2 3-Major   Wrong add suggestions created for already existing response URLs when differentiate HTTP and HTTPS is disabled
719005-1 3-Major   Login request may arrive corrupted to the backend server after CAPTCHA mitigation
717756-2 3-Major   High CPU usage from asm_config_server
716940-2 3-Major   Traffic Learning screen graphs shows data for the last day only
715128-1 3-Major   Simple mode Signature edit does not escape semicolon
713282-1 3-Major   Remote logger violation_details field does not appear when virtual server has more than one remote logger
712362-3 3-Major   ASM stalls WebSocket frames after legitimate websockets handshake with 101 status code, but without 'Switching Protocols' reason phrase
711405-1 3-Major K14770331 ASM GUI Fails to Display Policy List After Upgrade
710327-1 3-Major   Remote logger message is truncated at NULL character.
707147-1 3-Major   High CPU consumed by asm_config_server_rpc_handler_async.pl
706845-2 3-Major   False positive illegal multipart violation
706665-2 3-Major   ASM policy is modified after pabnagd restart
704643-1 3-Major   Backslashes preceding a forward slash are mistakenly doubled in regular expression within Signature rule
702008-1 3-Major   ASM REST: Missing DB Cleanup for some tables
700143-2 3-Major   ASM Request Logs: Cannot delete second 10,000 records of filtered event log messages
691897-3 3-Major   Names of the modified cookies do not appear in the event log
687759-1 3-Major   bd crash
686765-2 3-Major   Database cleaning failure may allow MySQL space to fill the disk entirely
674256-2 3-Major K60745057 False positive cookie hijacking violation
675232-6 4-Minor   Cannot modify a newly created ASM policy within an iApp template implementation or TMSH CLI transaction


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
710315-1 2-Critical   AVR-profile might cause issues when loading a configuration or when using config sync
698226-1 2-Critical   Filters under Security :: Reports :: DoS :: URL Latencies work incorrectly
696642-1 2-Critical   monpd core is sometimes created when the system is under heavy load.
721474-1 3-Major   AVR does not send all SSLO statistics to offbox machine.
715110 3-Major   AVR should report 'resolutions' in module GtmWideip
712118 3-Major   AVR should report on all 'global tags' in external logs
706361 3-Major   IPS stats tables are empty after upgrade from 13.1.0 to 14.0.0
696212-1 3-Major   monpd does not return data for multi-dimension query
648242-2 3-Major K73521040 Administrator users unable to access all partition via TMSH for AVR reports
649161-2 4-Minor K42340304 AVR caching mechanism not working properly


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
720214-1 2-Critical   NTLM Authentication might fail if Strict Update in iApp is modified
720189-1 2-Critical   VDI settings in APM Webtop have the incorrect URL for Citrix Receiver client download
719149-2 2-Critical   VDI plugin might hang while processing native RDP connections
716747-2 2-Critical   TMM my crash while processing APM or SWG traffic
715250-1 2-Critical   TMM crash when RPC resumes TCL event ACCESS_SESSION_CLOSED
713156-1 2-Critical   AGC cannot do redeploy in Exchange and ADFS use cases
710116-1 2-Critical   VPN clients experience packet loss/disconnection
694078-1 2-Critical   In rare cases, TMM may crash with high APM traffic
720695-1 3-Major   Export then import of APM access Profile/Policy with advanced customization is failing
719192 3-Major   In VPE Agent VMware View Policy shows no properties
715207-3 3-Major   coapi errors while modifying per-request policy in VPE
714961-1 3-Major   antserver creates large temporary file in /tmp directory
714700-2 3-Major   SSO for native RDP resources is not compatible with the 'Force Updated Clients' setting of 'Encryption Oracle Remediation' Group Policy
713111-1 3-Major   When ASM and APM are configured on the same virtual server, ASM might report some errors in its logging.
710305-1 3-Major   When ASM and APM are configured on the same virtual server, ASM could throw some errors in its logging.
709274-1 3-Major   RADIUS Accounting requests egress on different Self IP addresses than they did pre-v13.1.0
699267-2 3-Major   LDAP Query may fail to resolve nested groups
658278-1 3-Major   Network Access configuration with Layered-VS does not work with Edge Client


Service Provider Fixes

ID Number Severity Solution Article(s) Description
703515-3 2-Critical K44933323 MRF SIP LB - Message corruption when using custom persistence key
692310-2 3-Major K69250459 ICAP modified HTTP 1.1 request/response without content-length gets "chunked" even with no body


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
677473-3 2-Critical   MCPD core is generated on multiple add/remove of Mgmt-Rules


Policy Enforcement Manager Fixes

ID Number Severity Solution Article(s) Description
711570-3 3-Major   PEM iRule subscriber policy name query using subscriber ID, may not return applied policies
663874-2 3-Major K77173309 Off-box HSL logging does not work with PEM in SPAN mode.


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
719186-2 3-Major   Multipart/form-data requests may generate false positive 'missing strong integrity parameter' alerts
716318-2 3-Major   Engine/Signatures automatic update check may fail to find/download the latest update


Anomaly Detection Services Fixes

ID Number Severity Solution Article(s) Description
714334-1 2-Critical   admd stops responding and generates a core while under stress.
718772-2 3-Major   The generated signature creates incorrect predicate http.unknown_header (instead of http.unknown_header_exists)
718685-1 3-Major   The measured number of pending requests is two times higher than actual one
701288-1 3-Major   Server health significantly increases during DoSL7 TPS prevention


iApp Technology Fixes

ID Number Severity Solution Article(s) Description
693694-1 3-Major   tmsh::load within IApp template results in unpredicted behavior



Cumulative fixes from BIG-IP v13.1.0.7 that are included in this release


Functional Change Fixes

ID Number Severity Solution Article(s) Description
716392-1 1-Blocking   Support for 24 vCMP guests on a single 4450 blade
712429 1-Blocking   Serverside packets excluded from DoS stats
704552 3-Major   Support for ONAP site licensing


TMOS Fixes

ID Number Severity Solution Article(s) Description
707100 2-Critical   Potentially fail to create user in AzureStack
706688 2-Critical   Automatically add additional certificates to BIG-IP system in C2S and IC environments
709936 3-Major   Multiple NTP server/domain-searches in the DHCP lease can cause missing DNS/NTP configuration.
707585-1 3-Major   Use native driver for 82599 NICs instead of UNIC
703869 3-Major   Waagent updated to 2.2.21


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
713273 2-Critical   BIG-IP sys db variable avr.stats.internal.maxentitiespertable resets back to default value on services restart
715153-1 3-Major   AVR can generate many files in /var/avr/loader folder, which might cause disk-usage problem


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
716746 3-Major   Possible tmm restart when disabling single endpoint vector while attack is ongoing
712710 3-Major   TMM may halt and restart when threshold mode is set to stress-based mitigation


Traffic Classification Engine Fixes

ID Number Severity Solution Article(s) Description
699103-1 3-Major   tmm continuously restarts after provisioning AFM



Cumulative fixes from BIG-IP v13.1.0.6 that are included in this release


Vulnerability Fixes

ID Number CVE Solution Article(s) Description
709972-6 CVE-2017-12613 K52319810 CVE-2017-12613: APR Vulnerability
707186-1 CVE-2018-5514 K45320419 TMM may crash while processing HTTP/2 traffic
702232-1 CVE-2018-5517 K25573437 TMM may crash while processing FastL4 TCP traffic
693312-1 CVE-2018-5518 K03165684 vCMPd may crash when processing bridged network traffic
688516-1 CVE-2018-5518 K03165684 vCMPd may crash when processing bridged network traffic
686305-1 CVE-2018-5534 K64552448 TMM may crash while processing SSL forward proxy traffic
589233-2 CVE-2018-5518 K03165684 vCMPd may crash when processing bridged network traffic
714369 CVE-2018-5526 K62201098 ADM may fail when processing HTTP traffic
714350 CVE-2018-5526 K62201098 BADOS mitigation may fail
710314-1 CVE-2018-5537 K94105051 TMM may crash while processing HTML traffic
706176-1 CVE-2018-5512 K51754851 TMM crash can occur when using LRO
706086-3 CVE-2018-5515 K62750376 PAM RADIUS authentication subsystem hardening
703940-2 CVE-2018-5530 K45611803 Malformed HTTP/2 frame consumes excessive system resources
699346-3 CVE-2018-5524 K53931245 NetHSM capacity reduces when handling errors
688011-7 CVE-2018-5520 K02043709 Dig utility does not apply best practices
688009-7 CVE-2018-5519 K46121888 Appliance Mode TMSH hardening
677088-2 CVE-2018-15321 K01067037 Qkview does not follow current best practices
708653-1 CVE-2018-15311 K07550539 TMM may crash while processing TCP traffic
632875-5 CVE-2018-5516 K37442533 Non-Administrator TMSH users no longer allowed to run dig


Functional Change Fixes

ID Number Severity Solution Article(s) Description
708389 3-Major   BADOS monitoring with Grafana requires admin privilege
680850-2 3-Major K48342409 Setting zxfrd log level to debug can cause AXFR and/or IXFR failures due to high CPU and disk usage.


TMOS Fixes

ID Number Severity Solution Article(s) Description
694897-2 1-Blocking   Unsupported Copper SFP can trigger a crash on i4x00 platforms.
708054-1 2-Critical   Web Acceleration: TMM may crash on very large HTML files with conditional comments
706305-1 2-Critical   bgpd may crash with overlapping aggregate addresses and extended-asn-cap enabled
706087 2-Critical   Entry for SSL key replaced by config-sync causes tmsh load config to fail
703761-2 2-Critical   Disable DSA keys for public-key and host-based authentication in Common Criteria mode
696113-3 2-Critical   Extra IPsec reference added per crypto operation overflows connflow refcount
692683-1 2-Critical   Core with /usr/bin/tmm.debug at qa_device_mgr_uninit
690793-1 2-Critical K25263287 TMM may crash and dump core due to improper connflow tracking
689577-3 2-Critical K45800333 ospf6d may crash when processing specific LSAs
688911-1 2-Critical K94296004 LTM Policy GUI incorrectly shows conditions with datagroups
563661-1 2-Critical   Datastor may crash
704282-2 3-Major   TMM halts and restarts when calculating BWC pass rate for dynamic bwc policy
703298-2 3-Major   Licensing and phonehome_upload are not using the sync'd key/certificate
701626-2 3-Major K16465222 GUI resets custom Certificate Key Chain in child client SSL profile
698429-1 3-Major   Misleading log error message: Store Read invalid store addr 0x3800, len 10
693964-1 3-Major   Qkview utility may generate invalid XML in files contained in Qkview
691497-2 3-Major   tmsh save sys ucs <file> fails due to missing patch file in /config/.diffVersions
691210-1 3-Major   Traffic stops on tmm restart when guest VLAN tagging is used for BIG-IP VE.
687353-1 3-Major K35595105 Qkview truncates tmstat snapshot files
631316-2 3-Major K62532020 Unable to load config with client-SSL profile error
514703-3 4-Minor   gtm listener cannot be listed across partitions


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
709334-1 2-Critical   Memory leak when SSL Forward proxy is used and ssl re-negotiates
708114-1 2-Critical K33319853 TMM may crash when processing the handshake message relating to OCSP, after the SSL connection is closed
707447-1 2-Critical   Default SNI clientssl profile's sni_certsn_hash can be freed while in use by other profiles.
707246-1 2-Critical   TMM would crash if SSL Client profile could not load cert-key-chain successfully
706631-2 2-Critical   A remote TLS server certificate with a bad Subject Alternative Name should be rejected when Common Criteria mode is licensed and configured.
705611-2 2-Critical   The TMM may crash when under load when configuration changes occur when the HTTP/2 profile is used
704666-1 2-Critical   memory corruption can occur when using certain certificates
704435-1 2-Critical   Client connection may hang when NTLM and OneConnect profiles used together
703914-2 2-Critical   TMM SIGSEGV crash in poolmbr_conn_dec.
703191-2 2-Critical   HTTP2 requests may contain invalid headers when sent to servers
701244-1 2-Critical K81742541 An incorrect data manipulation in cipher encrypt and decrypt could cause TMM crash with SIGABRT
701202-3 2-Critical K35023432 SSL memory corruption
700393-3 2-Critical   Under certain circumstances, a stale HTTP/2 stream can cause a tmm crash
697259-2 2-Critical K14023450 Different versioned vCMP guests on the same chassis may crash.
694656-1 2-Critical K05186205 Routing changes may cause TMM to restart
686228-1 2-Critical K23243525 TMM may crash in some circumstances with VLAN failsafe
680074-2 2-Critical   TMM crashes when serverssl cannot provide certificate to backend server.
667770-1 2-Critical K12472293 SIGSEGV to tmm when performing multiple, repeated SSL profile updates, or during UCS restore
648320-5 2-Critical K38159538 Downloading via APM tunnels could experience performance downgrade.
705794-2 3-Major   Under certain circumstances a stale HTTP/2 stream might cause a tmm crash
701147-2 3-Major K36563645 ProxySSL does not work properly with Extended Master Secret and OCSP
700057-4 3-Major   LDAP fails to initiate SSL negotiation because client cert and key associated file permissions are not preserved
693910-4 3-Major   Traffic Interruption for MAC Addresses Learned on Interfaces that Enter Blocked STP State (2000/4000/i2800/i4800 series)
693244-2 3-Major   BIG-IP not sending RST for SYN,ACK packets when ASM is provisioned
690042-1 3-Major K43412307 Potential Tcl leak during iRule suspend operation
689561-1 3-Major   HTTPS request hangs when multiple virtual https servers shares the same ip address
686972-4 3-Major   The change of APM log settings will reset the SSL session cache.
685615-4 3-Major K24447043 Incorrect source mac for TCP Reset with vlangroup for host traffic
677525-2 3-Major   Translucent VLAN group may use unexpected source MAC address
663821-1 3-Major K41344010 SNAT Stats may not include port FTP traffic
653976-4 3-Major K00610259 SSL handshake fails if server certificate contains multiple CommonNames
594751-1 3-Major K90535529 LLDP VLAN Information not Transmitted to Neighbors When Interfaces are Added to a Trunk after the Trunk has Already Been Assigned to a VLAN


Global Traffic Manager (DNS) Fixes

ID Number Severity Solution Article(s) Description
710424-2 2-Critical K00874337 Possible SIGSEGV in GTMD when GTM persistence is enabled.
678861-1 2-Critical   DNS:: namespace commands in procs cause upgrade failure when change from Link Controller license to other


Application Security Manager Fixes

ID Number Severity Solution Article(s) Description
710870 2-Critical   Temporary browser challenge failure after installing older ASU
711011-2 3-Major   'API Security' security policy template changes
683241-1 3-Major K70517410 Improve CSRF token handling


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
710947-1 2-Critical   AVR does not send errdef for entity DosIpLogReporting.
710110-1 2-Critical   AVR does not publish DNS statistics to external log when usr-offbox is enabled.
711929-1 3-Major   AVR should not send statistic on hidden rows for module InterfaceTraffic and module InterfaceHealth


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
679221-2 1-Blocking   APMD may generate core file or appears locked up after APM configuration changed
708005-1 2-Critical K12423316 Users cannot use Horizon View HTML5 client to launch Horizon 7.4 resources
703208-1 2-Critical   PingAccessAgent causes TMM core
702278-2 2-Critical   Potential XSS security exposure on APM logon page.
700522-1 2-Critical   APMD may unexpectedly restart when worker threads are stuck
700090-2 2-Critical   tmm crash during execution of a per-request policy when modified during execution.
699686-1 2-Critical   localdbmgr can occasionally crash during shutdown
697452-1 2-Critical   Websso crashes because of bad argument in logging
712924-1 3-Major   In VPE SecurID servers list are not being displayed in SecurID authentication dialogue
703793-3 3-Major   tmm restarts when using ACCESS::perflow get' in certain events
703171-1 3-Major   High CPU usage for apmd, localdbmgr and oauth processes
702487-3 3-Major   AD/LDAP admins with spaces in names are not supported
684937-3 3-Major K26451305 [KERBEROS SSO] Performance of LRU cache for Kerberos tickets drops gradually with the number of users
683113-3 3-Major K22904904 [KERBEROS SSO][KRB5] The performance of memory type Kerberos ticket cache in krb5 library drops gradually with the number of users
681415-3 3-Major   Copying of profile with advanced customization or images might fail
678427-1 3-Major K03138339 Safari 11 displays F5 EPI and F5 VPN launch confirmation dialogs twice
675775-4 3-Major   TMM crashes inside dynamic ACL building session db callback
671597-3 3-Major   Import, export, copy and delete is taking too long on 1000 entries policy
673717-3 4-Minor   VPE loading times can be very long


Service Provider Fixes

ID Number Severity Solution Article(s) Description
701889-1 2-Critical   Setting log.ivs.level or log-config filter level to informational causes crash
679114-4 3-Major   Persistence record expires early if an error is returned for a BYE command


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
708888-1 2-Critical K79814103 Some DNS truncated responses may not be processed by BIG-IP
667353 2-Critical   Intermittent TMM crash when AFM dynamic (behavioral) signature is enabled and past attack signatures incorrectly exist in the correlation stats table


Policy Enforcement Manager Fixes

ID Number Severity Solution Article(s) Description
702705-2 2-Critical   Core observed during lease query testing when RADIUS Authentication is configured in DHCP profile
699531-1 2-Critical   Potential TMM crash due to incorrect number of attributes in a PEM iRule command
696294-1 2-Critical   TMM core may be seen when using Application reporting with flow filter in PEM
711093-1 3-Major   PEM sessions may stay in marked_for_delete state with Gx reporting and Gy enabled
709610-3 3-Major   Subscriber session creation via PEM may fail due to a RADIUS-message-triggered race condition in PEM
697718-1 3-Major   Increase PEM HSL reporting buffer size to 4K.
677494-1 3-Major   Flow filter with Periodic content insertion action could leak insert content record
677148-1 3-Major   Periodic content insertion creates duplicate records if same policy is added to Global high and subscriber specific
676346-2 3-Major   PEM displays incorrect policy action counters when the gate status is disabled.
648802-1 3-Major   Required custom AVPs are not included in an RAA when reporting an error.


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
710701-1 3-Major   "Application Layer Encryption" option is not saved in DataSafe GUI
709319-2 3-Major   Post-login client-side alerts are missing username in bigIQ
706835 3-Major   When cloning a profile, URL parameters are not shown
706771-1 3-Major   FPS ajax-mapping property may be set even when it should be blocked
706651-1 3-Major   Cloning URL does not clone "Description" field
706276-1 4-Minor   Unnecessary pop-up appears


Device Management Fixes

ID Number Severity Solution Article(s) Description
708305-2 3-Major   Discover task may get stuck in CHECK_IS_ACTIVE step
705593-5 4-Minor   CVE-2015-7940: Bouncy Castle Java Vulnerability



Cumulative fixes from BIG-IP v13.1.0.5 that are included in this release


Functional Change Fixes

None


TMOS Fixes

ID Number Severity Solution Article(s) Description
633441-1 3-Major   Datasync Background Tasks running even without features requiring it


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
708189 4-Minor   OAuth Discovery Auto Pilot is implemented


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
708840 3-Major   13.0.0-to-13.1.0 upgrade operation on VIPRION 2250 blades might fail if global whitelist is configured



Cumulative fixes from BIG-IP v13.1.0.4 that are included in this release


Vulnerability Fixes

ID Number CVE Solution Article(s) Description
700556-1 CVE-2018-5504 K11718033 TMM may crash when processing WebSockets data
699012-1 CVE-2018-5502 K43121447 TMM may crash when processing SSL/TLS data
698080-3 CVE-2018-5503 K54562183 TMM may consume excessive resources when processing with PEM
695901-1 CVE-2018-5513 K46940010 TMM may crash when processing ProxySSL data
691504-1 CVE-2018-5503 K54562183 PEM content insertion in a compressed response may cause a crash.
704580-1 CVE-2018-5549 K05018525 apmd service may restart when BIG-IP is used as SAML SP while processing response from SAML IdP
701447-1 CVE-2017-5754 K91229003 CVE-2017-5754 (Meltdown)
701445-1 CVE-2017-5753
CVE-2017-9074
CVE-2017-7542
CVE-2017-11176
K91229003 CVE-2017-5753 (Spectre Variant 1)
701359-4 CVE-2017-3145 K08613310 BIND vulnerability CVE-2017-3145
699455-4 CVE-2018-5523 K50254952 SAML export does not follow best practices
699451-3 CVE-2018-5511 K30500703 OAuth reports do not follow best practices
676457-5 CVE-2017-6153 K52167636 TMM may consume excessive resource when processing compressed data
640766-2 CVE-2016-10088
CVE-2016-9576
K05513373 Linux kernel vulnerability: CVE-2016-10088 CVE-2016-9576


Functional Change Fixes

ID Number Severity Solution Article(s) Description
686389-1 3-Major   APM does not honor per-farm HTML5 client disabling at the View Connection Server
678524-1 3-Major   Join FF02::2 multicast group when router-advertisement is configured
693007-1 4-Minor   Modify b.root-servers.net IPv4 address 192.228.79.201 to 199.9.14.201 according to InterNIC


TMOS Fixes

ID Number Severity Solution Article(s) Description
707226 1-Blocking   DB variables to disable CVE-2017-5754 Meltdown/PTI mitigations
700315-2 1-Blocking K26130444 Ctrl+C does not terminate TShark
667148-3 1-Blocking K02500042 Config load or upgrade can fail when loading GTM objects from a non-/Common partition
706998-3 2-Critical   Memory leak when OCSP is configured in clientSSL profile for C3D, or in serverSSL profile for server authentication
692890-3 2-Critical   Adding support for BIG-IP 800 in 13.1.x
685458-7 2-Critical K44738140 merged fails merging a table when a table row has incomplete keys defined.
665354-1 2-Critical K31190471 Silent reboot, identified with bad_tlp_status and completion_time_out in the sel log
703848-1 3-Major   Possible memory leak when reusing statistics rows in tables
702520-2 3-Major K53330514 Same AZ failover in AWS fails in some configurations with two or more objects sharing the same IP address.
694740-3 3-Major   BIG-IP reboot during a TMM core results in an incomplete core dump
692753-1 3-Major   shutting down trap not sent when shutdown -r or shutdown -h issued from shell
689691-2 3-Major   iStats line length greater than 4032 bytes results in corrupted statistics or merge errors
686029-2 3-Major   A VLAN delete can result in unrelated VLAN FDB entries being flushed on shared VLAN member interfaces
669462-2 3-Major   Error adding /Common/WideIPs as members to GTM Pool in non-Common partition
589083-6 3-Major   TMSH and iControl REST: When logged in as a remote user who has the admin role, cannot save config because of permission errors.
699281-1 4-Minor   Version format of hypervisor bundle matches Version format of ISO
685475-1 4-Minor K93145012 Unexpected error when applying hotfix


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
706534-1 1-Blocking   L7 connection mirroring may not be fully mirrored on standby BigIP
698424-1 1-Blocking K11906514 Traffic over a QinQ VLAN (double tagged) will not pass
700862-1 2-Critical K15130240 tmm SIGFPE 'valid node'
699298-2 2-Critical   13.0.0 Hotfix HF3 3.0.1679 TMM CORED due to SIGSEGV.
698461-1 2-Critical   tmm may crash in fastl4 TCP
692970-2 2-Critical   Using UDP port 67 for purposes other than DHCP might cause TMM to crash
691095-1 2-Critical   CA bundle manager loses certificates in the CA bundle if the serial number is longer than 4 bytes
687635-1 2-Critical K58002142 Tmm becomes unresponsive and might restart
687205-2 2-Critical   Delivery of HUDEVT_SENT messages at shutdown by SSL may cause tmm restart
681175-3 2-Critical K32153360 TMM may crash during routing updates
674576-3 2-Critical   Outage may occur with VIP-VIP configurations
452283-5 2-Critical   An MPTCP connection that receives an MP_FASTCLOSE might not clean up its flows
440620-1 2-Critical   New connections may be reset when a client reuses the same port as it used for a recently closed connection
704073-1 3-Major K24233427 Repeated 'bad transition' OOPS logging may appear in /var/log/ltm and /var/log/tmm
702439 3-Major K04964898 Non-default HTTP/2 header_table_size causes HTTP/2 streams to be reset
698916-1 3-Major   TMM crash with HTTP/2 under specific condition
698379-2 3-Major K61238215 HTTP2 upload intermittently is aborted with HTTP2 error error_code=FLOW_CONTROL_ERROR(
698000-3 3-Major K04473510 Connections may stop passing traffic after a route update
695707-5 3-Major   BIG-IP does not retransmit DATA_FIN when closing an MPTCP connection
691806-1 3-Major K61815412 RFC 793 - behavior receiving FIN/ACK in SYN-RECEIVED state
689449-1 3-Major   Some flows may remain indefinitely in memory with spdy/http2 and http fallback-host configured
688571-2 3-Major K40332712 Untrusted cert might be accepted by the server-ssl even though when 'untrusted-cert-response-control drop' is configured in the server-ssl profile.
688570-5 3-Major   BIG-IP occasionally sends MP_FASTCLOSE after an MPTCP connection close completes
686307-3 3-Major K10665315 Monitor Escaping is not changed when upgrading from 11.6.x to 12.x and later
686065-2 3-Major   RESOLV::lookup iRule command can trigger crash with slow resolver
682104-3 3-Major   HTTP PSM leaks memory when looking up evasion descriptions
680264-2 3-Major K18653445 HTTP2 headers frame decoding may fail when the frame delivered in multiple xfrags
677666-2 3-Major K60909141 /var/tmstat/blades/scripts segment grows in size.
664528-2 3-Major K53282793 SSL record can be larger than maximum fragment size (16384 bytes)
251162-1 3-Major K11564 The error message 'HTTP header exceeded maximum allowed size' may list the wrong profile name
685467-1 4-Minor K12933087 Certain header manipulations in HTTP profile may result in losing connection.


Global Traffic Manager (DNS) Fixes

ID Number Severity Solution Article(s) Description
699135-1 2-Critical   tmm cores with SIGSEGV in dns_rebuild_response while using host command for not A/AAAA wideip
692941-1 2-Critical   GTMD and TMM SIGSEGV when changing wide IP pool in GTMD
691287-1 2-Critical   tmm crashes on iRule with GTM pool command
682335-1 2-Critical   TMM can establish multiple connections to the same gtmd
580537-3 2-Critical   The GeoIP update script geoip_update_data cannot be used to install City2 GeoIP data
562921-5 2-Critical   Cipher 3DES and iQuery encrypting traffic between BIG-IP systems
705503-3 3-Major   Context leaked from iRule DNS lookup
703702 3-Major   Fixed iControl REST not listing GTM Listeners
700527-3 3-Major   cmp-hash change can cause repeated iRule DNS-lookup hang
699339-3 3-Major K24634702 Geolocation upgrade files fail to replicate to secondary blades
696808-1 3-Major   Disabling a single pool member removes all GTM persistence records
691498-3 3-Major   Connection failure during iRule DNS lookup can crash TMM
690166-1 3-Major   ZoneRunner create new stub zone when creating a SRV WIP with more subdomains
687128-1 3-Major   gtm::host iRule validation for ipv4 and ipv6 addresses
680069-1 3-Major K81834254 zxfrd core during transfer while network failure and DNS server removed from DNS zone config
679149-1 3-Major   TMM may crash or LB::server returns unexpected result due to reused lb_result->pmbr[0]
667469-3 3-Major K35324588 Higher than expected CPU usage when using DNS Cache
636997-1 4-Minor   big3d may crash
636994-1 4-Minor   big3d may crash
636992-1 4-Minor   big3d may crash
636986-1 4-Minor   big3d may crash
636982-1 4-Minor   big3d may crash


Application Security Manager Fixes

ID Number Severity Solution Article(s) Description
705774-1 3-Major   Add a set of disallowed file types to RDP template
703833-1 3-Major   Some bot detected features might not work as expected on Single Page Applications
702946-3 3-Major   Added option to reset staging period for signatures
701841-2 3-Major   Unnecessary file recovery_db/conf.tar.gz consumes /var disk space
701327-2 3-Major   failed configuration deletion may cause unwanted bd exit
700812-1 3-Major   asmrepro recognizes a BIG-IP version of 13.1.0.1 as 13.1.0 and fails to load a qkview
700726-2 3-Major   Search engine list was updated, and fixing case of multiple entries
698919-3 3-Major   Anti virus false positive detection on long XML uploads
697756-1 3-Major   Policy with CSRF URL parameter cannot be imported as binary policy file
697303-1 3-Major   BD crash
696265-5 3-Major K60985582 BD crash
696073-2 3-Major   BD core on a specific scenario
695563-1 3-Major   Improve speed of ASM initialization on first startup
694922-5 3-Major   ASM Auto-Sync Device Group Does Not Sync
693780-1 3-Major   Proactive Bot Defense sends CAPTCHA to the UC browser on iOS devices
693663-1 3-Major   Proactive Bot Defense sends CAPTCHA to the Firefox browser on iOS only in desktop mode
691477-2 3-Major   ASM standby unit showing future date and high version count for ASM Device Group
679384-3 3-Major K85153939 The policy builder is not getting updates about the newly added signatures.
678293-2 3-Major K25066531 Uncleaned policy history files cause /var disk exhaustion
665992-2 3-Major K40510140 Live Update via Proxy No Longer Works
608988-1 3-Major   Error when deleting multiple ASM Policies


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
703233 3-Major   Some filters don't work in Security->Reporting->URL Latencies page


Access Policy Manager Fixes

ID Number Severity Solution Article(s) Description
707676-1 2-Critical   Memory leak in Machine Certificate Check agent of the apmd process
700724-2 2-Critical   Client connection with large number of HTTP requests may cause tmm to restart
692557-1 2-Critical   When BIG-IP as SAML IdP processes signed authentication requests from external SAML SP, a block of memory may become corrupted.
690116-1 2-Critical   websso daemon might crash when logging set to debug
689591-2 2-Critical   When pingaccess SDK processes certain POST requests from the client, the TMM may restart
677368-2 2-Critical   Websso crash due to uninitialized member in websso context object while processing a log message
631286-3 2-Critical   TMM Memory leak caused by APM URI cache entries
703429-2 3-Major   Citrix Receiver for Android (v3.13.1) crashes while accessing PNAgent services
702263-1 3-Major   An access profile with large number of SAML Resources (greater than 200) causes APM error ERR_TOOBIG while loading.
702222-1 3-Major   RADIUS and SecurID Auth fails with empty password
701740-1 3-Major   apmd leaks memory when updating Access V2 policy
701737-1 3-Major   apmd may leak memory on destroying Kerberos cache
701736-1 3-Major   Memory leak in Machine Certificate Check agent of the apmd process
701639-1 3-Major   Session variables in Requested Authentication Context Class in SP do not get resolved when Authentication Request is generated by BIG-IP as SP.
697636-3 3-Major   ACCESS is not replacing headers while replacing POST body
695953-1 3-Major   Custom URL Filter object is missing after load sys config TMSH command
694624-1 3-Major   SSO enabled Native RDP resources can't be accessed via hardware BIG-IP with "Intel Cave Creek" coprocessor
693844-1 3-Major K58335157 APMD may restart continuously and cannot come up
692307-3 3-Major   User with 'operator' role may not be able to view some session variables
687937-1 3-Major   RDP URIs generated by APM Webtop are not properly encoded
685862-1 3-Major   BIG-IP as SAML IdP/SP may include last x509 certificate found in the configured bundle in signed SAML Response or single logout message
684583-1 3-Major   Buitin Okta Scopes Request object uses client -id and client-secret
684325-1 3-Major   APMD Memory leak when applying a specific access profile
683389-3 3-Major   Error #2134 when attempting to create local flash.net::SharedObject in rewritten ActionScript 3 file
683297-2 3-Major   Portal Access may use incorrect back-end for resources referenced by CSS
682500-2 3-Major   VDI Profile and Storefront Portal Access resource do not work together
678851-3 3-Major   Portal Access produces incorrect Java bytecode when rewriting java.applet.AppletStub.getDocumentBase()
675866-4 3-Major   WebSSO: Kerberos rejects tickets with 2 minutes left in their ticket lifetime, causing APM to disable SSO
671627-3 3-Major K06424790 HTTP responces without body may contain chunked body with empty payload being processed by Portal Access.
632646-1 3-Major   APM - OAM login with ObSSOCookie results in error page instead of redirecting to login page, when session cookie (ObSSOCookie) is deleted from OAM server.
629334-1 3-Major   Portal Access: JavaScript expressions in parentheses may be rewritten incorrectly
612792-1 3-Major   Support RDP redirection for connections launched from APM Webtop on iOS
612118-2 3-Major   Nexthop explicit proxy is not used for the very first connection to communicate with the backend.
536831-1 3-Major   APM PAM module does not handle local-only users list correctly


Service Provider Fixes

ID Number Severity Solution Article(s) Description
698338-1 2-Critical   Potential core in MRF occurs when pending egress messages are queued and an iRule error aborts the connection
689343-2 2-Critical   Diameter persistence entries with bi-directional flag created with 10 sec timeout
685708-4 2-Critical   Routing via iRule to a host without providing a transport from a transport-config created connection cores
700571-4 3-Major   SIP MR profile, setting incorrect branch param for CANCEL to INVITE
696049-1 3-Major   High CPU load on generic message if multiple responses arrive while asynchronous Tcl command is running
674747-4 3-Major K30837366 sipdb cannot delete custom bidirectional persistence entries.
656901-3 3-Major   MRF add 'existing_connection_only' and 'outgoing_connection_instance_seed' two iRule commands


Advanced Firewall Manager Fixes

ID Number Severity Solution Article(s) Description
704207-1 2-Critical   DNS query name is not showing up in DNS AVR reporting
703517 2-Critical K23520761 TMM may crash when processing TCP DNS traffic
692328-1 2-Critical   Tmm core due to incorrect memory allocation
705161-1 3-Major K23520761 TMM may crash when processing TCP DNS traffic
703959 3-Major   Manual thresholds for DoS Protection Dynamic Signatures are not accepted via TMUI
631418-1 3-Major   Packets dropped by HW grey list may not be counted toward AVR.


Policy Enforcement Manager Fixes

ID Number Severity Solution Article(s) Description
696383-1 2-Critical   PEM Diameter incomplete flow crashes when sweeped
694717-1 2-Critical   Potential memory leak and TMM crash due to a PEM iRule command resulting in a remote lookup.
616008-1 2-Critical K23164003 TMM core may be seen when using an HSL format script for HSL reporting in PEM
696789-1 3-Major   PEM Diameter incomplete flow crashes when TCL resumed
695968-1 3-Major   Memory leak in case of a PEM Diameter session going down due to remote end point connectivity issues.
694319-1 3-Major   CCA without a request type AVP cannot be tracked in PEM.
694318-1 3-Major   PEM subscriber sessions will not be deleted if a CCA-t contains a DIAMETER_TOO_BUSY return code and no request type AVP.
684333-1 3-Major   PEM session created by Gx may get deleted across HA multiple switchover with CLI command
678820-1 3-Major   Potential memory leak if PEM Diameter sessions are not created successfully.
642068-4 3-Major   PEM: Gx sessions will stay in marked_for_delete state if CCR-T timeout happens
624231-4 3-Major   No flow control when using content-insertion with compression
680729-1 4-Minor K64307999 DHCP Trace log incorrectly marked as an Error log.


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
697363-1 2-Critical   FPS should forward all XFF header values
705559-1 3-Major   FPS: false positive "no strong integrity param" when none of the configured data-integrity params are present in request
662311-1 3-Major   CS alerts should contain actual client IP address in XFF header


Protocol Inspection Fixes

ID Number Severity Solution Article(s) Description
671716-1 3-Major   UCS version check was too strict for IPS hitless upgrade



Cumulative fixes from BIG-IP v13.1.0.3 that are included in this release


Functional Change Fixes

None


TMOS Fixes

ID Number Severity Solution Article(s) Description
660239-6 4-Minor   When accessing the dashboard, invalid HTTP headers may be present


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
677919-4 3-Major   Enhanced Data Manipulation AJAX Support



Cumulative fixes from BIG-IP v13.1.0.2 that are included in this release


Vulnerability Fixes

ID Number CVE Solution Article(s) Description
681955-1 CVE-2017-3167 CVE-2017-3169 CVE-2017-7679 CVE-2017-9788 K23565223 Apache CVE-2017-9788
673595-9 CVE-2017-3167 CVE-2017-3169 K34125394 Apache CVE-2017-3167
694274-1 CVE-2017-3167 CVE-2017-3169 CVE-2017-7679 CVE-2017-9788 CVE-2017-9798 K23565223 [RHSA-2017:3195-01] Important: httpd security update - EL6.7
672124-6 CVE-2018-5541 K12403422 Excessive resource usage when BD is processing requests
673607-9 CVE-2017-3169 K83043359 Apache CVE-2017-3169
672667-6 CVE-2017-7679 K75429050 CVE-2017-7679: Apache vulnerability
641101-7 CVE-2016-8743 K00373024 httpd security and bug fix update CVE-2016-8743
684033-3 CVE-2017-9798 K70084351 CVE-2017-9798 : Apache Vulnerability (OptionsBleed)
661939-2 CVE-2017-2647 K32115847 Linux kernel vulnerability CVE-2017-2647


Functional Change Fixes

ID Number Severity Solution Article(s) Description
685056 3-Major   VE OVAs is not the supported platform to run VMware guest OS customization
670103-1 3-Major   No way to query logins to BIG-IP in TMUI
681385-2 4-Minor   Forward proxy forged cert lifespan can be configured from days into hours.


TMOS Fixes

ID Number Severity Solution Article(s) Description
700247 2-Critical K60053504 APM Client Software may be missing after doing fresh install of BIG-IP VE
693979 3-Major   Autoscale is not functional because of the change in file permission of the /shared/vadc/aws/iid-document document
683131-1 3-Major   Hotfix install fails on vCMP guest when both guest and host have the same base version ISO present
682213-1 3-Major K31623549 TLS v1.2 support in IP reputation daemon
669585-1 3-Major   The tmsh sys log filter is unable to display information in uncompressed log files.
668826-1 3-Major   File named /root/.ssh/bigip.a.k.bak is present but should not be
668276-1 3-Major   BIG-IP does not display failed login attempts since last login in GUI
668273-1 3-Major K12541531 Logout button not available in Configuration Utility when using Client Cert LDAP
471237-4 3-Major K12155235 BIG-IP VE instances do not work with an encrypted disk in AWS.


Local Traffic Manager Fixes

ID Number Severity Solution Article(s) Description
699624-1 2-Critical K93400155 Config with custom 'SIP' or 'Firepass' monitor fails to load after upgrade
463097-5 3-Major   Clock advanced messages with large amount of data maintained in DNS Express zones


Global Traffic Manager (DNS) Fixes

ID Number Severity Solution Article(s) Description
672504-2 2-Critical K52325625 Deleting zones from large databases can take excessive amounts of time.
667542-6 2-Critical   DNS Express does not correctly process multi-message DNS IXFR updates.
645615-6 2-Critical K70543226 zxfrd may fail and restart after multiple failovers between blades in a chassis.
655233-2 3-Major K93338593 DNS Express using wrong TTL for SOA RRSIG record in NoData response
648766-2 3-Major K57853542 DNS Express responses missing SOA record in NoData responses if CNAMEs present
646615-2 4-Minor   Improved default storage size for DNS Express database


Application Security Manager Fixes

ID Number Severity Solution Article(s) Description
699720-1 2-Critical   ASM crash when configuring remote logger for WebSocket traffic with response-logging:all
691670-5 2-Critical   Rare BD crash in a specific scenario
686108-1 2-Critical   User gets blocking page instead of captcha during brute force attack
684312-1 2-Critical K54140729 During Apply Policy action, bd agent crashes, causing the machine to go Offline
698940-1 3-Major   Add new security policy template for API driven systems - "API Security"
690883-1 3-Major   BIG-IQ: Changing learning mode for elements does not always take effect
686517-2 3-Major   Changes to a parent policy that has no active children are not synced to the secondary chassis slots.
686470-1 3-Major   Enable AJAX Response Page or Single Page Application support causes the part of the web page failed to load.
686452-1 3-Major   File Content Detection Formats are not exported in Policy XML
685964-1 3-Major   cs_qualified_urls bigdb does not cause configured URLs to be qualified.
685771-1 3-Major   Policies cannot be created with SAP, OWA, or SharePoint templates
685207-1 3-Major   DoS client side challenge does not encode the Referer header.
685164-1 3-Major   In partitions with default route domain != 0 request log is not showing requests
683508-1 3-Major K00152663 WebSockets: umu memory leak of binary frames when remote logger is configured
680353-1 3-Major   Brute force sourced based mitigation is not working as expected
674494-4 3-Major K77993010 BD memory leak on specific configuration and specific traffic
668184-2 3-Major   Huge values are shown in the AVR statistics for ASM violations
694073-3 4-Minor   All signature update details are shown in 'View update history from previous BIG-IP versions' popup
685193-1 4-Minor   If Inheritance is None in the Parent Policy and there are at least 1 child policy, the number of Comments shown in Inheritance Settings is equal to number of child policies


Application Visibility and Reporting Fixes

ID Number Severity Solution Article(s) Description
679861 1-Blocking   Weak Access Restrictions on the AVR Reporting Interface
697421 3-Major   Monpd core when trying to restart
688813-2 3-Major K23345645 Some ASM tables can massively grow in size.
686510-1 3-Major   If tmm was restarted during an attack, the attack might appear ongoing in GUI
683474 3-Major   The case-sensitive problem during comparison of 2 Virtual Servers
679088-1 3-Major   Avr reporting and analytics does not display statistics of many source regions


Fraud Protection Services Fixes

ID Number Severity Solution Article(s) Description
684852-1 2-Critical   Obfuscator not producing deterministic output
692123 3-Major   GET parameter is grayed out if MobileSafe is not licensed


Anomaly Detection Services Fixes

ID Number Severity Solution Article(s) Description
700320 2-Critical   tmm core under stress when BADOS configured and attack signatures enabled
691462-1 3-Major   Bad actors detection might not work when signature mitigation blocks bad traffic
687987 3-Major   Presentation of signatures in human-readable format
687986 3-Major   High CPU consumption during signature generation, not limited number of signatures per virtual server
687984 3-Major   Attacks with randomization of HTTP headers parameters generates too many signatures


Traffic Classification Engine Fixes

ID Number Severity Solution Article(s) Description
698396-1 2-Critical   Config load failed after upgrade from 12.1.2 to 13.x or 14.x



Cumulative fixes from BIG-IP v13.1.0.1 that are included in this release


Functional Change Fixes

None


TMOS Fixes

ID Number Severity Solution Article(s) Description
686190-1 2-Critical   LRO performance impact with BWC and FastL4 virtual server
667173-1 2-Critical   13.1.0 cannot join a device group with 13.1.0.1
683114-2 3-Major   Need support for 4th element version in Update Check


Performance Fixes

ID Number Severity Solution Article(s) Description
685628-1 1-Blocking   Performance regression on B4450 blade
673832-1 1-Blocking   Performance impact for certain platforms after upgrading to 13.1.0.
696525-1 2-Critical   B2250 blades experience degraded performance.

 

Cumulative fix details for BIG-IP v13.1.1.3 that are included in this release

745783-3 : Anti-fraud: remote logging of login attempts

Component: Fraud Protection Services

Symptoms:
There is no support for logging of login attempts to a remote service.

Conditions:
Using high speed logging (HSL) to log login attempts.

Impact:
There is no support for logging of login attempts.

Workaround:
None.

Fix:
FPS now uses HSL to report login attempts using configured templates, rate-limit, and publisher to a remote service.

To enable this feature:

# via tmsh only
tmsh modify sys db antifraud.riskengine.reportlogins value enable
 
# via tmsh or GUI
tmsh modify sys db antifraud.internalconfig.string1 value "<login attempt log template>"
tmsh modify sys db antifraud.internalconfig.string2 value "<log-rate-exceeded log template>"
tmsh modify sys db antifraud.internalconfig.number1 value "<log-rate-exceeded threshold>"
tmsh modify security anti-fraud profile <fps profile> risk-engine-publisher <publisher>
 
 
It is recommended that you use encoding when composing an HTTP template. The default encoding level is 0, meaning 'never encode'.

To change encoding level:

tmsh modify sys db antifraud.internalconfig.number2 value <0/1/2>

Behavior Change:
FPS now uses HSL to report login attempts using configured templates, rate-limit, and publisher to a remote service.

To enable this feature:

# via tmsh only
tmsh modify sys db antifraud.riskengine.reportlogins value enable
 
# via tmsh or GUI
tmsh modify sys db antifraud.internalconfig.string1 value "<login attempt log template>"
tmsh modify sys db antifraud.internalconfig.string2 value "<log-rate-exceeded log template>"
tmsh modify sys db antifraud.internalconfig.number1 value "<log-rate-exceeded threshold>"
tmsh modify security anti-fraud profile <fps profile> risk-engine-publisher <publisher>
 
 
You should probably use encoding when composing an HTTP template. The default encoding level is 0, meaning 'never encode'.

To change encoding level:

tmsh modify sys db antifraud.internalconfig.number2 value <0/1/2>


743810-1 : AWS: Disk resizing in m5/c5 instances fails silently.

Component: TMOS

Symptoms:
Resizing the disk in an m5/c5 instance in AWS has no impact on the disk size, and this operation fails silently.

Conditions:
BIG-IP system deployed in AWS with m5/c5 instance type that uses NVME disks instead of the usual native paravirt disks.

Impact:
Disk resizing fails; administrators cannot grow the instance disk in response to their requirements and instance utilization post-deployment.

Workaround:
There is no workaround.

Fix:
AWS: Disk resizing now works as expected.


743082-1 : Upgrades to 13.1.1.2, 14.0.0, 14.0.0.1, or 14.0.0.2 from pre-12.1.3 can cause configurations to fail with GTM Pool Members

Component: TMOS

Symptoms:
Upgrading to 13.1.1.2, 14.0.0, 14.0.0.1, or 14.0.0.2 might result a configuration that fails to load due to a stray colon-character that gets added to the configuration file.

Conditions:
-- Upgrade from pre-12.1.3 to 13.1.1.2, 14.0.0, 14.0.0.1, or 14.0.0.2.
-- GTM Pool Members.

Impact:
Configuration fails to load.

Workaround:
Remove stray colon-character from bigip_gtm.conf.

Fix:
Fixed an issue preventing configurations from loading into 13.1.1.2, 14.0.0, 14.0.0.1, or 14.0.0.2 from a version before 12.1.3.


741858-1 : TMM may crash while processing Portal Access requests

Solution Article: K52206731


740777-1 : Secondary blades mcp daemon restart when subroutine properties are configured

Component: Access Policy Manager

Symptoms:
Secondary blades' MCP daemons restart with the following error messages:
-- err mcpd[26818]: 01070734:3: Configuration error: Configuration from primary failed validation: 01020036:3: The requested Subroutine Properties (/Common/confirm_10 /Common/confirm_10) was not found.... failed validation with error 16908342.

Conditions:
When a subroutine is configured in the access policy.

Impact:
The secondary blade MCP daemon restarts. Cannot use subroutine in the access policy.

Workaround:
There is no workaround other than to not use subroutine in the access policy.

Fix:
You can now use subroutines in the access policy.


740719-2 : ASM CSP header parser does not honor unsafe-inline attribute within script-src directive

Component: Application Security Manager

Symptoms:
Browser reports Content-Security-Policy error when ASM modifies the 'Content-Security-Policy' (CSP) header.

Conditions:
1. ASM provisioned.
2. ASM policy attached to a virtual server.
3. CSRF or Ajax blocking page enabled within ASM policy
4. Backend server sends 'Content-Security-Policy' header with 'script-src' 'unsafe-inline' directive.

Impact:
Browser posts 'Content-Security-Policy' error and stops JavaScript execution.

Workaround:
Disable 'Content-Security-Policy' header parsing for ASM policies. To do so, follow these steps:

1. In /usr/share/ts/bin/add_del_internal, run the following command:
add csp_enabled 0

2. Restart ASM by running the following command:
bigstart restart asm

Fix:
ASM 'Content-Security-Policy' header parser no longer modifies the 'Content-Security-Policy' header when there is 'script-src' 'unsafe-inline' directive arriving from a backend server. This is correct behavior.


739947-1 : TMM may crash while processing APM traffic

Component: Access Policy Manager

Symptoms:
Under certain condition TMM may crash while processing APM traffic

Conditions:
APM enabled.

Impact:
TMM crash leading to a failover event.

Workaround:
None.

Fix:
TMM now correctly processes APM traffic.


739939-1 : Ping Access Agent Module leaks memory in TMM.

Component: Access Policy Manager

Symptoms:
Whenever a ping access request traverses through apm_ivs (internal virtual server) to the Ping Access Agent (e.g., Cache-Miss), a memory leak occurs.

Conditions:
Ping Access Agent request that is not served from local cache (e.g., in the Cache-Miss case).

Impact:
The slow memory leak might eventually cause TMM to restart. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Ping Access Agent Module no longer leaks memory in TMM.


739716-2 : APM Subroutine loops without finishing

Component: Access Policy Manager

Symptoms:
Some APM subroutines never finish. In authentication use cases, the end-users will continue to see logon pages, even after submitting correct credentials. In SWG confirm-and-continue use cases, the end-users will continue to see confirm boxes, even if they had already clicked "continue".

Conditions:
APM with a subroutine. Reproducibility will vary by platform and deployment.

Impact:
Subroutines never finish. End-users are not able to access resources.

Workaround:
TMM restart does resolve the issue.

Fix:
The memory corruption issue was resolved and the subroutines now correctly finish.


739674-1 : TMM might core in SWG scenario with per-request policy.

Component: Access Policy Manager

Symptoms:
TMM can core when Secure Web Gateway (SWG) is configured, and SSL is resumed through per-request policy.

Conditions:
-- SWG is configured.
-- SSL is resumed through per-request policy.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
TMM does not core now when using SWG scenario with per-request policy.


739507 : How to recover from a failed state due to FIPS integrity check

Component: TMOS

Symptoms:
After FIPS 140-2 license is installed on FIPS-certified hardware devices, and the device rebooted, the system halts upon performing FIPS integrity check.

Conditions:
[1] Some system applications, monitored by FIPS 140-2, get routinely changed.
[2] The device was containing a FIPS 140-2 enabled license installed.
[3] The device operator installs a FIPS 140-2 enabled license
[4] The device is rebooted

Impact:
The device is halted and cannot be used.

Workaround:
Workaround:
[1] The device needs to have serial console access (Telnet).
[2] From the Telnet console, enter the GRUB menu and boot into a different partition not having a FIPS 140-2 enabled license.
[3] Examine the contents of file /config/fipserr which will show the files that were changed, leading to failure of the FIPS 140-2 license-enabled partition.
[4] Restore those files to their original ones and reboot.

If system still halts, repeat from Step [1] above, until this no longer happens.

Fix:
Here are the steps, in summary form.

[1] Connect a terminal to the BIG-IP serial console port
[2] From the Telnet console, enter the GRUB menu.
[3] Before the countdown expires, use the Up Arrow and Down Arrow keys to stop the countdown, and select the appropriate boot image.
[4] Press the key 'E' to start the edit options. A new GRUB menu displays.
[5] Use the Up Arrow and Down Arrow keys to navigate to the line that contains the keyword "module".
[6] Add a space, followed by NO_FIPS_INTEGRITY=1. DO NOT press ENTER.
[7] Press the Ctrl-X sequence or the F10 key to restart the system using the modified options.

The machine will boot into the partition containing FIPS 140-2-enabled license.

[8] Examine the content of file /config/f5_public/fipserr to ascertain the cause of the FIPS module startup error.
[9] Fix the problem reported in the aforementioned error file.
[10] Run the test tool /usr/libexec/sys-eicheck.py to ensure that no fatal error is reported, such as:

Integrity Check Result: [ FAIL ]

If fatal error persists, DO NOT REBOOT (otherwise the system will go into the halt state, and the steps starting from Step [1] will need to be repeated). Instead, fix the problematic files reported. Re-run the test tool until no error is seen.


739505 : Automatic ISO digital signature checking not required when FIPS license active

Component: TMOS

Symptoms:
When the FIPS license is active, digital signature checking of the ISO is automatically performed. This requires that both the ISO and the digital signature (.sig) file are uploaded to the system. Installation will not complete if the .sig file is not present or not valid.

Conditions:
The .sig file corresponding to the ISO to be installed is either missing or invalid.

Impact:
Installation failure.

Workaround:
Follow the procedure in K2434114 (https://support.f5.com/csp/article/K24341140) to validate the ISO on the BIG-IP.

Fix:
The restriction of requiring automatic signature checking of the ISO is removed. The procedure in the KB article above to perform the checks on or off BIG-IP is still valid, but that checking is optional.


739285-1 : GUI partially missing when VCMP is provisioned

Component: TMOS

Symptoms:
GUI may be partially missing.

Conditions:
VCMP must be provisioned.

Impact:
GUI may be partially missing.

Workaround:
Use tmsh or deprovision VCMP.

Fix:
the GUI now works as expected when VCMP is provisioned.


739277 : TMM core when the virtual server is deleted during POST data and BADOS standard/aggressive mode

Component: Anomaly Detection Services

Symptoms:
TMM core when the virtual server is deleted during POST data and BADOS standard/aggressive mode

Conditions:
-- The virtual server is deleted during POST data.
-- BADOS configured in standard/aggressive mode.

Impact:
TMM core / traffic does not path through till TMM restarts.

Workaround:
You can work around this issue using either of the following options if you plan to delete a virtual server:

-- First detach the DoS protection profile.
-- Remove BADOS configuration or change BADOS mitigation to Conservative.

Fix:
This release corrects an TMM crash that occurred when the virtual server was deleted during POST data and BADOS was in standard/aggressive mode.


739190 : Policies could be exported with not patched /Common partition

Component: Access Policy Manager

Symptoms:
Policies could be exported with not patched /Common partition and it's heading to profiles that are not being imported.

Conditions:
Policy has objects outside of partition of the policy.

Impact:
Policy cannot be imported on the same system it was exported from.

Workaround:
There is no workaround.

Fix:
Proper naming of partitions has been restored, import is back to working.


739126 : Multiple VE installations may have different sized volumes

Component: TMOS

Symptoms:
When installing a 2nd, 3rd, (or more) version of BIG-IP to a Virtual Edition (VE) instance, the sizes of the non-shared volumes may be smaller than the first. This can be an issue if, for example, /var is smaller and fills up due to UCS archives, data gathered during troubleshooting, etc.

Conditions:
Install an additional version of BIG-IP to an existing VE instance.

Impact:
Disk volumes may run out of space sooner than expected, leading to issues when that space is needed for other operations.

Workaround:
Provision additional disk space to expand the available storage.

Fix:
In this release, the installer handles this condition without issue.


738669-2 : Login validation may fail for a large request with early server response

Component: Fraud Protection Services

Symptoms:
in case of large request/response, if FPS needs to store ingress and ingress chunks in buffer for additional processing (ingress :: for parameter parsing, egress :: for login validation's banned/mandatory strings check or scripts injection), if the server responds fast enough, the buffer may contain mixed parts of request/response. This may have several effects, from incorrectly performing login-validation to generating a tmm core file.

Conditions:
-- Login validation is enabled and configured to check for banned/mandatory string.
-- A username parameter is configured.
-- There are no parameters configured for encrypt/HTML Field Obfuscation (HFO), and no decoy parameters.
-- There is a large request and response.
-- The system response very quickly.

Impact:
This results in one or more of the following:
-- Login validation failure/skip.
-- Bad response/script injection.
-- tmm core. In this case, traffic is disrupted while tmm restarts.

Workaround:
None.

Fix:
FPS now handles ingress/egress buffers separately, so this issue no longer occurs.


738582-1 : Ping Access Agent Module leaks memory in TMM.

Component: Access Policy Manager

Symptoms:
While handling Ping Access Request, if the internal events passing between modules fail, it might cause a memory leak inside TMM.

Conditions:
Internal events passing between Ping Access Request processing modules fail.

Impact:
Ping Access Agent Module leaks memory in TMM.

Workaround:
None.


738397-1 : SAML -IdP-initiated case that has a per-request policy with a logon page in a subroutine fails.

Component: Access Policy Manager

Symptoms:
For a BIG-IP system configured as IdP, in the SAML-IdP-initiated case: If the IdP has a Per-Request policy (in addition to a V1 policy), such that the Per-Request policy has a subroutine or a subroutine macro with a logon page, the system reports access failure on running the policy.

The error logged is similar to the following: '/Common/sso_access:Common:218f759e: Authorization failure: Denied request for SAML resource /Common/saml_resource_obj1&state=000fffff032db022'.

Conditions:
-- BIG-IP system configured as IdP.
-- SAML IdP initiated case in which the following is true:
  + The IdP has a Per-Request policy (in addition to a V1 policy).
  + That Per-Request policy has a subroutine or a subroutine macro with a logon page.

Impact:
Access is denied. The request URI has '&state=<per-flow-id>' appended to the SAML Resource name.

Workaround:
None.

Fix:
The system now checks for additional query parameters in the request_uri while extracting Resource name so it is not incorrectly set to 'ResourceName&state=<per-flow-id>' which causes the access failure.


737867-1 : Scheduled reports are being incorrectly displayed in different partitions

Component: Application Visibility and Reporting

Symptoms:
Scheduled analytics reports are being displayed in the list regardless of selected partition. However, you can only open reports belonging to the selected partition, even if they belong to 'Common'.

Conditions:
System configured with multiple partitions.

Impact:
It makes it difficult to modify reports from different partitions.

Workaround:
Switch to the report's partition before editing it.

Fix:
Report's partition is now indicated in the list and correct handling is performed according to standard partition rules.


737550 : State Mirroring between BIG-IP 13.0.x and 13.1.x systems may cause TMM core on standby system during upgrade

Component: Local Traffic Manager

Symptoms:
BIG-IP devices running 13.0.x (13.0.x or a 13.0.x point release) and 13.1.x software versions in a High-Availability (HA) configuration with state mirroring enabled may cause a standby system to produce a TMM core file.

Conditions:
-- State mirroring configured on two or more BIG-IP systems (state mirroring is enabled by default).
-- For VIPRION clusters or VIPRION-based vCMP guests, the systems are configured to mirror 'Between Clusters'.
-- The active system is running v13.0.x, and the standby system is running v13.1.x, e.g., as a result of an in-progress upgrade.

Impact:
TMM may crash on a standby system during upgrade.

This issue should not disrupt traffic, because the TMM is coring only on the standby unit.

Workaround:
To workaround this issue, disable State Mirroring prior to upgrading, and re-enable it once both devices are running v13.1.x, or complete the upgrade of both devices to v13.1.x.

1. You can disable mirroring using either the GUI or the command line.

1a. In the GUI:
-- Set the Primary and Secondary Local Mirror Address configurations to 'None' under Device Management :: Devices :: [Self] device :: Mirroring Configuration.

1b. From the command-line:
-- Run the following command:
tmsh modify cm device <name-of-self-device> mirror-ip any6 mirror-secondary-ip any6 && tmsh save sys config


Important: This action results in connection state loss on failover.


2. Once all devices are running the same software version, re-enable state mirroring by re-adding the device mirror IPs removed previously.

Note: F5 recommends that BIG-IP systems run with the same software version on all devices.

Fix:
TMM on standby no longer cores during upgrade.


737500-2 : Apply Policy and Upgrade time degradation when there are previous enforced rules

Component: Application Security Manager

Symptoms:
When there are previously enforced rules present in the system, the time to apply changes made to a policy, and the time to upgrade the configuration to a new version suffers.

Conditions:
-- Signature Staging is enabled.
-- Updated Signature Enforcement is set to 'Retain previous rule enforcement and place updated rule in staging'.
-- Signatures are enforced on a policy.
-- A new ASM Signature Update is installed, which modifies the matching rule for some enforced signatures.

Impact:
The time to apply changes made to a policy, and the time to upgrade the configuration to a new version, suffers from a inefficiently performing query related to the existence of previously enforced rules.

Workaround:
There is no workaround at this time.

Fix:
Query indexing and performance is fixed: Apply Policy executes in the same time whether there are previously enforced rules in the system or not.

Enforcing all signatures in a set now correctly removes the previously enforced rule from the signature.


737445-2 : Use of TCP Verified Accept can disable server-side flow control

Component: Local Traffic Manager

Symptoms:
Unexpectedly high memory usage, aggressive sweeper messages, TCP memory pressure messages in logs.

Conditions:
Verified Accept is enabled in TCP profiles that are associated with virtual servers.

Impact:
Excessive memory usage.

Workaround:
There is no workaround other than disabling Verified Accept.

Fix:
Fixed server-side flow control.


737368-1 : Fingerprint cookie large value may result in tmm core.

Component: Fraud Protection Services

Symptoms:
The JavaScript component calculates part of the fingerprint cookie value. In some rare cases, large values are generated and may result in tmm core.

Conditions:
A large value (more than 35 characters) inside the fingerprint cookie.

Impact:
Memory overrun, tmm core in some cases.

Workaround:
N/A

Fix:
FPS will check the value and truncate it if it exceeds the maximum length.


737355-1 : HTTP Strict-Transport-Security (HSTS) headers not being added to all APM generated files

Component: Access Policy Manager

Symptoms:
HTTP Strict-Transport-Security (HSTS) headers are missing for some APM-generated files.

Conditions:
This occurs when the following conditions are met:
-- HTTP profile is configured with HSTS enabled.
-- HTTP GET requests for APM renderer files, including CSS, JS, and image files from the webtop.

Impact:
Without these headers, the user agent (browser) may switch to non-secure communication.

Workaround:
None.

Fix:
When the HTTP profile is configured with HSTS enabled, all APM renderer files are now sent with HSTS headers.


737064-2 : ACCESS::session iRule commands may not work in serverside events

Component: Access Policy Manager

Symptoms:
-- 'ACCESS::session sid' will returns empty.
-- 'ACCESS::session data get <varname>' returns empty.

Conditions:
-- Using IP-based sessions
-- ACCESS::session iRule commands in a serverside iRule event, such as HTTP_PROXY_CONNECT.
-- Using SSL Bypass. Note: This issue is known to occur when using SSL Bypass, but this is not the only configuration that exhibits the problem.

Impact:
iRules may not work as expected.

Workaround:
There is no workaround at this time.

Fix:
The ACCESS::session iRules now work in serverside events when doing IP-based sessions.


734822-3 : TMSH improvements

Solution Article: K77313277


734595-2 : sp-connector is not being deleted together with profile

Component: Access Policy Manager

Symptoms:
If a profile is connected to an SSO SAML IdP configuration with an SP connector, the sp-connector is not available for delete when the profile is deleted.

Conditions:
-- Profile is connected to an SSO SAML IdP configuration with an SP connector.
-- Deleting the profile and attempting to delete the sp-connector.

Impact:
The SP connector is not listed for delete when the profile is deleted.

Workaround:
To delete the SP connector, run the following command:
tmsh delete apm sso saml-sp-connector NAME

Fix:
SP connectors are now available for delete when profile is deleted.


734527-1 : BGP 'capability graceful-restart' for peer-group not properly advertised when configured

Component: TMOS

Symptoms:
BGP neighbor using peer-group with 'capability graceful-restart' enabled (which is the default). However, on the system, peer-group defaults to disable. Because there is no config statement to enable 'capability graceful-restart' after restart, it gets turned off after config load, which also turns it off for any BGP neighbor using it.

Conditions:
- Neighbor configured to use peer-group.
- Peer-group configured with capability graceful-restart, but with the default: disabled.

Impact:
Because peer-group is set to disable by default, and there is no operation to enable it after restart, 'capability graceful-restart' gets turned off after config load, and neighbors using it from peer-group get turned off as well.

Workaround:
Re-issue the commands to set peer-group 'capability graceful-restart' and 'clear ip bgp *' to reset BGP neighbors.

Fix:
The peer-group now has 'capability graceful-restart' enabled by default, so this issue no longer occurs.

Behavior Change:
In this release, peer-group has 'capability graceful-restart' enabled by default.


734446-2 : TMM crash after changing LSN pool mode from PBA to NAPT

Component: Carrier-Grade NAT

Symptoms:
TMM crashes after changing LSN pool mode from PBA to NAPT when long lived connections are killed due to the PBA block lifetime and zombie timeout expiring.

Conditions:
An LSN pool using PBA mode with a block lifetime and zombie timeout set and long lived connections.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Instead of changing the LSN pool mode from PBA to NAPT, create a new LSN pool configured for NAPT and change the source-address-translation pool on the virtual servers that use the PBA pool.

The PBA pool can be deleted after the virtual servers are no longer using it.

Fix:
TMM no longer crashes after changing LSN pool mode from PBA to NAPT.


733585-3 : Merged can use %100 of CPU if all stats snapshot files are in the future

Component: TMOS

Symptoms:
Merged uses %100 of CPU if it cannot remove the oldest snapshot file, due to all snapshot files having timestamps in the future.

Conditions:
All stats snapshot file having timestamps in the future, release has the fix for issue 721740, but not this issue.

Impact:
Merged using %100 of the CPU.

Workaround:
Remove snapshot stats files that have timestamps in the future and restart merged.

Fix:
Correctly exit cleanup logic when all stats snapshot files have timestamps in the future.


727467-1 : Some iSeries appliances can experience traffic disruption when the HA peer is upgraded from 12.1.3 and earlier to 13.1.0 or later.

Component: TMOS

Symptoms:
-- CPU core 0 can be seen utilizing 100% CPU.
-- Other even cores may show a 40% increase in CPU usage.
-- Pool monitors are seen flapping in /var/log/ltm.
-- System posts the following messages:
   + In /var/log/ltm:
     - err tmm4[21025]: 01340004:3: HA Connection detected dissimilar peer: local npgs 1, remote npgs 1, local npus 8, remote npus 8, local pg 0, remote pg 0, local pu 4, remote pu 0. Connection will be aborted.
    + In /var/log/tmm:
      - notice DAGLIB: Invalid table size 12
      - notice DAG: Failed to consume DAG data

Conditions:
-- Active unit on a pre-12.1.3.1 release.
-- Standby peer upgraded to a 13.1.0 or later release.
-- Device is an iSeries device (i5600 or later).

Important: This issue may also affect iSeries HA peers on the same software version if the devices do not share the same model number.

Note: Although this also occurs when upgrading to 12.1.3.8 and 13.0.x, the issue is not as severe.

Impact:
- High CPU usage.
- Traffic disruption.

Workaround:
Minimize impact on affected active devices by keeping the upgraded post-13.1.0 unit offline as long as possible before going directly to Active.

For example, on a 12.1.3 unit to be upgraded (pre-upgrade):
-- Run the following command: tmsh run sys failover offline persist
-- Run the following command: tmsh save sys config
-- Upgrade to 13.1.0.8.
-- Unit comes back up on 13.1.0.8 as 'Forced Offline' and does not communicate with the active unit running 12.1.3 at all.
-- Set up HA group and make sure the 12.1.3 Active unit's HA score is lower than 13.1.0.8.
-- To cause the 13.1.0.8 unit to go directly to Active and take over traffic, run the following command on the unit running 13.1.0.8:
tmsh run sys failover online

At this point, the 12.1.3 unit starts to show symptoms of this issue, however, because it is no longer processing traffic, there is no cause for concern.

Fix:
This release introduces a new bigdb variable DAG.OverrideTableSize. To prevent the issue on an upgraded post-13.1.0 unit, set DAG.OverrideTableSize to 3.

In order to return the system to typical CPU usage, you must set the db variable, and then restart tmm by running the following command:
bigstart restart tmm

(Restarting tmm is required for 13.1.1.2 and newer 13.1.1.x releases.)

Note: Because the restart is occurring on the Standby unit, no traffic is disrupted while tmm restarts.


727297-3 : GUI TACACS+ remote server list should accept hostname

Component: TMOS

Symptoms:
Cannot add hostnames to the Remote - TACACS+ server list in the GUI.

Conditions:
-- On the System :: Users : Authentication page with Remote - TACACS+ specified.
-- Add hostname to the server list.

Impact:
Validation does not accept a hostname. Cannot add hostname as a server.

Workaround:
Use tmsh to add a hostname.

Fix:
The system now allows hostname to be added with proper validation in this case.


727222-1 : 206 Partial Content responses from ramcache have malformed Content-Range header

Component: Local Traffic Manager

Symptoms:
When ramcache serves a 206 Partial Content response from cache, the Content-Range header repeats the name:

  Content-Range: Content-Range: bytes 0-5/28

Conditions:
Request from client for partial document (Range header) against a virtual server with a web-acceleration profile having no applications (ramcache), where the requested document is present in ramcache.

Impact:
The client may mishandle the response, as the Content-Range header is malformed. This may cause additional traffic as the client may retrieve the entire document in a subsequent request due to the malformed response.

Workaround:
Remove the duplicate portion of the Content-Range header using an iRule at HTTP_RESPONSE_RELEASE time.

Fix:
The Content-Range header is now correctly formed for 206 Partial Content responses served from ramcache.


727044-2 : TMM may crash while processing compressed data

Component: Local Traffic Manager

Symptoms:
Under certain conditions, TMM may crash while processing compressed data.

Conditions:
Compression enabled
Hardware compression disabled

Impact:
TMM crash leading to a failover event.

Workaround:
No workaround.

Fix:
TMM now correctly processes compressed traffic


726895 : VPE cannot modify subroutine settings

Component: Access Policy Manager

Symptoms:
Open per-request policy in Visual Policy Editor (VPE) that has a subroutine. Click 'Subroutine Settings / Rename.

Numeric values like the inactivity timeout are displayed as 'NaN. Attempts to modify the values results in MCP validation errors.

Conditions:
-- Per-request policy in the VPE.
-- Subroutine in the per-request policy.
-- Attempt to change the values.

Impact:
All fields say 'NaN', and error when trying to modify properties. Subroutine settings like the Inactivity Timeout and Gating Criteria cannot be modified through the VPE

Workaround:
Use tmsh to modify these values, for example:

tmsh modify apm policy access-policy <policy_name> subroutine properties modify { all { inactivity-timeout 301 } }

Fix:
The issue has been fixed; it is now possible to view and modify subroutine settings in the VPE.


726872-2 : IApp LX directory disappears after upgrade or restoring from ucs

Component: iApp Technology

Symptoms:
After BIG-IP version upgrade or restoring from UCS, iApps/Application Services/Applications LX screen displays instances of iApps LX, but their UI is not functional. That is caused by the software defect causing iApp LX directory to disappear from /var/config/rest/iapps.

Conditions:
This issue can only happen during initial start after BIG-IP version upgrade or restoring from UCS. The more iApps LX instances and the more configuration they use, the more likely this issue to happen. We observed this issue with 90+ instances of f5-ddos-hybrid-defender iApp LX.

Impact:
The code of iAppLX is removed from the system because of the defect. That makes iAppLX UI unusable. The configuration deployed by the iApp LX instances remains in effect. The iApp LX configuration data remain intact and UI can be completely restored after manual installation of iApp LX code.

Workaround:
1. Copy iAppLX code from an unaffected BIG-IP to the BIG-IP impacted by this defect. For example,
/var/config/rest/iapps/f5-ddos-hybrid-defender.
2. Create a symlink to UI code for UI to work. For example,
ln -sfvn /var/config/rest/iapps/f5-ddos-hybrid-defender/presentation /var/iapps/www/f5-ddos-hybrid-defender
3. Restart restjavad:
bigstart restart restjavad
4. Restart restnoded:
bigstart restart restnoded

Fix:
The issue should not happen when upgrading to the BIG-IP version with the fix or restoring from UCS on the BIG-IP version with the fix.


726647-3 : PEM content insertion in a compressed response may truncate some data

Component: Policy Enforcement Manager

Symptoms:
HTTP compressed response with content insert action can truncate data.

Conditions:
PEM content insertion action with compressed HTTP response.

Impact:
Data might be truncated.

Workaround:
There is no workaround other than disabling compression accept-encoding attribute in the HTTP request.

Fix:
HTTP compressed response with content insert action no longer truncates data.


726616-1 : TMM crashes when a session is terminated

Component: Access Policy Manager

Symptoms:
TMM crashes when it is releasing the memory used by a session that is terminated. Either one of the following error messages might appear in the TMM log:

-- notice panic: ../kern/umem.c:4166: Assertion "valid type" failed.

-- notice panic: ../kern/page_alloc.c:705: Assertion "vmem_hashlist_remove not found" failed.

Conditions:
The memory used by a session is corrupted during the lifetime of the session. This issue happens infrequently.

Impact:
TMM restarts, causing service interruption if the BIG-IP system is not part of a high-availability setup.

Workaround:
There is no workaround at this time.

Fix:
TMM no longer crashes when removing an access session.


726592-1 : Invalid logsetting config messaging between our daemons could send apmd, apm_websso, localdbmgr into infinite loop

Component: Access Policy Manager

Symptoms:
Invalid logsetting config messaging between our daemons could send apmd, apm_websso, localdbmgr into infinite loop. This could be triggered by invalid state of our control plane daemons.

Conditions:
This is an extremely rare situation that can be caused by invalid logsetting config messaging between our daemons. However, once it happens it can impact multiple daemons at the same time causing all of them to hang.

Impact:
Once this happens it can impact multiple daemons (apmd, apm_websso, localdbmgr) at the same time causing all of them to hang.

Workaround:
There is no workaround at this time, you can recover by restarting the daemons that hang.

Fix:
We have fixed a memory corruption that can break the linkages in our data structure which would cause certain traversals to loop indefinitely.


726409-4 : Kernel Vulnerabilities: CVE-2017-8890 CVE-2017-9075 CVE-2017-9076 CVE-2017-9077

Solution Article: K61429540


726303-1 : Unlock 10 million custom db entry limit

Component: Traffic Classification Engine

Symptoms:
Cannot add more than 10 million custom db entries.

Conditions:
This happens when you try to add more than 10 million custom db entries.

Impact:
Not able to add more than 10 million entries.

Workaround:
There is no workaround at this time.

Fix:
This release provides a sys db var, tmm.urlcat.no_db_limit, to allow growth beyond the existing limit of 10 million custom db entries.


726239-4 : interruption of traffic handling as sod daemon restarts TMM

Component: Local Traffic Manager

Symptoms:
When the receiving host in a TCP connection has set its send window to zero (stopping the flow of data), following certain unusual protocol sequences, the logic in the TMM that persists in probing the zero window may enter an endless loop.

Conditions:
When the TCP implementation is probing a zero-window connection under control of a persist timer.

Impact:
Lack of stability on the device. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
This fix handles a rare TMM crash when TCP persist timer is active.


726154-2 : TMM restart when there are virtual server and route-domains with the same names and attached firewall or NAT policies

Component: Advanced Firewall Manager

Symptoms:
TMM might restart when changing firewall or NAT configurations on a virtual server with the same name as a route-domain.

Conditions:
- There is a virtual server and a route-domain that have the same name in the configuration.
- A firewall or NAT policy is being attached or removed from one or both of the virtual server or route-domain.

Impact:
TMM might halt and traffic processing temporarily ceases. TMM restarts and traffic processing resumes.

Workaround:
There is no workaround other than not to use the same name for virtual server and route-domain objects.

Fix:
TMM no longer crashes under the conditions described. Firewall and NAT configurations are applied correctly on virtual servers with the same names as route-domains.


726090-1 : No Bot Defense Logs when ASM Policy is used without Proactive Bot Defense

Component: Application Security Manager

Symptoms:
Device ID challenges are not logged in the Bot Defense Request Log (local and remote) when associating the Bot Defense logging profile to the virtual server.

Conditions:
-- Device ID is enabled on the ASM Policy.
-- DoS profile with Proactive Bot Defense is not associated with the virtual server.

Impact:
There are no logs to show the browser challenges and the requests from clients that do not support JavaScript.

Workaround:
There is no workaround at this time.

Fix:
Requests are now logged to the Bot Defense Request Log with Device ID enabled on the ASM Policy and no associated DoS profile.


726089-2 : Modifications to AVR metrics page

Solution Article: K44462254


725867-2 : ADFS proxy does not fetch configuration for non-floating virtual servers

Component: Access Policy Manager

Symptoms:
If the virtual address of a virtual server has a non-floating traffic group (e.g., traffic-group-local-only), ADFS proxy does not perform periodic configuration updates, including listing allowed URIs from ADFS, and thus blocks all the requests (by responding with 404).

Conditions:
-- Virtual address of virtual server has non-floating traffic group.

-- ADFS proxy feature is enabled on the virtual server.

Impact:
All the requests to ADFS are blocked.

Workaround:
Set floating traffic group for virtual address (e.g., traffic-group-1).

Fix:
ADFS proxy now fetches configuration from ADFS for non-floating virtual servers.


725815-1 : vlangroup usage may cause a excessive resource consumption

Solution Article: K72442354


725801-4 : CVE-2017-7889: Kernel Vulnerability

Solution Article: K80440915


725696-1 : A timer loop might occur when OCSP Stapling is enabled resulting in tmm getting aborted

Component: TMOS

Symptoms:
When OCSP Stapling is enabled on a client SSL profile, certain uncommon operations might result in a tmm timer queue getting into a loop, which results in tmm being aborted by sod. tmm restart

Conditions:
-- There are SSL handshakes waiting for an OCSP response, and one of the following:
  + There is a CMP transition.
  + There are changes made to the OCSP object.

Impact:
tmm restarts. Traffic interrupted while tmm restarts.

Workaround:
There is no workaround other than disabling OCSP stapling.

Fix:
The timer issue has been corrected.


725635-2 : CVE-2018-3665: Intel Lazy FPU Vulnerability

Solution Article: K21344224


725612-1 : syslog-ng remote destination needs unique name that changes on address change.

Component: TMOS

Symptoms:
Changing syslog server IP address requires syslog-ng restart to take effect. Syslog operations do not use the new remote destination address on syslog service reconfiguration.

Conditions:
1. Add Remote Syslog server (Server A) using System :: Logs :: Configuration :: Remote Syslog.
2. Click Update (Syslog messages go out toward Server A).
3. Remove Server A from configuration.
4. Add new Server B (different IP address).
5. Click Update.

Impact:
Syslog messages still go out toward Server A.

Workaround:
Restart the syslog service using the following command:
bigstart restart syslog-ng

Messages now properly go out toward Server B (the new IP address).

Fix:
Syslog operations now use the new remote destination address on syslog service reconfiguration.


725545-1 : Ephemeral listener might not be set up correctly

Component: Local Traffic Manager

Symptoms:
When ephemeral listeners are set up across a cluster, the transaction might fail.

Conditions:
When using Network Access tunnel with proxy ARP and no SNAT.

Impact:
The client-assigned IP address might intermittently fail to be resolved via ARP on the serverside/leasepool VLAN.

Workaround:
None.

Fix:
The ephemeral listener is now set up correctly.


725412-1 : APM does not follow current best practices for HTTP headers

Component: Access Policy Manager

Symptoms:
APM does not follow current best practices for HTTP headers

Conditions:
APM enabled

Impact:
HTTP headers not generated as intended

Workaround:
None.

Fix:
APM now follows current best practices for HTTP headers


724906-1 : sasp_gwm monitor leaks memory over time

Component: Local Traffic Manager

Symptoms:
The Server/Application State Protocol (SASP) monitor, sasp_gwm memory usage slowly increases over time.

Conditions:
The Server/Application State Protocol (SASP) monitor is configured.
The leakage occurs during normal operation of the monitor and is not tied to any specific user or traffic action.

Impact:
sasp_gwm grows over time and eventually the system may be pushed into swap. Once swap is exhausted, the Linux OOM killer might terminate critical system processes, disrupting operation.

Workaround:
Periodic restarts of sasp_gwm processes avoids the growth affecting the system.

Fix:
The sasp_gwm monitor no longer leaks memory when processing messages.


724868-1 : dynconfd memory usage increases over time

Solution Article: K11662998

Component: Local Traffic Manager

Symptoms:
dynconfd memory usage slowly increases over time as it processes various state-related messages.

Conditions:
No special conditions as dynconfd is a core LTM process. Large numbers of pool members combined with flapping might increase the rate of memory usage increase.

Impact:
dynconfd grows over time and eventually the system is pushed into swap. Once swap is exhausted, the Linux OOM killer might terminate critical system processes, disrupting operation.

Workaround:
Periodic restarts of dynconfd avoids the growth affecting the system.

Fix:
dynconfd no longer leaks memory when processing messages.


724680-4 : OpenSSL Vulnerability: CVE-2018-0732

Solution Article: K21665601


724571-1 : Importing access profile takes a long time

Component: Access Policy Manager

Symptoms:
It takes a long time for the 'Apply Access Policy' link to show up on the admin UI after importing an access profile.

Conditions:
-- Access policy with many macros.
-- Import exported profile multiple times with Reuse Existing Objects checked
-- As the number of imports increases, so does the latency.

Impact:
The imported access policy takes a long time to be imported and ready to use.

Workaround:
None.


724532-2 : SIG SEGV during IP intelligence category match in TMM

Component: Advanced Firewall Manager

Symptoms:
TMM restart while matching traffic from source IP to IP Intelligence category.

Conditions:
Multiple configuration changes that include deleting the custom IP intelligence categories.

Impact:
TMM restart. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
TMM no longer restarts while matching traffic from source IP to IP Intelligence category.


724414-2 : ASM UMU memory leak when WebSocket url has json profile with parse parameters enabled

Component: Application Security Manager

Symptoms:
When ASM inspects WebSocket frames, the bd daemon memory is leaking; eventually ASM stops inspecting traffic and sends resets or bypass requests.

Conditions:
ASM module is provisioned.
-- ASM policy and WebSocket profile are attached to a virtual server.
-- WebSocket URL has JSON profile attached to it.
-- JSON profile has parse parameters flag enabled (this is the default).

Impact:
ASM may reset connections; failover might occur.

Workaround:
There are two workarounds:
-- Remove JSON profile from WebSocket URL in the ASM policy.

-- Disable parse parameters flag in the json profile.

Fix:
The system now frees the allocated memory when it finishes the inspect of a WebSocket frame.


724339-1 : Unexpected TMUI output in AFM

Solution Article: K04524282


724335-1 : Unexpected TMUI output in AFM

Solution Article: K21042153


724213-1 : Modified ssl_profile monitor param not synced correctly

Component: Local Traffic Manager

Symptoms:
After modifying the ssl_profile attribute on an HTTPS monitor on a device in a high availability (HA) configuration, and after performing a full ConfigSync, the corresponding monitor on the peer-sync unit does not have the updated value.

Conditions:
-- An HTTPS monitor is used on BIG-IP systems in an HA configuration.
-- The ssl_profile field is modified on an HTTPS monitor.
-- A sync-to-peer (full ConfigSync, not incremental sync) is attempted to propagate the modified ssl_profile value to the peer units.

Impact:
The ssl_profile value for the HTTPS monitor on the peer unit is set to none, resulting in the two devices reporting themselves as in-sync, but having potentially different HTTPS monitor configurations.

Workaround:
-- Do not run HTTPS monitors using in-tmm monitors,
-- Use the traditional HTTPS monitor configuration for SSL-attributes (cipherlist, key, cert, and compatibility attributes on HTTPS monitor).

Note: Using these attributes generates deprecation warnings, but the configuration still takes effect.

Fix:
After modifying the ssl_profile attribute on an HTTPS monitor on a system within an HA configuration, and after performing a full ConfigSync, the corresponding monitor on the peer unit now receives the updated monitor attribute, as expected.


724032-1 : Searching Request Log for value containing backslash does not return expected result

Component: Application Security Manager

Symptoms:
Searching within Request Log for a value containing backslash does not return the expected result. This happens because escaping of backslashes in REST API access is handled differently with 'eq' and 'contains' filters.

Conditions:
Searching within Request Log for a value containing backslash.

Impact:
Search within Request Log record containing backslash does not return the expected result.

Workaround:
Backslashes in REST API requests should be doubled when searching using 'contains' filter, but must not be doubled when searching with 'eq' filter.

Fix:
Searching within Request Log for a value containing backslash returns the expected result.


723792-2 : GTM regex handling of some escape characters renders it invalid

Component: Global Traffic Manager (DNS)

Symptoms:
The memory footprint of big3d increases.

Conditions:
GTM monitor recv string such as the following: ^http\/\d\.\d[23]\d\d

Impact:
Escaped characters are mishandled. This causes the regex compilation to fail and leak memory. The monitor marks the server it is monitoring UP as long as the server returns any response.

Workaround:
If you notice big3d memory footprint increase and you have a monitor recv string that has escaped characters such as '^http\/\d\.\d[23]\d\d', try changing the recv string to something similar to the following: ^HTTP/[0-9][.][0-9] [23][0-9]{2}

Fix:
Fixed handling of escape characters. Recv strings that previously failed to compile will now compile.


723722-2 : MCPD crashes if several thousand files are created between config syncs.

Component: TMOS

Symptoms:
If more than several thousand (typically 20,000, but number varies by platform) files, for example SSL certificates or keys, are created between config syncs, the next config sync operation will take too long and mcpd will be killed by sod.

Conditions:
Creation of several thousand SSL certificates or keys followed by a config sync operation.

Impact:
Traffic is disrupted while the MCPD process restarts.

Workaround:
Run a config sync operation after every ~5000 files created.

Fix:
MCPD no longer crashes if several thousand files are created between config sync operations.


723300-2 : TMM may crash when tracing iRules containing nameless listeners on internal virtual servers

Component: Local Traffic Manager

Symptoms:
TMM may crash when tracing iRules containing nameless listeners on internal virtual servers.

Conditions:
-- Using iRule tracing.
-- Internal virtual servers.
-- Listener iRule, where the listener has no name.

Impact:
TMM crash. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
TMM no longer crashes when tracing iRules containing nameless listeners on internal virtual servers.


723130-1 : Invalid-certificate warning displayed when deploying BIG-IP VE OVA file

Component: TMOS

Symptoms:
The OVA signing certificate that signs BIG-IP Virtual Edition (VE) OVA files expired. When deploying a BIG-IP VE from an OVA file, an invalid-certificate warning might be displayed due to the expired OVA signing certificate.

Conditions:
This issue may be encountered during the creation of new instances of BIG-IP VE in clients that check the validity of the OVA signing certificate (e.g., VMware).

Note: Existing BIG-IP VE instances are not subject to this issue.

Impact:
There might be questions about the integrity of the OVA file, and in some cases, might not be able to deploy a new instance from an OVA file.

Workaround:
The expired OVA signing certificate has been replaced with a valid signing certificate.

Fix:
The expired OVA signing certificate has been replaced with a valid signing certificate.


722969-2 : Access Policy import with 'reuse' enabled instead rewrites shared objects

Component: Access Policy Manager

Symptoms:
If a policy is exported and then imported with 'reuse' it rewrites objects on the server instead of simply reusing them.

Conditions:
-- Exporting profile A.
-- Importing profile B, with 'reuse' configured.
-- The objects in both profiles are named the same and are the same type.

Impact:
-- 'Reused' objects on the system are modified.
-- Policy A requires 'apply', because objects shared with Policy B are overwritten.

Workaround:
None.

Fix:
Access policy import with 'reuse' option enabled no longer rewrites shared objects


722893-1 : The TMM - host interface may stall when the kernel memory is fragmented

Solution Article: K30764018

Component: Local Traffic Manager

Symptoms:
MCP logs 'Removed publication with publisher id TMMx' and the affected TMM restarts.

Conditions:
This occurs when the following conditions are met:
-- Linux kernel memory fragmentation exists.
-- Another operation is occurring, including (among others):
  + Config-sync with full reload is initiated.
  + Running tcpdump.

Impact:
Degraded performance and unexpected failover when tmm restarts. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
The internal driver has been improved, allowing it to work in low- and/or fragmented-memory conditions.


722682-2 : Fix of ID 615222 results in upgrade issue for GTM pool member whose name contains a colon; config failed to load

Component: TMOS

Symptoms:
Loading configuration process failed after upgrade when pool member names contain colons. The system posts errors similar to the following:
01070226:3: Pool Member 443:10.10.10.10 references a nonexistent Virtual Server
Unexpected Error: Loading configuration process failed.

Conditions:
-- GTM pool member has colon in its name.
-- Upgrade to any of the following versions:
  + 12.1.3.x
  + Any 13.0.x
  + All 13.1.x earlier than 13.1.1.2
  + 14.0.x earlier than 14.0.0.3

Impact:
The system removes the part of the pool member name that precedes the colon, and configuration load fails. For example, when the configuration contains a pool member named example_pm:443:10:10:10:10, the system removes the part of the pool member name that precedes the first colon, in this case, example_pm, leaving the pool member name 443:10.10.10.10.

Workaround:
After upgrade, add two backslash characters, \\ before the first colon, : character.

1. Upgrade the BIG-IP DNS system to the new version of software.
2. SSH to log into the bash prompt of the BIG-IP DNS system.
3. Run the following command to add the string \\ in front of the first colon character, : in the name of each affected GTM pool member:

  for j in $(find /config/ -name bigip_gtm.conf); do for i in $(awk '$0 ~ /^gtm server.*:/ {gsub("[/:]"," ",$0); print $4}' ${j}); do sed -i "s, /Common/${i}, /Common/${i}\\\\\\\,g" ${j} | grep " /Common/${i}"; done; done

4. Run the following command: load sys config gtm-only

Fix:
Configuration load no longer fails when upgrading with a configuration that contains BIG-IP DNS pools with members whose names contain colon characters.


722677-4 : High-Speed Bridge may lock up

Component: Local Traffic Manager

Symptoms:
Under certain conditions, hardware systems with a High-Speed Bridge and using Layer 2 forwarding configurations may experience a lockup of the High-Speed Bridge.

Conditions:
Hardware platform with High-Speed Bridge.
Layer 2 forwarding enabled.
vlangroup.flow.allocate disabled.

Impact:
High-Speed Bridge lockup, leading to a failover event.

Workaround:
The vlangroup.flow.allocate DB variable is enabled by default.

Ensure that vlangroup.flow.allocate is enabled with the command:
modify /sys db vlangroup.flow.allocate value enable


722594-2 : TCP flow may not work as expected if double tagging is used

Solution Article: K91300169

Component: Local Traffic Manager

Symptoms:
TCP flow may have an incorrect ACK number, and the flow may stall or reset. The BIG-IP system sends an ACK that is higher than it should be based on the data received from the client.

Conditions:
Double tagging is used.

Impact:
TCP connection fails.

Workaround:
Change the db variable tm.tcplargereceiveoffload value to disable.

Fix:
TCP flow now has the correct ACK number when double tagging is used.


722423-1 : Analytics agent always resets when Category Lookup is of type custom only

Component: Access Policy Manager

Symptoms:
Analytics errors out with a log that says you cannot use custom lookup only along with analytics (/var/log/apm). Even if the Analytics agent is set to disable RST on failure, the agent will send a RST.

Conditions:
In the per-request policy, Category Lookup agent is set to custom lookup only.

Impact:
This configuration leads to an Analytics agent with 'RST on Failure' disabled. The BIG-IP Admin cannot make the choice to disable RST on failure for this particular setup (even though it is a misconfiguration).

Workaround:
You can avoid the RST and error by modifying the configuration to use a standard category lookup before the analytics agent.

Using a custom-only 'Category Lookup' immediately followed by 'Response Analytics' is not a valid configuration. In this configuration, Analytics cannot run the way it is supposed to because it does not have the right information. 'Response Analytics' gives the option to choose what to do on error: RST, or continue on, implying that the non-RST option is valid. If the admin chooses to not reset, the system should log the error and continue. In this case, the system actually does send a RST, even though the RST is not what the BIG-IP admin intended to configure.

Fix:
Disabling RST on failure now works properly in this scenario now. The configuration is still technically incorrect, but now the system takes the correct specified action-upon-error.


722387-3 : TMM may crash when processing APM DTLS traffic

Component: Local Traffic Manager

Symptoms:
When processing DTLS traffic for APM, TMM may crash.

Conditions:
APM provisioned and configured.
DTLS enabled in APM configuration.

Impact:
TMM crash, leading to a failover event.

Workaround:
None.

Fix:
DTLS traffic is now processed as expected.


722363-2 : Client fails to connect to server when using PVA offload at Established

Component: Local Traffic Manager

Symptoms:
A client can fail to connect to the server on subsequent attempts if using FastL4 with hardware (HW) acceleration.

When this issue occurs, the profile_bigproto_stat/rxsynatestablished stat is non-zero.

Conditions:
A FastL4 virtual server is configured with offload_state = EST.

Impact:
Clients fail to connect to the server.

Workaround:
There is no workaround other than to disable PVA acceleration.


722091-3 : TMM may crash while processing HTTP traffic

Solution Article: K64208870


722013 : MCPD restarts on all secondary blades post config-sync involving APM customization group

Component: Access Policy Manager

Symptoms:
After performing a series of specific config-sync operations between multi-blade systems, the MCPD daemon restarts on all secondary blades of one of the systems.

Each affected blade will log an error message similar to the following example:

-- err mcpd[5038]: 01070711:3: Caught runtime exception, Failed to collect files (snapshot path () req (7853) pmgr (7853) not valid msg (create_if { customization_group { customization_group_name "/Common/Test-A_secure_access_client_customization" customization_group_app_id "" customization_group_config_source 0 customization_group_cache_path "/config/filestore/files_d/Common_d/customization_group_d/:Common:Test-A_secure_access_client_customization_66596_1" customization_group_system_path "" customization_group_is_system 0 customization_group_access 3 customization_group_is_dynamic 0 customization_group_checksum "SHA1:62:fd61541c1097d460e42c50904684def2794ba70d" customization_group_create_time 1524643393 customization_group_last_update_time 1524643393 customization_group_created_by "admin" customization_group_last_updated_by "admin" customization_group_size 62 customization_group_mode 33188 customization_group_update_revision 1

Conditions:
This issue occurs when all of the following conditions are met:

- You have configured a device-group consisting of multi-blade systems (such as VIPRION devices or vCMP guests).

- Systems are provisioned for APM.

- The device-group is configured for incremental manual synchronizations.

- On one system (for example, source_system), you create an APM configuration object (for example, a connectivity profile) that causes the automatic creation of an associated customization group.

- You synchronize the configuration from the source_system to the device-group.

- On the source_system, you create a new configuration object of any kind (for example, an LTM node).

- Instead of synchronizing the newest configuration to the device-group, you synchronize the configuration from another device in the device-group to the source_system (which effectively undoes your recent change).

- The MCPD daemon restarts on all secondary blades of the source_system.

Impact:
While the MCPD daemon restarts, the blade is offline, and many other BIG-IP daemons need to restart along with MCPD.

-- If the source_system is Active, the load capacity of the system temporarily decreases proportionally to the number of blades restarting. Additionally, if the systems use the min-up-members or HA-Group features, losing a certain number of secondary blades may trigger a failover.

-- If the source_system is Standby, some of the previously mirrored information may become lost as TMM on secondary blades restarts along with MCPD. Should this system later become the Active system, traffic may be impacted by the lack of previously mirrored information.

Workaround:
None.

Fix:
The MCPD daemon no longer restarts on secondary blades after config-sync of APM.


721985 : PAYG License remains inactive as dossier verification fails.

Component: TMOS

Symptoms:
- BIG-IP is deployed in a cloud environment (AWS/Azure/GCE) with PAYG licenses. The license won't activate on the startup.

Conditions:
- There are multiple ways this can happen but all of those come down to user networking issue where the http calls to the cloud metadata service fails.

- This can be a simple routing issue to the metadata service or a firewall issue.

Impact:
As license activation fails, the instance becomes unusable.

Workaround:
User should look at /var/log/ltm to determine the networking issue that is causing the dossier verification failure. This would be typically printed in the following way:

Curl request to metadata service failed with error(<error-code>): '<error-message>'

By resolving this networking error, license activation should succeed.

Fix:
PAYG License remains inactive as dossier verification fails.


721924-2 : bgpd may crash processing extended ASNs

Component: TMOS

Symptoms:
Under certain conditions bgpd may crash while processing extended ASNs.

Conditions:
Dynamic routing enabled.
Extended ASP capabilities enabled: bgp extended-asn-cap enabled

Impact:
Dynamic routing disrupted while bgpd restarts.

Fix:
bgpd now processes extended ASNs as expected.


721895 : Add functionality to configure the minimum TLS version advertised and accepted by big3d (iQuery)

Component: Global Traffic Manager (DNS)

Symptoms:
big3d advertises a TLSv1.0 version. Even though big3d requires previously exchanged certificates to validate a connection request, the TLSv1.0 advertisement triggers various vulnerability scanners and is flagged.

Conditions:
Running a vulnerability scanner or other SSL test tool.

Impact:
The scanner or tool reports that big3d might potentially accept a TLSv1.0 connection request (which is considered insecure). Vulnerability scanners then flag the BIG-IP system as vulnerable.

Workaround:
Although there is no workaround, because big3d accepts connections only from clients that match the certificates on the BIG-IP system, the risk is minimal.

In addition, you can deploy firewall rules to accept connections only on port 4353 from know BIG-IP systems.

Fix:
This version adds a db variable for the big3d
big3d.minimum.tls.version. By default the value is 'TLSv1'. You can also specify TLSV1.1 or TLSV1.2 (the setting is case insensitive).

After changing the DB variable, restart big3d. Change the value on all BIG-IP systems that are subject to scans. This includes GTM as well as LTM configurations.


721741-3 : BD and BD_Agent out-of-sync for IP Address Exception, false positive/negative

Component: Application Security Manager

Symptoms:
bd log spits this error.
-------
ECARD_POLICY|NOTICE|May 24 04:49:42.035|4143|table.h:2408|IPTableList::del_object key not found in table
ECARD|ERR |May 24 04:49:42.035|4143|table.h:0398|KEY_UPDATE: Failed to REMOVE data will continue to add
-------

Conditions:
Configuring IP Address Exceptions in certain order - w/ and w/o route domain.

Impact:
BD and BD_Agent out-of-sync for IP Address Exception, causes false positives / false negatives

Workaround:
There is no workaround at this time.

Fix:
System no longer generates these false positive/negative log entries.


721740-2 : CPU stats are not correctly recorded when snapshot files have timestamps in the future

Component: TMOS

Symptoms:
One symptom is that a message similar to the following comes out in the log files frequently.

May 24 16:31:53 lusia_60.F5.COM warning merged[6940]: 011b0914:4: No individual CPU information is available.

Merged CPU stats will be 0.

Conditions:
If all of the snapshot stats files have timestamps in the future, CPU stats will not be correctly merged.

Impact:
Frequent error messages in the logs, and incorrect merged CPU stats.

Workaround:
Remove all of the stats snapshot files that have timestamps in the future and restart merged.

Fix:
Merged has been update to correctly deal with the case where all of the stats snapshot file have timestamps in the future, and will correctly merge the CPU stats.


721704-1 : UDP flows are not deleted after subscriber deletion

Component: Policy Enforcement Manager

Symptoms:
UDP flows continue to live till UDP idle time occurs, even after the subscriber is gone and the option for immediate deletion of the flow is enabled.

Conditions:
-- The option to delete flows upon subscriber deletion is enabled.
-- The UDP flow is established with an idle time greater than the re-evaluate timeout.

Impact:
The UDP flows continue to be alive after the required time, but only act to drop the traffic.

Workaround:
To work around this issue:
1. Modify the UDP idle timer to a suitable value.
2. Force delete the UDP flow from CLI.

Fix:
UDP flows are now deleted after subscriber deletion.


721571-1 : State Mirroring between BIG-IP 12.1.3.* and 13.* or 14.* systems may cause TMM core on standby system during upgrade

Component: Local Traffic Manager

Symptoms:
BIG-IP devices running 12.1.3.x (12.1.3 or a 12.1.3 point release) and 13.x or 14.x software versions in a high-availability (HA) configuration with state mirroring enabled may cause a standby system to produce a TMM core file.

Conditions:
-- State mirroring configured on two or more BIG-IP systems (state mirroring is enabled by default).
-- The active system is running v12.1.3.x, and the standby system is running v13.x or v14.x, as a result of an in-progress upgrade.
-- For VIPRION clusters or VIPRION-based vCMP guests, the systems are configured to mirror 'Between Clusters'.

Impact:
TMM may crash on a standby system during upgrade.

This issue should not disrupt traffic, because the TMM is coring only on the standby unit.

Workaround:
To workaround this issue, disable State Mirroring prior to upgrading, and re-enable it once both devices are running v13.x or v14.x, or complete the upgrade of both devices to v13.x or v14.x.

1. You can disable mirroring using either the GUI or the command line.

1a. In the GUI: -- Set the Primary and Secondary Local Mirror Address configurations to 'None' under Device Management :: Devices :: [Self] device :: Mirroring Configuration.

1b. From the command-line: -- Run the following command:
tmsh modify cm device <name-of-self-device> mirror-ip any6 mirror-secondary-ip any6 && tmsh save sys config

Important: This action results in connection state loss on failover.

2. Once all devices are running the same software version, re-enable state mirroring by re-adding the device mirror IP addresses removed previously.

Note: F5 recommends that BIG-IP systems in HA configurations run with the same software version on all devices.


721570-1 : TMM core when trying to log an unknown subscriber

Solution Article: K20285019

Component: Carrier-Grade NAT

Symptoms:
Using CGNAT or FW-NAT with subscriber-id logging enabled can cause a TMM core when the subscriber ID is unknown.

Conditions:
-- A LSN pool or FW-NAT source translation that has a logging profile with subscriber-id enabled.
-- A PEM profile that allows unknown subscribers.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Config PEM to deny connections from unknown subscribers.

Fix:
The system no longer crashes. It logs 'unknown' for unknown subscribers.


721512 : Config tool fails to configure management-ip when default routes exist for both IPv4 and IPv6.

Component: TMOS

Symptoms:
Configuration tool fails to configure management-ip when default routes exist for both IPv4 and IPv6.

Conditions:
This can happen in following two scenarios:
-- A configured IPv4 management-ip that is switched to IPv6.
-- A configured IPv6 management-ip that is switched to IPv4.

Impact:
Cannot successfully change an IPv4 or IPv6 management-ip address using config.

For either of the above cases, if the IP addresses are switched back to IPv4/IPv6, the config tools fails to configure management-ip with this error:

ERROR: route_mgmt_entry count is 2

Workaround:
Manually delete the default6 (if current management-ip is IPv4) or default (if current management-ip is IPv6) management-route by running the following command:

  tmsh delete sys management-route <default/default6>

Fix:
Config tool now works to configure management-ip when default routes exist for both IPv4 and IPv6, so you can switch back and forth between IPv4 and IPv6 IP addresses without error.


721474-1 : AVR does not send all SSLO statistics to offbox machine.

Component: Application Visibility and Reporting

Symptoms:
When using the 'use-offbox' option, AVR does not send SSLO statistics to the offbox system.

Conditions:
-- AVR provisioned.
-- Use-offbox is enabled.

Impact:
SSLO statistics are not available for BIG-IQ analytics.

Workaround:
There is no workaround.

Fix:
AVR now sends SSLO statistics to offbox systems when the 'use-offbox' option is enabled.


721375-1 : Export then import of config with RSA server in it might fail

Component: Access Policy Manager

Symptoms:
If an exported policy configuration contains both an RSA server as well as the RSA-provided sdconf.rec and sdstatus.12 config files, policy import might fail.

Conditions:
-- RSA server and access profile are in the same, non-Common partition.
-- Exported policy contains an RSA server as well as both the RSA-provided sdconf.rec and sdstatus.12 config files.

Impact:
Unable to import the exported configuration. This occurs because of how the names for the files are resolved in the exported configuration.

Workaround:
Although there is no actual workaround, you can avoid this issue if the profile is outside of the partition. That case uses a different name resolution during the export, so import works as expected.

Fix:
You can now successfully import an exported policy containing an RSA server as well as both sdconf.rec and sdstatus.12 files.


721364 : BIG-IP per-application VE BYOL license does not support three wildcard virtual servers

Component: TMOS

Symptoms:
BIG-IP Virtual Edition (VE) admin can create only 1 wildcard virtual servers for per-application BYOL license.

Note: If you are using BIG-IQ 6.0.0 or 6.0.1 to manage AWS-based, per-app BIG-IP VE systems running versions 13.1.0.5 through 13.1.0.8, you must use the following AWS templates:

-- Default-AWS-f5-HTTPS-WAF-lb-template
-- Default-AWS-f5-HTTPS-offload-lb-template

For these two templates, ports 443 and 80 (for HTTP redirect) are hard-coded in an iRule, which enable this functionality.

Conditions:
Per-app VE with BYOL license.

Impact:
Per-app VE with BYOL license does not work as expected.

Workaround:
N/A

Fix:
Per-app VE BYOL license now supports three wildcard virtual servers.


721350-2 : The size of the icrd_child process is steadily growing

Component: TMOS

Symptoms:
The resident segment size (rss) for the icrd_child process continues to increase on issuing iCR GET requests. This increase is permanent and the rss does not drop back to its original size.

Conditions:
The following is an example config under which this issue occurs. Note this is an illustrative example; the actual issue might occur for various config objects.

GET tm/ltm/virtual?expandSubcollections=true wherein the virtual has some profiles associated with it.

ltm pool p-http { }
ltm virtual novel-1000 {
...
    pool p-http
    profiles {
        analytics { }
        http { }
        tcp { }
    }
....
}


# tmctl proc_pid_stat 'proc_name=icrd_child' -s proc_name,ppid,vsize,rss

On subsequent GET requests the rss size continues to increase.

Impact:
Increase in the rss of the icrd_child process. This increase could lead to icrd_child core or swapping of processes into the rss due to limited memory left for other processes.

Workaround:
There is no workaround.

Fix:
The memory leak was identified and fixed.


721342 : No support for Per-App VE LTM and Per-App VE WAF images for various cloud environments.

Component: TMOS

Symptoms:
No Per-App Virtual Edition (VE) LTM or Per-App VE WAF images for various cloud environments.

Conditions:
Various Cloud environments (i.e., Azure, Google, AWS).

Impact:
No options to use various Per-App VE features.

Workaround:
None.

Fix:
There are now offerings for various cloud environments (Azure, Google, AWS). There are two types of Per-App VE LTM and two of Per-App VE WAF images.


721261-1 : v12.x Policy rule names containing slashes are not migrated properly

Component: Local Traffic Manager

Symptoms:
When migrating from v12.x to v13.0.x, 13.1.x, or 14.0.x, LTM polices that have rules containing the slash character will not migrate properly, and roll-forward migration will fail.

Conditions:
-- BIG-IP systems running v12.x.
-- Configuration contains LTM Policy rules whose names include the slash (/) character.
-- Upgrading to v13.0.x, 13.1.x, or 14.0.x.

Impact:
Roll-forward migration fails with the error: illegal characters in rule name.

Workaround:
Edit the rule names within LTM Policies, replacing the now-illegal slash (/) character with a legal alternative, such as the underscore (_).

Alternately, prior to migration, rename the rules on the v12.x system, and then perform the upgrade.

Fix:
BIG-IP software v12.x Policy rule names containing slashes are properly migrated.


720961-1 : Upgrading in Intelligence Community AWS environment may fail

Component: TMOS

Symptoms:
After importing an AWS instance into the Intelligence Community (IC) AWS environment you cannot upgrade to 13.1.0.2 (or later) because their license fails to validate afterwards.

Conditions:
-- Manually imported BIG-IP image into AWS IC.
-- Upgraded to 13.1.0.2 or newer.

Impact:
The system fails to validate the license after rebooting into the new software. Licenses are inoperative after the upgrade.

Workaround:
To work around this issue, perform the following procedure:
1. Redeploy a BYOL image from the IC marketplace.
2. Move over the license.
3. Configure the system.

Fix:
This release provides improved license validation to address this situation by detecting the IC environment, but allowing prior licenses to function.


720757-1 : Without proper licenses Category Lookup always fails with license error in Allow Ending

Component: Access Policy Manager

Symptoms:
When a per-request policy's Category Lookup agent queries Standard Categories with an expired SWG/URLF license (and has 'RST on failure' disabled), the execution always ends with the following log messages:

Error: Global concurrent url filter session limit reached

The connection is aborted.

Conditions:
Category Lookup agent set to query the urldb engine (standard categories) with either no SWG/URLF license or an expired license.

Impact:
Even if 'RST on failure' is disabled, traffic cannot reach the allow ending successfully.

Workaround:
Remove the Category Lookup with standard category lookup from the per-request policy if the licenses do not allow for its use.

Fix:
The allow ending is now reached successfully and does not error out if Category Lookup fails due to licensing errors but is set to disable 'RST on failure'.


720756-1 : SNMP platform name is unknown on i11400-DS/i11600-DS/i11800-DS

Component: TMOS

Symptoms:
The SNMP query for platform marketing name is 'unknown' for i11400-DS/i11600-DS/i11800-DS.

Conditions:
-- On licensed system, check OID marketing name.
-- Using i11400-DS/i11600-DS/i11800-DS platforms.

Impact:
Cannot tell the actual platform name in the SNMP query.

Workaround:
There is no workaround at this time.

Fix:
SNMP platform name is now reported correctly on BIG-IP i11400-DS/i11600-DS/i11800-DS platforms.


720713-2 : TMM traffic to/from a i10600/i10800 device in vCMP host mode may fail

Component: TMOS

Symptoms:
When a i10600/i10800 device is configured as a vCMP host and at least one vCMP guest is running (or ever ran), TMM traffic to the vCMP host may fail.

Note: Management port traffic to/from the device is unaffected.

Note: TMM traffic to/from the vCMP guests running on the device is unaffected. This issue affects only the vCMP host.

The most visible symptom is that TMM on the vCMP host fails to answer ARP requests that it receives for its own Self-IP addresses.

Conditions:
This issue occurs when all of the following conditions apply:

- i10600/i10800 device in vCMP host mode.

- At least one vCMP guest is deployed or was deployed, at some point.

Impact:
Inability to communicate to/from the vCMP host via its Self-IP addresses.

Workaround:
None.

Fix:
The vCMP host continues to handle traffic correctly once a guest is started.


720695-1 : Export then import of APM access Profile/Policy with advanced customization is failing

Component: Access Policy Manager

Symptoms:
An exported policy containing advanced customization fails to import.

Conditions:
-- Export policy containing advanced customization.
-- Import the same policy.

Impact:
Import fails.

Workaround:
None.

Fix:
Access policy import containing advanced customization now succeeds.


720651-2 : Running Guest Changed to Provisioned Never Stops

Component: TMOS

Symptoms:
After changing a vCMP guest status to provisioned the guest remains running showing a status of stopping.

Conditions:
-- Guests deployed on a BIG-IP VCMP host.
-- Changing the state from deployed to provisioned.

Impact:
Guests do not stop and change status until vcmpd process is restarted.

Workaround:
There is no workaround.

Fix:
The guest now stops when the state is changed from deployed to provisioned.


720585-1 : Signatures generated by Behavioral DOS algorithm can create false-positive signatures

Component: Anomaly Detection Services

Symptoms:
There is probability that the generated signatures will block unknown traffic (the traffic that was not presenting before the attack) even if it's not necessary from service health perspective

Conditions:
Run attack traffic.
In parallel run unknown traffic.
It should exceed the learned baseline together with the good traffic.

Impact:
The signatures may block unknown traffic even if it's not necessary from S/H perspective

Workaround:
There is no workaround at this time.

Fix:
Implement adaptive ratio threshold for covering current bad traffic samples. The ratio increases as long as the health is not good.
If the health returns to good levels (below one) the ratio is restarted to the initial value.


720461-2 : qkview prompts for password on chassis

Component: TMOS

Symptoms:
Qkview prompts for a password when executing getnodedates from the chassis module.

Conditions:
SSH auth keys are missing or corrupted.

Impact:
This blocks collecting qkview.

Workaround:
Edit the /user/bin/getnodedates script and add -oBatchMode=yes for the SSH command as shown in the following example:

        $date = `/usr/bin/ssh -q -oBatchMode=yes -oStrictHostKeyChecking=no $ip /bin/date`;

Fix:
The qkview is no longer blocked with a password prompt.


720460-1 : Compression traffic still goes to hardware accelerator when compression.strategy is set to softwareonly

Component: Local Traffic Manager

Symptoms:
The sys db compression.strategy is used to control the compression-provider selection. When it is set to 'softwareonly', compression somehow still selects a hardware accelerator.

Conditions:
This always happens when compression.strategy is set to 'softwareonly'.

Impact:
Compression.strategy set to 'softwareonly' does not work, and if there is a problem in a hardware accelerator, compression will still select the hardware accelerator, even though software compression is desired.

Workaround:
There is no workaround.

Fix:
The system now supports software compression when the sys db variable compression.strategy is set to 'softwareonly'.


720391-2 : BIG-IP i7800 Dual SSD reports wrong marketing name under 'show sys hardware'

Component: TMOS

Symptoms:
BIG-IP i7800 Dual SSD reports wrong marketing name under 'show sys hardware' and 'unknown' when OID sysObjectID is queried.

Conditions:
1. Execute 'tmsh show sys hardware'.
2. Query for OID sysObjectID using snmpwalk.

Impact:
System is reported as 'none', when it should be i7800-D. Platform with dual SSD not correctly identified using tmsh command or snmpwalk.

Workaround:
None.

Fix:
Added new SNMP OID for BIG-IP i7800 Dual SSD, reported as BIG-IP i7800-D, which is the correct designation.


720293-3 : HTTP2 IPv4 to IPv6 fails

Component: Local Traffic Manager

Symptoms:
An IPv4-formatted virtual server with an IPv6-formatted pool member cannot establish the connection to the pool member if HTTP2 is configured.

Conditions:
-- IPv4 virtual Server.
-- IPv6 pool member.
-- HTTP2 profile.

Impact:
Traffic connection does not establish; no traffic passes.

Workaround:
None.

Fix:
In this release, an IPv4-formatted virtual server with an IPv6-formatted pool member works with HTTP2.


720214-1 : NTLM Authentication might fail if Strict Update in iApp is modified

Component: Access Policy Manager

Symptoms:
Exchange Proxy NTLM Authentication failure when iApp strict updates is disabled initially and then turned on. NTLM authentication fails with STATUS_NO_LOGON_SERVERS.

Conditions:
The Strict Update option in the iApp is modified.

Impact:
Any service using NTLM authentication will be disrupted.

Workaround:
Restart ECA and NLAD modules to work correctly again. To do so, run the following commands:

bigstart restart nlad
bigstart restart eca

Fix:
NTLM authentication now works as expected when Strict Update in the iApp is modified.


720189-1 : VDI settings in APM Webtop have the incorrect URL for Citrix Receiver client download

Component: Access Policy Manager

Symptoms:
VDI settings have HTML5 package URL instead of Citrix Receiver download link. Hyperlink directs to HTML5 package link.

Conditions:
-- Citrix VDI is configured in Replacement mode.
-- HTML5 package is configured using Citrix client bundle.
-- Citrix HTML5 client bundle is used with Connectivity profile attached to the virtual server.

Impact:
The incorrect package is downloaded to the APM Webtop user.

Workaround:
None.

Fix:
Fixed the hyperlink for Citrix Receiver download in VDI settings of Webtop.


720104-1 : BIG-IP i2800/i2600 650W AC PSU has marketing name missing under 'show sys hardware'

Component: TMOS

Symptoms:
BIG-IP i2800/i2600 650W AC PSU is missing marketing name under 'show sys hardware' and displays 'unknown' when OID sysObjectID is queried.

Conditions:
-- Execute 'tmsh show sys hardware'.
-- Query for OID sysObjectID using snmpwalk.
-- Using BIG-IP i2800/i2600 platforms.

Impact:
System reports an empty marketing name when queried. Platform with 650W AC PSU not identified using tmsh command or snmpwalk.

Workaround:
There is no workaround at this time.

Fix:
Added new SNMP OID for BIG-IP i2800/i2600 650W AC PSU.


720045-1 : IP fragmented UDP DNS request and response packets dropped as DNS Malformed

Component: Advanced Firewall Manager

Symptoms:
AFM/DHD treats the IP fragmented UDP DNS packet (request or response) as DNS Malformed packet and drops these packets.

Conditions:
-- AFM/DHD is enabled (provisioned and licensed).
-- DNS Malformed vector is enabled at Device context (by default, it's always enabled).
-- AFM/DHD receives fragmented IP packet for UDP DNS request or response.

Impact:
AFM/DHD incorrectly treats such packets as DNS malformed and drops them.

If AFM/DHD receives any DNS request/response UDP packet that is fragmented at the IP layer, the system drops the packet, interrupting DNS service between client/servers through BIG-IP systems.

Workaround:
None.

Fix:
This issue is now fixed, as follows:

a) For IP fragmented UDP DNS request packet, AFM parses the first IP fragment to retrieve the L4 (UDP) header and DNS header and Question section.

  - If this information is available in the first IP fragment, AFM processes the packet for further DoS checks.
  - If this information is not available in first IP fragment, AFM drops the fragment as DNS Malformed.

b) For IP fragmented UDP DNS response packet, AFM parses the first IP fragment to retrieve the L4 (UDP) header and DNS header and Question section.

  - If this information is available in the first IP fragment, AFM processes the packet for further DOS checks.
  - If, however, this information is not available in first IP fragment, default behavior is that AFM allows such fragmented DNS response packet through. This behavior can be changed to drop incomplete IP fragmented DNS response packet by setting db variable 'Dos.dns.respfrag.allow' to 'false'.


719770-2 : tmctl -H -V and -l options without values crashed

Component: TMOS

Symptoms:
When the -H, -V or -l options were passed to tmctl without a following value, then tmctl crashed.

Conditions:
Use one of these options without the required value.

Impact:
Core file. No other impact.

Workaround:
Be sure to pass the required value with these options.

Fix:
The missing value is now reported as an error.


719644-2 : If auto-discovery is enabled, GTM configuration may be affected after an upgrade from 11.5.x to later versions

Component: Global Traffic Manager (DNS)

Symptoms:
When an LTM configuration has virtual server names containing period (.), a GTM configuration with auto-discovery enabled is upgraded from v11.5.x to a later version might lose consistency.

Conditions:
-- Upgrading GTM from v11.5.x to later versions (e.g., BIG-IP DNS v12.x).
-- Auto-discovery is enabled.
-- LTM configuration containing virtual server names containing the period character (.).
-- The same virtual server name is used in multiple partitions.

Impact:
The configuration after the upgrade might be affected in the following ways:
-- Missing virtual servers.
-- Deletion of some pool members.
-- Virtual servers with incorrect addresses.

Workaround:
There is no workaround at this time.

Fix:
This release supports GTM upgrades from 11.x to BIG-IP DNS 12.x and later versions for cases where:
-- Auto-discovery is enabled.
-- LTM virtual server name a period character (.).
-- The same virtual server name is used in multiple partitions.


719600-2 : TCP::collect iRule with L7 policy present may result in connection reset

Component: Local Traffic Manager

Symptoms:
If an iRule utilizing TCP::collect and HTTP_REQUEST is on a virtual server with an L7 policy, the policy engine may cause the connection to be unexpectedly reset with a 'policy execution error' reset cause, and 'Unable to resume pending policy event on connflow' will be logged to /var/log/ltm.

Conditions:
TCP::collect and HTTP_REQUEST iRule with L7 policy on virtual server.

Impact:
Connections may be unexpectedly reset and errors logged to /var/log/ltm.

Workaround:
At the start of the HTTP_REQUEST event, issue an 'after 1' command to allow the policy engine to reach a consistent state before proceeding with the remainder of the iRule.


719597 : HA between VIPRION chassis with B2250 blades might not work when running v12.1.1 and v13.1.0

Component: TMOS

Symptoms:
When two VIPRION chassis with B2250 blades try to form a high availability (HA) connection, with one running v12.1.1 and the other running v13.1.0, HA connections between TMMs on Active and TMMs on Standby might fail.

Conditions:
-- Two VIPRION chassis with B2250 blades.
-- One blade running v12.1.1 and the other running v13.1.0.
-- Try to configure HA connection.

Impact:
Fail to form HA connection.

Workaround:
There is no workaround other than installing the same software on both blades.

Fix:
A new sys db mhdag.bitshift is added to support HA configurations using B2250 blades running v12.1.1 and v13.1.x. Set mhdag.bitshift to 5 on VIPRION chassis with B2250 blades running v13.1.x by running the following command: tmsh modify sys db mhdag.bitshift value 5

HA between VIPRION chassis running v12.1.1 and v13.1.x with B2250 blades now works as expected.


719554-2 : Linux Kernel Vulnerability: CVE-2018-8897

Solution Article: K17403481


719459-2 : Wrong add suggestions created for already existing response URLs when differentiate HTTP and HTTPS is disabled

Component: Application Security Manager

Symptoms:
Policy builder suggests adding already existing URLs (as HTTPS) learned from response.

Conditions:
-- Differentiate HTTP and HTTPS URLs is disabled.
-- Learn-from response is enabled.
-- Server responses contain HTTPS links for HTTP requests instead of using protocol-relative URLs.

Impact:
This is a cosmetic issue in that incorrect suggestions are created. There is no impact to functionality.

Workaround:
Add the incorrect suggestions to the 'ignore' list.

Fix:
Policy builder no longer creates suggestions to add already existing URLs.


719396-1 : DHCP Client on BIG-IP sends host-name 'localhost' in DISCOVER packet after first boot.

Solution Article: K34339214

Component: TMOS

Symptoms:
DHCP Client on the BIG-IP system sends host-name 'localhost' in DISCOVER packet after first boot.

Note: The problem goes away after the first boot.

Conditions:
-- DHCP is enabled (this is the default on BIG-IP systems).
-- Initial startup.

Impact:
This can affect the lease given out by the DHCP server in the given network if the DHCP server is configured in such a way as to dynamically update the DNS based on the host-name option in the lease.

Workaround:
-- Empty the /var/lib/dhclient/dhclient.leases file.
-- Run the following command: bigstart restart dhclient

Fix:
DHCP Client on the BIG-IP system no longer sends 'localhost' as the host-name in DISCOVER packet after first boot.


719192 : In VPE Agent VMware View Policy shows no properties

Component: Access Policy Manager

Symptoms:
When opened in Visual Policy Editor (VPE) VMware View, the policy shows an empty properties page instead of the expected policy options.

Conditions:
Open a policy in VPE VMware View.

Impact:
Unable to configure VMware view policy from VPE.

Workaround:
Use tmsh to configure VMware View policies.

Fix:
Properties are now displayed correctly in Visual Policy Editor (VPE) VMware View.


719186-2 : Multipart/form-data requests may generate false positive 'missing strong integrity parameter' alerts

Component: Fraud Protection Services

Symptoms:
Multipart/form-data requests are not supported in FPS. FPS-protected pages which have the enhanced data-manipulation feature enabled, may generate a false-positive 'missing strong integrity parameter' alert for multipart/form-data requests.

Conditions:
-- FPS profile attached to a virtual server.
-- Multipart/form-data requests to a URL.
-- Enhanced data-manipulation feature enabled on a protected URL.

Impact:
False-positive 'missing strong integrity parameter' alert.

Workaround:
Use the ANTIFRAUD::disable_alert iRule command to drop the alert:

(set the static::drop_alert variable, probably matching URL name and checking that Content-Type header starts with 'multipart/form-data')

when ANTIFRAUD_ALERT {
    if {$static::drop_alert eq 1 &&
            [ANTIFRAUD::alert_type] eq "vtoken" &&
            [ANTIFRAUD::alert_component] eq "no_strong_integrity_param" } {
        ANTIFRAUD::disable_alert
        set static::drop_alert 0
    }
}

Fix:
FPS no longer sends automatic-transaction alerts for unsupported requests, so multipart/form-data requests no longer generate false positive 'missing strong integrity parameter' alerts.


719149-2 : VDI plugin might hang while processing native RDP connections

Component: Access Policy Manager

Symptoms:
Rarely, during processing of native RDP connections, the VDI plugin might hang, which prevents launch of VDI resources (Native RDP, Citrix, VMware View) from the APM Webtop.

Conditions:
APM Webtop is configured with native RDP resource.

Impact:
VDI resources (Native RDP, Citrix, VMware View) cannot be launched from APM Webtop.

Workaround:
None.

Fix:
Fixed rare VDI plugin hang caused by processing of native RDP connections.


719005-1 : Login request may arrive corrupted to the backend server after CAPTCHA mitigation

Component: Application Security Manager

Symptoms:
Login request arrives corrupted (specifically, some of the characters in the payload are changed).

Conditions:
-- A brute force CAPTCHA mitigation happens.
-- Specific traffic conditions.

Impact:
Login request fails.

Workaround:
None.

Fix:
CAPTCHA request-handling now works as expected.


718885-3 : Under certain conditions, monitor probes may not be sent at the configured interval

Solution Article: K25348242

Component: Global Traffic Manager (DNS)

Symptoms:
When sufficient resources are configured with the same monitor interval, and the monitor timeout value is no more than two times the monitor interval, the monitored resource may be marked as unavailable (due to a timeout) then immediately changed to available.

Conditions:
Monitor interval is lower than the number of resources configured with the same monitor interval.

Impact:
Monitor probes are not consistently performed at the configured interval.

Workaround:
Set the monitor interval to a value greater than the number of resources monitored using that same interval value.

The key to this workaround is to ensure that the sum of the resources monitored at a given interval is not greater than the interval. That can be accomplished several ways depending on your configuration and requirements.

For example, if the monitoring interval is 30 seconds and there are 40 different monitors with a 30-second interval and each monitor is assigned to exactly one resource, there are at least two options:

-- Change the interval for 10 of the monitors to a different value.

-- Set the monitor interval to 40.

Note: If you change the monitor interval, make sure to also change the timeout value to follow best practices:
timeout value = (3 x interval) + 1.

Fix:
Monitoring is now consistently performed at the configured interval regardless of the number of resources with the same monitor interval.


718772-2 : The generated signature creates incorrect predicate http.unknown_header (instead of http.unknown_header_exists)

Component: Anomaly Detection Services

Symptoms:
The generated signature creates incorrect predicate http.unknown_header (instead of http.unknown_header_exists).

Conditions:
Attack with traffic with 'unknown' header, for example
'Upgrade-Insecure-Requests: 1'.

Impact:
In the GUI, when the signature with the predicate 'unknown_header' is edited, this predicate is empty (instead of exists / does not exist).

Workaround:
There is no workaround.

Fix:
1. Change 'http.unknown_header' predicate into 'http.unknown_header_exists'.
2. Keep supporting the old format 'http.unknown_header'.


718685-1 : The measured number of pending requests is two times higher than actual one

Component: Anomaly Detection Services

Symptoms:
The measured number of pending requests is two times higher than actual.

Conditions:
Virtual server configured with a Behavioral DoS profile.

Impact:
Server stress mechanism is more sensitive than planned. A temporary traffic spike can cause unnecessary DoS mitigation start.

Note: The admd process provides stress-based DoS detection and mitigation control. When admd is not running, there is no stress-based anomaly detection or behavioral statistics aggregation.

Workaround:
Modify the adm.health.sensitivity value.

For example, to change health sensitivity from 50 to 500, run the following command:
tmsh modify sys db adm.health.sensitivity value 500

Fix:
Fixed initial adm flow sampling, so that the measured number of pending requests now equals actual.


718525-1 : PostGreSQL errors seen on startup after removing the mcpd binary database and rebooting

Component: TMOS

Symptoms:
After removing the mcpd binary database and restarting, occasionally there are errors similar to the following in /var/log/ltm:

warning postgres[7663]: [3-1] ERROR: duplicate key value violates unique constraint "vlan_pkey"

(The object type may be something other than 'vlan_pkey'.)

Conditions:
This occurs when you remove the mcpd binary database and reboot the system.

Impact:
The configuration does not load until 'bigstart restart' is executed.

Workaround:
None.

Fix:
The 'duplicate key value' errors are no longer seen when removing the mcpd binary database and restarting.


718152 : ASM GUI request log does not load on cluster

Solution Article: K14591455

Component: Application Security Manager

Symptoms:
The ASM Request Log fails to load, and it keeps reading 'Loading Requests Log...'.
'Security :: Event Logs :: Application :: Requests'.

Conditions:
-- Any cluster device (vCMP or not), even if there is a single blade in use.
-- Running BIG-IP v13.1.0.4, v13.1.0.5, v13.1.0.6, or v13.1.0.7. (Other releases are not affected.)

Impact:
Cannot view the Request Log in the GUI.

Workaround:
None

Fix:
The ASM request log can now be loaded correctly on cluster devices.


718071-1 : HTTP2 with ASM policy not passing traffic

Component: Local Traffic Manager

Symptoms:
HTTP2 traffic does not pass traffic if an ASM policy is applied.

Conditions:
-- HTTP2 virtual server.
-- ASM policy applied.

Impact:
Traffic does not pass.

Workaround:
No workaround.

Fix:
HTTP2 and ASM now work correctly together.


717909 : tmm can abort on sPVA flush if the HSB flush does not succeed

Component: Advanced Firewall Manager

Symptoms:
When the BIG-IP system comes up, or when tmm/dwbld/iprepd restarts, tmm does a flush of sPVA. If the operation does not succeed, the system can wait for 10 seconds, which might cause an abort due to heartbeat failure. tmm crash

Conditions:
-- BIG-IP system comes up, or tmm/dwbld/iprepd restart.
-- HSB flush does not succeed within ~3 seconds (it is supposed to succeed within ~3 seconds unless something is wrong with the HSB).

Impact:
tmm will have to be restarted. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
The system now checks asynchronously to determine whether or not the flush sPVA has succeeded.


717900-2 : TMM crash while processing APM data

Solution Article: K27044729


717888 : TMM may leak memory when a virtual server uses the MQTT profile.

Solution Article: K26583415


717832 : Remove unneeded files from UCS backup directories

Component: TMOS

Symptoms:
When using auto scale cloud formation templates, the system creates large bigip.ucs files that require additional storage space.

Conditions:
Deploy BIG-IP as part of auto scale cloud formation template (CFT).

Impact:
Large bigip.ucs file created requires additional storage space and might increase network traffic. (Size greater than 100 MB.)

Workaround:
Delete /config/cloud/* directories from the bigip.ucs file.

Fix:
This system no longer saves /config/cloud/ directories in UCS files, so the issue no longer occurs.


717785-1 : Interface-cos shows no egress stats for CoS configurations

Component: TMOS

Symptoms:
No egress packet counts per CoS queue are reported on B21x0 and 5x00/7x00 platforms. The issue affects the egress packet statistics reporting per CoS queue only. The operation of the CoS feature is not affected and the egress packets statistics reporting per external interface is also not affected.

Conditions:
-- Valid 8 HW CoS feature configuration has been enabled and passing traffic.
-- BIG-IP 5x00/7x00 platforms and VIPRION 21x0 blades.
-- Running the show net interface-cos command.

Impact:
Egress packet statistics reported per CoS queue shows no counts.

Workaround:
None.

Fix:
This release supports per egress CoS queue packet count statistics reporting for BIG-IP 5x00/7x00 platforms and VIPRION 21x0 blades.


717756-2 : High CPU usage from asm_config_server

Component: Application Security Manager

Symptoms:
Use of Automatic policy builder might result in high CPU usage of asm_config_server (and ASM slowdown).

Conditions:
- Automatic policy builder.
- Several entity types learning in 'Add all entities' configuration.

Impact:
ASM availability impacted.

Workaround:
-- Switch to Manual policy builder.
-- Set entity types learning to compact / selective / never.

Fix:
Policy builder no longer puts unnecessary load on ASM configurations.


717742-5 : Oracle Java SE vulnerability CVE-2018-2783

Solution Article: K44923228


717346-2 : [WebSocket ] tmsh show /ltm profile WebSocket current and max numbers far larger than total

Solution Article: K13040347

Component: Local Traffic Manager

Symptoms:
WebSocket profile statistics for current and maximum connections are always very large, even right after restarting the system. The numbers are several orders of magnitude larger than the statistics for total connections.

Conditions:
Rarely occurring, unstable network could be one of the reasons.

Impact:
Cannot use stats for troubleshooting.

Workaround:
Reset the stats using the following command:
# tmsh reset-stats ltm profile websocket


716992-2 : The ASM bd process may crash

Solution Article: K75432956


716940-2 : Traffic Learning screen graphs shows data for the last day only

Component: Application Security Manager

Symptoms:
Traffic Learning screen graphs shows data for the last day only.

Conditions:
Visit Learning screen 1 hour after policy creation.

Impact:
Statistics shown are for the last day. No statistics for longer or for shorter periods of time.

Workaround:
There is no workaround.

Fix:
Statistics are shown for the correct time interval, at most 2 weeks/policy creation date. Possible statistics intervals are as follows: 1 hour, 1 day, 2 weeks.


716922-2 : Reduction in PUSH flags when Nagle Enabled

Component: Local Traffic Manager

Symptoms:
When Nagle is enabled in the TCP profile, the number of PUSH flags generated by the BIG-IP system drops substantially compared to the Nagle-disabled case, or to the Nagle-enabled case prior to v12.1.2-HF1. This matters most when there is a single outstanding unsent segment in the send buffer awaiting acknowledgment of all other data.

Conditions:
-- Nagle is enabled.
-- Running BIG-IP software versions later than v12.1.2-HF1.

Note: The problem is only impactful when the client withholds ACKs when there is no PUSH flag.

Impact:
If the client withholds ACKs, this can save handset power, but it also causes Nagle's algorithm to withhold the last bit of data, increasing latency.

Workaround:
Set Nagle to the 'Auto' setting or 'Disabled'.

Mote: To take advantage of some of the Nagle benefits, use 'Auto'.

Fix:
Revised PUSH flag setting logic to set the flag in cases where sending is Nagle-limited.


716900-2 : TMM core when using MPTCP

Component: Local Traffic Manager

Symptoms:
In some cases TMM may crash when processing MPTCP traffic.

Conditions:
A TCP profile with 'Multipath TCP' enabled is attached to a virtual server.

Impact:
Traffic interrupted while TMM restarts.

Workaround:
There is no workaround other than to disable MPTCP.

Fix:
TMM no longer produces a core.


716788-2 : TMM may crash while response modifications are being performed within DoSL7 filter

Component: Application Security Manager

Symptoms:
TMM might crash when a response is modified by the DoSL7 filter while injecting an HTML response from a backend server.

Conditions:
-- ASM provisioned.
-- DoS application profile is attached to a virtual server.

Impact:
Traffic disrupted while tmm restarts, failover may occur.

Workaround:
There is no workaround other than to remove the DoS application profile from the virtual server.

Fix:
Response modification handler has been modified so that this issue no longer occurs.


716747-2 : TMM my crash while processing APM or SWG traffic

Component: Access Policy Manager

Symptoms:
Under certain circumstances, TMM may crash when processing APM or SWG.

There will be a log message in /var/log/apm near the time of crash with this:

err tmm[20598]: 01490514:3: (null):Common:00000000: Access encountered error: ERR_ARG. File: ../modules/hudfilter/access/access.c, Function: access_initiate_swgt_policy_done, Line: 17000.

Conditions:
APM or SWG enabled.

Impact:
TMM crash, leading to a failover event.

Workaround:
There is no workaround at this time.

Fix:
TMM now processes APM and SWG traffic as expected.


716746 : Possible tmm restart when disabling single endpoint vector while attack is ongoing

Component: Advanced Firewall Manager

Symptoms:
tmm restarts.

Conditions:
-- AFM DoS single endpoint (sweep or flood) vector is configured.
-- The attack is ongoing.
-- The attack vector is being mitigated in hardware (HW).
-- The vector is manually disabled.

Impact:
tmm can crash and restart. Traffic disrupted while tmm restarts.

Workaround:
If you do not want to mitigate, set the mitigation_threshold to infinite.

Note: Do not disable the single endpoint vectors when an attack is ongoing and the vector is being mitigated in HW.

Fix:
tmm no longer restarts when disabling single endpoint vector while an attack is ongoing.


716469 : OpenSSL 1.0.1l fails with 512 bit DSA keys

Component: TMOS

Symptoms:
In certain cases with FIPS enabled the box would fail to boot because of attempts to use 512 bit DSA keys.

Conditions:
During BIG-IP booting and fips is enabled.

Impact:
BIG-IP failed to boot.

Workaround:
There is no workaround at this time.

Fix:
Boot will no longer fail with OpenSSL and 512 bit DSA keys.


716392-1 : Support for 24 vCMP guests on a single 4450 blade

Component: TMOS

Symptoms:
Cannot create more than 12 vCMP guests per blade.

Conditions:
-- Using vCMP.
-- VIPRION blades.

Impact:
Cannot configure more than 12 vCMP guests.

Workaround:
None.

Fix:
This release supports 24 vCMP guests on 4450 blades. Increased vCMP density ensures that all guests can access the hardware capabilities available from the BIG-IP platforms and VIPRION blades.

Behavior Change:
This release supports 24 vCMP guests on VIPRION 4450 blades. Increased vCMP density ensures that all guests can access the hardware capabilities available from BIG-IP platforms and VIPRION blades.


716391-2 : High priority for MySQL on 2 core vCMP may lead to control plane process starvation

Solution Article: K76031538

Component: TMOS

Symptoms:
vCMP guest with only 2 cores (or 2 cores per blade for multi-blade guests) may undergo control plane process starvation, which could lead to failover due to CPU starvation of sod.

Conditions:
-- A device using Intel Hyper-Threading Technology is configured with only 2 cores (or 2 cores per blade for multi-blade vCMP guests).
-- A module using MySQL is provisioned, MySQL, for example BIG-IP ASM and BIG-IP Analytics (AVR). These other modules also implicitly provision AVR: ASM, AFM, DOS, APM, PEM, and vCMP.

Impact:
Control plane processes may experience CPU starvation, including failover due to CPU starvation of sod. This is a rarely occurring issue.

Workaround:
Revert to pre-11.5.1 HF4 behavior by setting the scheduler.splitplanes.asmopt database key to false.
 
IMPORTANT: You should not revert to pre-11.5.1 HF4 behavior unless requested by F5 Support. However, if required, you can disable this new behavior and revert to pre-11.5.1-HF4 behavior. For instructions on how to do so, see K16469: Certain BIG-IP ASM control plane processes are now pinned to the highest-numbered logical CPU core :: https://support.f5.com/csp/article/K16469.


716318-2 : Engine/Signatures automatic update check may fail to find/download the latest update

Component: Fraud Protection Services

Symptoms:
Engine/Signatures automatic update check may fail to find/download the latest update if the timestamp of the factory update file is later than the timestamp of the update file in downloads.f5.com.

Note: This issue is relevant only for engineering hotfixes.

Conditions:
This occurs when the following conditions are met:
-- Engineering hotfix.
-- The factory update file timestamp is later than the timestamp of the update file in downloads.f5.com.

Impact:
Automatic update check will detect the wrong update file.

Workaround:
Manually check, and then download and install the update file from downloads.f5.com, if needed.

Fix:
The factory update file timestamp is now set to 0, unless there is a reason to install the factory file over the current update file in downloads.f5.com.


716213-1 : BIG-IP system issues TCP-Reset with error 'SSID error (Out of Bounds)' in traffic

Component: Local Traffic Manager

Symptoms:
When APM connects to Active Directory Federation Services (ADFS), a blank page is observed. The log file /var/log/ltm exhibits a TCP reset issued by the BIG-IP system. The TCP capture informs that the issue is due to an out-of-bounds error that occurs when traffic gets parsed by the SSL Persistence Parser (SSID).

Conditions:
-- SSL persistence (SSID) is enabled for a virtual host.
-- APM connects to ADFS.

Impact:
A blank page is observed due to the TCP reset.

Workaround:
No workaround is available.

Fix:
The SSL persistence parser now validates the handshake message data-bytes are available prior to fetching them, so the bounds error does not get raised.


715923-1 : When processing TLS traffic TMM may reset connections

Solution Article: K43625118


715883 : tmm crash due to invalid cookie attribute

Component: Local Traffic Manager

Symptoms:
tmm crash due to invalid request-side cookie attribute.

Conditions:
While processing an HTTP request, an iRule tries to add a cookie attribute to a cookie that is only valid in an HTTP response (for instance, an 'expires' attribute).

Impact:
TMM cored. Traffic disrupted while tmm restarts.

Workaround:
None.


715820-1 : vCMP in HA configuration with VIPRION chassis might cause unstable data plane

Solution Article: K61422392

Component: TMOS

Symptoms:
When multiple vCMP guests are deployed in a high availability (HA) with VIPRION chassis, the data plane cluster might become unstable. When this issue occurs, the system posts repeated log messages in /var/log/ltm similar to the following:

-- CDP: exceeded 1/2 timeout for PG 3

Conditions:
-- Multiple vCMP guests are deployed.
-- HA configured.
-- Using VIPRION chassis.

Impact:
Unstable data plane might cause traffic disruption/packet drops.

Workaround:
None.

Fix:
This issue no longer occurs.


715785-2 : Incorrect encryption error for monitors during sync or upgrade

Component: Local Traffic Manager

Symptoms:
The system logs an error message similar to the following in /var/log/ltm:

err mcpd[6823]: 01071768:3: Encryption of the field (password) for object (<monitor name>) failed.

This may cause a configuration sync to fail, or an upgrade to fail.

Conditions:
The exact conditions are unknown, however it may occur under these circumstances:

-- Performing a config sync operation.
-- Performing an upgrade.

Impact:
Inability to sync peer devices, or an inability to upgrade.

Workaround:
There is no workaround at this time.

Fix:
This error is no longer triggered erroneously.


715756-2 : Clusterd may not trigger primary election when primary blade has critical filesystems mounted read-only

Component: Local Traffic Manager

Symptoms:
When filesystems critical to TMOS functional operation are mounted read-only, clusterd on that blade should trigger a primary election if it was primary. Either way, the cluster should be informed of this critical error state.

Conditions:
A critical filesystem (e.g., '/', '/var', '/var/run', '/config', '/shared') has been mounted read-only.

Impact:
The blade with read-only filesystems and degraded functionality might stay primary and claim to be passing traffic.

Workaround:
There is no workaround other than to avoid mounting filesystems read-only on a BIG-IP system.

Fix:
The blade with read-only filesystems and degraded functionality now yields primaryship to a more healthy cluster member.


715750-2 : The BIG-IP system may exhibit undesirable behaviors when receiving a FIN midstream an SSL connection.

Component: Local Traffic Manager

Symptoms:
Upon receiving a FIN midstream in an SSL connection, the BIG-IP system will immediately proxy the FIN to the remote host on the peer side. At this point, depending on the specific configuration of the remote host on the peer side, the BIG-IP system may exhibit behaviors that may be deemed undesirable.

For instance, if the remote host on the peer side acknowledges the FIN but ignores it and keeps sending data to the BIG-IP system, the BIG-IP system will drop all ingress data (i.e., not proxy it) while keeping the side that transmitted the original FIN open indefinitely.

Once the remote host on the peer side completes transmitting data and sends a FIN of its own, the BIG-IP system will finally release the side that sent the original FIN by allowing it to close.

Conditions:
This issue occurs when the following conditions are met:

-- A standard virtual server with the clientssl and serverssl profiles in use.

-- As part of a connection handled by the virtual server, one side sends a FIN midstream to the BIG-IP system.

Impact:
There is no real impact to the BIG-IP system because of this issue. However, both the client and the server can be negatively impacted by this issue.

For example, if the original FIN was received by the BIG-IP system on the clientside:

-- The client connection is not allowed to close for potentially a very long time. During this time, the client does not receive anything other than Keep-Alive packets from the BIG-IP system. This can delay or adversely impact applications on the client.

-- The server continues to send data to the BIG-IP system unnecessarily. After accounting for the fact this could be a lot of data and many connections could be in this state, the server and/or its surrounding network could become impacted/congested. At a minimum, this might cause a waste of resources.

Workaround:
There is no workaround at this time.

Fix:
SSL shutdown behavior is controlled with profile option 'Alert Timeout', which specifies the duration in time for the system to try to close an SSL connection before resetting the connection. The default is 'Indefinite'.

Behavior Change:
SSL shutdown behavior is controlled with profile option 'Alert Timeout', which specifies the duration in time for the system to try to close an SSL connection before resetting the connection. The default is 'Indefinite'.


715747 : TMM may restart when running traffic through custom SSLO deployments.

Component: Local Traffic Manager

Symptoms:
TMM restarts with a SIGSEGV signal and dumps core.

Conditions:
This issue is known to happen when passing traffic through some custom SSLO deployments (e.g., iRule-based configurations).

Impact:
TMM restarts. If the system is in a high availability configuration, a failover occurs. Traffic disrupted while TMM restarts.

Workaround:
There is no workaround at this time.

Fix:
TMM no longer restarts.


715467-2 : Ephemeral pool member and node are not updated on 'A' record changes if pool member port is ANY

Component: Local Traffic Manager

Symptoms:
If the 'A' record for an FQDN node is removed or replaced on the DNS server, then the BIG-IP system does not remove the old ephemeral pool member or create the new pool member and node.

Conditions:
-- FQDN nodes.
-- Pool members with a service port of ANY.

Impact:
Ephemeral pool member and nodes are not removed or updated when the DNS 'A' record is removed or changed.

Workaround:
There is no workaround at this time.

Fix:
Upon a DNS 'A' record being removed or changed, the ephemeral pool members and nodes based on the old value are removed or changed.


715448-2 : Providing LB::status with a GTM Pool name in a variable caused validation issues

Component: Global Traffic Manager (DNS)

Symptoms:
When utilizing an LB::status iRule where the GTM Pool name was provided in a variable, instead of directly written into the command, the system posts the following error message: Can't read 'monkey': no such variable.

Conditions:
LB::status pool a <Variable containing string>.

Impact:
Unable to use LB::status iRule.

Workaround:
There is no workaround at this time.

Fix:
Can now use LB::status iRule command to display the status of a GTM Pool when the name of the pool is provided in a variable.


715250-1 : TMM crash when RPC resumes TCL event ACCESS_SESSION_CLOSED

Component: Access Policy Manager

Symptoms:
TMM generates a core and restarts when RPC resumes iRule event ACCESS_SESSION_CLOSED.

Conditions:
Plugin RPC call has iRule event ACCESS_SESSION_CLOSED configured.

Impact:
System instability, failover, traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.


715207-3 : coapi errors while modifying per-request policy in VPE

Component: Access Policy Manager

Symptoms:
System reports errors about coapi in /var/log/ltm when modifying a per-request policy in the Visual Policy Editor (VPE).

err coapi: PHP: requested conversion of uninitialized member.

Conditions:
-- The per-request policy being modified is attached to the virtual server.
-- Using VPE.

Impact:
The impact can vary by version.
-- In some versions it represents a stray error log.
-- In other versions the 'Accepted Languages' in the Logon Page agent does not have the full list.
-- In other versions, there are server 500 errors when modifying agents.

Workaround:
To work around this issue, perform the following procedure: 1. Remove the per-request policy from the virtual server.
2. Modify the policy.
3. Add the policy back to the virtual server.

Fix:
Now per-request access policies can be simultaneously used and edited without causing spurious 'coapi' log errors.


715153-1 : AVR can generate many files in /var/avr/loader folder, which might cause disk-usage problem

Component: Application Visibility and Reporting

Symptoms:
-- The folder /var/avr/loader contains many files (e.g., more than 1500 files).
-- monpd is not running.

Conditions:
This occurs when the following conditions are met:
-- Avrd is running.
-- monpd is down.

Impact:
AVR writes many files to /var/avr/loader. Depending on disk usage, this might cause disk-usage problems.

Workaround:
There are two possible workarounds:
-- Restart monpd. When monpd starts up, it deletes the files under /var/avr/loader.
-- Delete all files under /var/avr/loader.

Fix:
There is now a limit for the /var/avr/loader folder, so that it can contain no more than 1100 files. This prevents disk-usage problems.


715128-1 : Simple mode Signature edit does not escape semicolon

Component: Application Security Manager

Symptoms:
When creating a user-defined ASM Signature in Simple mode, semicolon followed by space cannot be used in a keyword.

Conditions:
-- Using a user-defined ASM Signature in Simple mode.
-- There is a semicolon followed by a space in a keyword.

Impact:
The signature cannot be created.

Workaround:
The signature can be created in Advanced Mode by escaping semicolon as "|3B|".


715110 : AVR should report 'resolutions' in module GtmWideip

Component: Application Visibility and Reporting

Symptoms:
AVR does not report 'resolutions' in GtmWideip module.

Conditions:
One of the following modules is provisioned: AVR, AFM, or DNS/GTM.

Impact:
There are no statistics reported on 'resolutions' in GtmWideip module.

Workaround:
There is no workaround.

Fix:
AVR now reports 'resolutions' in GtmWideip module.


714974-2 : Platform-migrate of UCS containing QinQ fails on VE

Component: TMOS

Symptoms:
If you are doing platform-migrate from a hardware device to a Virtual Edition (VE) guest, and the UCS that came from the hardware device has a QinQ VLAN configuration, then the upgrade will fail.

Conditions:
-- VLAN with a QinQ configuration exists on the hardware configuration.
-- UCS file is saved on the hardware device.
-- UCS is copied to a VE guest.
-- Attempt to load the configuration using the following command: load sys ucs <UCS filename> no-license platform-migrate.

Impact:
The UCS load will fail and generate an error:

01071bd8:3: The tag-mode for requested member <VLAN name> has to be 'none' on platforms that do not support QinQ. Unexpected Error: Loading configuration process failed.

Workaround:
None.

Fix:
The configuration now loads successfully, disables QinQ on the associated VLAN, and warns that this action was automatically taken.


714961-1 : antserver creates large temporary file in /tmp directory

Component: Access Policy Manager

Symptoms:
SWG Analytics (running through the antserver daemon) creates a large temporary file in the /tmp directory due to a lack of write permissions on the appropriate directory.

Conditions:
-- SWG provisioned.
-- Viewing SWG Analytics.

Impact:
/tmp is temporarily populated with a large file that might fill up the directory if it is already close to capacity.

Workaround:
There is no workaround at this time.

Fix:
System now writes to /shared/tmp/ant_server so that it no longer writes to /tmp, so the issue no longer occurs.


714903-2 : Errors in chmand

Component: TMOS

Symptoms:
VIPRION cluster does not form; only primary blade stays online. Chassis manager chmand generated a core file. System logs the following error in ltm log: err clusterd[11162]: 013a0004:3: Error adding cluster mgmt addr, HAL error 7.

Conditions:
-- Restoring UCS.
-- Chassis starting up.
-- System in stressed condition, where there is very little or no memory available.

Impact:
Cluster does not form.

Workaround:
None.

Fix:
These errors in chmand are fixed.


714879-3 : APM CRLDP Auth passes all certs

Solution Article: K34652116


714749-2 : cURL Vulnerability: CVE-2018-1000120

Component: TMOS

Symptoms:
It was found that libcurl did not safely parse FTP URLs when using the CURLOPT_FTP_FILEMETHOD method. An attacker, able to provide a specially crafted FTP URL to an application using libcurl, could write a NULL byte at an arbitrary location, resulting in a crash, or an unspecified behavior.

Conditions:
BIG-IP systems are not affected by this vulnerability.

Impact:
None.

Workaround:
None.

Fix:
Patched CVE-2018-1000120


714700-2 : SSO for native RDP resources is not compatible with the 'Force Updated Clients' setting of 'Encryption Oracle Remediation' Group Policy

Component: Access Policy Manager

Symptoms:
To address a vulnerability in their CredSSP implementation Microsoft released set of updates for all versions of Windows (https://aka.ms/credssp). Although the APM implementation is not affected by this vulnerability, the Microsoft Windows Server fix introduces compatibility issues. The update adds new Group Policy 'Encryption Oracle Remediation', which, if set to 'Force Updated Clients' on the server might break SSO for APM's native RDP resources.

Conditions:
-- RDP server has https://aka.ms/credssp update installed.
-- 'Encryption Oracle Remediation' Group Policy on the RDP server is set to 'Force Updated Clients'.

Impact:
SSO for native RDP resources does not work.

Workaround:
Set 'Encryption Oracle Remediation' Group Policy on the RDP server to 'Mitigated'.

Fix:
SSO for native RDP resources is now compatible with the 'Force Updated Clients' setting of 'Encryption Oracle Remediation' Group Policy.


714654-2 : Creating static route for advertised dynamic route might result in dropping the tmrouted connection to TMM

Component: TMOS

Symptoms:
While creating a static route, tmrouted disconnects with error: existing entry is not a dynamic.

Conditions:
Creating a static route for a network that already has an advertised dynamic route.

Impact:
TMM's connection to tmrouted goes down, which results in loss of the dynamic route for TMM.

Workaround:
There is no workaround other than not creating a static route for routes received through dynamic routing.

Fix:
Creating static routes for advertised dynamic route no longer causes the tmrouted-to-TMM connection to drop.


714384-3 : DHCP traffic may not be forwarded when BWC is configured

Component: Local Traffic Manager

Symptoms:
DHCP traffic may not be forwarded when BWC is configured on the system.

Conditions:
-- DHCP virtual server configured.
-- BWC policy configured and attached to the route-domain.

Impact:
DHCP traffic may not be forwarded.

Workaround:
There is no workaround other than to remove the BWC policy.

Fix:
DHCP traffic is now forwarded when BWC is configured,


714369 : ADM may fail when processing HTTP traffic

Solution Article: K62201098


714350 : BADOS mitigation may fail

Solution Article: K62201098


714334-1 : admd stops responding and generates a core while under stress.

Component: Anomaly Detection Services

Symptoms:
admd stops responding and generates a core while under stress.

Conditions:
-- The BIG-IP system is under stress.
-- CPU starvation is occurring.

Impact:
admd core and restart.

Note: The admd process provides stress-based DoS detection and mitigation control. When admd is not running, there is no stress-based anomaly detection or behavioral statistics aggregation.

Workaround:
None.

Fix:
This issue no longer occurs.


714303-1 : X520 virtual functions do not support MAC masquerading

Component: TMOS

Symptoms:
MAC masquerading is not supported when using X520 virtual functions via SR-IOV in Virtual Edition (VE). This is due to a lack of functionality in the underlying hardware.

Conditions:
-- Use SR-IOV virtual functions as interfaces within VE.
-- Configure MAC masquerading with a shared MAC that should follow the active system after traffic group failovers.

Impact:
MAC masquerading will not function in this environment.

Workaround:
There is no workaround other than not to use MAC masquerading, as conventional failover works for this environment.

Fix:
MAC masquerading is now supported when using X520 virtual functions via SR-IOV in Virtual Edition (VE).


713951-5 : tmm core files produced by nitrox_diag may be missing data

Component: Local Traffic Manager

Symptoms:
When the nitrox_diag utility generates a tmm core file, that file might include data for only one tmm thread instead of all tmm threads.

Conditions:
-- Running the nitrox_diag utility.
-- Using devices with the Cavium Nitrox crypto card.
-- The nitrox_diag utility generates a tmm core file.

Impact:
The resulting core file might include data for only one tmm thread instead of all tmm threads, making it more difficult for F5 to diagnose reported problems with the Cavium Nitrox crypto card. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
When the nitrox_diag utility generates a tmm core file, that file now includes data for all tmm threads instead of only one.


713947-1 : stpd repeatedly logs "hal sendMessage failed"

Component: TMOS

Symptoms:
On non-primary clustered blades in a BIG-IP chassis environment, stpd may repeatedly log "hal sendMessage failed"

Conditions:
Two or more blades clustered in a chassis with STP enabled on one or more ports.

Impact:
All BIG-IP blades

Workaround:
No workaround except to ignore log messages - they are spurious and have no ill effect on the system besides log spam.


713934-2 : Using iRule 'DNS::question name' to shorten the name in DNS_REQUEST may result malformed TC response

Component: Local Traffic Manager

Symptoms:
Received malformed Truncated DNS response.

Conditions:
-- Using iRule 'DNS::question name' to shorten the name.
-- DNS response is longer than 4096 UDP limit.

Impact:
DNS request might not be resolved correctly.

Workaround:
There is no workaround at this time.

Fix:
In this release, when DNS::query name changes the query name, the software ensures that the request is updated with proper lengths.


713932-1 : Commands are replicated to PostgreSQL even when not in use.

Component: TMOS

Symptoms:
In MCP, PostgreSQL is used only by AFM. Even when AFM is not provisioned, however, commands are being replicated to PostgreSQL.

Conditions:
AFM is not provisioned.

Impact:
Unnecessary CPU cycles spent by MCP replicating commands to PostgreSQL.

Workaround:
None.

Fix:
Prevented replication of commands to PostgreSQL when it is not in use.


713820-1 : Pass in IP to urldb categorization engine

Component: Access Policy Manager

Symptoms:
Category lookup results might be less accurate. In some cases, the system returns 'uncategorized' when the reference (Forcepoint) returns a specific category.

Conditions:
Category Lookup agent is in per-request policy using the categorization engine to lookup up a website's classification.

Impact:
Actions leveraging categorization results will be applied incorrectly.

Workaround:
None.

Fix:
This release can now pass in more information to the urldb categorization engine, which supports finger-grained categorization.


713813-2 : Node monitor instances not showing up in GUI

Component: TMOS

Symptoms:
Navigating to Local Traffic :: Monitors :: <some_monitor> should show a list of nodes with some_monitor assigned to them. GUI does not list related nodes under Instances tab.

Conditions:
-- At Local Traffic :: Monitors :: <some_monitor>.
-- Under the Instances tab.

Impact:
No instances listed. Cannot use the GUI to determine which nodes are associated with a monitor.

Workaround:
Use tmsh to list nodes associated with a monitor.

Fix:
The GUI now lists all associated nodes under Local Traffic ›› Monitors :: <some_monitor> :: Instances tab.


713655-2 : RouteDomainSelectionAgent might fail under heavy control plane traffic/activities

Component: Access Policy Manager

Symptoms:
If per-session access policy requests route domain selection during policy evaluation, the agent might hang and fail to retrieve route domain information. APMD might restart and produce a core file.

Conditions:
-- Policy sync operation.
-- System receives a query for route domain information.
-- There is a heavy load of control plane traffic to process.

Impact:
Poor performance and/or APMD restarts with a core file, causing brief disruption of the authentication service.

Workaround:
None.

Fix:
Route domain selection operations no longer fail under heavy control-plane traffic/activities.


713612-1 : tmm might restart if the HTTP passthrough on pipeline option is used

Component: Local Traffic Manager

Symptoms:
The TMM may crash if the HTTP profile's 'passthrough_pipeline' field is set to 'passthrough'.

Conditions:
-- HTTP profile is configured as a transparent proxy.
-- HTTP profile has the 'passthrough_pipeline' field is set to 'passthrough'.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
tmm no longer crashes when HTTP switches to passthrough mode in some cases.


713533-2 : list self-ip with queries does not work

Component: Local Traffic Manager

Symptoms:
"list net self" command always returns all Self IPs, regardless of the regex patterns.

Conditions:
list net self always returns all Self IPs

Impact:
You are unable to filter the Self IP list using a regex pattern.

Fix:
You can now use pattern matching to list Self IPs


713390-1 : ASM Signature Update cannot be performed on hourly billing cloud instance

Component: Application Security Manager

Symptoms:
ASM Signature Update cannot be performed on hourly billing cloud (AWS) instance. Licenses on these devices cannot be updated and have a fixed Service Check Date (SCD), which must be more recent to allow ASM Signature Update.

Conditions:
Attempt to perform ASM Signature Update on hourly billing cloud (AWS) instance.

Impact:
Performing ASM Signature Update fails.

Workaround:
There is no workaround at this time.

Fix:
ASM Signature Update can now be performed on hourly billing cloud instance.


713282-1 : Remote logger violation_details field does not appear when virtual server has more than one remote logger

Component: Application Security Manager

Symptoms:
Remote logger violation_details field appears empty.

Conditions:
-- More than one remote logger is attached to the virtual server.
-- A violation occurs.
-- Viewing the associated log.

Impact:
Violation_details field appears empty in logs.

Workaround:
There is no workaround at this time.

Fix:
Remote logger violation_details field is now populated as expected when the virtual server has more than one remote logger.


713273 : BIG-IP sys db variable avr.stats.internal.maxentitiespertable resets back to default value on services restart

Component: Application Visibility and Reporting

Symptoms:
After a BIG-IP system reset, a modified setting for the BIG-IP sys db variable avr.stats.internal.maxentitiespertable returns to the default value.

Conditions:
1. avr.stats.internal.maxentitiespertable value is modified from the default.
2. The BIG-IP system restarts.

Impact:
avr.stats.internal.maxentitiespertable returns to its default value.

Workaround:
After BIG-IP system reset, specify the value of avr.stats.internal.maxentitiespertable again.

Fix:
A modified avr.stats.internal.maxentitiespertable value no longer returns to the default value after BIG-IP system restart.


713156-1 : AGC cannot do redeploy in Exchange and ADFS use cases

Component: Access Policy Manager

Symptoms:
In AGC exchanges or Active Directory Federation Services (ADFS) configurations, the system creates an SSL HTML form and SSO HTML form control object. Because of the limitation of ICRD, the system cannot directly delete SSO HTML form control objects.

Conditions:
-- Redeploy occurs in an AGC exchange ADFS configuration.
-- Modifying existing configurations.

Impact:
Redeploy fails, configuration remain unmodified.

Workaround:
Do a undeploy, followed by a deploy.

Fix:
Redeploy now succeeds when using AGC with Exchange and ADFS use cases.


713111-1 : When ASM and APM are configured on the same virtual server, ASM might report some errors in its logging.

Component: Access Policy Manager

Symptoms:
When APM (SSO feature) and ASM are configured on the same virtual server, WebSSO recreates requests on 401 responses. Such requests have the same support ID, so ASM logs errors.

Conditions:
APM (WebSSO) and ASM are configured on same virtual server.

Impact:
ASM might potentially block such requests, so APM SSO functionality may not work.

Workaround:
There is no workaround except to not configure APM (WebSSO) and ASM on same virtual server.

Fix:
This issue has been fixed.


713066-1 : Connection failure during DNS lookup to disabled nameserver can crash TMM

Solution Article: K10620131

Component: Global Traffic Manager (DNS)

Symptoms:
TMM restart as a result of a DNS lookup to disabled nameserver.

Conditions:
DNS lookup that tries to connect to a nameserver that is not reachable for some reason.

This could be an explicit 'RESOLV::lookup' command in iRule, or it could be DNS lookups internally triggered by APM or HTTP.

Impact:
TMM restarts. Traffic disrupted while tmm restarts.

Workaround:
Verify connectivity to nameserver.

As an alternative, refrain from using RESOLV::lookup in iRules.

Fix:
This issue is now fixed.


712924-1 : In VPE SecurID servers list are not being displayed in SecurID authentication dialogue

Component: Access Policy Manager

Symptoms:
In VPE SecurID servers list are not being displayed in SecurID authentication dialogue. List is displaying only None.

Conditions:
Always when adding SecureID authentication action.

Impact:
Inability to (re)configure SecureId via VPE.

Workaround:
Manually modify RSA AAA server in SecurID Auth Agent via tmsh:

tmsh modify apm policy agent aaa-securid <aaa agent> server <securid server name>


712876-2 : CVE-2017-8824: Kernel Vulnerability

Solution Article: K15526101


712819-2 : 'HTTP::hsts preload' iRule command cannot be used

Component: Local Traffic Manager

Symptoms:
Using the 'HTTP::hsts preload' iRule command in the GUI, you might see error message similar to the following:
01070151:3: Rule [/Common/test] error: /Common/test:3: error: ["invalid argument preload; expected syntax spec: "][HTTP::hsts preload enable].

The message is incorrect: the command has the correct format. However, the system does not run it.

Conditions:
Using the 'HTTP::hsts preload' iRule command in the GUI.

Impact:
The command does not run. Cannot use the 'HTTP::hsts preload' iRule command in the GUI.

Workaround:
None.

Fix:
'HTTP::hsts preload' iRule command now works as expected.


712710 : TMM may halt and restart when threshold mode is set to stress-based mitigation

Component: Advanced Firewall Manager

Symptoms:
When auto-DoS vector's threshold mode is set to stress-based mitigation, but the vector is in disabled state, TMM may halt and restart.

Conditions:
-- Threshold mode is set to stress-based mitigation.
-- Vector is disabled.

Impact:
TMM restarts. Traffic disrupted while TMM restarts.

Workaround:
There is no workaround other than not setting threshold mode to stress-based mitigation if the vector is disabled.

Fix:
TMM no longer restarts when threshold mode is set to stress-based mitigation and the vector is in disabled state.


712637-2 : Host header persistence not implemented

Component: Local Traffic Manager

Symptoms:
Online documentation describes a persistence mechanism which uses the HTTP Host: header. However, this mechanism is not implemented, and attempts to use it generates an error.

Conditions:
Online help for ltm persistence says that Host persistence can be enabled via iRule, but attempting to do so generates an error 'invalid argument host'.

Impact:
Although this does not impact any existing functionality, the documented function is not available.

Workaround:
There is no workaround at this time.

Fix:
LTM Host: header persistence is implemented.


712475-3 : DNS zones without servers will prevent DNS Express reading zone data

Solution Article: K56479945

Component: Local Traffic Manager

Symptoms:
DNS Express does not return dig requests.

Conditions:
DNS Express is configured a zone without a server.

Impact:
DNS Express does not return dig requests.

Workaround:
You can use either of the following workarounds:
-- Remove the zone without the server.
-- Configure a server for the zone.

Fix:
DNS zones without servers no longer prevent DNS Express reading zone data.


712437-3 : Records containing hyphens (-) will prevent child zone from loading correctly

Solution Article: K20355559

Component: Local Traffic Manager

Symptoms:
The dig command returns no record with SOA from the parent zone even though dnsxdump shows that the record exists.

Conditions:
-- Parent zone has a record containing - (a hyphen) and the preceding characters match the child zone. For example,
 myzone.com -- parent
 foo.myzone.com -- child
 
-- In myzone.com, you have a record of the following form:
foo-record.myzone.com

Impact:
DNS can not resolve records correctly.

Workaround:
None.

Fix:
DNS now handles the case in which the parent zone has a record containing - (a hyphen) and the preceding characters match the child zone.


712429 : Serverside packets excluded from DoS stats

Component: Advanced Firewall Manager

Symptoms:
BIG-IP systems configured with L4 DoS Protection might not provide sufficiently granular DDoS detection and mitigation to ensure that legitimate traffic is not impacted.

Conditions:
Configured for DDoS detection and mitigation.

Impact:
Legitimate traffic might be impacted.

Workaround:
None.

Fix:
The following DoS vectors no longer count serverside packets.

-- Single-Endpoint Flood
-- Global-Device level aggregate vectors
-- Bad-actor/attacked-dst for all vectors

Additionally, hardware-accelerated, device-level (global) aggregate DoS vectors are now programmed dynamically when traffic is detected, rather than at configuration time.

These behavior changes provide greater granularity in DDoS detection and mitigation to ensure that legitimate traffic is not impacted.

Behavior Change:
The following DoS vectors no longer count serverside packets.

-- Single-Endpoint Flood
-- Global-Device level aggregate vectors
-- Bad-actor/attacked-dst for all vectors

Additionally, hardware-accelerated, device-level (global) aggregate DoS vectors are now programmed dynamically when traffic is detected, rather than at configuration time.

These behavior changes provide greater granularity in DDoS detection and mitigation to ensure that legitimate traffic is not impacted.


712401-1 : Enhanced administrator lock/unlock for Common Criteria compliance

Component: TMOS

Symptoms:
The Network Device and Firewall collaborative Protection Profiles v2.0 require certain behavior for locking and unlocking administrative-user accounts on the BIG-IP system. BIG-IP software needs to be enhanced for compliance with those requirements.

Conditions:
The ccmode script must be run to activate these enhancements. Also, see the Common Criteria Guidance document (published when the certificate is obtained) for more details.

Impact:
Without these enhancements activated, the BIG-IP system is not compliant with Common Criteria requirements.

Workaround:
Risk acceptance for Common Criteria non-compliance.

Fix:
To meet Common Criteria requirements, the BIG-IP system is enhanced in two areas:

1. The primary administrative user account (generally 'admin') can be locked out, as any other administrative-user account can be. However, it is never locked out when signing in from the serial console.

2. Locked out administrative-users are unlocked only after an administrator-specified time period has passed. The default is 10 minutes, and is set in the ccmode script.


712362-3 : ASM stalls WebSocket frames after legitimate websockets handshake with 101 status code, but without 'Switching Protocols' reason phrase

Component: Application Security Manager

Symptoms:
When the WebSocket HTTP handshake response comes without 'Switching Protocols' reason phrase at the first line, the ASM does not follow up WebSocket frames on the WebSocket's connection.

The system posts the following messages in /ts/log/bd.log:
-- IO_PLUGIN|ERR |Mar 28 09:16:15.121|30539|websocket.c:0269|101 Switching Protocols HTTP status arrived, but the websocket hanshake failed.
-- IO_PLUGIN|ERR |Mar 28 09:16:15.121|30539|websocket.c:0270|Possible reasons are websocket profile isn't assigned on a virtual server or handshake is illegal.

Conditions:
-- ASM provisioned.
-- ASM policy and WebSocket profile attached to a virtual server.
-- WebSocket backend server sends 101 response without the 'Switching Protocols' phrase.

Impact:
WebSocket frames stalls.

Workaround:
#1 Change the backend server:
Change WebSocket backend server to return 101 response to include the 'Switching Protocols' reason phrase:

HTTP/1.1 101 Switching Protocols


#2 Use an irRule:
when SERVER_CONNECTED {
    TCP::collect 15
}
when SERVER_DATA {
    if { [TCP::payload 15] contains "HTTP/1.1 101 \r\n" } {
        TCP::payload replace 0 12 "HTTP/1.1 101 Switching Protocols"
    }
}

Fix:
This release correctly handles 101 responses even without the 'Switching Protocols' reason phrase.


712266-1 : Decompression of large buffer might fail with comp_code=11 with Nitrox 3 hardware

Component: TMOS

Symptoms:
Messages like the following may show up in /var/log/ltm:

-- crit tmm5[28908]: 01010025:2: Device error: n3-compress0 Zip engine ctx eviction (comp_code=11): ctx dropped.

This occurs because the decompression of large compressed data failed.

Conditions:
This issue occurs when compressed data cannot be decompressed in a single request to the Nitrox 3 hardware accelerator.

Impact:
Requests fail with a connection reset.

Workaround:
Use zlib software decompression.

Fix:
This release fixes this decompression issue in the Nitrox 3 driver.


712118 : AVR should report on all 'global tags' in external logs

Component: Application Visibility and Reporting

Symptoms:
AVR reports only 'ssgName' from the global tags.

Conditions:
-- A BIG-IQ operation configures the 'tag file' (/var/config/rest/downloads/app_mapping.json) on the BIG-IP system.
-- Statistics are sent to the BIG-IQ system.

Impact:
Not all the tags are sent to the BIG-IQ system.

Workaround:
There is no workaround at this time.

Fix:
AVR now reports statistics on all tags to the BIG-IQ system.


712102-2 : customizing or changing the HTTP Profile's IPv6 field hides the field or the row

Solution Article: K11430165

Component: TMOS

Symptoms:
Customizing or changing the HTTP Profile's IPv6 field hides the field or the row.

Conditions:
Customizing or changing the HTTP Profile's IPv6 field.

Impact:
The HTTP Profile's IPv6 field cannot be customized or changed.

Workaround:
Use tmsh to set the HTTP Profile's IPv6 field.

Fix:
Customizing or changing the HTTP Profile's IPv6 field doesn't hide the field or the row.


711929-1 : AVR should not send statistic on hidden rows for module InterfaceTraffic and module InterfaceHealth

Component: Application Visibility and Reporting

Symptoms:
AVR sends data on all interfaces, hidden and not hidden. It should send information only on not-hidden interfaces.

Conditions:
-- Tmstat table interface_stat exists.
-- Viewing statistics for module InterfaceTraffic and module InterfaceHealth.

Impact:
Irrelevant data is sent.

Workaround:
None.

Fix:
AVR now sends data only on not-hidden interfaces.


711683-2 : bcm56xxd crash with empty trunk in QinQ VLAN

Component: TMOS

Symptoms:
When an empty trunk (no members) is configured 'tagged' in a QinQ VLAN, bcm56xxd will continuously crash.

Conditions:
Trunk with no members, configured as 'tagged' in a QinQ VLAN.

Impact:
bcm56xxd continuously crashes.

Workaround:
Use either of the following workarounds:
-- Add members to the trunk.

-- Remove the trunk from the QinQ VLAN.

Fix:
Do not program QinQ switch hardware if the trunk has no members.


711570-3 : PEM iRule subscriber policy name query using subscriber ID, may not return applied policies

Component: Policy Enforcement Manager

Symptoms:
PEM::subscriber config policy get <subscriber id> does not return policy names

Conditions:
PEM iRule using subscriber ID to get policy name.

Impact:
Subscriber policy names are not returned.

Workaround:
Use PEM::subscriber config policy get <IP address> instead.

Fix:
Subscriber policy names are now returned when running PEM iRules using subscriber ID to get policy names.


711427-2 : Edge Browser does not launch F5 VPN App

Component: Access Policy Manager

Symptoms:
On Microsoft Windows v10, use Edge Browser to establish VPN. Edge Browser does not launch F5 VPN App.

Conditions:
On Windows 10, use Edge Browser to establish VPN.

Impact:
APM end user cannot establish VPN tunnel using Edge Browser.

Workaround:
Use Mozilla Firefox or Google Chrome.

Fix:
You can now use Windows 10 to launch Edge Browser to establish VPN connections.


711405-1 : ASM GUI Fails to Display Policy List After Upgrade

Solution Article: K14770331

Component: Application Security Manager

Symptoms:
After upgrade to version 13.1.0 or later, ASM GUI fails to display the Policy List.

Conditions:
Using an 11.5.0 to 11.5.4-HF1 version that is affected by the issue of policies with duplicate REST IDs that occurred after restoring from policy history or using Policy Diff to compare policies.

Impact:
ASM GUI starting using the REST API from version 13.1.0, so any systems that have been affected by this issue on a previous version now cause the GUI to fail to display the Policy List.

Workaround:
On the device run the following script:
---------------------------------
perl -MF5::DbUtils -MF5::Utils::Rest -MF5::ASMConfig::Entity::Policy -e '$dbh = F5::DbUtils::get_dbh();
 $dbh->begin_work();
 $dbh->do("UPDATE PLC.PL_POLICIES a SET a.rest_uuid = \"\" WHERE a.rest_uuid IN ( SELECT rest_uuid FROM (SELECT rest_uuid FROM PLC.PL_POLICIES GROUP BY rest_uuid having count(*) > 1) b)");
 F5::Utils::Rest::populate_uuids(dbh => $dbh);
 $dbh->commit();'
---------------------------------
Note: In a management environment, this change must be pushed to the peer as a full sync. The easiest way to do so is to ensure that the ASM sync device group is in manual sync mode, and to push the config to the peers from the device where the script was run.

Fix:
This data inconsistency is now repaired on upgrade, and the GUI loads the policy list successfully.


711281-5 : nitrox_diag may run out of space on /shared

Component: Local Traffic Manager

Symptoms:
Running nitrox_diag may lose collected data if there is insufficient free space for the tar file to be created.

Conditions:
-- Running nitrox_diag.
-- Insufficient free space available on /shared.

Impact:
Might lose data required to diagnose problems with Cavium Nitrox chips.

Workaround:
The only workaround is to ensure there is enough free space for the files to be created.

In general, planning enough space for two copies of a tmm core file and two copies of a qkview works. That might require approximately one gigabyte. Though more might be needed for systems with a large amount of RAM.

Fix:
nitrox_diag now clears the older data before gathering new data, instead of after. Note, however, that if there is insufficient free space on /shared to collect the raw data, the operation still cannot succeed.


711249-1 : NAS-IP-Address added to RADIUS packet unexpectedly

Component: TMOS

Symptoms:
RADIUS authentication request contains NAS-IP-Address attribute with value 127.0.0.1.

Conditions:
VIPRION (single or multiblade) with no cluster member IP addresses configured.

Impact:
Authentication does not work as it has invalid/unrecognized source NAS-IP-Address.

Workaround:
Configure cluster member IP address or define hostname and address via global-settings remote-host.


711093-1 : PEM sessions may stay in marked_for_delete state with Gx reporting and Gy enabled

Component: Policy Enforcement Manager

Symptoms:
PEM sessions may stay in the marked-for-delete state after session deletion.

Conditions:
This occurs if Gx final usage reporting response comes before Gy delete response (CCA).

Impact:
PEM sessions remain in marked-for-delete state.

Workaround:
None.

Fix:
PEM sessions no longer stay in the marked-for-delete state after session delete


711011-2 : 'API Security' security policy template changes

Component: Application Security Manager

Symptoms:
Parameter-related Learn/Alarm/Block settings in 'API Security' security policy template should be 'ON' by default.

Conditions:
Learn/Alarm/Block settings in 'API Security' security policy template.

Impact:
Settings not active.

Workaround:
None.

Fix:
Parameter-related Learn/Alarm/Block settings in 'API Security' security policy template are now 'ON' by default.


710996-2 : VIPRION outgoing management IPv6 traffic from primary blade uses the cluster member IP

Component: Local Traffic Manager

Symptoms:
The behavior of outgoing IPv6 management and IPv4 management traffic from the primary blade differs:
IPv4 traffic is sourced from the cluster IP
IPv6 traffic is sourced from the cluster member IP

Conditions:
IPv6 configured on the 'cluster' address and 'cluster member' address.

Impact:
The blade IP address, rather than the cluster floating IP, will be used as the source IP when querying the RADIUS server for remote-auth login against the management port.

Workaround:
There is no workaround at this time.


710976-1 : Network Map might take a long time to load

Component: TMOS

Symptoms:
Network Map loading time is directly proportional to the number of profiles/policies associated with a virtual server. When there are a significant number, Network Map might take many tens of seconds to load; when there are fewer associated profiles/policies, loading time is shorter.

For example, consider a configuration that contains virtual servers with a large number of associated profiles/policies.

In the following example config, the virtual server has two associated profiles. The time to load the Network Map page increases with each additional profile/policy.

# tmsh list ltm virtual vs
ltm virtual vs {
    creation-time 2018-03-06:18:27:53
    destination 0.0.0.0:any
    ip-protocol tcp
    last-modified-time 2018-03-06:18:27:53
    mask any
    profiles {
        myhttp { }
        tcp { }
    }
    rules {
        myrule
    }
    source 0.0.0.0/0
    translate-address disabled
    translate-port disabled
    vlans {
        socks-tunnel
    }
    vs-index 5
}

Conditions:
-- Navigate to Local Traffic :: Network Map in the BIG-IP Configuration Utility.

-- The configuration contains a significantly large number of profiles/policies associated with the virtual server.

Impact:
It takes tens of seconds to load the Network Map page.

Workaround:
None.

Fix:
The data loading performance was improved to load the page faster.


710947-1 : AVR does not send errdef for entity DosIpLogReporting.

Component: Application Visibility and Reporting

Symptoms:
AVR does not send errdef for entity DosIpLogReporting.

Conditions:
-- AVR is configured.
-- View the DosIpLogReporting report.

Impact:
There is no errdef for module DosIpLogReporting

Workaround:
None.

Fix:
Added errdef for module DosIpLogReporting.


710884-1 : Portal Access might omit some valid cookies when rewriting HTTP request.

Component: Access Policy Manager

Symptoms:
Portal Access is not sending certain cookies to the backend application.

Conditions:
-- Request path and cookie path are equal.
-- The last character of request path is not a forward slash (/).

Impact:
Backend application does not receive some cookies and may decide that APM end users are not authenticated, when they are.

Workaround:
There is no workaround at this time.

Fix:
Fixed an issue in Portal Access which could cause web-applications to lose some valid cookies.


710870 : Temporary browser challenge failure after installing older ASU

Component: Application Security Manager

Symptoms:
After installing an older ASM Signature Update (ASU) may cause the browser challenge to fail for the first few minutes after provisioning ASM.

Conditions:
-- Using BIG-IP version 13.1.0.5.
-- Installing an ASU from before April 2018.

Impact:
Browsers remain on whitepage after receiving a browser challenge.

Note: The problem should go away after 10-to-15 minutes of provisioning ASM, when more versions of JavaScript are generated.

Workaround:
Install the latest ASU.

Fix:
The browser challenges will succeed even after installing an older ASU.


710827-2 : TMUI dashboard daemon stability issue

Component: TMOS

Symptoms:
Some dashboard requests may cause a crash of TMUI dashboard daemons, affecting the TMUI dashboard.

Conditions:
Request sent to BIG-IP dashboard.

Impact:
Only the TMUI dashboard goes offline. Other TMUI functionality is not affected by this issue.

Workaround:
None available.

Fix:
Setup a correct exception handling prevented TMUI dashboard service failure.


710755-1 : Crash when cached route information becomes stale and the system accesses the information from it.

Solution Article: K30572159

Component: Advanced Firewall Manager

Symptoms:
The crash happens intermittently when the cached route information becomes stale and the system accesses the information from it.

Conditions:
Use stale cached route information.

Impact:
This condition can lead to a crash as the route information is no longer valid, and can lead to illegal memory access.

Workaround:
None.

Fix:
The system now fetches the latest egress route/interface information before accessing it.


710705-2 : Multiple Wireshark vulnerabilities

Solution Article: K34035645


710701-1 : "Application Layer Encryption" option is not saved in DataSafe GUI

Component: Fraud Protection Services

Symptoms:
"Application Layer Encryption" checkbox will remain enabled if un-checked via DataSafe GUI and will not be saved.

Conditions:
- Install DataSafe license
- Provision FPS
- Create URL

Impact:
Cannot enable/disable "Application Layer Encryption" via DataSafe GUI.

Workaround:
Application Layer Encryption can be enabled or disabled via TMSH command line or REST API.

Fix:
"Application Layer Encryption" option is saved if changed via DataSafe GUI.


710424-2 : Possible SIGSEGV in GTMD when GTM persistence is enabled.

Solution Article: K00874337

Component: Global Traffic Manager (DNS)

Symptoms:
When GTM persistence is enabled, GTMD may occasionally crash and restart.

Conditions:
GTM persistence is enabled.

Impact:
GTMD may occasionally restart.

Workaround:
Disable GTM persistence.

Fix:
GTMD will no longer crash and restart when persistence is enabled.


710327-1 : Remote logger message is truncated at NULL character.

Component: Application Security Manager

Symptoms:
When a request including binary null character has been sent to an remote logger, configured for ASM, the request will arrive to the remote destination and will be truncated exactly at binary NULL character.

Conditions:
-- ASM provisioned.
-- ASM policy attached to a virtual server.
-- ASM remote logger attached to a virtual server.
-- Request including binary NULL character is sent to the remote logger destination.

Impact:
Partial request is logged at the remote logger destination.

Workaround:
None.

Fix:
Binary NULL character is now escaped before being sent to the remote logger destination, so the request is no longer truncated.


710315-1 : AVR-profile might cause issues when loading a configuration or when using config sync

Component: Application Visibility and Reporting

Symptoms:
Some fields in AVR-profile contain lists of items. Those lists can be set only if the relevant flag is set to 'true'. In case of a flag configuration change, the system must keep the lists as they were and not reset them, so they can be available in case the flag changes back again.

Validation settings were created such that the lists flag is set to 'true' by default, but this can cause the load/merge process to break if the list was set, and afterwards the flag was set to 'false'.

Conditions:
Setting the relevant flag to 'false' after creating a list of items.

The relevant fields in AVR-profile that have that logic are:
-- IPs-list.
-- Subnets-list.
-- Countries-list.
-- URLs-list.

Impact:
Management load and sync process may not work as expected.

Workaround:
None.

Fix:
Validation for those fields when the associated flag is set to 'false' will be skipped in a load/merge scenario.


710314-1 : TMM may crash while processing HTML traffic

Solution Article: K94105051


710305-1 : When ASM and APM are configured on the same virtual server, ASM could throw some errors in its logging.

Component: Access Policy Manager

Symptoms:
When ASM and APM WebSSO are on same virtual server, WebSSO might generate a new request. When that happens, ASM might see multiple requests with same support ID. This can cause issues with ASM and log errors.

Conditions:
When APM WebSSO is configured (only for Basic, NTLM, Kerberos).

Impact:
ASM stops processing the HTTP requests that have duplicate support IDs, causing an issue to ASM/APM end users.

Workaround:
None.

Fix:
When ASM and APM WebSSO are on same virtual server, WebSSO no longer generates a new request, so duplicate support IDs are no longer created.


710244-3 : Memory Leak of access policy execution objects

Solution Article: K27391542


710232-2 : platform-migrate fails when LACP trunks are in use

Component: TMOS

Symptoms:
Use of the platform-migrate option might fail due to some configurations being invalid on the new system. The error message generated appears as follows:

01070687:3: Link Aggregation Control Protocol (LACP) is not supported on this platform.

Conditions:
You have a trunk with LACP enabled.
-- You are migrating from a hardware platform to a Virtual Edition (VE).

Impact:
Configuration fails to migrate.

Workaround:
After taking the old system offline, but before saving the UCS, disable LACP on all trunks.


710221-2 : Ephemeral node stays Unavailable on 'standby' BIG-IP system after a forced-offline FQDN node is re-enabled

Solution Article: K67352313

Component: Local Traffic Manager

Symptoms:
In an high availability (HA) configuration where an FQDN template-node on the 'active' unit is 'Force Offline', and the configuration is synchronized to the 'standby' unit, upon 'Enable' on that FQDN template-node in the 'active' unit (and the configuration is synchronized to the 'standby' unit), the ephemeral nodes associated with that FQDN template node remain Unavailable on the 'standby' unit.

Conditions:
HA configuration is configured with 'active' sync.
-- FQDN template-node is configured.
-- After ephemeral nodes are refreshed on the 'active' unit, it transitions to the 'standby' unit.
-- The FQDN template node on the (new-'active') unit is 'Force Offline' by the user.
-- The user explicitly 'Enable' that FQDN template node on the 'active' unit.

Impact:
The ephemerals on the 'standby' unit remain unavailable, even though the associated FQDN template node is 'available'. However, as no traffic is routed to the 'standby' unit, traffic is unaffected. Upon transition from 'standby' to' active', the ephemeral nodes will be refreshed and traffic should continue.

Workaround:
You can use either of these configuration options to prevent this issue from occurring:
-- Select 'Disabled' (rather than 'Force Offline') for the FQDN template node on the 'active' unit.
-- Configure the node without using FQDN.

Fix:
Upon 'Force Offline' and then 'Enable' for an FQDN template node on an 'active' unit in an HA configuration, the ephemerals on the 'standby' unit will reflect the re-enabled status based on the associated FQDN template-node on the 'active' unit.


710148-2 : CVE-2017-1000111 & CVE-2017-1000112

Solution Article: K60250153


710140-1 : TMM may consume excessive resources when processing SSL Intercept traffic

Solution Article: K20134942


710116-1 : VPN clients experience packet loss/disconnection

Component: Access Policy Manager

Symptoms:
VPN clients experience packet loss/disconnection.

Conditions:
In certain scenarios, the tunnel establishment procedure might leak a small memory. If the tmm is running for a longer duration, this small leak can accumulate and result in out-of-memory condition

Impact:
Connections start to drop as tmm runs out of memory. TMM will eventually run out of memory and connections could be terminated. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
A rare memory leak during APM VPN establishment has been corrected.


710110-1 : AVR does not publish DNS statistics to external log when usr-offbox is enabled.

Component: Application Visibility and Reporting

Symptoms:
AVR does not send DNS statistics to external logs when analytics global setting usr-offbox is enabled, if the following security analytics settings are set to disable:
-- collected-stats-internal-logging.
-- collected-stats-external-logging.

Conditions:
-- Security analytics settings collected-stats-internal-logging is disabled.
-- Security analytics settings collected-stats-external-logging is disabled.
-- Analytics global settings usr-offbox is enabled.

Impact:
DNS statistic are not sent to external log.

Workaround:
To work around this issue, perform the following procedure:
1. Provision ASM or AFM.
2. Run the tmsh command to set to enabled the security analytics setting collected-stats-external-logging.
2. Deprovision ASM/AFM.

Fix:
AVR now publishes DNS statistics to external logs when usr-offbox is enabled, as expected.


710032-1 : 'No Access' error when viewing GSLB Server's virtual server that has a name indicating a partition that does not exist on that BIG-IP system.

Component: Global Traffic Manager (DNS)

Symptoms:
The BIG-IP system posts a 'No Access' error message instead of the virtual server's properties.

Conditions:
-- At least two BIG-IP systems in a sync group.
-- One of the BIG-IP systems has a partition that does not exist on the other with an LTM virtual server on that partition.
-- The issue occurs when a GSLB Server discovers that LTM virtual server and displays it on its virtual server page.

Note: This same error message displays for GSLB pool member properties accessed by navigating to GSLB :: pools :: [pool] :: members :: Member : Address. The differing issues have been fixed in differing releases. You can find information about the similar issue in the AskF5 Bug Tracker article for 722734.

Impact:
It makes the GSLB Server's virtual server's properties page unavailable in this case.

Workaround:
Use either of the following workaroudns:
-- Use TMSH to view or edit the properties of that virtual server.

-- Create partitions on the GTM device to match those appearing to be referenced in the object names.

Fix:
Partition checking has been disabled for virtual servers on the GTM side, since a virtual server owned by a server is always in the partition of that server (/Common).


709972-6 : CVE-2017-12613: APR Vulnerability

Solution Article: K52319810


709936 : Multiple NTP server/domain-searches in the DHCP lease can cause missing DNS/NTP configuration.

Component: TMOS

Symptoms:
/var/log/boot.log has this error:
dhcp_config: save error: 01070911:3: The requested host (ntp-servers-ip) is invalid for servers in ntp (/Common/ntp).

Conditions:
- The BIG-IP system has DHCP enabled for management interface.
- Multiple NTP servers or domain-search are specified in the dhclient lease (stored at /var/lib/dhclient/dhclient.leases).

Impact:
- DNS and NTP servers are not configured in MCPD.
- Name resolution can break.

Workaround:
None.

Fix:
Multiple NTP server/domain-searches in the DHCP lease can cause missing DNS/NTP configuration.


709828-2 : fasthttp can crash with Large Receive Offload enabled

Component: Local Traffic Manager

Symptoms:
fasthttp and lro can lead to a tmm crash.

Conditions:
fasthttp and lro enabled. (lro is enabled by default >= 13.1.0)

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Do not use fasthttp

Fix:
fasthttp with lro enabled no longer causes tmm to crash.


709688-3 : dhcp Security Advisory - CVE-2017-3144, CVE-2018-5732, CVE-2018-5733

Solution Article: K08306700


709670-2 : iRule triggered from RADIUS occasionally fails to create subscribers.

Solution Article: K44067891

Component: Policy Enforcement Manager

Symptoms:
The subscriber can not be added with the same IP address after retries. This is a rarely occurring issue: if the daily average of subscriber additions is ~30000 subscribers, 5 to 6 subscribers face this problem (0.02%).

Conditions:
Running iRule to create PEM sessions from RADIUS packets using the 'PEM::subscriber create' command.

Impact:
Intermittently (a few operations per day), the command silently fails. No session is created, and no TCL error is produced. Subscriber addition requires manual intervention.

Workaround:
Trigger a RADIUS START using a different IP address for the subscriber.


709610-3 : Subscriber session creation via PEM may fail due to a RADIUS-message-triggered race condition in PEM

Component: Policy Enforcement Manager

Symptoms:
Session replacement fails as a result of a race condition between multiple RADIUS Accounting Start and Stop messages for the same subscriber.

Conditions:
This occurs when the following conditions are met:
-- These SysDB variables are set as follows:
sys db tmm.pem.session.radius.retransmit.timeout {
    value "0"
}
sys db tmm.pem.session.provisioning.continuous {
    value "disable"
}

-- Actions occur in the following order:
 1. PEM receives RADIUS START with subscriber ID1 and IP1.
 2. PEM receives RADIUS STOP with subscriber ID1 and IP1.
 3. PEM receives RADIUS START with subscriber ID1 and IP2.
 4. PEM receives RADIUS STOP with subscriber ID1 and IP2.

-- The time interval between steps 1 and 2 is very small (less than ~1ms).

Impact:
Subscriber session creation via PEM may fail.

Workaround:
There is no workaround other than to leave SysDbs variables set to the default values.

Fix:
The system now handles PEM subscriber session updates properly, so this issue no longer occurs.


709444-2 : "NTP not configured on device" warning seen when NTP symmetric key authentication is configured

Component: TMOS

Symptoms:
When NTP symmetric key authentication is configured, as per https://support.f5.com/csp/article/K14120, and the BIG-IP is part of a device service cluster, then a warning similar to this will be seen periodically in /var/log/ltm:

warning mcpd[6317]: 01071af0:4: NTP not configured on device 1.2.3.4 - within a trust

Conditions:
- NTP symmetric key authentication must be configured.
- The BIG-IP must be part of a device service cluster.
- Non-authenticated NTP servers must not be in-use.

Impact:
Incorrect NTP warnings are seen periodically in /var/log/ltm.

Workaround:
There is no workaround at this time.

Fix:
Prevented this warning from being emitted when NTP symmetric key authentication is in-use in a device service cluster.


709383-2 : DIAMETER::persist reset non-functional

Component: Service Provider

Symptoms:
When the iRule DIAMETER::persist reset is called, it is meant to remove existing persistence records from diadb. Without this bug fix, calling DIAMETER::persist reset has no effect and the persistence record remains.

Conditions:
Diameter session profile has persistence enabled and an iRule attempts to remove a persistence record.

Impact:
not provided by ENE

Workaround:
none

Fix:
DIAMETER::persist reset now functions properly. You can delete diameter persistence records with this iRule.


709334-1 : Memory leak when SSL Forward proxy is used and ssl re-negotiates

Component: Local Traffic Manager

Symptoms:
When looking at tmsh show sys memory you will see
ssl_compat continue to grow and fails to release memory.

Conditions:
-SSL Forward Proxy In use
-SSL Re-negotiations happening

Impact:
Eventually memory reaper will kick in.

Workaround:
There is no workaround at this time.

Fix:
ssl_compat now properly releases connections on re-negotiation.


709319-2 : Post-login client-side alerts are missing username in bigIQ

Component: Fraud Protection Services

Symptoms:
A client-side alert that contains a FPS-Username header with a value, but an empty fpm_username parameter - will be reported with "Unknown" username in bigIQ.

Conditions:
1. post login (alert is sent after submitting username parameter) client side alerts
2. alert-pool points to bigIQ IP (not Alert-Server)

Impact:
Post login client side alerts are missing username (will show as "Unknown" in bigIQ, works well with Alert-Server).

Workaround:
Route all client-side alerts to another virtual server and strip of the empty fpm_username parameter from payload/query-string.

Fix:
FPS will always send username in the fpm_username parameter in case it was empty and FPS has username value.


709274-1 : RADIUS Accounting requests egress on different Self IP addresses than they did pre-v13.1.0

Component: Access Policy Manager

Symptoms:
RADIUS Accounting requests egress different self IP addresses.
* START accounting message egresses floating self IP addresses.
* STOP accounting message egresses local self IP addresses.

Some RADIUS messages will come from floating IP addresses, some from self IP addresses. The RADIUS server should be configured to accept all self- and floating-IP addresses of all the devices in the high availability (HA) group, to ensure all messages are received.

Conditions:
RADIUS server configured with pool option.

Impact:
Causes RADIUS server to be unable to reconcile accounting messages.

Workaround:
You can reconcile accounting messages by tracking them through the Acct-Session-Id in RADIUS AVP's message, which is the same for the corresponding START and STOP messages to uniquely identify the session.

Fix:
RADIUS START and STOP messages now egress the same interface.


709256-2 : CVE-2017-9074: Local Linux Kernel Vulnerability

Solution Article: K61223103


709192-1 : GUI/iControl SOAP - KeyManagement export_all_to_archive_file failing immediately after HTTPD restart

Component: TMOS

Symptoms:
The export operation of SSL keys and certificate into a .tgz archive file fails when the httpd daemon is restarted.

Conditions:
-- Export SSL Keys and certificates into archive file.
-- httpd is restarted.

Impact:
Export of SSL Keys and certificates into archive file does not work immediately after httpd restart via GUI/iControl.

Workaround:
After httpd restart, perform other operation like creating a key/cert before exporting archive file.

Fix:
The system allows exporting SSL keys and certificate to be archived into .tgz files after httpd restart.


709133-2 : When the BigDB variable tmm.ssl.loggingcreatedcerts is set, a double-free error might occur

Component: Local Traffic Manager

Symptoms:
When the BigDB variable tmm.ssl.loggingcreatedcerts is set, a double-free error can occur.

Conditions:
-- The BigDB variable tmm.ssl.loggingcreatedcerts is set.
-- Forward proxy is being used.
-- An error occurs when reading the certificate serial number.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround other than not setting the tmm.ssl.loggingcreatedcerts BigDB variable.

Fix:
Double-free removed.


709132-1 : When the BigDB variable tmm.ssl.loggingcreatedcerts is set a buffer overflow can occur

Component: Local Traffic Manager

Symptoms:
When the BigDB variable tmm.ssl.loggingcreatedcerts is set, a buffer overflow can occur.

Conditions:
-- The BigDB variable tmm.ssl.loggingcreatedcerts is set.
-- Forward proxy is being used.
-- A malformed certificate with a serial number length equal to 256 bytes is parsed during forging.

Impact:
A off-by-one error causes one byte to write off the end of an array.

Workaround:
There is no workaround other than to not set tmm.ssl.loggingcreatedcerts BigDB variable.

Fix:
Buffer no longer overflows.


708956-1 : During system boots up, error message displays: 'Dataplane INOPERABLE - only 1 HSBes found on this platform'

Solution Article: K51206433

Component: TMOS

Symptoms:
During system bootup, the system occasionally posts the following error message and the system does not come up:
 Dataplane INOPERABLE - only 1 HSBes found on this platform.

Conditions:
This occurs occasionally at system bootup.
-- Using the following platforms:
i4600, i4800, i10600, i10800, i2600, i2800, i7600, i7800, i5600, i5800, i15600, i15800, i12600, i12800, HERCULON i2800, HERCULON i5800, HERCULON i10800, i11600, i11800, i11800-DS, i5820-DF, i7820-DF.

Impact:
System does not come up.

Workaround:
Reboot system.

Because this condition only happens occasionally, rebooting typically corrects the issue.

Fix:
This release fixes a race condition that might have occurred while reading FPGA firmware loading status.


708888-1 : Some DNS truncated responses may not be processed by BIG-IP

Solution Article: K79814103

Component: Advanced Firewall Manager

Symptoms:
On 13.1.x DNS responses with truncated bit set are dropped when AFM DNS DoS is enabled.

Conditions:
-- AFM DNS DoS is enabled.
-- Using 13.1.x.

Impact:
Clients do not receive truncated DNS responses.

Workaround:
Disable DNS DoS protection by changing the dos.dnsport variable to another port for which there is no valid traffic. For instance:

tmsh modify sys db dos.dnsport value 54


708840 : 13.0.0-to-13.1.0 upgrade operation on VIPRION 2250 blades might fail if global whitelist is configured

Component: Advanced Firewall Manager

Symptoms:
Upgrading from 13.0.0 to 13.1.0 on VIPRION 2250 blades might fail if global whitelist is configured. After the upgrade, the system will stay offline.

Conditions:
-- Global whitelist configured.
-- Running on VIPRION 2250 blades.

Impact:
System fails to run normally.

Workaround:
Remove global whitelist before upgrading to 13.1.0, add it back after upgrading.

Fix:
This issue no longer occurs in fixed versions, so you can upgrade from 13.0.0 to a post-13.1.0 version of the software without encountering this issue.


708830-2 : Inbound or hairpin connections may get stuck consuming memory.

Component: Carrier-Grade NAT

Symptoms:
When inbound or hairpin connections require a remote Session DB lookup and the lookup request or response messages get lost, the connections can get stuck in an embryonic state. They will be stuck in this state until they timeout and expire. In this state UDP connections will queue inbound packets. If the client application continues to send packets, the connection may never expire. The queued packets will accumulate consuming memory. If the memory consumption becomes excessive, connections may be killed and “TCP: Memory pressure activated” and “Aggressive mode activated” messages will appear in the logs.

Conditions:
A LSN pool with inbound and/or hairpin connections enabled. Lost Session DB messages due to heavy load or hardware failure. Remote lookups are more likely when using PBA mode or NAPT mode with default DAG.

Impact:
Excessive memory consumption that leads to dropped connections.

Workaround:
There is no workaround at this time.

Fix:
When Session DB messages are lost, the connection will be killed and any queued packets will be discarded. If the client application resends packets they will be treated as a new connection.


708653-1 : TMM may crash while processing TCP traffic

Solution Article: K07550539


708484-2 : Network Map might take a long time to load

Component: TMOS

Symptoms:
Network Map loading time is directly proportional to the number of profiles/policies associated with a virtual server. When there are a significant number, Network Map might take many tens of seconds to load; when there are fewer associated profiles/policies, loading time is shorter.

For example, consider a configuration that contains virtual servers with a large number of associated profiles/policies.

In the following example config, the virtual server has two associated profiles. The time to load the Network Map page increases with each additional profile/policy.

# tmsh list ltm virtual vs
ltm virtual vs {
    creation-time 2018-03-06:18:27:53
    destination 0.0.0.0:any
    ip-protocol tcp
    last-modified-time 2018-03-06:18:27:53
    mask any
    profiles {
        myhttp { }
        tcp { }
    }
    rules {
        myrule
    }
    source 0.0.0.0/0
    translate-address disabled
    translate-port disabled
    vlans {
        socks-tunnel
    }
    vs-index 5
}

Conditions:
-- Navigate to Local Traffic :: Network Map in the BIG-IP Configuration Utility.

-- The configuration contains a significantly large number of profiles/policies associated with the virtual server.

Impact:
It takes tens of seconds to load the Network Map page.

Workaround:
None.


708389 : BADOS monitoring with Grafana requires admin privilege

Component: Anomaly Detection Services

Symptoms:
Current Grafana monitoring requires admin privilege.
Grafana stores its internal database in unencrypted format, so the admin password can be extracted from a compromised computer.

Conditions:
Monitoring using Grafana.

Impact:
Guest user cannot access data needed for Grafana.

Workaround:
None.

Fix:
There is now a REST call to pool the Grafana statistics. This allows any user (including guest), not just admin or root, to access data needed for Grafana.

Behavior Change:
This release introduces the following tmsh commands:
-- tmsh run util admdb - for help
   + list-element path_folder - lists folder
   + view-element path_file - view file contents
   + list-metrics path vs
   + table-query base_path db sRate tsfiles ts metric_columns_aliases

The path must be under /shared/admdb, for example:

-- run util admdb list-element /shared/admdb/default/_a_l_l

-- run util admdb view-element /shared/admdb/default/_a_l_l/info.sysinfo/1000/1522229248000.txt

-- run util admdb table-query /shared/admdb default 1000 '[1522233344000]' '[1522234774492,1522235074492]' '[["info.attack",["v0"],"Attack"],["sig.health",["v0"],"Health"],["info.learning",["v0"],"Learning"],["info.learning",["v2"],"Learned samples"]]'


708305-2 : Discover task may get stuck in CHECK_IS_ACTIVE step

Component: Device Management

Symptoms:
The discover tasks is running periodically after user creates the task. But it may get stuck in the middle steps and fail to run periodically.

Conditions:
When HA failover group is set up and a discover task is created on one of the devices.

Impact:
The discover task will periodically pull the OpenID information and update oauth jwt and jwk configurations in MCP. If the task sticks, the jwt and jwk configuration will not sync to the latest version and may cause access policy fail.

Workaround:
If the task is stuck in any step that is not SLEEP_AND_RUN_AGAIN for more than one minute, manually cancel and delete the task and create the same task again.

Fix:
Discover task no longer gets stuck in CHECK_IS_ACTIVE step.


708249-2 : nitrox_diag utility generates QKView files with 5 MB maximum file size limit

Component: Local Traffic Manager

Symptoms:
When nitrox_diag generates a QKView file, the utility does not use the -s0 flag for the qkview command. That means there is a 5 MB file-size limit for the resulting QKView file nitrox_diag generates.

Conditions:
Run the nitrox_diag command.

Impact:
QKView files generated in response to running the nitrox_diag command might not contain all necessary information, for example, the result might contain truncated log files.

Workaround:
After running nitrox_diag, run the following command to generate a complete QKView file: qkview -s0

Fix:
Nitrox_diag utility now uses the -s0 command to generate QKView files, so there is no longer a 5 MB maximum file size limit, and the full QKView file is created.


708189 : OAuth Discovery Auto Pilot is implemented

Component: Access Policy Manager

Symptoms:
This now adds a new capability to allow user to select a period to have OAuth auto discovery automatically pull down JWT keys.

Conditions:
Follow the new added UI and configure frequency to start.

Impact:
No impact, it has usability improvement over manual discovery.

Workaround:
There is no workaround.

Fix:
New auto pilot capability is added for usability.


708114-1 : TMM may crash when processing the handshake message relating to OCSP, after the SSL connection is closed

Solution Article: K33319853

Component: Local Traffic Manager

Symptoms:
TMM crashes when receiving the HUDEVT_SSL_OCSP_RESUME_CLNT_HS after the SSL connection is closed.

Conditions:
-- The SSL connection has been closed.
-- SSL receives the HUDEVT_SSL_OCSP_RESUME_CLNT_HS message.

Impact:
TMM crash. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
The system now ensures that SSL can still properly process the messages, even when the SSL connection is closed.


708054-1 : Web Acceleration: TMM may crash on very large HTML files with conditional comments

Component: TMOS

Symptoms:
Web Acceleration feature may not handle big HTML file with improper conditional comments inside in some cases. TMM may crash processing such files if Web Acceleration profile is attached to VIP.

Conditions:
- HTML file with conditional comments inside:
  <!--[if condition...]> ... <![endif]-->

- The size of "condition" from the pattern above is comparable with RAM size of BIG-IP box, i.e. ~8-10GB.

Impact:
TMM crash interrupts all active sessions.

Workaround:
There is no workaround at this time.

Fix:
Now Web Acceleration feature can handle long HTML conditional comments correctly.


708005-1 : Users cannot use Horizon View HTML5 client to launch Horizon 7.4 resources

Solution Article: K12423316

Component: Access Policy Manager

Symptoms:
When using VMware View HTML5 client, end users are able to authenticate and see available View resources on the APM webtop. However, any attempt to launch the resource (desktop or application) in HTML5 mode momentarily appears to function, but then redirects to the initial APM login page.

Conditions:
This occurs when the following conditions are met:
-- BIG-IP APM is protecting VMware Horizon View 7.4 resources.
-- End user tries to launch a View resource from the APM webtop using the Horizon View HTML5 client.

Impact:
End user cannot launch VMware View resources with View HTML5 client.

Workaround:
You can use the following workarounds:

-- If you are already running Horizon 7.4, use native View clients instead.

-- If you have not upgraded to Horizon 7.4, stay on an older Horizon release until this issue is resolved.

-- If you are running BIG-IP APM release 13.1.0, you can add the following iRule to the virtual server that handles HTML5 client connections:

when HTTP_REQUEST {
    if { ([info exists tmm_apm_view_uuid]) &&
         ([HTTP::method] == "GET") &&
         ([HTTP::uri] ends_with "/portal/webclient/sessiondata")} {
        HTTP::cookie remove "sessionDataServiceId"
    }
}

when HTTP_RESPONSE {
    if { ([info exists tmm_apm_view_uuid]) } {
        set cookieNames [HTTP::cookie names]
        foreach aCookie $cookieNames {
            set path [HTTP::cookie path $aCookie]
            if {[string length $path] > 0} {
                HTTP::cookie path $aCookie "/f5vdifwd/vmview/$tmm_apm_view_uuid$path"
            }
        }
    }
}

Important:
-- After applying the iRule and before attempting a connection, be sure to clear all cache and cookies from the client systems. Otherwise, the test operation may need to be executed before exhibiting successful behavior.
-- The iRule workaround is for BIG-IP APM release 13.1.0. It is not supported for older BIG-IP releases.

Fix:
Horizon View version 7.4 in HTML5 mode now functions correctly with APM.


707990-2 : Unexpected TMUI output in SSL Certificate Instance page

Component: TMOS

Symptoms:
Under certain conditions, the TMUI for SSL Certificate Instances does not produce the expected HTML output.

Conditions:
Authenticated TMUI user browsing the SSL Certificate Instance page.

Impact:
Unexpected HTML output.

Workaround:
None.

Fix:
SSL Certificate Instance now produces the expected HTML output.


707961-2 : Unable to add policy to virtual server; error = Failed to compile the combined policies

Solution Article: K50013510

Component: Local Traffic Manager

Symptoms:
LTM policy does not compile if it has more than one similar condition, and some of them refer to datagroup and others to non-datagroup types. The policy compiles if similar conditions all refer to datagroups or to non-datagroups.

010716d8:3: Compilation of ltm policies for virtual server /Common/vs_name failed; Failed to compile the combined policies.

Conditions:
-- LTM policy that has more than one similar condition.
-- Some conditions refer to datagroup types.
-- Some conditions refer to non-datagroup types.

Following is an example of such an LTM policy (i.e., referring to both datagroup and non-datagroup types):

ltm policy /Common/example_ltm_policy {
      published-copy /Common/block_URI
      requires { http }
      rules {
          example_Rule {
              conditions {
                  0 {
                      http-host
                      host
                      datagroup /Common/example_datagroup <------ Datagroup
                  }
                  1 {
                      http-host
                      host
                      values { example.com } <------ Non-Datagroup
                  }
              }
          }
      }
      strategy /Common/first-match
 }

Impact:
LTM policy does not compile. Cannot use the policy.

Workaround:
Create a datagroup for the non-datagroup type and use it in policy instead.

Fix:
LTM Policy compiles if it has more than one similar condition, and some of them refer to datagroup and others to non-datagroup types.


707951-2 : Stalled mirrored flows on HA next-active when OneConnect is used.

Component: Local Traffic Manager

Symptoms:
-- 'tmctl -d blade tmm/umem_usage_stat | grep xdata' may show increased memory usage.
-- 'tmsh show sys connect' shows idle flows.

Conditions:
- OneConnect profile is configured.
- High availability (HA) mirroring is enabled.

Impact:
- Some flows are mirrored incorrectly.
- Stalled flows may occupy a lot of memory.

Workaround:
Disable OneConnect.

Fix:
Stalled mirrored flows no longer appear when OneConnect is used.


707740-4 : Failure deleting GTM Monitors when used on mulitple Virtual Servers with the same ip:port combination

Component: TMOS

Symptoms:
User would get "monitor is in use" when attempting to delete a GTM Monitor, even after removing that monitor from all GTM Virtual Servers

Conditions:
Attach a gtm monitor to multiple gtm virtual servers in the same transaction, where both of the virtual servers are monitoring the same ip:port

Impact:
User will not be able to ever delete the un-used gtm monitor

Workaround:
Remove monitor from VSs
Reload GTM configuration with tmsh load sys config gtm-only
Delete monitor

Fix:
Fixed issue preventing users from deleting an unused GTM monitor, if that monitor was attached to two different GTM virtual servers of the same ip+port combination.


707676-1 : Memory leak in Machine Certificate Check agent of the apmd process

Component: Access Policy Manager

Symptoms:
The apmd process leaks a small amount of memory in Machine Certificate Check agent

Conditions:
- Machine Certificate Check agent is configured in an Access Policy
- inspected machine certificate is revoked by CRL

Impact:
The apmd process may grow in size. This may lead to high memory utilization and instability in BIG-IP.

Workaround:
There is no workaround

Fix:
A memory leak in the APM Machine Certificate check agent has been corrected.


707631-2 : The 'SYN Challenge Handling' setting of a TCP profile can revert to defaults when updating the profile using the GUI

Component: TMOS

Symptoms:
The 'SYN Challenge Handling' setting of a TCP profile can be reverted to defaults when a TCP profile is updated using the BIG-IP management GUI.

Conditions:
The SYN Challenge Handling settings of a TCP profile have previously been set to non-default values, and a configuration change is later made to the same TCP profile using the GUI.

Impact:
Loss of TCP profile syn challenge configuration settings

Workaround:
In the GUI, specifically set the SYN Challenge Handling fields after making an update to other TCP profile fields, or use tmsh to make the changes instead

SYN Challenge

GUI Setting: Nominal
TMSH:
    syn-cookie-enable enabled
    syn-cookie-whitelist disabled

GUI Setting: Challenge and Remember
TMSH:
    syn-cookie-enable enabled
    syn-cookie-whitelist enabled


GUI Setting: Disable Challenges:
    syn-cookie-enable disabled
    syn-cookie-whitelist disabled

Fix:
Now syn challenge handling setting isn't overwritten when tcp profile is updated


707585-1 : Use native driver for 82599 NICs instead of UNIC

Component: TMOS

Symptoms:
-- Lower performance in TPS.
-- Increased CPU consumption for throughput.

Conditions:
-- Running BIG-IP Virtual Edition (VE) (UNIC is the default Virtual Function Network Driver for VE).
-- High data rates, perhaps particularly when connection mirroring is configured.

Impact:
Lower TPS performance numbers and higher CPU usage in throughput tests.

Workaround:
In case of UNIC, extra CPUs must be allocated to handle SoftIRQs for higher performance.

Fix:
This release provides a native driver based on F5's physical platforms.


707509-1 : Initial vCMP guest creations can fail if certain hotfixes are used

Component: TMOS

Symptoms:
vCMP guest fails to enter the 'provisioned' or 'deployed' states and similar messages can be seen in /var/log/ltm:

-- vcmpd[14254]: 01510003:2: Guest (guest_name): Install failed.
-- vcmpd[14254]: 01510004:3: Guest (guest_name): Install to VDisk /shared/vmdisks/guest_name.img FAILED: Child exited with non-zero exit code: 255

Conditions:
Creating vCMP guest using certain hotfix images, such as BIG-IP HF software released as partial .iso software images, and engineering hotfixes.

Impact:
vCMP guest cannot be created.

Workaround:
1. Create and deploy the vCMP guest using the full .iso base software image.
2. Log in to the vCMP guest once it has finished starting up.
3. Apply the hotfix image or engineering hotfix.

Fix:
Guest creation succeeds.


707447-1 : Default SNI clientssl profile's sni_certsn_hash can be freed while in use by other profiles.

Component: Local Traffic Manager

Symptoms:
SSL SNI mechanism creates a hash containing a mapping between SAN entries in a given profile's certificate and the profile. This hash is owned by the default NI profile, however is held by the other profiles on the VIP without a reference. If a connection utilizes SNI to use a non-default SNI profile *and* the default SNI profile reinitializes its state for any reason, the prf->sni_certsn_hash can be cleared and freed, leaving the existing connection(s) with profiles that refer to the freed hash. If then the connection attempts a renegotiation that causes the hash to be used, the freed hash can cause a fault should it have been reused in the meantime (i.e. the contents are invalid).
       
The fix: When the default SNI profile is initializing, and SSL handshake is searching SAN/COMMON cert from
non-default SNI profile, stop the search and return NULL.

Conditions:
SNI configured with one default SNI profile, one or multiple SNI profiles. The default SNI profile is changed and renegotiation with SNI(with non-default SNI profile) is issued.

Impact:
Traffic is disrupted while TMM restarts.

Workaround:
There is no workaround at this time.

Fix:
When the default SNI profile is initializing, and SSL handshake is searching SAN/COMMON cert from non-default SNI profile, stop the search and return NULL.


707445-3 : Nitrox 3 compression hangs/unable to recover

Solution Article: K47025244

Component: TMOS

Symptoms:
LTM logs show the following message:

    Nitrox 3, Hang Detected: compression device was reset

When the error manifests, there will be three error messages sent to the log over a period of several seconds. The device is then considered unrecoverable and marked down, and will no longer accept compression requests.

Conditions:
This applies only to vCMP guests. Some compression requests can stall the device after a bad compression request is made.

Note: Traffic volume and concurrence, along with the type of error have to occur together in order to result in this issue, so the issue is not easily reproduced.

Impact:
Once the device is marked down, compression will be sent to the software compression provider, until tmm on the device is restarted. This can cause local CPU utilization to climb.

Workaround:
There is no complete workaround without a software fix. However, compression will always default to the software compression provider when hardware cannot be recovered.

There are three recovery options available if the TMM-internal reset fails to recover the compression device automatically. These should be employed in this order:

A. Restart tmm using the command: bigstart restart tmm.
B. Restart the vCMP guest.
C. Restart the host (which restarts all guests).

Note: Because of the traffic volume, timing, and error type that cause this condition, this error might recur. This issue appears to be caused by a particular compression request. So regardless of the recovery method you execute, the problem may recur in a short time, or months later.

Fix:
Compression device reset recovery made more robust for some compression failures.


707391-2 : BGP may keep announcing routes after disabling route health injection

Component: TMOS

Symptoms:
As a result of a known issue BIG-IP with BGP configured may continue to announce routes even after disabling the virtual address or disabling route announcement on the virtual address.

Conditions:
BGP configured with multiple routes announced via Virtual-address route announcements.
Configuration changes made on the BGP configuration itself.
Virtual address or its route health announcement disabled.

Impact:
Prefixes continue to exist in the BGP table even after disabling the virtual address (visible via imish with the "show ip bgp" command); which will continue to announce the prefix to configured peers.

Workaround:
Workaround would be to restart the dynamic routing process.

Fix:
BGP may no longer keeps announcing routes after disabling route health injection


707310-2 : DNSSEC Signed Zone Transfers of Large Zones Can Be Incomplete (missing NSEC3s and RRSIGs)

Component: Global Traffic Manager (DNS)

Symptoms:
It is possible that when performing a sufficiently large DNSSEC signed zone transfer (e.g., ~1 KB records) that the final packet (or several packets) of NSEC3s and RRSIGs could be missing from the zone. This issue can be detected using a free third-party tool such as ldns-verify-zone or some other DNSSEC validating tool.

Conditions:
-- Dynamic DNSSEC signing of zone transfers is configured for a given zone from an authoritative DNS Server (this could be on-box BIND, off-box BIND, etc.), or from DNS Express.

Impact:
The DNSSEC signed zone transfer could be incomplete, that is missing some NSEC3s and RRSIGs. An incomplete DNSSEC zone is considered an invalid zone.

Workaround:
There is no workaround at this time.

Fix:
DNSSEC signed zone transfers from general authoritative DNS Servers as well as from DNS Express are now complete regardless of the number of records in the zone.


707267 : REST Framework HTTP header limit size increased to 8 KB

Component: TMOS

Symptoms:
HTTP Header size limit causing 'Error getting auth token from login provider' GUI messages.

Conditions:
A client uses an HTTP Header larger than 4 KB to make a request to the REST framework.

Impact:
Users cannot login or access certain pages in the GUI.

Workaround:
Clear browser cookies, or otherwise reduce the size of the HTTP headers, so that the entire content HTTP headers is smaller than 4 KB.

Fix:
The HTTP header size limit for the REST Framework has been increased to 8 KB to match the limit set by Apache.


707246-1 : TMM would crash if SSL Client profile could not load cert-key-chain successfully

Component: Local Traffic Manager

Symptoms:
TMM would crash if SSL Client profile could not load cert-key-chain successfully, and SSL is working in the fwd-proxy-mode.

Conditions:
1. SSL is working in the fwd-proxy-mode.
2. SSL could not load the cert-key-chain in the clientssl profile successfully. There could be couple of reasons:

2.1.We fail to configure the password required by the cert-key-chain.
2.2.Configured cert-key-chain type is not supported.
2.3.cert-key-chain name is incorrect.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Configure the cert-key-chain in the clientssl profile correctly.

Fix:
If we fail to load the cert-key-chain in the clientssl profile, and ssl is working in the fwd-proxy-mode, we will mark the corresponding ssl clientssl profile as invalid, then we will not accept the incoming SSL handshake destined to this profile.


707244-3 : iRule command clientside and serverside may crash tmm

Component: Local Traffic Manager

Symptoms:
Using clientside and serverside command in iRules may crash tmm.

Conditions:
Using such HTTP commands as HTTP::password in clientside and serverside nesting script.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this point.

Fix:
Fix clientside and serverside command do not work with certain HTTP commands.


707226 : DB variables to disable CVE-2017-5754 Meltdown/PTI mitigations

Component: TMOS

Symptoms:
Mitigations might CVE-2017-5754 Meltdown/PTI (Page Table Isolation) can negatively impact performance.

Please see https://support.f5.com/csp/article/K91229003 for additional Spectre and Meltdown information.

Conditions:
Mitigations for CVE-2017-5754 Meltdown/PTI (Page Table Isolation) enabled.

Impact:
Meltdown/PTI mitigations may negatively impact performance.

Workaround:
Disable CVE-2017-5754 Meltdown/PTI mitigations.

To turn off mitigations for CVE-2017-5754 Meltdown/PTI, run the following command:

tmsh modify sys db kernel.pti value disable

Note: Turning off these mitigations renders the system vulnerable to CVE-2017-5754 Meltdown; but in order to take advantage of this vulnerability, they must already possess the ability to run arbitrary code on the system. Good access controls and keeping your system up-to-date with regards to security fixes will mitigate this risk on non-VCMP systems. vCMP systems with multiple tenants should leave these mitigations enabled.

Please see https://support.f5.com/csp/article/K91229003 for additional Spectre and Meltdown information.

Fix:
On releases that provide mitigations for CVE-2017-5754 Meltdown/PTI, the protection is enabled by default, but can be controlled using db variables.

Please see https://support.f5.com/csp/article/K91229003 for additional Spectre and Meltdown information.


707207-1 : iRuleLx returning undefined value may cause TMM restart

Component: Local Traffic Manager

Symptoms:
When an iRulesLX rule returns an undefined value, TMM may restart. An example of an undefined value is one where
its jsonrpc v2.0 representation is missing required fields, such as "result".

Conditions:
iRulesLX is licensed, and a rule is run that returns an undefined value.

Impact:
Traffic is interrupted.

Workaround:
There is no workaround at this time.

Fix:
A TMM restart resulting from an iRulesLX rule returning an undefined value was fixed.


707186-1 : TMM may crash while processing HTTP/2 traffic

Solution Article: K45320419


707147-1 : High CPU consumed by asm_config_server_rpc_handler_async.pl

Component: Application Security Manager

Symptoms:
During a period of heavy learning from Policy Builder, typically due to Collapse Common elements, the backlog of events can cause a process to consume high CPU.

Conditions:
1) Automatic Policy Builder is enabled
2) An element is configured to "Always" learn new entities, and collapse common elements
3) An extended period of heavy learning due to traffic is enountered

Impact:
A process may consume high CPU even after the high traffic period is finished.

Workaround:
Kill asm_config_server.pl (This will not affect traffic)

Optionally modify one of the following
A) Learn "Always" to another mode
B) Turn off Collapse common URLs
C) Change from Automatic Learning to Manual


707109-1 : Memory leak when using C3D

Component: Local Traffic Manager

Symptoms:
When using the Client Certificate Constrained Delegation Support (C3D) feature, memory can leak.

Conditions:
Traffic passes through a virtual server with C3D enabled.

Impact:
Memory is leaked.

Workaround:
There is no workaround.

Fix:
When using C3D memory no longer leaks.


707100 : Potentially fail to create user in AzureStack

Component: TMOS

Symptoms:
Using Microsoft Azure Stack on the BIG-IP Virtual Edition (VE) configured with password authentication, the system sometimes fails to create the provided user, which means that the VE admin cannot ssh in after deployment. Note that an unexpected reboot causes this to occur during provisioning, and that Microsoft might have already fixed this issue.

Conditions:
Azure Stack VE provisioned with password authentication.

Impact:
Admin loses provisioned VE instance because there is no way to ssh in.

Workaround:
Deploy VE with key authentication.

Fix:
Extra handling was added to make user creation work even with unexpected reboots happening during Azure Stack provisioning.


707054-1 : SYN Cookie MSS for Fast L4 Profiles is limited to 256-9162

Component: Advanced Firewall Manager

Symptoms:
User cannot configure SYN Cookie MSS values for Fast L4 Profiles outside of 256-9162.

Conditions:
Under syncookie activated condition, SYN Cookie MSS value is set to < 256.

Impact:
Under syncookie activated condition, server side MSS lower than SYN Cookie MSS lower limit cannot be honored on client side. As a result, server side MSS discards data segment if larger than SYN Cookie MSS lower limit.

Fix:
This ID allows to configured 128-9162.


707003-3 : Unexpected syntax error in TMSH AVR

Component: TMOS

Symptoms:
The following tmsh command does not work: tmsh show analytics http report view-by virtual measures { transactions } drilldown

It fails with the following error message: 'Syntax Error: "drilldown" property requires at least one of (device device-list) to be specified before using.'

Conditions:
Whenever the affected tmsh command is run.

Impact:
The following tmsh command will not run: tmsh show analytics http report view-by virtual measures { transactions } drilldown

Workaround:
There is no workaround besides not running the affected command.

Fix:
The following command now works as expected: tmsh show analytics http report view-by virtual measures { transactions } drilldown


706998-3 : Memory leak when OCSP is configured in clientSSL profile for C3D, or in serverSSL profile for server authentication

Component: TMOS

Symptoms:
There is a memory leak when OCSP is configured in the clientSSL profile for C3D feature, or in serverSSL profile for server certificate authentication.

Conditions:
OCSP is configured in the clientSSL profile for C3D feature, or in serverSSL profile for server certificate authentication.

Impact:
TMM will run out of memory.

Workaround:
There is no workaround at this time.

Fix:
The memory leak has been fixed.


706845-2 : False positive illegal multipart violation

Component: Application Security Manager

Symptoms:
A false positive multipart violation.

Conditions:
Uploading a file with a filename value that is encoded in non utf-8 encoding.

Impact:
A false positive violation, request rejected.

Workaround:
Might be workaround using an irule

Fix:
Corrected ASM multipart parsing.


706835 : When cloning a profile, URL parameters are not shown

Component: Fraud Protection Services

Symptoms:
In Fraud Protection Service GUI, cloning a profile and then navigating to a URL, its parameters are not shown.

Conditions:
Provision and license Fraud Protection Service.

Impact:
Fraud Protection Service GUI.

Workaround:
Navigating again from Profiles will show the parameters.

Fix:
Parameters are now shown on first attempt after cloning a profile.


706804-1 : SNMP trap destination configuration of network option is missing "default" keyword

Component: TMOS

Symptoms:
When SNMP trap destinations are configured, the user can specify the network that the traps are transmitted out from. By default, the routing table is consulted. Use the network keyword to overwrite this with either "management" or "other". There is also a "default" keyword, which was removed since it was confusing. However, this broke backward compatibility of the REST API; so, it was put back.

Conditions:
Including the "network default" keywords in trap configuration reports an error with version 13.0.0 where the "default" keyword was removed.

Impact:
Existing scripts may encounter errors if they used this keyword.

Workaround:
Don't use the "default" keyword with the snmp trap destination network configuration.

Fix:
The "default" keyword was put back.


706771-1 : FPS ajax-mapping property may be set even when it should be blocked

Component: Fraud Protection Services

Symptoms:
Ajax mapping may be set only when 1) ajax-encryption is enabled OR 2) ajax-integrity AND strong-integrity are enabled.

The bug allows to set ajax-mapping even for the following (invalid) configuration:

  ajax-encryption: disabled
  ajax-integrity: enabled
  strong-integrity: disabled

Conditions:
1)
  ajax-encryption: disabled
  ajax-integrity: enabled
  strong-integrity: disabled

2)
  non-empty ajax-mapping

Impact:
System will set the ajax-mapping field when it should have been blocked.

Workaround:
There is no workaround at this time.

Fix:
FPS should block ajax-mapping configuration when the pre-conditions weren't met.


706750-1 : Changing CGNAT SIP ALG profile log settings while handling traffic may cause tmm crash.

Component: Service Provider

Symptoms:
Altering the router profile log settings (log publisher and logging profile) may cause the tmm to crash when handling traffic.

Conditions:
-- CGNAT SIP ALG.
-- Changing log settings while handling traffic.

Impact:
TMM may crash. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Changing CGNAT SIP ALG profile log settings while handling traffic no longer causes tmm core.


706688 : Automatically add additional certificates to BIG-IP system in C2S and IC environments

Component: TMOS

Symptoms:
In order to function properly, the BIG-IP system's failover and autoscale features need extra certificates (which are used inside C2S environments) and endpoint URL. The extra certificates must be pointed to by the $AWS_CA_BUNDLE environment variable. The BIG-IP system's failover and autoscale feature use AWS CLI commands. Those AWS CLI commands need to have the $AWS_CA_BUNDLE pointing to the certificate files location and endpoint URL. These must be configured manually.

Conditions:
-- BIG-IP is running in AWS C2S (Commercial Cloud Services) where the domain is either c2s.ic.gov or sc2c.sgov.gov.

-- The BIG-IP system is configured to do failover or autoscale in those environments.

Impact:
Requires manual copying of the extra certificates, which are pointed to by AWS_CA_BUNDLE, to the BIG-IP system.

Workaround:
None.

Fix:
In this release, you can add all the needed certificates' URLs in the AWS user-data. The BIG-IP system then automatically downloads and stores the certificate and sets the $AWS_CA_BUNDLE environment variable and endpoint-url.

To use the functionality, when launching the BIG-IP system from the AWS web console, specify the C2S certificate URL and test URL in the following format:
 
c2s-keys-urls=url1,url2,url3;c2s-cert-test-url=ec2.us-iso-east-1.c2s.ic.gov:443;

Where the syntax explanation is as follows:
1. The string literal : c2s-keys-urls=
2. List of comma separated URLs.
3. A semicolon (;).
4. The string literal : c2s-cert-test-url=
5. A URL which is of following format
    <A service name (e.g., ec2)>.<The region where it is running (e.g., us-iso-east-1)>.<the domain name :443 (e.g., c2s.ic.gov:443)>
     
Example: ec2.us-iso-east-1.c2s.ic.gov:443;


706665-2 : ASM policy is modified after pabnagd restart

Component: Application Security Manager

Symptoms:
ASM policy modifications might occur after the the pabnagd daemon is restarted. Modifications include the following:

-- Length attributes might change from 'any' to a low auto learning value.
-- Check signature / metachars might change from unchecked to checked.

This applies for the following entity types:
filetypes, URLs, parameters, cookies, WS URLs, content profiles.

Conditions:
-- Configuration containing a policy in which automatic learning mode is configured.
-- Restart of pabnagd (the automated policy-building operations daemon).

Impact:
ASM policy is modified.

Workaround:
Switch policy builder to manual learning mode.

Fix:
Prevent unwanted adjust operations from being called on policy-catchup complete.


706651-1 : Cloning URL does not clone "Description" field

Component: Fraud Protection Services

Symptoms:
When cloning URL using the "Clone URL" feature in FPS/DataSafe GUI, description field is not cloned to new URL.

Conditions:
Provision and license FPS/DataSafe.

Impact:
Not all expected configuration values of the URL are cloned.

Workaround:
There is no workaround.

Fix:
Description field is now cloned to the new URL.


706642-2 : wamd may leak memory during configuration changes and cluster events

Component: WebAccelerator

Symptoms:
wamd memory consumption increases over time.

Conditions:
-- AAM is provisioned so wamd is running.
-- User-initiated configuration change and/or other internal configuration or cluster events.

Impact:
wamd grows slowly over time, eventually crashing due to lack of memory. Temporary outage of services provided by wamd such as PDF linearization, invalidation, etc.

Workaround:
No workaround available.

Fix:
wamd n longer leaks memory during configuration changes and cluster events.


706631-2 : A remote TLS server certificate with a bad Subject Alternative Name should be rejected when Common Criteria mode is licensed and configured.

Component: Local Traffic Manager

Symptoms:
According to RFC2818, if the certificate sent by the TLS server has a valid Common Name, but the Subject Alternative Name does not match the Authenticate Name in the server-ssl profile, the connection should be terminated.

Conditions:
-- A server-ssl profile is enabled on a virtual server and has the 'authenticate-name' property set.

-- The TLS server presents a certificate in which the Subject Alternative Name does not match the configured authenticate-name.

-- Common Criteria mode licensed and configured.

Impact:
A TLS connection succeeds which should fail.

Workaround:
There is no workaround at this time.

Fix:
A remote TLS server certificate with a bad Subject Alternative Name is now rejected when Common Criteria mode is licensed and configured.


706534-1 : L7 connection mirroring may not be fully mirrored on standby BigIP

Component: Local Traffic Manager

Symptoms:
As a result of a known issue L7 connection mirroring may not be fully mirrored on standby BigIP

Conditions:
L7 VIP with mirroring configured
Connections with transfer of substantial size.

Impact:
Connections may be mirrored initially but removed after some time.
If there is a failover these connections may not be correctly handled.

Workaround:
Disabling LRO via
tmsh modify sys db tm.tcplargereceiveoffload value disable

May workaround this issue

Fix:
BIG-IP now fully mirrors all L7 connections


706423-1 : tmm may restart if an IKEv2 child SA expires during an async encryption or decryption

Component: TMOS

Symptoms:
TMM may discard an existing child-SA via timer during the moment it is in use for encryption or decryption. And in another spot, an error aborting negotiation can cause multiple timers associated with one child-SA, which fight one another.

Conditions:
Errors in IPsec config that fail negotiation can happen when a child-SA is in a state that does not manage timers correctly.

A config with short SA lifetime, causing frequent re-keying, can have the effect of searching for a race condition when expire happens during active use in crypto.

Impact:
TMM restarts, disrupting traffic and causing HA failover.

Workaround:
Ensure that the IPsec IKEv2 configuration in the IPsec policy is correct (the same) on both IPsec peers. Also, ensure SA lifetime has longer duration, instead of merely seconds. (The default is one day.)

Fix:
Now we ensure only one timer can be associated with a child-SA related to short-term progress toward maturity during negotiation, even if an error happens.

Now we also ensure a child-SA is safe during async crypto operations, even if they expire while currently in use.


706361 : IPS stats tables are empty after upgrade from 13.1.0 to 14.0.0

Component: Application Visibility and Reporting

Symptoms:
The IPS stats tables are empty after upgrade from 13.1.0 to 14.0.0.

Conditions:
-- Upgrade from 13.1.0 to 14.0.0.
-- AVR is NOT provisioned.
-- Viewing IPS stats tables.

Impact:
All statistics that relate to IPS are lost.

Workaround:
Before upgrading, run the following SQL command:

update AVR_CONF_FACT_TABLES set export_dir='/shared/avr_afm' where fact_name="AVR_STAT_IPS";

Fix:
The IPS stats tables are now saved in the '/shared_avr_afm' export directory.


706354-2 : OPT-0045 optic unable to link

Component: TMOS

Symptoms:
The OPT-0045 optical transceiver when inserted into a 40G port would not function. The following error appeared in /var/log/ltm:
Invalid module for bundle configuration of interface <portNumber>.0

Conditions:
OPT-0045 in a 40G port

Impact:
Optic was not usable, interface would not come up.

Workaround:
There is no workaround at this time.

Fix:
BIG-IP now supports the OPT-0045 optical transceiver.


706305-1 : bgpd may crash with overlapping aggregate addresses and extended-asn-cap enabled

Component: TMOS

Symptoms:
bgpd may crash on a BIG-IP system configured for dynamic routing announcing multiple overlapping aggregate-addresses with extended asn capabilities enabled.

Conditions:
- BGP enabled on route-domain
- BGP configured to announce several overlapping aggregate-addresses
- BGP configured with extended-asn-cap enabled.

Impact:
Inability for the unit to use BGP

Workaround:
Disabling extended-asn-cap or not announcing multiple overlapping aggregate addresses may allow to workaround this issue.

Fix:
bgpd no longer crashes with overlapping aggregate addresses and extended-asn-cap enabled


706276-1 : Unnecessary pop-up appears

Component: Fraud Protection Services

Symptoms:
A pop-up dialog box appears when 'Enhanced Data Integrity Check' is clicked.

Conditions:
-- Provision and license FPS.
-- Add URL.
-- Disable 'Check Full AJAX for Data Manipulation'.

Impact:
Unnecessary dialog box appears.

Workaround:
None.

Fix:
The pop-up does not appear.


706176-1 : TMM crash can occur when using LRO

Solution Article: K51754851


706169-3 : tmsh memory leak

Component: TMOS

Symptoms:
tmsh is leaking approximately 24 KB for every 'save sys config' call.

Conditions:
this occurs under the following conditions:
-- In tmsh, run the following command repeatedly:
 save sys config
-- In the GUI, modify the configuration and click Update, change pages, or otherwise cause a save to the configuration.

Impact:
This results in a memory leak, and a possible out-of-memory condition.

Workaround:
None.

Fix:
tmsh no longer leaks memory when performing configuration-save operations.


706128-2 : DNSSEC Signed Zone Transfers Can Leak Memory

Component: Global Traffic Manager (DNS)

Symptoms:
TMM might leak memory when performing DNSSEC signed zone transfers. You can detect this issue by examining system memory before and after performing zone transfers.

For example:

tmsh show sys memory raw | grep dnssec

Conditions:
-- Dynamic DNSSEC signing of zone transfers is configured for a given zone from an authoritative DNS Server (this could be on-box BIND, off-box BIND, etc) or from DNS Express.

Impact:
TMM leaks memory related to the signed zone transfer.

Workaround:
There is no workaround at this time.

Fix:
TMM no longer leaks DNSSEC zone transfer related memory.


706087 : Entry for SSL key replaced by config-sync causes tmsh load config to fail

Component: TMOS

Symptoms:
After config-sync, the secondary unit's key file does not match the passphrase stored for the key. This is a generic problem where config-sync is not synchronizing any differing file-objects on the secondary unit that happen to have the same cache_path as the primary.

Conditions:
If the cache_path of the encrypted key happens to be the same on the HA-pair, but the keys are different and have different passphrases.

Impact:
Secondary unit will fail to load the config during boot-up, so it will be offline. Other file-objects that had the same cache_path but where different files do not sync. The latter may not be noticed since nothing fails on the secondary unit.

Workaround:
Check if the cache_path of the encrypted key is the same on both systems prior to config-sync and that the sha1sum are different. If this is the case, remove the key on one of the systems and re-install the key and make sure the cache_path name is different.

Fix:
The key files (in the cache_path) will sync despite having the same name. The problem goes away. The same goes for any file-object that happened to have the same cache_path prior to sync.


706086-3 : PAM RADIUS authentication subsystem hardening

Solution Article: K62750376


705818-1 : GUI Network Map Policy with forward Rule to Pool, Pool does not show up

Component: TMOS

Symptoms:
When a Virtual Server has a Policy with a rule to forward request to a Pool, the Pool should be associated to the Virtual Server on the Network Map.

Conditions:
Create a Virtual Server with a Policy to forward requests to a Pool.

Impact:
The relationship of the Virtual Server to the Pool via the indirect Policy Rule is not visible in the network map.

Workaround:
No workaround to the visual.

Fix:
Associate Virtual Server with Policy that forwards requests to a Pool on the Network Map.


705799-2 : TMSH improvements

Solution Article: K77313277


705794-2 : Under certain circumstances a stale HTTP/2 stream might cause a tmm crash

Component: Local Traffic Manager

Symptoms:
A HTTP/2 stream is getting overlooked when cleaning up a HTTP/2 flow.

Conditions:
The only known condition is that the closing_stream is not empty. Exact entrance conditions are not clear.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
HTTP/2 flows are now properly cleaned up to prevent a tmm crash.


705774-1 : Add a set of disallowed file types to RDP template

Component: Application Security Manager

Symptoms:
Universally dangerous filetypes are not included in RDP policy template.

Conditions:
The user creates a new policy using the RDP template.

Impact:
Universally dangerous filetypes are not disallowed.

Workaround:
Dangerous filetypes can be added to policies created from RDP template.

Fix:
Universally dangerous filetypes are now included in RDP policy template.


705611-2 : The TMM may crash when under load when configuration changes occur when the HTTP/2 profile is used

Component: Local Traffic Manager

Symptoms:
The TMM may later crash after a configuration change done under load if the HTTP/2 profile is used.

Conditions:
A configuration change occurs. The configuration change is large enough, or the TMM load large enough that the change is not synchronous. The change involves a HTTP/2 profile.

Data traffic progresses through a partially initialized HTTP/2 virtual server, potentially causing a later crash.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
Large configuration changes to the TMM involving a HTTP/2 profile will no longer potentially cause a TMM crash.


705593-5 : CVE-2015-7940: Bouncy Castle Java Vulnerability

Component: Device Management

Symptoms:
An attacker could extract private keys used by Bouncy Castle in elliptic curve cryptography with a few thousand queries.

Conditions:
No specific conditions.

Impact:
None. BIG-IP software does not use the impacted library features.

Fix:
Version 1.59 of the library is installed on the BIG-IP system at the following paths:
/usr/share/java/rest/libs/bcprov-1.59.jar
/usr/share/java/rest/libs/bcpkix-1.59.jar


705559-1 : FPS: false positive "no strong integrity param" when none of the configured data-integrity params are present in request

Component: Fraud Protection Services

Symptoms:
A false positive "no strong integrity param" is sent when none of the configured data-integrity parameters are present in the request.

Conditions:
1. a protected URL has at least one parameter configured with data0integrity check enabled
2. enhanced data manipulation is enabled
3. a request without any of the data-integrity parameters is sent to the protected URL

Impact:
A false positive "no strong integrity param" alert is sent.

Workaround:
There is no workaround at this time.

Fix:
"No strong integrity param" alert should be suppressed in case that none of the data-integrity parameters were sent.

In case that forcing all data-integrity parameters was enabled (tmsh modify sys db antifraud.autotransactions.parameternameintegrity value enable) - the alert will be sent.


705503-3 : Context leaked from iRule DNS lookup

Component: Global Traffic Manager (DNS)

Symptoms:
The memory usage increases, and stats are inaccurate.

Conditions:
Call RESOLV::lookup from an iRule.

Impact:
Memory leak that accumulates over time and inaccurate stats.

Workaround:
There is no workaround other than not using RESOLV::lookup in an iRule.

Fix:
Memory leak no longer occurs.


705476-2 : Appliance Mode does not follow design best practices

Solution Article: K28003839


705456-1 : VCMP Guests unable to install block-device-image ISOs when http->https redirection is enabled

Component: TMOS

Symptoms:
ISOs of type block-device-image do not show up on VCMP Guests and are not available for installation when http->https redirection is enabled.

Conditions:
VCMP Guest has http->https redirection enabled.

Impact:
Not all available images are installable.

Workaround:
User must manually copy images to VCMP guest.

Fix:
Configured iControl REST to allow appropriate daemons access when http->https is enabled.


705161-1 : TMM may crash when processing TCP DNS traffic

Solution Article: K23520761

Component: Advanced Firewall Manager

Symptoms:
Under certain conditions, when processing TCP DNS traffic, TMM may crash

Conditions:
DNS profile enabled
TCP profile enabled
AVR enabled
ASM enabled

Impact:
TMM crash, leading to a failover event.

Fix:
TMM processes TCP DNS traffic as expected


704804-1 : The NAS-IP-Address in RADIUS remote authentication is unexpectedly set to the loopback address

Component: TMOS

Symptoms:
The NAS-IP-Address in RADIUS remote authentication requests is set to the loopback address, not the management IP.

Conditions:
This applies to remote authentication for the control plane, not APM.

Impact:
Login may be impacted.

Workaround:
There is no workaround at this time.

Fix:
NAS-IP-Address Attribute is now set correctly when the the BIG-IP system is configured to use RADIUS authentication.


704764-3 : SASP monitor marks members down with non-default route domains

Component: Local Traffic Manager

Symptoms:
The LTM SASP monitor marks pool members down, whose address include non-default route domains.

Conditions:
1. Using the SASP health monitor to monitor a pool or pool member.
2. The address of the pool member includes a non-default route domain, such as:

ltm pool rd_test {
    members {
        test_1:http {
            address 12.34.56.78%99
        }
    }
    monitor my_sasp
}

Impact:
Pool members with non-default route domains will never be marked up by the SASP monitor.

Workaround:
Do not specify non-default route domains in the addresses of pool members monitored by the SASP health monitor.

The SASP protocol does not support the use of route domains in member addresses. Thus, the SASP GWM (Global Workload Manager) will ignore route domain information included in the member addresses when registered by the SASP monitor in BIG-IP, and will report the status of members using addresses without route domains.

Therefore, even with a fixed version of the SASP monitor, BIG-IP LTM administrators must be careful to avoid configuring multiple pool members with the same address except for different route domains, to be monitored by the SASP monitor. If case of such a configuration error, the status of the member reported by the SASP GWM may not accurately reflect the status of one or more of the pool members with matching IP addresses.

Fix:
The LTM SASP monitor will correctly report the status of pool members configured with addresses that include a non-default route domain, if the SASP GWM is monitoring members with matching addresses (minus route-domain information).


704755-1 : EUD_M package could not be installed on 800 platforms

Component: TMOS

Symptoms:
Attempts to install the EUD_M EUD package fail on 800 platforms even though they are supposed to be supported.

Conditions:
Attempt to install EUD_M package on 800 platforms.

Impact:
Cannot install EUD_M package on a platform that is claimed to support it.

Workaround:
None.

Fix:
EUD_M package can now be installed on 800 platforms as expected.


704733-1 : NAS-IP-Address is sent with the bytes in reverse order

Component: TMOS

Symptoms:
The NAS-IP-Address has the address of the local device sent with the bytes in reverse order (e.g., 78.56.30.172, where 172.30.56.78 is expected).

Conditions:
-- This affects IPv4 addresses only.
-- BIG-IP system is configured for RADIUS authentication.

Impact:
The server may be configured to check the NAS-IP-Address before allowing logins, in which case it would fail.

Workaround:
There is no workaround at this time.

Fix:
This has been corrected.


704666-1 : memory corruption can occur when using certain certificates

Component: Local Traffic Manager

Symptoms:
If a certificate has an extremely long common name, or an extremely long alternative name and is attached to an SSL profile, memory corruption can occur when loading the profile.

Conditions:
An SSL profile is loaded with a certificate containing an extremely long common name or subject alternative name.

Impact:
TMM could crash.

Workaround:
Do not use certificates with extremely long common names

Fix:
A length check has been added to avoid corruption when using extremely long common names.


704643-1 : Backslashes preceding a forward slash are mistakenly doubled in regular expression within Signature rule

Component: Application Security Manager

Symptoms:
When adding or modifying a user-defined Attack Signature in Simple Edit Mode, backslashes preceding a forward slash are mistakenly doubled in regular expression keywords within the Signature.

Conditions:
-- Add or modify a user-defined Attack Signature with a regular expression keyword containing an escaped forward slash in Simple Edit Mode.

Impact:
The backslash escaping the forward slash is doubled, which changes the enforced signature rule.

Workaround:
Create or modify the Signature rule using Advanced Edit Mode.

Fix:
When adding or modifying a user-defined Attack Signature in Simple Edit Mode, backslashes preceding a forward slash are handled correctly in regular expression keywords within the Signature rule.


704580-1 : apmd service may restart when BIG-IP is used as SAML SP while processing response from SAML IdP

Solution Article: K05018525


704552 : Support for ONAP site licensing

Component: TMOS

Symptoms:
ONAP site licensing not supported.

Conditions:
-- Attempting to use ONAP site licensing

Impact:
BIG-IP system does not license.

Workaround:
None.

Fix:
Ported ONAP site licensing support to this version of the software.

Behavior Change:
This version of the software supports ONAP site licensing.


704528-2 : tmm may run out of memory during IP shunning

Component: Advanced Firewall Manager

Symptoms:
If no AppIQ is configured on an AFM-provisioned system, over time the system can run out of memory causing tmm to crash/restart.

Conditions:
-- Blacklist profile is configured with blacklist categories.
-- AppIQ is not configured.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
If no AppIQ is configured, the system now handles the shunned IP's that are to be sent to ECM server.


704512-1 : Automated upload of qkview to iHealth can time out resulting in error

Component: TMOS

Symptoms:
The automated upload of qkview files to iHealth via the support page of the BIG-IP GUI can time out waiting for an analysis from iHealth. Sometimes, iHealth can take several minutes to complete analysis, and this is a realistic scenario.

If the BIG-IP system times-out waiting for completion of the analysis, the link to the iHealth record is not stored.

Conditions:
iHealth takes longer than three minutes to complete analysis of a qkview file after uploading.

Impact:
Support history will not contain links to completed qkviews.

Workaround:
Run qkview from the command line and upload to iHealth manually.

Fix:
The iHealth link is now stored immediately after the qkview is successfully uploaded, and the timeout is not considered an error.


704435-1 : Client connection may hang when NTLM and OneConnect profiles used together

Component: Local Traffic Manager

Symptoms:
In deployments where a NT LanManager (NTLM) authentication profile and a OneConnect profile are used together in a LTM virtual server to label an authenticated connection to a Domain Controller (DC), if the persisted connection to the DC is re-used, the connection may hang. A connection in this state may not be cleaned up by the sweeper, resulting in a memory leak.

Conditions:
The NTLM and OneConnect profiles are associated with a LTM virtual server.

Impact:
A client connection is not serviced, and TMM memory will leak. Over a long time period, this may result in more widespread service disruptions.

Workaround:
Avoid the use of OneConnect profiles on virtual servers that use NTLM profiles. The connections to the Domain Controller are not pooled, but all other features are retained.

Fix:
Fixed a problem that prevented NTLM and OneConnect profiles from working properly on the same LTM virtual server.


704381-5 : SSL/TLS handshake failures and terminations are logged at too low a level

Component: Local Traffic Manager

Symptoms:
SSL/TLS handshake failures and terminations are being logged at too low of a level (INFO).

Conditions:
-- SSL/TLS connections are received or sent.
-- Handshake failures are logged.

Impact:
Sometimes other errors are produced, and the cause is a SSL/TLS handshake failure, but this failure is not being logged.

Workaround:
There is no workaround.

Fix:
SSL/TLS handhake failures and terminations are now logged at a higher level (WARNING).


704369-2 : TMM on BIG-IP restarts if a dos profile is attached to a virtual with sip-routing enabled

Component: Advanced Firewall Manager

Symptoms:
TMM restarts on a BIG-IP if a dos profile is attached to a virtual with sip-routing enabled

Conditions:
1. A virtual with sip-routing enabled.
2. A dos profile is attached to this virtual

Impact:
Traffic is disrupted while TMM restarts.

Workaround:
There is no workaround at this time.

Fix:
After fix, tmm is not restarting any more.


704336-1 : Updating 3rd party device cert not copied correctly to trusted certificate store

Component: TMOS

Symptoms:
When a BIG-IP admin updates the Device Certificate which also includes multiple CA intermediate and root certificates, it's expected that the new Device Certificate and its trust chain certificates are written to /config/big3d/client.crt and /config/gtm/server.crt. However, if the new Device Certificate is signed by a third party, only the Device Certificate is copied to client.crt and server.crt, even though root and intermediate certificates are written to /config/httpd/conf/ssl.crt/server.crt.

Conditions:
Updating Device certificate which also includes multiple intermediate and root certificates.

Impact:
The Trusted Device and Trusted server Certificates do not include intermediate CA and root certificates.

Workaround:
Manually copy/append the missing intermediate and root certificate to /config/big3d/client.crt and /config/gtm/server.crt.

Fix:
The fix will now add all the intermediate and root certificate including device certificate to Trusted Server and Trusted Device certificate bundle.


704282-2 : TMM halts and restarts when calculating BWC pass rate for dynamic bwc policy

Component: TMOS

Symptoms:
Rarely, TMM halts and restarts when calculating the BWC pass rate for dynamic bwc policy.

Conditions:
-- Using BWC dynamic bwc policy.
-- Fair share rate is low.

For this to happen, the fair share rate of the instance of dynamic policy has to be less than than 32 times the average packet size of the instance of policy.

For example, if the average packet size is 1 KB, then the fair share would be under 32 KB.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
F5 does not recommend running the BWC under 64Kbps.

Either decrease the number of subscribers or increase the max-rate of dynamic policy.

Fix:
TMM no longer halts and restarts when calculating the BWC pass rate for dynamic bwc policy when fair share rate is low.


704247-2 : BIG-IP software may fail to install if multiple copies of the same image are present and have been deleted

Component: TMOS

Symptoms:
BIG-IP .iso images may fail to install if multiple copies of the same image are present and have been deleted.

Conditions:
-- Have multiple copies of the same image .iso in the /shared directory on a BIG-IP system.
-- Delete one of them.

Impact:
Installation attempt of the remaining image(s) might fail.

Workaround:
Restart the lind process, so the installation can continue.

Fix:
The BIG-IP system now installs an image if multiple copies of the same image are present and have been deleted


704207-1 : DNS query name is not showing up in DNS AVR reporting

Component: Advanced Firewall Manager

Symptoms:
DNS query name is not showing up in DNS AVR reporting.

Conditions:
Sending traffic to Virtual with DNS profile.

Impact:
No query information for DNS is reported in AVR.

Workaround:
There is no workaround at this time.

Fix:
After fix, the query name is now showing up in AVR reporting.


704184-6 : APM MAC Client create files with owner only read write permissions

Solution Article: K52171282


704143-1 : BD memory leak

Component: Application Security Manager

Symptoms:
A BD memory leak.

Conditions:
websocket traffic with specific configuration

Impact:
Resident memory increases, swap getting used.

Workaround:
Make sure the web socket violations are not due to bad websocket URL configuration, that they are turned on in Learn/Alarm or block and that there is a local logger configured and attached.


704073-1 : Repeated 'bad transition' OOPS logging may appear in /var/log/ltm and /var/log/tmm

Solution Article: K24233427

Component: Local Traffic Manager

Symptoms:
'bad transition' OOPS messages may be repeatedly logged to /var/log/ltm and /var/log/tmm over time, polluting the log files.

Conditions:
Although there are no definitive user-discernible conditions, use of SSL functionality might increase the likelihood of logging.

Impact:
Log pollution and potential for performance degradation.

Workaround:
Suppress the logging using the following command:
tmsh modify sys db tmm.oops value silent

Fix:
The 'bad transition' OOPS logging has been demoted to internal, debug builds only.


703959 : Manual thresholds for DoS Protection Dynamic Signatures are not accepted via TMUI

Component: Advanced Firewall Manager

Symptoms:
Manual thresholds for DoS Protection Dynamic Signatures are not accepted via TMUI. The 'Infinite' values for detection and mitigation are retained. No error message is returned.

Conditions:
Attempting to configure manual AFM detection and mitigation threshold for DoS Protection Dynamic Signatures using the Management GUI.

Impact:
The BIG-IP system Administrator is not aware that config change failed to be applied.

Workaround:
Manual thresholds for Dynamic Signatures can be configured using TMSH.

Fix:
You can now change manual detection and mitigation threshold via TMUI.


703940-2 : Malformed HTTP/2 frame consumes excessive system resources

Solution Article: K45611803


703914-2 : TMM SIGSEGV crash in poolmbr_conn_dec.

Component: Local Traffic Manager

Symptoms:
TMM cores in poolmbr_conn_dec function.

Conditions:
A connection limit is reached on a VS having an iRule using LB::reselect and a pool using request queueing.

Impact:
TMM core, traffic interruption, possible failover.

Workaround:
Do not use LB::reselect, disable request queueing on a pool associated with the VS.

Fix:
TMM correctly handles LB:reselect on virtual servers with pools using request queueing.


703869 : Waagent updated to 2.2.21

Component: TMOS

Symptoms:
Microsoft updates waagent via an opensource process, but it is not compatible with BIG-IP software, and so cannot be upgraded outside of BIG-IP releases. Without this, Microsoft will not support older releases of BIG-IP systems in their environment.

Conditions:
Using Microsoft Azure.

Impact:
Microsoft does not support the version of waagent shipped as part of BIG-IP software.

Workaround:
None.

Fix:
Waagent was updated to 2.2.21 from Microsoft along with F5 changes for compatibility with BIG-IP software.


703848-1 : Possible memory leak when reusing statistics rows in tables

Component: TMOS

Symptoms:
The handling of the pointers to memory in the statistics tables includes a path that zeros out a pointer to more memory that should be free'd. This means the memory is not free'd for that case.

Conditions:
This condition is usually only hit when the entire file is being deleted and so it doesn't matter that the list is not fully traversed.

Impact:
When slabs are being reused this bug may cause a memory leak.

Workaround:
There is no workaround at this time.

Fix:
The code has been fixed to properly follow the list.


703833-1 : Some bot detected features might not work as expected on Single Page Applications

Component: Application Security Manager

Symptoms:
Some client side features do not work correctly when enabling single page application.

Conditions:
Enabling single page application (on DoS or ASM), and Web Scraping-> Persistent Client Identification

Impact:
Captcha challenge causes a loop of ajax requests.

Workaround:
There is no workaround at this time.

Fix:
Fixing Persistent Client Identification for Single Page Applications.


703793-3 : tmm restarts when using ACCESS::perflow get' in certain events

Component: Access Policy Manager

Symptoms:
tmm will restart when using 'ACCESS::perflow get' if the per-request policy has not been started yet.

Conditions:
Use 'ACCESS::perflow get' iRule in an event that runs before the per-request policy (e.g., CLIENT_ACCEPTED).

Impact:
tmm cores and traffic flow will be interrupted while it restarts.

Workaround:
None.

Fix:
Initialization of certain variables was reworked so that the iRule command will not cause a core anymore if the per-flow value is unavailable due to the per-request policy not having been started yet.


703761-2 : Disable DSA keys for public-key and host-based authentication in Common Criteria mode

Component: TMOS

Symptoms:
On a BIG-IP system configured for Common Criteria (CC) mode, DSA keys can still be used for key-based authentication to the BIG-IP system.

Conditions:
-- Running a CC release with CC Mode enabled.
-- Attempt to SSH-connect to the BIG-IP system from another system with a DSA key.
-- That DSA key is present in the BIG-IP system's authorized_keys file.

Impact:
If the key is in the BIG-IP system's authorized_keys file, public-key authentication succeeds, when DSA authentication should be disabled in CC mode.

Workaround:
There is no workaround at this time.

Fix:
DSA SSH keys are now disabled for public-key and host-based authentication in Common Criteria mode, as expected.


703702 : Fixed iControl REST not listing GTM Listeners

Component: Global Traffic Manager (DNS)

Symptoms:
When using iControl REST to get a list of GTM Listeners, no listeners will be returned.

Conditions:
Use iControl REST to get a list of GTM Listeners

Impact:
Cannot get a list of GTM Listeners by iControl REST

Workaround:
Use iControl REST to get a list of all LTM Virtual Servers, and then look for virtual-servers with a DNS Profile

Fix:
Fixed issue preventing iControl REST from returning a list of GTM Listeners


703669-2 : Eventd restarts on NULL pointer access

Component: TMOS

Symptoms:
The loop that reads /config/eventd.xml to load configuration data processes data in 1024-byte chunks. If the size of the file is a multiple of 1024 bytes, the end of file condition that terminates the file read does not occur until the subsequent read. The subsequent read returns zero (0) bytes and passes an empty buffer to the parser, which causes eventd to restart repeatedly.

Conditions:
The length of the file /config/eventd.xml is a multiple of 1024 bytes.

Impact:
Causes eventd to crash.

Workaround:
To work around the problem, insert whitespace in /config/eventd.xml to pad the file to a size that is not a multiple of 1024 bytes.


703580-1 : TLS1.1 handshake failure on v12.1.3 vCMP guest with earlier BIG-IP version on vCMP host.

Component: Local Traffic Manager

Symptoms:
TLS1.1 handshake failure on guest. The following error appears in /var/log/ltm:
warning tmm[11611]: 01260009:4: Connection error: ssl_hs_cn_vfy_fin:2339: corrupt Finished (20)

Conditions:
-- Using the VIPRION 42xx/43xx and B21xx blades.
-- Running BIG-IP software earlier than v12.1.3 (for example v12.1.2-hf2) on the vCMP host system.
-- Deploying vCMP guest running v12.1.3.
-- Using TLS1.1.

Impact:
TLS1.1 handshake fails on the guest.

Workaround:
Use the same software version on the vCMP host and vCMP guests.

Fix:
TLS1.1 handshake no longer fails running v12.x/v13.x vCMP guest with earlier BIG-IP software version on vCMP host.


703545-1 : DNS::return iRule "loop" checking disabled

Component: Global Traffic Manager (DNS)

Symptoms:
In ID 517347, checking was added to attempt to detect infinite loops caused by improper use of the DNS::return iRule command.

This is occasionally catching false positive loops resulting in connections being dropped incorrectly.

Conditions:
A virtual with a DNS profile that is using the udp profile instead of the udp_gtm_dns profile. An iRule that uses the DNS::return command.

Impact:
If a loop is erroneously detected, the connection will be dropped.

Workaround:
Where possible use the udp_gtm_dns profile instead of udp on virtuals with a DNS profile.

Where possible, use a "return" command immediately after the "DNS::return" command to prevent accidentally calling DNS::return multiple times.

Fix:
The loop detection logic has been removed.


703517 : TMM may crash when processing TCP DNS traffic

Solution Article: K23520761

Component: Advanced Firewall Manager

Symptoms:
Under certain conditions, when processing TCP DNS traffic, TMM may crash

Conditions:
DNS profile enabled
TCP profile enabled
AVR enabled
ASM enabled

Impact:
TMM crash, leading to a failover event.

Fix:
TMM processes TCP DNS traffic as expected


703515-3 : MRF SIP LB - Message corruption when using custom persistence key

Solution Article: K44933323

Component: Service Provider

Symptoms:
If the custom persistence key is not a multiple of 3 bytes, the SIP request message may be corrupted when the via header is inserted.

Conditions:
Custom persistence key is not a multiple of 3 bytes

Impact:
The SIP request message may be corrupted when the via header is inserted.

Workaround:
Pad the custom persistence key to a multiple of 3 bytes in length.

Fix:
All persistence key lengths work as expected.


703429-2 : Citrix Receiver for Android (v3.13.1) crashes while accessing PNAgent services

Component: Access Policy Manager

Symptoms:
Citrix Receiver for Android (v3.13.1) crashes while accessing PNAgent services through F5 BIG-IP APM virtual server. The application closes just after entering the credentials.

Conditions:
-- Citrix Receiver for Android (v3.13.1) is used.
-- PNAgent replacement mode is configured for BIG-IP APM virtual server.

Impact:
No access to published Applications and Desktops through Citrix Receiver for Android.

Workaround:
None.

Fix:
System now provides valid data to Citrix Receiver for Android client.


703298-2 : Licensing and phonehome_upload are not using the sync'd key/certificate

Component: TMOS

Symptoms:
After config-sync, the secondary unit's key passphrase does not decrypt the cached key file.

Conditions:
The original file for f5_api_com.key is used instead of the cached file.

Impact:
phonehome_upload will fail on the secondary unit because the passphrase doesn't match the key file.

Workaround:
After sync, copy the file /config/filestore/files_d/Common_d/certificate_key_d/:Common:f5_api_com.key_xxxx over to /config/ssl/ssl.key/f5_api_com.key using the following commands:

# cd /config/filestore/files_d/Common_d/certificate_key_d
# cp -a :Common:f5_api_com.key_xxxx /config/ssl/ssl.key/f5_api_com.key :Common:f5_api_com.key_xxxx

Once the /config/ssl/ssl.key file is in sync, then loading the config with either cached or un-cached file will work fine.

Fix:
The system now removes the source-path files and only keeps the cache-path files. phonehome_upload now will work on the standby unit after a config-sync. Without the source-path files which do not get sync'd, there is no danger of re-loading them.


703266-2 : Potential MCP memory leak in LTM policy compile code

Component: Local Traffic Manager

Symptoms:
Failure in processing LTM policy may result in MCP memory leak

Conditions:
When Centralized Policy Management (CPM) fails to process LTM policy

Impact:
MCP memory leak

Workaround:
There is no workaround at this time.

Fix:
This fix handles rare MCP memory leak which may occur if CPM fails to process LTM policy


703233 : Some filters don't work in Security->Reporting->URL Latencies page

Component: Application Visibility and Reporting

Symptoms:
If a filter by Virtual Servers or URLs in Security->Reporting->URL Latencies page, the data is not filtered.

Conditions:
No special condition.

Impact:
It it impossible to filter data in the aforementioned page.

Workaround:
There is no workaround at this time.

Fix:
An incorrect SQL query was applied to the statistics database upon such data request. The SQL query is fixed.


703208-1 : PingAccessAgent causes TMM core

Component: Access Policy Manager

Symptoms:
PingAccessAgent can cause TMM to core due to accessing freed memory.

Conditions:
It happens in edge case situation. Exact steps are still under investigation. Suspicion is that the client aborts the connection while TMM/PingAccessAgent module is still awaiting response from the PingAccessAgent back-end server.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.


703191-2 : HTTP2 requests may contain invalid headers when sent to servers

Component: Local Traffic Manager

Symptoms:
HTTP requests handled by an HTTP/2 virtual server may have blank header names when proxied through to the server or when handled via iRules.

Conditions:
-- Virtual server has the HTTP/2 profile assigned.
-- Client and the BIG-IP system negotiate/use HTTP/2.

Impact:
HTTP/2 applications may generate CSRF-related errors. Alternately, the server may return intermittent (and from the client's perspective, spurious) 400 Bad Request responses.

Workaround:
There is no workaround other than to remove the HTTP/2 profile from the virtual server.


703171-1 : High CPU usage for apmd, localdbmgr and oauth processes

Component: Access Policy Manager

Symptoms:
High CPU Usage for apmd, localdbmgr, and oauthd with large configurations.

Conditions:
-- APM provisioned.
-- BIG-IP has a large configuration (i.e., a large number of virtual servers).
-- One of the following:
  + A full config sync happens from one device (with a large configuration) to another device.
  + When loading BIG-IP configurations that contain a large number of virtual servers.

Impact:
Depending on the operation:
  + The process on the second device exhibits high CPU usage
  + The loading device exhibits high CPU usage.

APM end user traffic might not be processed by APM until it is done processing all the config changes. The amount of time service is down depends on how large the configuration is.

Workaround:
None.

Fix:
Startup processing of apmd, localdbmgr, and oauthd have been optimized to reduce the CPU usage.


703045-1 : If using TMSH commands with deprecated attributes in iAPP, the upgrade will fail.

Component: TMOS

Symptoms:
TMSH commands with deprecated attributes will fail if used in iAPP.

Conditions:
TMSH commands with deprecated attributes will fail if used in iAPP. This is so whether the iAPP is activated during the upgrade process or simply run under iAPP service at the user display.

Impact:
TMSH commands will not execute like create command will result in no objects (eg monitor, virtual server etc) being created.

Workaround:
Try to avoid deprecated attributes of the object in the iAPP.

Fix:
All TMSH commands should handle deprecated attributes of objects consistently across TMSH command line, CLI Script and iAPP and like so:

- run TMSH commands to full execution with only warning message.
- full execution means objects action should be executed without error and no something amiss silently either.


702946-3 : Added option to reset staging period for signatures

Component: Application Security Manager

Symptoms:
In cases where a staging period was started, but no traffic passed through security policy, you might want to reset the staging period when traffic starts, but there is no option to do so.

Conditions:
Staging enabled for signatures in policy.

Impact:
There is a suggestion to enforce the signature before any traffic can influence this decision.

Workaround:
If all signatures are staged, you can enforce them all, and then enable staging again.

Note: Apply policy is required between actions.

Fix:
Added option to reset the staging period for all or specific signatures. In modal windows shown after clicking 'Change properties...' on the Policy Signatures screen, when 'No' is not selected for 'Perform Staging', the system presents a checkbox: Reset Staging Period.


702738-1 : Tmm might crash activating new blob when changing firewall rules

Solution Article: K32181540

Component: Advanced Firewall Manager

Symptoms:
TMM crashes with core when changing firewall rules. TMM can enter a crash-loop, so it will crash again after restarting.

Conditions:
Updating, removing, or adding firewall rules.

Specific characteristics of change that can cause this issue are unknown; this issue occurs rarely.

Impact:
Data traffic processing stops.

Workaround:
There are two workaround options:
Option A
1. Delete all policies.
2. Create them again without allowing blob compilation.
3. Repeat steps 1 and 2 until all the policies have been created (enable on-demand-compilation).

Option B
Modify all the rules simultaneously.

For example, the following steps will resolve this issue:
1. Enable on-demand-compilation
2. Select an IP address that is not used in any rules, e.g., 1.1.1.1.
3. Add that IP address to all the rules/source in all of the policies. To do so, run the following command for each policy:
tmsh modify security firewall policy POLICY_NAME rules modify { all { source { addresses add { 1.1.1.1 } } } }

4. Delete the IP address (restore rules) in all of the policies. To do so, run the following command for each policy:
tmsh modify security firewall policy POLICY_NAME rules modify { all { source { addresses delete { 1.1.1.1 } } } }
5. Disable on-demand-compilation. Doing so starts new blob compilation.

Fix:
TMM no longer crashes when changing firewall rules.


702705-2 : Core observed during lease query testing when RADIUS Authentication is configured in DHCP profile

Component: Policy Enforcement Manager

Symptoms:
Tmm may halt and restart when RADIUS Authentication is configured in DHCP profile.

Conditions:
1. RADIUS Authentication is configured in a DHCP profile.
2. DHCP response does not have proper info.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
This version handles these conditions, so tmm does not halt and restart.


702520-2 : Same AZ failover in AWS fails in some configurations with two or more objects sharing the same IP address.

Solution Article: K53330514

Component: TMOS

Symptoms:
BIG-IP fails to reattach floating addresses to local interfaces during failover, when two or more objects are configured with the same IP address in a given traffic group.

Failover fails with the following error in /var/log/ltm: err logger: /usr/libexec/aws/aws-failover-tgactive.sh (traffic-group-1): Failed to reassign some or all address(es): <IP address> <the same IP address> on interface <eni address>.

Conditions:
-- AZ AWS failover.
-- Same IP address is used for two or more virtual addresses, self IPs, NAT, SNAT translation.

Note: Having two virtual servers with the same IP address (but different ports) does not cause the problem. Also, there is no conflict when using the same IP address for different traffic groups.

Impact:
Failover will fail; some or all IP addresses will not be transferred to the active BIG-IP system.

Workaround:
The only workaround is to change the configuration to use unique IP addresses for conflicting objects.

Fix:
This issue has been resolved.


702487-3 : AD/LDAP admins with spaces in names are not supported

Component: Access Policy Manager

Symptoms:
If admin is imported from Active Directory (AD) or Lightweight Directory Access Protocol (LDAP) and the name contains spaces, Visual Policy Editor (VPE), import/export/copy/delete, etc., fail, and post an error message similar to the following: sourceMCPD::loadDll User 'test user' have no access to partition 'Common'.

Note: Names containing spaces are not supported on BIG-IP systems.

Conditions:
-- Admin name containing spaces is imported from AD or LDAP.
-- Attempt to perform operations in VPE.

Impact:
VPE, import/export/copy/delete do not work.

Workaround:
There is no workaround other than to not use admin names containing spaces.

Fix:
VPE and utilities operations are now supported, even for admins with spaces in names. However, F5 recommends against using spaces in admin names, and recommends that you modify the admin name to remove the spaces.


702450-1 : The validation error message generated by deleting certain object types referenced by a policy action is incorrect

Component: Local Traffic Manager

Symptoms:
When deleting certain objects that are referenced by policy actions, you may see a validation error like this:

# tmsh delete ltm virtual test-vs
01071726:3: Cannot delete policy action '/Common/test-vs'. It is in use by ltm policy '/Common/test-policy'.

The referenced object is not a "policy action" in this case, but is a virtual server.

Conditions:
LTM policies must be in use, and at least one policy action must forward to an object. The user must attempt to delete that object.

Impact:
Possible confusion at the error message.

Workaround:
There is no workaround at this time.

Fix:
Made the error message accurately reflect what the user was attempting to delete.


702439 : Non-default HTTP/2 header_table_size causes HTTP/2 streams to be reset

Solution Article: K04964898

Component: Local Traffic Manager

Symptoms:
If the HTTP/2 configuration header_table_size is changed from the default value of 4096, then streams will be reset with a RST_STREAM error.

Conditions:
The header_table_size field in the HTTP/2 profile is changed from the default.

Impact:
HTTP/2 connections will be unusable.

Workaround:
Set the header table size argument back to its default.

Fix:
The HTTP/2 filter correctly handles the dynamic header table resize notifications triggered by a non-default header table size. Streams will not be reset with a RST_STREAM error.

Additionally, the BIG-IP system will now send the correct number of dynamic header table resize notifications when the table is resized by the client multiple times between header blocks.


702278-2 : Potential XSS security exposure on APM logon page.

Component: Access Policy Manager

Symptoms:
Potential XSS security exposure on APM logon page.

Conditions:
-- A LTM virtual server with an Access Policy assigned to it.
-- The Access Policy is configured to use a 'Logon Page' VPE agent, followed by AAA agent with 'Max Logon Attempts Allowed' set to a value greater than 1.

Impact:
Potential XSS security exposure.

Workaround:
1. Using APM Advance Customization UI, remove setOrigUriLink(); call in logon.inc from the next JS function:

369 function OnLoad()
370 {
371 var header = document.getElementById("credentials_table_header");
372 var softTokenHeaderStr = getSoftTokenPrompt();
373 if ( softTokenHeaderStr ) {
374 header.innerHTML = softTokenHeaderStr;
375 }
376 setFormAttributeByQueryParams("auth_form", "action", "/subsession_logon_submit.php3");
377 setFormAttributeByQueryParams("v2_original_url", "href", "/subsession_logon_submit.php3");
378 // ===> REMOVE THIS ONE setOrigUriLink();
----

Fix:
Potential security exposure has been removed from APM logon page.


702263-1 : An access profile with large number of SAML Resources (greater than 200) causes APM error ERR_TOOBIG while loading.

Component: Access Policy Manager

Symptoms:
Using a SAML SP-initiated use case (APM is IdP) and having a large Access Policy with 200 or more SAML resources assigned to users. Each time a new SAML resource is added to that Access Policy, the entire SSO service becomes unusable. No new sessions can be established. The system generates internal metadata that consists of the names of all the SAML resources along with its SSO name. This has a limit of size 4 KB. When this limit is reached, the system logs errors similar to the following:

-- err tmm3[15840]: 01490514:3: (null):Common:00000000: Access encountered error: ERR_TOOBIG. File: ../modules/hudfilter/access/access.c, Function: access_create_saml_meta_data, Line: 21001


-- err tmm3[15840]: 014d1014:3: /Common/SAML_SSO_access:Common:7a34161c:SAML SSO: Error(16) Unable to find SAML SSO/SP Connector object matching SAML Authn Request.

Conditions:
A SAML SSO Access policy has a large number of SAML Resources assigned to it (such that the combined length of their names is greater-than-or-equal-to 4 KB).

Impact:
The system logs an error on running the Access Policy, the whole SSO service becomes unusable. No new sessions can be established.

Workaround:
Delete any unused SAML resources from SAML SSO access policy so that the combined length of the names of all SAML resources assigned to Access policy is less than 4 KB.

Fix:
The system now allocate memory dynamically for the internally stored metadata, so it can handle large lists of assigned SAML resource objects.


702232-1 : TMM may crash while processing FastL4 TCP traffic

Solution Article: K25573437


702227-3 : Memory leak in TMSH load sys config

Component: TMOS

Symptoms:
When loading a configuration, either through tmsh load sys config or the corresponding iControl REST command, the TMSH or icrd_child processes can leak memory.

Conditions:
When configuration is loaded via TMSH or iControl REST.

Impact:
The TMSH process memory or icrd_child process memory continues to grow every time the config is loaded.

Workaround:
If memory consumption grows in the TMSH process, restart the TMSH process.

If memory consumption grows in iControl REST, restart the REST service using the following command: bigstart restart restjavad.

Fix:
There is no longer a memory leak when loading configuration in TMSH or iControl REST.


702222-1 : RADIUS and SecurID Auth fails with empty password

Component: Access Policy Manager

Symptoms:
If password value is empty, the following error message will be logged in /var/log/apm:

err apmd[14259]: 014902f0:3: /Common/profile_name:Common:eb69a5gd: RADIUS Agent: Failed to read Password Source session variable:

Conditions:
This occurs only when following conditions are met:
- RADIUS or SecurID auth agent is included in the access policy.
- Empty password value is used for authentication.

Impact:
User may not be authenticated.

Workaround:
- Add variable assignment agent before RADIUS/SecurID auth agent in the access policy.
- Set 'session.logon.last.password' (or whatever password source is used for authentication) to a random value.

Fix:
RADIUS/SecurID auth agent allows empty password value for authentication.


702151-1 : HTTP/2 can garble large headers

Component: Local Traffic Manager

Symptoms:
The HTTP/2 filter may incorrectly encode large headers.

Conditions:
A header that encodes to larger than 2048 bytes may be incorrectly encoded.

Impact:
The garbled header may no longer conform to the HPACK spec, and cause the connection to be dropped. The garbled header may be correctly formed, but contain incorrect data.

Fix:
The HTTP/2 filter correctly encodes large HTTP headers.


702008-1 : ASM REST: Missing DB Cleanup for some tables

Component: Application Security Manager

Symptoms:
Finished REST tasks that are not deleted by the client that initiated them are meant to be cleaned periodically. Certain tasks are not included in this cleanup job.

Conditions:
The following tasks are not reaped automatically if left uncleaned by the REST client that initiated them:

From 13.0.x:
-- /mgmt/tm/asm/tasks/apply-server-technologies
-- /mgmt/tm/asm/tasks/bulk
-- /mgmt/tm/asm/tasks/export-policy-template
-- /mgmt/tm/asm/tasks/export-requests
-- /mgmt/tm/asm/tasks/import-policy-template

From 13.1.0:
-- /mgmt/tm/asm/tasks/export-data-protection
-- /mgmt/tm/asm/tasks/import-data-protection
-- /mgmt/tm/asm/tasks/import-certificate
-- /mgmt/tm/asm/tasks/policy-diff
-- /mgmt/tm/asm/tasks/policy-merge
-- /mgmt/tm/asm/tasks/update-enforcer

Impact:
DB space usage grows with each ASM REST task that is not cleaned up.

Workaround:
REST Clients that initiate tasks can delete them after verifying the task has reached a final state.

Fix:
REST tasks left behind are now be pruned by the DB Cleanup process.


701889-1 : Setting log.ivs.level or log-config filter level to informational causes crash

Component: Service Provider

Symptoms:
Certain log messages for internal virtual server (IVS) at 'informational' log level, cause TMM to crash when they are logged. The messages are logged at the end of an HTTP transaction to or from an IVS.

Conditions:
Information level logging enabled:

- sys db log.ivs.level informational or
- log-config filter level set to info

A transaction that passes HTTP to/from an internal virtual server.

Impact:
TMM crashes and restarts, causing loss of connections.

Workaround:
Avoid setting log.ivs.level to 'informational' or higher level and/ log-config filter level to 'info' or higher. By default the level is 'error' which does not trigger the bug.

Fix:
Informational messages for internal virtual server (IVS) are logged as expected and TMM does not crash.


701856-1 : Memory leak in ASM-config Event Dispatcher upon continuous Policy Builder restart

Component: Application Security Manager

Symptoms:
In rare circumstance, when Policy Builder restarts continuously (in response to a different issue in which there are many shmem segments allocated and used by tmm), ASM-config Event Dispatcher memory usage grows uncontrollably.

Conditions:
In rare circumstances, Policy Builder restarts continuously (in response to a different issue in which there are many shmem segments allocated and used by tmm).

Impact:
ASM-config Event Dispatcher memory usage grows continuously until the device eventually fails over.

Workaround:
Restart asm_config_server on all devices using the following command:
 killall asm_config_server.pl

Fix:
ASM-config Event Dispatcher memory usage remains stable even upon multiple Policy Builder restarts.


701841-2 : Unnecessary file recovery_db/conf.tar.gz consumes /var disk space

Component: Application Security Manager

Symptoms:
A file, /ts/var/install/recovery_db/conf.tar.gz ,is saved unnecessarily during UCS file save, and consumes /var disk space.

Conditions:
UCS file is saved.

Impact:
The /var filesystem can become full; this may degrade system performance over time, and can eventually lead to traffic disruptions.

Workaround:
Manually delete /ts/var/install/recovery_db/conf.tar.

Fix:
Unnecessary file recovery_db/conf.tar.gz is no longer written.


701826 : qkview upload to ihealth fails or unable to untar qkview file

Component: TMOS

Symptoms:
qkview upload to ihealth fails unable to untar qkview file.

Conditions:
When qkview file is untarred, it creates a same directory name in loop as below and fails to untar successfully.

.../dir1/
.../dir1/dir1/
.../dir1/dir1/dir1/
...

This happens due to dangling symlink dir1 which points to nothing.

[root@localhost:Active:Standalone] config # ls -l /config/bigip/auth/pam.d/dir1
lrwxrwxrwx. 1 root root 64 2018-01-30 08:56 /config/bigip/auth/pam.d/dir1 ->
[root@localhost:Active:Standalone] config # stat /config/bigip/auth/pam.d/dir1
  File: `/config/bigip/auth/pam.d/dir1' -> `'
  Size: 64 Blocks: 8 IO Block: 4096 symbolic link
Device: fd16h/64790d Inode: 112045 Links: 1
Access: (0777/lrwxrwxrwx) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2018-01-30 08:56:20.000000000 -0800
Modify: 2018-01-30 08:56:20.000000000 -0800
Change: 2018-01-31 08:39:35.000000000 -0800
[root@localhost:Active:Standalone] config #

Impact:
Unable to untar qkview or qkview upload to ihealth fails.

Workaround:
Identify the dangling symlink and delete. Then generate qkview or use ihealth to generate qkview and upload to ihealth.

Fix:
Qkview tool will identify dangling symlink and handle safely to avoid looping.


701800-2 : SSO-enabled native RDP resources cannot be launched from APM Webtop with Mac RDP client 10.x

Component: Access Policy Manager

Symptoms:
Mac RDP client 10.x shows credentials prompt when SSO-enabled native RDP resource is launched from the APM Webtop. If the user enters credentials manually, the RDP client will display an error message.

Conditions:
-- Mac RDP client is of version 10.x (e.g., 10.1.6).
-- Native RDP resource is assigned to the webtop.
-- SSO is enabled on the RDP Resource.

Impact:
RDP resource cannot be launched.

Workaround:
Add the following to 'Custom Parameters' of SSO-enabled RDP resources:
full address:s:%{session.server.network.name}
promptcredentialonce:i:1

Fix:
SSO-enabled native RDP resources now can be launched from APM Webtop with Mac RDP client 10.2.0.


701740-1 : apmd leaks memory when updating Access V2 policy

Component: Access Policy Manager

Symptoms:
A small leak occurs in the apmd process when processing mcp notifications about configuration updates.

Conditions:
-- Changing an Access Policy configurations.
-- apmd receives a notification about it.

Impact:
apmd grows in size very slowly. The issue does not have any immediate and significant impact on BIG-IP system functionality.

Workaround:
There is no workaround at this time.

Fix:
apmd no longer leaks a small amount when processing MCP notifications.


701737-1 : apmd may leak memory on destroying Kerberos cache

Component: Access Policy Manager

Symptoms:
ampd leaks memory in AD Query agent.

Conditions:
The leak happens in response to any of the following conditions:
-- A Kerberos cache reset is requested (any of the caches - GROUP/PSO/KERBEROS).
-- Change to associated AAA AD Server were made and new Access Policy is applied.
-- AD Query was not able to make ldap_bind to KDC and the error is NOT a timeout (e.g., invalid administrator password).

Impact:
The ampd leaks memory and might cause unstable behavior.
The apmd process, or some other daemon may be killed by OOM killer when it tries to allocate memory.

Workaround:
There is no workaround at this time.

Fix:
AD Query agent no longer causes apmd memory leak during group cache update.


701736-1 : Memory leak in Machine Certificate Check agent of the apmd process

Component: Access Policy Manager

Symptoms:
apmd process leaks memory in Machine Certificate Check agent

Conditions:
Machine Certificate Check agent is configured in an Access Policy.

Impact:
apmd may grow in size. This may lead to the apmd process or another process, to be killed by OOM-killer

Workaround:
There is no workaround at this time.

Fix:
An apmd memory leak in the Machine Certificate Check agent has been fixed.


701690-1 : Fragmented ICMP forwarded with incorrect icmp checksum

Solution Article: K53819652

Component: Local Traffic Manager

Symptoms:
Large fragmented ICMP packets that traverse the BIG-IP might have their source or destination IP addresses changed and transmitted with the checksum incorrectly calculated.

Conditions:
A FastL4 virtual server able to transmit ICMP frames (ip-protocol ICMP or any), or SNAT/NAT only when a fragmented ICMP packet is received and is expected to be passed through the virtual server (or SNAT or NAT).

Impact:
Large ICMP echo packets (greater than MTU) will be dropped by the recipient due to the checksum error. No echo response will be seen.

Workaround:
Turn on IP fragment reassembly on the FastL4 profile associated with the virtual server.


701639-1 : Session variables in Requested Authentication Context Class in SP do not get resolved when Authentication Request is generated by BIG-IP as SP.

Component: Access Policy Manager

Symptoms:
Session variables in Requested Authentication Context Class in SP do not get resolved when Authentication Request is generated by the BIG-IP system as SP. They are sent as is. This is a behavior change from v12.1.2/v12.1.3/v13.0.0, where, the value gets substituted in the SP's AuthnRequest sent to IDP.

Conditions:
On configuring Requested Authentication Context Class in SP to define a session variable similar to the following:
%{session.client.type}

Impact:
The generated Authentication Request does not have the session variable resolved. The string is sent as is. The Authentication Request fails and the session cannot be established.

Workaround:
None.

Fix:
The system now resolves the session variable in the configured Authentication Context Class for SP while generating the Authentication Request.


701637 : Crash in bcm56xxd during TMM failover

Component: Advanced Firewall Manager

Symptoms:
During a TMM failover, such as after an upgrade to a later version of software, bcm56xxd might crash.

Conditions:
TMM failover.

Impact:
Restart of bcm56xxd; temporary loss of network connectivity.

Workaround:
There is no workaround at this time.

Fix:
Bcm56xxd no longer crashes and restarts on a TMM failover.


701626-2 : GUI resets custom Certificate Key Chain in child client SSL profile

Solution Article: K16465222

Component: TMOS

Symptoms:
In the GUI, editing a client SSL profile or selecting a different parent profile changes the Certificate Key Chain to default (i.e., /Common/default.crt and /Common/default.key).

Conditions:
This happens in the following scenario:

1. Using the GUI, create a client SSL profile.
2. Configure the new profile to inherit from a client SSL profile other than the default, clientssl.
3. Click the Custom box for Certificate Key Chain and select a different cert and key from the default.
4. Click Update.
5. In the GUI, change any setting in the newly created profile, or select a different parent profile (but not the clientssl profile).
6. Click Update again.

Impact:
The system resets Certificate Key Chain to default, even though the Custom box is checked.

Workaround:
To work around this issue in the GUI, click the Custom checkbox next to the 'Certificate Key Chain' option in the parent profile. This will set the value of inherit-certkeychain to false , preventing the issue from occurring.

You can also use tmsh to update parent profile settings to avoid the occurrence of this issue..

Fix:
GUI no longer resets custom Certificate Key Chain in child client SSL profiles.


701538-2 : SSL handshake fails for OCSP, TLS false start, and SSL hardware acceleration configured

Component: Local Traffic Manager

Symptoms:
SSL handshake fails if the client initiates the handshake with TLS false start (specifically, client SSL sends the SSL record data to the server before the Server sends out the CSS is FINISHED notification).

Conditions:
1. Client initiates the SSL handshake with False Start.
2. The BIG-IP system has SSL hardware acceleration enabled (which is default for for non-virtual edition (VE) versions).

Impact:
The BIG-IP system sends the RST to tear down the connection in TLS false start.

Workaround:
There are no true workarounds. You must disable one of the conditions to workaround the issue:

-- Disable TLS False Start. (Note: this might not be feasible because it needs to be done on all clients.)
-- Disable SSL acceleration.
-- Disable AES-GCM ciphers in the client SSL profile. Without AES-GCM, clients do not try to use TLS false start and still be able to use (EC)DHE.

Fix:
The system no longer processes application data before verifying that the finished message arrives and handshake is complete.


701447-1 : CVE-2017-5754 (Meltdown)

Solution Article: K91229003


701445-1 : CVE-2017-5753 (Spectre Variant 1)

Solution Article: K91229003


701359-4 : BIND vulnerability CVE-2017-3145

Solution Article: K08613310


701327-2 : failed configuration deletion may cause unwanted bd exit

Component: Application Security Manager

Symptoms:
Immediately after the deletion of a configuration fails, bd exists.

Conditions:
When deleting a configuration fails.

Impact:
Unwanted bd restart.

Workaround:
None.

Fix:
bd will exit upon a failed configuration only when configured to exit on failure.


701288-1 : Server health significantly increases during DoSL7 TPS prevention

Component: Anomaly Detection Services

Symptoms:
Mitigation of DoSL7 TPS affects server health value.

Conditions:
-- DoSL7 TPS configured together with BADOS.
-- DoSL7 TPS is active.

Impact:
-- Incorrect Server Health reporting.
-- Might activate Behavioral DoS (BADoS) false-attack detection when attacks mitigated by DoSL7 TPS are stopped.

Workaround:
None.

Fix:
Server health now displays the actual backend server state, and does not incorrectly grow.


701253-5 : TMM core when using MPTCP

Solution Article: K16248201


701249-1 : RADIUS authentication requests erroneously specify NAS-IP-Address of 127.0.0.1

Component: TMOS

Symptoms:
RADIUS requests from BIG-IP have attribute NAS-IP-Address = 127.0.0.1, which might cause authentication to fail.

The NAS-IP-Address is essentially the resource an end user client is trying to authenticate to. This is typically the management IP address of the BIG-IP system, but the BIG-IP system always sends 127.0.0.1 instead. That might fail or it might work, depending on how the server is configured.

Conditions:
This is an issue for all RADIUS authentication requests that use the attribute NAS-IP-Address.

Note: This affects remote control plane authentication only, not APM or other uses of RADIUS.

Impact:
BIG-IP system always sends 127.0.0.1 instead of the BIG-IP system's management IP address. RADIUS server might not service the request, so authentication fails.

Workaround:
There is no workaround.


701244-1 : An incorrect data manipulation in cipher encrypt and decrypt could cause TMM crash with SIGABRT

Solution Article: K81742541

Component: Local Traffic Manager

Symptoms:
TMM receives SIGABRT from failover daemon, sod, due to heartbeat failure shortly after TMM starts up.

Conditions:
In some rare scenarios, TCP fast open encrypt/decrypt key may not be properly initialized when traffic comes into the BIG-IP system.

Impact:
Multiple TMM threads can get into a loop, causing heartbeat failure. TMM restarts, Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
The incorrect data manipulation in cipher encrypt and decrypt has been fixed.


701202-3 : SSL memory corruption

Solution Article: K35023432

Component: Local Traffic Manager

Symptoms:
In some instances random memory can be corrupted causing TMM core.

Conditions:
SSL is configured (either client-ssl or server-ssl) and SNI is used to select a non-default profile.

Impact:
TMM crash, disrupting traffic.

Workaround:
There is no workaround at this time.

Fix:
The memory corruption issue has been fixed.


701147-2 : ProxySSL does not work properly with Extended Master Secret and OCSP

Solution Article: K36563645

Component: Local Traffic Manager

Symptoms:
SSL handshake fails if the BIG-IP system is operating in ProxySSL mode, while client and server negotiate to use the Extended Master Secret and OCSP features together.

Conditions:
1. Virtual server is configured to work in ProxySSL mode.
2. Client and server negotiate the SSL handshake with the Extended Master Secret.
3. Client and Server negotiate to use the OCSP.

Impact:
ProxySSL does not work properly with Extended Master Secret and OCSP simultaneously.

Workaround:
None.

Fix:
Included the certificate status message in the calculation of Extended Master Secret.


701056-1 : User is not able to reset their Active Directory password

Component: Access Policy Manager

Symptoms:
When Active Directory is used for authenticating APM users and the user is required to change password on next APM logon, APM fails to update the password.

Conditions:
- APM is licensed and provisioned
- Active Directory is used for authenticating the users
- When logging on to APM, user is asked to change the password

Impact:
User is not able to change the password.

Workaround:
There is no workaround.

Fix:
APM end users can now successfully reset the password.


700895-1 : GUI Network Map objects in subfolders are not being shown

Solution Article: K34944451

Component: TMOS

Symptoms:
Objects created in subfolders under a partition are not showing up in the GUI Network Map when selecting the partition.

Conditions:
-- Create a virtual server under a subfolder.
-- View Network Map while /Common is the active partition.

For example:

1. Create a subfolder such as /Common/subfolder.
2. In that subfolder, create a virtual server such as /Common/subfolder/virtualserver1.
3. Select /Common as the partition.
4. View the Network Map.

The virtual server /Common/subfolder/virtualServer1 is not shown on the Network Map.

Impact:
Cannot see the objects in the subfolder.

Workaround:
Select the partition 'All[Read Only]' to see all objects in subfolders.


700889-3 : Software syncookies without TCP TS improperly include TCP options that are not encoded

Solution Article: K07330445

Component: Local Traffic Manager

Symptoms:
When sending a software syncookie and there is no TCP timestamp option, tmm sends back TCP options like window scaling (WS), sackOK, etc. The values for these options are encoded in the timestamp field which is not sent. When the final ACK of the 3WHS arrives (without a timestamp), there is no way to know that the BIG-IP system negotiated the use of SACK, WS and other options that were encoded in the timestamp. This will leave the client believing that options are enabled and the BIG-IP believing that they are not.

Conditions:
TCP timestamps are disabled by the client, or in the TCP profile.

Impact:
In one known case, the client was Windows 7 which apparently disables timestamps by default. Users might experience poor connection performance because the client believed it was using WS, and that the BIG-IP system would scale up the advertised window. However, the BIG-IP system does not using WS in this case, and used the window size from the TCP header directly, causing the BIG-IP system to send small packets (believing it had filled the window) and wait for a response.

Workaround:
Specifically prevent the WS issue by lowering the send_buffer_size and receive_window_size to less than or equal to 65535.

Fix:
Added dependency between the window scale option and the timestamp option in a SYN/ACK response.


700862-1 : tmm SIGFPE 'valid node'

Solution Article: K15130240

Component: Local Traffic Manager

Symptoms:
A rare TMM crash with tmm SIGFPE 'valid node' may occur if the host is unreachable.

Conditions:
The host is unreachable.

Impact:
Lack of stability on the device. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
This fix handles a rare TMM crash when the host is unreachable.


700812-1 : asmrepro recognizes a BIG-IP version of 13.1.0.1 as 13.1.0 and fails to load a qkview

Component: Application Security Manager

Symptoms:
asmrepro recognizes a BIG-IP version of 13.1.0.1 as 13.1.0 and fails to load a qkview.

Conditions:
Try to deploy a qkview on a BIG-IP using the asmrepro tool
while BIG-IP has a 4 element version like 13.1.0.1.

Impact:
Cannot deploy a 4 element version qkview on a BIG-IP using the asmrepro tool.

Workaround:
n/a

Fix:
asmrepro now handles the version number properly.


700757-1 : vcmpd may crash when it is exiting

Component: TMOS

Symptoms:
vcmpd may crash when it is exiting. The system logs an error message similar to the following in /var/log/ltm:

err vcmpd[14604]: 01510000:3: Uncaught exception: basic_string::_S_create

It's possible that vcmpd will then not start up, logging errors similar to the following in /var/log/ltm:

umount(/var/tmstat-vcmp/<guest name>) failed: (16, Device or resource busy

Conditions:
vCMP must be in use.

Impact:
vcmpd cores, but does so when already exiting. It is possible that vcmpd will then be unable to restart itself, and will need to be manually restarted.

Workaround:
If vcmpd cannot restart itself, you can manually restart it by running the following command:

tmsh restart sys service vcmpd

Fix:
Prevented vcmpd from crashing when exiting.


700726-2 : Search engine list was updated, and fixing case of multiple entries

Component: Application Security Manager

Symptoms:
Default search engine list does not identify current search engines, and blocks traffic unnecessarily. Part of the issue is that when adding custom search engines, there may be multiple search engines which match the User-Agent header, and this causes the match to fail.

Conditions:
Site accessed by search engines.

Impact:
Traffic from search engines is blocked unnecessarily.

Workaround:
Manually add search engines.

Fix:
Search engine list has been updated to reflect current common search engine usage. Also, this version removes the check of multiple search engines, so that now when multiple Search Engines are matched, the Search Engine bypasses the challenges.


700724-2 : Client connection with large number of HTTP requests may cause tmm to restart

Component: Access Policy Manager

Symptoms:
tmm may restart while processing client request

Conditions:
- PingAccess profile is configured on the virtual server.

- Client connection sends over 64k HTTP requests that result in BIG-IP's connection to the PingAccess policy server.

Impact:
Traffic will be disrupted while TMM restarts.

Workaround:
Modify HTTP profile used by affected virtual to specify the limit of HTTP requests per connection "maximum requests per connection" to be less then 64k, e.g. 63000 or less.

Fix:
Traffic will no longer be disrupted when client sends over 64k uncached requests on the same TCP connection.


700696-1 : SSID does not cache fragmented Client Certificates correctly via iRule

Component: Local Traffic Manager

Symptoms:
The last few bytes of a very large-sized Client Certificate (typically greater than 16,384 bytes) are not cached correctly if the certificate is received fragmented by the SSL Session ID (SSID) parser.

Conditions:
-- Client Authentication is enabled.
-- A very large Client Certificate is supplied (typically greater than 16,384 bytes).
-- SSL Session ID Persistence is enabled.
-- The iRule CLIENTSSL_CLIENTCERT is enabled.

Impact:
The client certificate is not stored on the BIG-IP device correctly. The last few bytes are missing.

Workaround:
Disable the CLIENTSSL_CLIENTCERT iRule when SSL Session ID (SSID) persistence is in use. Even though the Client Certificate does not get cached, that is preferable to caching an incorrect client certificate.

Fix:
This release supports caching of fragmented client certificates in the SSL Session ID (SSID) persistence feature to properly cache very large-size client certificates (typically exceeding 16,384 bytes).


700597-1 : Local Traffic Policy on HTTP/2 virtual server no longer matches

Component: Local Traffic Manager

Symptoms:
Local Traffic Policies may not match properly when a virtual server is handling HTTP/2 traffic.

Conditions:
Virtual server with Local Traffic Policy and HTTP/2 profile.

Impact:
Traffic fails to pass through the virtual server, or fails to be processed as expected.

Workaround:
If able, use HTTP rather than HTTP/2. Or disable the policy. Otherwise there is no workaround.

Fix:
Traffic now processed as expected.


700576-1 : GUI - Server SSL Profile shows irrelevant options when "Server Certificate" is set to "Ignore"

Component: TMOS

Symptoms:
In the GUI, the ServerSSL Profile options "Expire Certificate Response Control" and "Untrusted Certificate Response Control" are shown as stand alone options, yet those settings are not honored when the "Server Certificate" option is set to "Ignore" (default).

Conditions:
Create server SSL profile with "Server Certificate" option is set to "Ignore" (default).
It shows "Expire Certificate Response Control" and "Untrusted Certificate Response Control" options, yet those settings are not honored.

Impact:
No functional Impact, it may cause confusion allowing view/modify for irrelevant options.

Workaround:
No functional Impact, Expire Certificate Response Control" and "Untrusted Certificate Response Control" options can be ignored when "Server Certificate" option is set to "Ignore" (default).

Fix:
"Expire Certificate Response Control" and "Untrusted Certificate Response Control" server SSL profile options are hidden when "Server Certificate" option is set to "Ignore" (default).


700571-4 : SIP MR profile, setting incorrect branch param for CANCEL to INVITE

Component: Service Provider

Symptoms:
BIG-IP SIP profile MR does not maintain the Via 'branch parameter' ID when the Via header insertion is enabled for INVITE and CANCEL for the same INVITE.

Conditions:
This happens only when the following conditions are both met:
-- The transport connection that issued INVITE has been terminated.
-- A new transport is used to issue CANCEL

Impact:
The result is different branch IDs for the BIG-IP system-generated Via header. INVITE is only cancelled on the calling side, while on the called side, the line will ring until time out.

Workaround:
None.

Fix:
The branch parameter value calculation now remains consistent throughout the connection.


700556-1 : TMM may crash when processing WebSockets data

Solution Article: K11718033


700527-3 : cmp-hash change can cause repeated iRule DNS-lookup hang

Component: Global Traffic Manager (DNS)

Symptoms:
An iRule that uses RESOLV::lookup can hang repeatedly when cmp-hash configuration is changed.

Conditions:
-- iRule is in the middle of a call to RESOLV::lookup.
-- A change is made to VLAN cmp-hash configuration.

Impact:
The iRule call can hang repeatedly.

Workaround:
Restart the TMM. This will interrupt client traffic while TMM restarts.

Fix:
The iRule connection is reestablished when the pending query expires, so subsequent RESOLV::lookup calls do not hang per TMM.


700522-1 : APMD may unexpectedly restart when worker threads are stuck

Component: Access Policy Manager

Symptoms:
APMD restarts and logs a message about all threads being stuck.

Conditions:
A race condition allows the busy thread count to remain higher than the actual value. If it reaches the maximum thread count, APMD will restart.

Impact:
APMD can restart unexpectedly.

Workaround:
There is no workaround.

Fix:
A rare APM timing condition leading to an unexpected restart of services has been corrected.


700433-1 : Memory leak when attaching an LTM policy to a virtual server

Solution Article: K10870739

Component: Local Traffic Manager

Symptoms:
BIG-IP LTM policies may cause an mcpd process memory leak.

As a result of this issue, you may encounter one or more of the following symptoms:

-- Latency when configuring the BIG-IP system.
-- Error messages logged in /var/log/ltm similar to the following example:
01140029:4: HA daemon_heartbeat mcpd fails action is restart.
-- The mcpd process may generate a core file in the /var/core directory.

Conditions:
This issue occurs when all of the following conditions are met:

-- Your configuration includes one or more virtual servers with an associated BIG-IP LTM policy.
-- The BIG-IP LTM policy has at least one rule.
Note: Rules with actions or conditions can leak increased amounts of memory.

-- You delete and add BIG-IP LTM policies that are associated with the virtual server.
Note: This modification causes the memory leak to increase over time.

Impact:
The mcpd process might run slower as memory is consumed, and can fail when all system memory is exhausted. Devices in a high availability (HA) configuration may experience a failover event.

Workaround:
None.

Fix:
The system now prevents MCP from leaking memory when attaching an LTM policy to a virtual server.


700426 : Switching partitions while viewing objects in GUI can result in empty list

Solution Article: K58033284

Component: TMOS

Symptoms:
In LTM Pool List, Node List, and Address Translations pages, switching partitions while viewing objects in the GUI can result in empty list.

Conditions:
This issue is present when all of the following conditions are met:
-- One partition contains multiple pages of objects.
-- The page count in one partition is greater than the page count in another.
-- The active page number is greater than 1.
-- You switch to a partition whose max number of pages is lower than the active page number.

For example, in the GUI:
1. Create two non-Common partitions.
2. In one partition, create enough pools so that they do not fit on one page.
3. In the second partition, create only enough pools for one page.
4. On the Local Traffic :: Pools list page in the first partition, navigate to the second page of objects.
5. Switch to the other partition.
6. Note that the displayed page contains no objects.

Impact:
The list of pools is empty despite the fact that there are pools available.

Workaround:
Return to the first page of objects before switching to any other partition.

Fix:
The system now resets to the first page if the page number is greater than the page count, so the partition's objects display correctly.


700393-3 : Under certain circumstances, a stale HTTP/2 stream can cause a tmm crash

Component: Local Traffic Manager

Symptoms:
Tmm might crash due to a stale/stalled HTTP/2 stream.

Conditions:
HTTP/2 profile in use.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
Stale/stalled HTTP2 streams are handled correctly to prevent a tmm crash.


700386-2 : mcpd may dump core on startup

Component: TMOS

Symptoms:
mcpd may generate a core file upon startup, with a message being logged that overdog restarted it because mcpd failed to send a heartbeat for five minutes.

Conditions:
This can happen only at startup.

Impact:
mcpd restarts, but resumes normal operation.

Workaround:
None.

Fix:
mcpd no longer generates a core on startup.


700320 : tmm core under stress when BADOS configured and attack signatures enabled

Component: Anomaly Detection Services

Symptoms:
Tmm core under stress. Note: This issue has a very low probability of occurring.

Conditions:
-- Out of memory.
-- BADOS configured.
-- Attack signatures enabled.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
None, except to not configure attack signatures.

Fix:
Added protection for the case when context adm_filters allocation is failed.


700315-2 : Ctrl+C does not terminate TShark

Solution Article: K26130444

Component: TMOS

Symptoms:
A running TShark process on the BIG-IP system has problems exiting when the user is finished capturing and presses CTRL+C while listening on a tmm VLAN or 0.0.

Conditions:
-- Using TShark on a tmm VLAN or 0.0.
-- Pressing Ctrl+C to exit.

Impact:
TShark does not exit as expected when pressing CTRL+C.

Workaround:
To recover, you can move the process to the background (CTRL+Z) and then halt the process, by running a command similar to the following: kill -9 'pidof tshark'

Fix:
Ctrl+C now terminates TShark as expected.


700250-3 : qkviews for secondary blade appear to be corrupt

Solution Article: K59327012

Component: TMOS

Symptoms:
Normally, a qkview created on a multi-blade system will include a qkview for every blade on the system. Those qkview files have an error that makes the tar-ball appear corrupt.

Conditions:
Running a qkview on a system with blades.
Attempt to access the tar-ball using the following command: tar -tvzf 127.3.0.2.qkview | tail.

Impact:
The system posts the following messages:
    gzip: stdin: unexpected end of file
    tar: Child returned status 1
    tar: Error is not recoverable: exiting now


Confuses troubleshooters into thinking that the blade qkview files are corrupt, which they are not.

Workaround:
None.

Fix:
By not always writing an errant newline, the problem is solved.


700247 : APM Client Software may be missing after doing fresh install of BIG-IP VE

Solution Article: K60053504

Component: TMOS

Symptoms:
apm client software checks is broken in VM created with BIG-IP-13.1.0.1.0.0.8.ALL-scsi.ova.

Conditions:
Any software instance created by deployment of any OVA for the affected software versions.

Impact:
APM endpoint inspection feature (for Mac, windows and Linux clients). [Users affected]
Configuration of APM client software check APM Visual policy editor. [Admin UI]
APM Client package @ Connectivity / VPN : Connectivity : Profiles if you select "Web Browser Add-ons for BIG-IP Edge Client" option. [Admin UI]

Workaround:
Try the "epsec refresh" commands again after removing all environment locks on the shared RPM database using the following command:

rm /shared/lib/rpm/__db.*
epsec refresh

Fix:
After deployment of a new OVA for the fixed version(s), the problem no longer occurs.


700143-2 : ASM Request Logs: Cannot delete second 10,000 records of filtered event log messages

Component: Application Security Manager

Symptoms:
Attempting to delete request logs by a filter (10,000 at a time), only works once. Subsequent delete by filter actions do not remove additional earlier logs.

Conditions:
System has over 10,000 ASM request logs by a selected filter, and delete all by filter is used multiple times.

Impact:
Only the latest 10,000 events are deleted.

Workaround:
No work around for deleting by a filter.
Delete all (without filter) works, and deleting selected requests works.

Fix:
Deletion by filter correctly deletes subsequent sets of 10,000 rows per action.


700090-2 : tmm crash during execution of a per-request policy when modified during execution.

Component: Access Policy Manager

Symptoms:
Modify/delete of per-request policy during heavy traffic flow causes tmm to crash.

Conditions:
While a per-request policy (macro) is getting executed.
Admin deletes the parent policy item (at the same time).

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround other than not deleting per-request policies during heavy traffic flow.

Fix:
Per-request access policies edited during execution are now held until not in use, so this issue no longer occurs.


700086-1 : AWS C5/M5 Instances do not support BIG-IP VE

Component: TMOS

Symptoms:
BIG-IP Virtual Edition (VE) does not function on AWS C5/M5 instances.

Conditions:
BIG-IP VE on AWS C5/M5 instances.

Impact:
Cannot use BIG-IP VE on AWS C5/M5 instances.

Workaround:
None.

Fix:
BIG-IP VE now functions on AWS C5/M5 Instances.


700061-4 : Restarting service MCPD or rebooting BIG-IP device adds 'other' file read permissions to key file

Component: Local Traffic Manager

Symptoms:
Restarting service MCPD or rebooting the BIG-IP device adds Unix 'other' read file permissions for key files.
Key files Unix permission changes from '-rw-r-----' to
'-rw-r--r--'

Conditions:
1. Restarting the service MCPD
2. Rebooting BIG-IP device.

Impact:
Key file Unix read permission changes from '-rw-r-----' to
'-rw-r--r--'

Workaround:
There is no workaround at this time.

Fix:
Fix made sure restart of service MCPD and reboot of device does not change key files Unix read permissions from '-rw-r-----' to
'-rw-r--r--'


700057-4 : LDAP fails to initiate SSL negotiation because client cert and key associated file permissions are not preserved

Component: Local Traffic Manager

Symptoms:
After upgrading to an affected build, the default key will have incorrect group ownership.

Conditions:
Upgrade or load a .ucs with SSL keys configured.

Impact:
File permissions are not preserved in the .ucs file. The httpd process will not be able to use the default key, so anything using it will fail.

Workaround:
Run the following two commands:
tmsh save /sys config
tmsh load /sys config

Fix:
The system now preserves correct permissions for default.key across upgrade and ucs load.


700056-1 : MCPD process may lock up and restart when applying Local Traffic Policy to virtual server

Component: Local Traffic Manager

Symptoms:
The mcpd process becomes unresponsive while attempting to load the configuration, or attempting to apply a configuration change that associates a particular local traffic policy to a virtual server. The mcpd process may also periodically restart every 5 minutes.

Conditions:
- Virtual servers configured.
- Specific policies configured and attached used by virtual servers.

Only policies configured in a specific way will trigger this situation. The specifics are not well understood, however, the vast majority of Local Traffic Policies are unaffected by this issue.

Impact:
Traffic disrupted while mcpd restarts.

Workaround:
There is no workaround.

Fix:
mcpd is responsive when an LTM Policy is attached to a virtual server, or when an already-attached policy is updated.


699720-1 : ASM crash when configuring remote logger for WebSocket traffic with response-logging:all

Component: Application Security Manager

Symptoms:
ASM may crash when configuring remote logger for WebSocket traffic virtual server.

Conditions:
-- Virtual server handling WebSocket traffic.
-- ASM remote logger on the same virtual server.

Impact:
ASM crash; system goes offline.

Workaround:
Use either of the following workarounds:

-- Remove remote logger.
-- Have response logging for illegal requests only.

Fix:
The system now handles memory correctly and avoids crashing in this specific scenario.


699686-1 : localdbmgr can occasionally crash during shutdown

Component: Access Policy Manager

Symptoms:
When localdbmgr process is restarted, occasionally, the process crashes and a core file will be generated.

Conditions:
-- APM is provisioned.
-- localdbmgr process is restarted.

Impact:
Although the process restarts, there is no impact to the APM functionality.

Workaround:
None.

Fix:
localdbmgr no longer crashes during shutdown.


699624-1 : Config with custom 'SIP' or 'Firepass' monitor fails to load after upgrade

Solution Article: K93400155

Component: Local Traffic Manager

Symptoms:
A configuration that contains custom 'SIP' or 'FirePass' monitors that is upgraded from a version earlier than v13.1.0 may either fail to load, or may result in a configuration that loads the first time after the upgrade, but cannot be re-loaded from the text config files.

If the BIG-IP system has partitions other than 'Common', the initial configuration load may fail with an error such as:

01070726:3: monitor /Common/sip-monitor in partition Common cannot reference SSL profile monitor parameter /Common/sip-monitor 1 SSL_PROFILE_NAME= in partition name-of-other-partition

If the BIG-IP system only has a 'Common' partition, the initial configuration load will succeed, but subsequent attempts to load the configuration (e.g., 'tmsh load sys config') may fail with this error:

Syntax Error:(/config/bigip.conf at line: 63) "user-defined" unknown property

Which corresponds to a SIP or FirePass monitor in the configuration such as:

ltm monitor sip /Common/test_sip_monitor {
    cipherlist DEFAULT:+SHA:+3DES:+kEDH
    compatibility enabled
    debug no
    defaults-from /Common/sip
    destination *:*
    filter 488
    interval 5
    mode tcp
    time-until-up 0
    timeout 16
    user-defined SSL_PROFILE_NAME /Common/test_sip_monitor_ssl_profile
}

Conditions:
Custom 'SIP' or 'FirePass' monitor is configured, and the config is upgraded from a version earlier than v13.1.0 to version v13.1.0.

Impact:
After upgrade, the configuration fails to load with an error such as:

01070726:3: monitor /Common/sip-monitor in partition /Common cannot reference SSL profile monitor parameter /Common/sip-monitor 1 SSL_PROFILE_NAME= in partition name-of-other-partition.

Alternatively, the configuration loads after upgrade, but the config file is corrupted, and will fail to load (such as after a system restart, or upon explicit 'tmsh load sys config'), with an error such as:

Syntax Error:(/config/bigip.conf at line: 63) "user-defined" unknown property

Workaround:
Remove custom 'SIP' and 'FirePass' monitors from the configuration, and re-create them manually after upgrade is complete.

Fix:
In this release, a configuration that contains a custom 'SIP' or 'FirePass ' monitor from a version earlier than v13.1.0 now loads correctly and continues to load as expected.


699598-2 : HTTP/2 requests with large body may result in RST_STREAM with FRAME_SIZE_ERROR

Component: Local Traffic Manager

Symptoms:
HTTP/2 requests with a large body may result in the BIG-IP system sending RST_STREAM with FRAME_SIZE_ERROR.

Conditions:
-- HTTP/2 profile is configured on the virtual server.
-- Request has body size greater than 16 KB.

Impact:
HTTP/2 stream is reset with FRAME_SIZE_ERROR and transfer does not complete.

Workaround:
None.

Fix:
Large HTTP/2 requests are now processed as expected.


699531-1 : Potential TMM crash due to incorrect number of attributes in a PEM iRule command

Component: Policy Enforcement Manager

Symptoms:
TMM crash.

Conditions:
An iRule with a PEM iRule command that does not have the minimum number of attributes.

Impact:
Loss of service. Traffic disrupted while tmm restarts.

Workaround:
Make sure the PEM iRule command is called with at least the minimum number of parameters.

For example, make sure to call 'PEM::subscriber config policy referential set' with referential policies.

Fix:
The system now provides the correct number of attributes, preventing a NULL pointer dereference.


699455-4 : SAML export does not follow best practices

Solution Article: K50254952


699454-4 : Web UI does not follow current best coding practices

Component: Advanced Firewall Manager

Symptoms:
The web UI does not follow current best coding practices while processing URL DB updates.

Conditions:
ASM provisioned.

Impact:
UI does not respond as intended.

Workaround:
None.

Fix:
The web UI now follows current best coding practices while processing URL DB updates.


699453-4 : Web UI does not follow current best coding practices

Solution Article: K20222812


699452-4 : Web UI does not follow current best coding practices

Component: Advanced Firewall Manager

Symptoms:
The web UI does not follow current best coding practices while processing PEM configuration updates.

Conditions:
PEM provisioned.

Impact:
UI does not respond as intended.

Workaround:
None.

Fix:
The web UI now follows current best coding practices while processing PEM configuration updates.


699451-3 : OAuth reports do not follow best practices

Solution Article: K30500703


699346-3 : NetHSM capacity reduces when handling errors

Solution Article: K53931245


699339-3 : Geolocation upgrade files fail to replicate to secondary blades

Solution Article: K24634702

Component: Global Traffic Manager (DNS)

Symptoms:
Geolocation upgrade files fail to replicate to secondary blades.

Conditions:
-- Multiblade VIPRION platforms/vCMP guests.
-- Upgrading geolocation files on the primary blade.
-- Viewing the geolocation files on the secondary blades.

Impact:
Geoip database is not updated to match primary blade.

Workaround:
Use either of the following workarounds:

-- Use the root account to manually create the /shared/GeoIP/v2 directory on secondary blades, and then run geoip_update_data on primary blade.

-- On primary blade:
1. Edit /etc/csyncd.conf as shown below.
2. Restart csyncd.
3. Run geoip_update_data.

To edit /etc/csyncd.conf:

Merge the following two terms:
 monitor dir /shared/GeoIP {...)
 monitor dir /shared/GeoIP/v2 {...}

into one term, as follows:
monitor dir /shared/GeoIP {
        queue geoip
        pull pri2sec
        recurse yes
        defer no
        lnksync yes
        md5 no
        post "/usr/local/bin/geoip_reload_data"
}

Fix:
Geolocation upgrade files now correctly replicate to secondary blades.


699298-2 : 13.0.0 Hotfix HF3 3.0.1679 TMM CORED due to SIGSEGV.

Component: Local Traffic Manager

Symptoms:
TMM may crash when woodside congestion-control is in use.

Conditions:
When woodside congestion-control is in use.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Other congestion control algorithms can be used as a workaround.

Fix:
This fix handles a rare TMM crash when woodside congestion-control is in use.


699281-1 : Version format of hypervisor bundle matches Version format of ISO

Component: TMOS

Symptoms:
Recently F5 incorporated 4th element into versioning scheme. 4th and 5th are separated by dash (instead of dot) in ISO name. This change/bug insures that names of hypervisor bundles also use dash between 4th and 5th elements.

Conditions:
Applies to hypervisor bundles (for example ova files for vmware).

Impact:
Version format in names of hypervisor bundles matches version format of ISO file

Workaround:
Version format in names of hypervisor bundles matches version format of ISO file

Fix:
Version format in names of hypervisor bundles matches version format of ISO file (usage of dash between 4th and 5th elements).


699273-1 : TMM Core During FTP Monitor Use

Component: Local Traffic Manager

Symptoms:
TMM Cores.

Conditions:
When the FTP monitor is used.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Turn off FTP monitoring.

Fix:
The tmm no longer cores when using a FTP monitor.


699267-2 : LDAP Query may fail to resolve nested groups

Component: Access Policy Manager

Symptoms:
LDAP Query agent may fail to resolve nested groups for a user.
/var/log/apm logfile contains the following error messages when 'debug' log level is enabled for Access Profile:
err apmd[17159]: 014902bb:3: /Common/ldap_access:Common:254fdc14 Failed to process the LDAP search result while getting group membership down with error (No such object.).
err apmd[17159]: 014902bb:3: /Common/ldap_access:Common:254fdc14 Failed to process the LDAP search result while querying LDAP with error (No such object.).

Conditions:
LDAP Query agent is configured in an Access Policy.
'Fetch groups to which the user or group belong' option is enabled

Impact:
LDAP Query agent fails.
unable to get user identity.
unable to finalize Access Policy.

Fix:
after fix, LDAP Query resolves all nested groups as expected and session.ldap.last.attr.memberOf attributes contains user's groups


699135-1 : tmm cores with SIGSEGV in dns_rebuild_response while using host command for not A/AAAA wideip

Component: Global Traffic Manager (DNS)

Symptoms:
tmm cores with SIGSEGV in dns_rebuild_response.

Conditions:
1. Create a wideip of type other than A/AAAA.
2. Create a iRule with a host command for the previously created wideip.
3. Create a related zonerunner record with the same name and type.
4. Dig against the wideip with type any.
5. Observe that tmm cores.

Impact:
tmm cores.

Workaround:
Don't use host command for non type A/AAAA wideips.


699103-1 : tmm continuously restarts after provisioning AFM

Component: Traffic Classification Engine

Symptoms:
tmm continuously restarts when the Webroot database is getting downloaded to a BIG-IP system with less than 16 GB RAM and AFM provisioned.

Conditions:
-- Webroot URL categorization configured for Traffic Classification.
-- BIG-IP system with less than 16 GB RAM.
-- AFM is provisioned.

Impact:
tmm restarts. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround other than to ensure that more than 16 GB RAM is available when AFM is provisioned.

Fix:
The BIG-IP system with less than 16 GB RAM and AFM provisioned now prevents downloading the Webroot database or any updates if it is not already downloaded.

Note: If the Webroot database already exists before upgrade to this release, Webroot lookup will continue to work.


699012-1 : TMM may crash when processing SSL/TLS data

Solution Article: K43121447


698992-1 : Performance degraded

Component: Performance

Symptoms:
Portal access performance had a slight performance degradation. This was identified to be due to a new queuing strategy implemented to improve per-request policy auth use-case performance for higher end platforms in the 13.0 release. The nature of the problem is such that overall system degradation may be observed if APM is provisioned and per-request policy is not used.

Conditions:
APM is provisioned, but functionality is not related to per-request policy.

Impact:
Performance will be slightly lower under load.

Workaround:
None.

Fix:
The queuing strategy was altered to take minimal CPU resources when idle.


698984-1 : Auto-disable TMM.HTTP.TCL.Validation when APM is provisioned

Component: Access Policy Manager

Symptoms:
The db variable Tmm.HTTP.TCL.Validation is enabled by default. This db variable should be disabled when APM is provisioned/enabled, and when ACCESS::restrict_irule_event is disabled and HTTP_RESPONSE_RELEASE events are detected with the assigned iRules.

Conditions:
Steps to Reproduce:
1. Define the following iRule in the virtual server.

when CLIENT_ACCEPTED {
    ACCESS::restrict_irule_events disable
}
when HTTP_REQUEST {
    set u [ HTTP::uri ]
    log local0. "XXX: [ HTTP::uri ]"
}
when HTTP_RESPONSE_RELEASE {
    log local0. "XXX: [ HTTP::status ] [ HTTP::header Location ]"
    set l [ HTTP::header Location ]
    if { $l starts_with {/my.policy} } {
       append l {?modified_by_irule=1}
       HTTP::header replace Location $l
    } elseif { $l starts_with {/renderer/agent_logon_page_form.eui} } {
        # Next response will be the real response to the client.
        ACCESS::log "XXX: lp_seen"
        set lp_seen 1
    }
    if { [ HTTP::status ] == 200 && [ info exists lp_seen ] && $lp_seen == 1 } {
        unset lp_seen
        HTTP::header insert X-MyAppSpecialHeader 1
    }
}
2. Configure START :: LOGON PAGE :: ALLOW policy.
3. Access the virtual server.

Impact:
TCP reset triggered when it should not. With respect to the specific condition described, the system should post the logon page.

Workaround:
Manually disable Tmm.HTTP.TCL.Validation.

Fix:
Tmm.HTTP.TCL.Validation is now disabled automatically when APM provisioned during the upgrades. This is correct behavior.


698940-1 : Add new security policy template for API driven systems - "API Security"

Component: Application Security Manager

Symptoms:
No security policy template for API Security for API driven systems.

Conditions:
-- Using API.
-- Attempting to define REST API protection, Web Socket protection.

Impact:
No policy template.

Workaround:
None.

Fix:
Added new security policy template for API driven systems - 'API Security'.


698919-3 : Anti virus false positive detection on long XML uploads

Component: Application Security Manager

Symptoms:
A false positive virus-detected violation. The description of the violation explains that the ICAP server was not contacted.

Conditions:
-- A long XML upload or payload.
-- The assigned XML profile is configured to be inspected by the ICAP server.

Impact:
Violation is detected where no violation has occurred (false positive violation).

Workaround:
Increase the internal parameter max_raw_request_len to the required length of the XML.

Note: This workaround will affect the amount of logged data from ASM.

Fix:
Fixed a false positive virus-detected violation related to long XML uploads.


698916-1 : TMM crash with HTTP/2 under specific condition

Component: Local Traffic Manager

Symptoms:
TMM may crash when upgrading an HTTP/1.1 connection to an HTTP/2 connnection.

Conditions:
-- HTTP/2 gateway enabled.
-- Pool member supports protocol switching.

Impact:
TMM crash, leading to a failover event.

Workaround:
There is no workaround other than removing the HTTP/2 profile from the virtual server.

Fix:
TMM properly handles protocol upgrade requests from pool members when HTTP/2 is enabled, so no crash occurs.


698875-1 : Qkview Security Hardening

Component: TMOS

Symptoms:
Qkview does not follow best practices for sanitizing and anonymizing collected data

Conditions:
Qkview created

Impact:
Under certain conditions, Qkviews may include sensitive information, which may in turn be uploaded to iHealth

Workaround:
None.

Fix:
Qkview now follows best practices for sanitizing and anonymizing collected data


698813-2 : When processing DNSX transfers ZoneRunner does not enforce best practices

Solution Article: K45435121


698461-1 : tmm may crash in fastl4 TCP

Component: Local Traffic Manager

Symptoms:
tmm crash and BIGIP fail over.

Conditions:
Virtual with fastl4 and TCP profile configured and used.
LRO is used.

Impact:
tmm may crash

Fix:
the crash is fixed.


698429-1 : Misleading log error message: Store Read invalid store addr 0x3800, len 10

Component: TMOS

Symptoms:
On rare occasions, messages like these can appear in /var/log/ltm:

Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Loading keys from the file.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Read the unit key file if exists.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Unit key hash from key header: 9d:e6:90:...
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Unit key hash computed from read key:9d:e6:90:...
Oct 20 14:21:04 localhost err mcpd[4139]: 010713d0:3: Symmetric Unit Key decrypt failure - decrypt failure
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071029:5: Symmmetric Unit Key decrypt
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071027:5: Master key OpenSSL error: 1506270960:error:06065064:digital envelope routines:EVP_DecryptFinal_ex:bad decrypt:evp_enc.c:587:
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Attempting Master Key migration to new unit key.
Oct 20 14:21:04 localhost warning chmand[5559]: 012a0004:4: Store Read invalid store addr 0x3800, len 10
Oct 20 14:21:04 localhost warning chmand[5559]: 012a0004:4: Store Read invalid store addr 0x3800, len 10
Oct 20 14:21:04 localhost warning mcpd[4139]: 012a0004:4: halStorageRead: unable to read storage on this platform.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071029:5: Cannot open unit key store
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Reloading the RSA unit to support config roll forward.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Loading keys from the file.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Read the unit key file if exists.
Oct 20 14:21:04 localhost notice mcpd[4139]: 01071038:5: Unit key hash from key header: 9d:e6:90:...

These messages are due to a transient failure reading the system's unit key from the virtual EEPROM on a vCMP guest. The error handling from this failure causes misleading messages, but it does successfully read the unit key during this error recovery.

Conditions:
These messages can only happen on a vCMP guest with SecureVault, and are very rare.

Impact:
None. These messages do not indicate an actual problem with the system.


698424-1 : Traffic over a QinQ VLAN (double tagged) will not pass

Solution Article: K11906514

Component: Local Traffic Manager

Symptoms:
Traffic on a QinQ VLAN will not pass.

Conditions:
This issue exists when a VLAN is configured as a QinQ VLAN (i.e., a double-tagged VLAN).

Impact:
Traffic on a QinQ VLAN will not pass.

Workaround:
Disabling LRO may workaround this issue.

Fix:
Traffic on a QinQ VLAN now passes successfully.


698396-1 : Config load failed after upgrade from 12.1.2 to 13.x or 14.x

Component: Traffic Classification Engine

Symptoms:
Sys load fails with following errors,
....
Loading schema version: 14.0.0
0107153e:3: Application id out of the valid range of [8192-16384).
Unexpected Error: Loading configuration process failed.

Conditions:
When an CEC IM is applied to 12.1.2 and then when we upgrade to 13.x or 14.x, sys load will fail.

Impact:
System will fail to come to Active state after upgrade.

Workaround:
It can be fixed by manually deleting /var/libdata/dpi/conf/classification_update.conf


698379-2 : HTTP2 upload intermittently is aborted with HTTP2 error error_code=FLOW_CONTROL_ERROR(

Solution Article: K61238215

Component: Local Traffic Manager

Symptoms:
Uploads for the HTTP2 virtual server fail intermittently with HTTP2 error error_code=FLOW_CONTROL_ERROR.

Conditions:
HTTP2 virtual server configured.

Impact:
Uploads for the HTTP2 virtual server might fail intermittently.

Workaround:
None.

Fix:
Uploads for the HTTP2 virtual server do not fail intermittently anymore.


698338-1 : Potential core in MRF occurs when pending egress messages are queued and an iRule error aborts the connection

Component: Service Provider

Symptoms:
The system may core.

Conditions:
-- Egress messages are queued waiting for MR_EGRESS event to be raised.
-- Current MR event exits with an error, thus aborting the connection.

Impact:
The system cores and will restart.

Workaround:
None.

Fix:
The system now returns pending messages back to the originator if the connection aborts due to an iRule error.


698333-1 : TMM core upon matching a past (saved) dynamic signature of a specific family (network or dns) after the dynamic signature is disabled for that family on the parent context (but still enabled for other families)

Solution Article: K43392052

Component: Advanced Firewall Manager

Symptoms:
TMM would core upon matching a past (saved) dynamic signature of a specific family (network or dns) after the dynamic signature is disabled for that family on the parent context (but still enabled for other families).

Conditions:
This occurs in the following scenario:
-- Enable Network and DNS BDOS simultaneously (on DoS Device config).
-- Generate dynamic signature that has both network and DNS metrics.
-- Wait for signature to be moved to 'past' (persist) state.
-- Disable either network or DNS BDOS (but not both).
-- TMM cores if the traffic matches this signature.

Impact:
Traffic interruption due to TMM restart. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
In this release, if the dynamic signature is disabled for a specific family on a parent context (but not disabled for other family on that context), any past attack signature for the context is now deleted from the system.


698226-1 : Filters under Security :: Reports :: DoS :: URL Latencies work incorrectly

Component: Application Visibility and Reporting

Symptoms:
When filtering data by a field in the 'Security :: Reports :: DoS :: URL Latencies' form, the filtering fails and the monpd process crashes.

Conditions:
There is some statistical data for DoS.

Impact:
Reports based on GUI filters are not complete.

Workaround:
No workaround.

Fix:
The system now creates the correct query for this filter, so the issue no longer occurs.


698182 : Upgrading from 13.1.1 to newer release might cause config to not be copied over

Component: Advanced Firewall Manager

Symptoms:
Upgrading from 13.1.1 to newer release might cause config to not be copied over. This is due to the UUID being available on the older release but not on the newer one.

Conditions:
Upgrade or loading a UCS from 13.1.1 to newer release.

Impact:
Config cannot be loaded or fails.

Workaround:
Copy config and remove UUID-specific schema before loading the config.

Fix:
When upgrading to a version in which UUID is not supported, the system now automatically copies the config and removes UUID-specific schema before loading it.


698084-3 : IPsec log messages in /var/log/ltm missing module ID to reach bigiq logs

Solution Article: K03776801

Component: TMOS

Symptoms:
Some groups of messages logged by tmipsecd are missing the errdefs annotation that identifies IPsec as the module. Messages reported when tunnels go up and down, or problems with listeners, go only to ltm logs, with no visibility to bigiq logs.

Conditions:
Missing the IPsec module subset ID.

Impact:
Missing IPsec messages in the bigiq logs.

Workaround:
No workaround at this time.

Fix:
The IPsec module subset ID has been added to tmipsecd log messages, so those messages will reach bigiq logs. Some log messages previously appearing only in /var/log/ltm now also appear in ipsec.log and also reach bigiq logs.


698080-3 : TMM may consume excessive resources when processing with PEM

Solution Article: K54562183


698000-3 : Connections may stop passing traffic after a route update

Solution Article: K04473510

Component: Local Traffic Manager

Symptoms:
When a pool is used with a non-translating virtual server, routing updates may lead to an incorrect lookup of the nexthop for the connection.

Conditions:
-- Pool on a non-translating virtual server.
-- Routing update occurs.

Impact:
Connections may fail after routing updates. New connections will not be affected.

Workaround:
Use a route to direct traffic to the ultimate destination rather than using a pool to indicate the nexthop.

Fix:
Routing updates no longer interrupts traffic to connections using a pool member to reach the nexthop.


697988-3 : During config sync, if many client-ssl profiles are on a virtual server the CPU may briefly spike to 100%

Component: Local Traffic Manager

Symptoms:
During config sync, if many (hundreds) of client-ssl profiles are attached to a virtual server, the CPU may spike to 100%.

Conditions:
-- Many (hundreds) of client-ssl profiles are attached to a virtual server.
-- Config sync is executed.

Impact:
If enough client-ssl profiles are attached, the watchdog could fire, crashing tmm and causing service disruption. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround other than not attaching hundreds of client-ssl profiles to a virtual server, or disabling config sync.

Fix:
Issue no longer occurs when there are 2000+ client-ssl profiles attached to a virtual server and config sync is executed.


697756-1 : Policy with CSRF URL parameter cannot be imported as binary policy file

Component: Application Security Manager

Symptoms:
A policy with at least 1 CSRF URL parameter defined cannot be imported as a binary policy file.

Conditions:
A policy has at least 1 CSRF URL parameter defined.

Impact:
The policy cannot be imported as a binary policy file.

Workaround:
There is no workaround at this time.

Fix:
A policy with CSRF URL parameters defined can now be imported as a binary policy file.


697718-1 : Increase PEM HSL reporting buffer size to 4K.

Component: Policy Enforcement Manager

Symptoms:
Before this fix, PEM HSL reporting buffer size is limited by 512 bytes.

Conditions:
If any PEM HSL flow reporting stream goes beyond 512 bytes, it will be truncated.

Impact:
Part of PEM HSL flow reporting information will be lost.

Fix:
We increase PEM HSL reporting buffer size to 4K, which should be enough for uses cases we currently know.


697636-3 : ACCESS is not replacing headers while replacing POST body

Component: Access Policy Manager

Symptoms:
If the first request for a session is a POST, APM will save the POST to replay after the policy completes. When the POST is restored after policy completion and released to the backend, the headers are the same as the most recent client request, not the original POST. In particular, the Content-Length header will not match the original POST.

Conditions:
First request for the session is a POST.

Impact:
Backend servers may complain of an incomplete HTTP POST due to a mismatching Content-Length header.

Workaround:
None.

Fix:
Now, the system takes all headers from the original POST, except the Authorization header that Kerberos RBA needs, which is taken from the most recent client request.


697616-2 : Report Device error: crypto codec qat-crypto0-0 queue is stuck on vCMP guests

Component: TMOS

Symptoms:
Failure in SSL traffic in vCMP configurations. The system logs the following device error:
-- crit tmm[17083]: 01010025:2: Device error: crypto codec qat-crypto0-0 queue is stuck.
-- warning sod[7759]: 01140029:4: HA crypto_failsafe_t qat-crypto0-0 fails action is failover.

Conditions:
-- vCMP guests when performing crypto operations.
-- i5600, i5800, i7600, i7800, i10600, i10800, i12600, i12800, i15600, i15800 platforms.

Impact:
The 'crypto queue stuck' message is reported, and failover will be triggered.

Workaround:
None.

Fix:
The 'crypto queue stuck' issue on vCMP platforms no longer occurs.


697615-1 : Neurond may restart indefinitely after boot, with neurond_i2c_config message

Solution Article: K65013424

Component: TMOS

Symptoms:
The neurond daemon may continually restart after a reboot. The problem may persist even after a reboot of the BIG-IP system. Manually stopping and starting neurond will not resolve the problem.

Conditions:
- This occurs only on BIG-IP platforms that contain a specific hardware part running v13.1.0.
- The issue happens only after a reboot of the BIG-IP system.

Impact:
The BIG-IP system constantly logs messages similar to the following:

emerg logger: Re-starting neurond


The /var/log/neurond logfile contains messages similar to the following:

-- neurond_i2c_config_steps: STEP 20 Checking for Lane Alignment
-- neurond_i2c_config_steps: Timeout waiting for good rx_align for ILK1 of NSP
-- neurond_i2c_config: neurond_i2c_config_steps failed.

Workaround:
If you are not using FIX features, disabling the neurond service is a safe option.

If your configuration relies on the FIX feature, a cold reboot by removing the BIG-IP system from the power may resolve the problem. However, multiple retries are sometimes necessary to get the part to initialize.

Fix:
This release increases the number of initialization retries to handle this condition, so continual restarts no longer occur.


697516 : Upgrading using a ucs or scf file does not autogenerate uuids when current config has the uuid-default-autogenerate flag enabled

Component: Advanced Firewall Manager

Symptoms:
Upgrading using a UCS or SCF file does not autogenerate uuids when the current config has the uuid-default-autogenerate flag enabled. This might cause issues when upgrading from older versions where uuids need to be quickly generated for existing firewall policies, rule lists, and management rules.

Conditions:
Upgrading from an older version with an existing security policy which has no uuids configured.

Impact:
Requires manually configuration of uuids for rules that come in from the older config.

Workaround:
Generate uuids for all policies, rule-lists, and management rules using the following three tmsh commands:

-- tmsh modify sec fire policy all rules modify { all { uuid auto-generate}}
-- tmsh modify sec fire rule-list all rules modify { all { uuid auto-generate}}
-- tmsh modify sec fire management-ip-rules rules modify { all { uuid auto-generate}}

Optionally, to ensure rules created in the future have uuids autogenerated issue the following tmsh command:

-- tmsh modify sec firewall uuid-default-autogenerate mode enabled

Fix:
No fix provided, Current behavior causes the uuid-default-autogenerate flag to be overwritten to disabled by the ucs load process. Workaround has been provided to mitigate against this behavior.


697452-1 : Websso crashes because of bad argument in logging

Component: Access Policy Manager

Symptoms:
Websso would crash because of bad argument in logging

Conditions:
Only when kerberos sso is configured

Impact:
Websso would crash and so single sign on may fail.

Workaround:
The workaround is not configure kerberos SSO

Fix:
This issue has been fixed.


697421 : Monpd core when trying to restart

Component: Application Visibility and Reporting

Symptoms:
Monpd tries to restart and tries to access a non-initiated variable

Conditions:
Monpd tries to restart due to change of primary blade

Impact:
Monpd cores

Workaround:
N/A

Fix:
Adding sanity check to the non-initiated variable before trying to access it


697363-1 : FPS should forward all XFF header values

Component: Fraud Protection Services

Symptoms:
For BIG-IP alerts, FPS will insert a single XFF with the client IP and discard all XFF values/headers in the original request (the request which triggered the alert)

Conditions:
Alert generated on BIG-IP side.

Impact:
Original XFF information will be lost: only a single XFF header (containing client IP) will be present.

Workaround:
None.

Fix:
FPS now copies all original XFF headers to the generated alert.


697303-1 : BD crash

Component: Application Security Manager

Symptoms:
BD crashes.

Conditions:
-- The internal parameter relax_unicode_in_json is set to 1.
-- Specific traffic scenario.

Impact:
BD crash, failover, and traffic disturbance.

Workaround:
Turn off the internal parameter relax_unicode_in_json.

Fix:
BD no longer crashes under these conditions.


697259-2 : Different versioned vCMP guests on the same chassis may crash.

Solution Article: K14023450

Component: Local Traffic Manager

Symptoms:
The vCMP guest TMM crashes soon after startup.

Conditions:
-- You are using BIG-IP software versions 12.1.0-12.1.2.
-- vCMP guests are deployed in a chassis.
-- You configure a new guest running unaffected software alongside an existing or new guest running affected software. In other words, the issue occurs if you mix guests running affected and non-affected versions in a single vCMP host.

Impact:
vCMP guests running older versions of the software might crash. Blades continuously crash and restart. Traffic cannot pass. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Different versioned vCMP guests on the same chassis no longer crash.


696808-1 : Disabling a single pool member removes all GTM persistence records

Component: Global Traffic Manager (DNS)

Symptoms:
Disabling a single pool member removes all GTM persistence records.

Conditions:
1. WideIP with persistence enabled.
2. drain-persistent-requests no.
3. GTM pool member toggled from enabled to disabled.

Impact:
All GTM persistence records are accidently cleared.

Workaround:
Set drain-persistent-requests yes.

Fix:
When drain-persistent-requests is set to no, only the persistence records associated with the affected pool members are cleared when a resource is disabled.


696789-1 : PEM Diameter incomplete flow crashes when TCL resumed

Component: Policy Enforcement Manager

Symptoms:
If a PEM Diameter flow is not fully created, for example suspended by an iRule, and the irule is resumed because of timeout.

Conditions:
PEM Diameter flow not fully created, suspended by iRule, and the iRule is resumed by timeout.

Impact:
The tmm will restart and all flows will reset.

Fix:
The PEM diameter flow is now created in such a way to prevent any crash by iRule resumed by timeout.


696755 : HTTP/2 may truncate a response body when served from cache

Component: Local Traffic Manager

Symptoms:
BIG-IP provides a client side HTTP/2 Gateway protocol implementation in conjunction with HTTP 1.x on a server side. A response can be cached in BIG-IP with a web acceleration profile. Sometimes a response served from cache is prematurely marked with END_STREAM flag causing the client to ignore the rest of the response body.

Conditions:
BIG-IP has a virtual where HTTP/2 and Web Acceleration profiles are configured.

Impact:
Some clients' browsers do not retry a resource causing incorrect rendering of an HTML page.

Workaround:
Adding the following iRule causes the body to be displayed:

when HTTP_RESPONSE_RELEASE {
    set con_len [string trim [HTTP::header value Content-Length]]
    HTTP::header remove Content-Length
    HTTP::header insert Content-Length "$con_len"
}

Fix:
With provided fix HTTP/2 users no longer experience the problem of incorrect page rendering due to this issue.


696732-3 : tmm may crash in a compression provider

Solution Article: K54431534

Component: TMOS

Symptoms:
TMM may crash with the following panic message in the log files:

panic: ../codec/compress/compress.c:1161: Assertion "context active" failed.

Conditions:
-- Running on the following platforms: BIG-IP i-series or VIPRION 4400 blades.
-- Virtual servers are performing compression.

Impact:
TMM crashes, Traffic disrupted while tmm restarts.

Workaround:
Do not use compression, or use software compression instead of hardware compression. To configure for software compression, run the following command:

tmsh modify sys db compression.strategy value softwareonly


696731-3 : The standard LinkUp/LinkDown traps may not be issued in all cases when an interface is administratively enabled/disabled

Solution Article: K94062594

Component: TMOS

Symptoms:
The standard LinkUp/LinkDown traps are issued when a port's status changes; that is, the link goes up or down. Additionally, the network manager can administratively enable and disable ports. For some platforms, there is no reliably issued standard LinkUp/LinkDown trap when a port is administratively enabled/disabled.

Conditions:
Administrative disabling an interface on BIG-IP

Impact:
May issue only the F5 Link change trap and not the additional standard MIB trap. The standard LinkUp/LinkDown traps may not be issued.

Workaround:
Use the F5-proprietary MIB: 1.3.6.1.4.1.3375.2.4.0.37 F5-BIG-IP-COMMON-MIB::BIG-IPExternalLinkChange trap to track administrative changes to links.


696669-1 : Users cannot change or reset RSA PIN

Component: Access Policy Manager

Symptoms:
User is not able to reset the PIN when RSA SecurID or RADIUS Auth agent is included in access policy.

Conditions:
- APM is licensed and provisioned.
- RSA SecurID or RADIUS Auth agent is included in an access policy.
- APM end user is challenged to reset the PIN or reenter the PIN/token.

Impact:
APM end users cannot reset the PIN or do not get authenticated.

Workaround:
There is no workaround.

Fix:
APM users can now successfully reset the PIN or reenter the token.


696642-1 : monpd core is sometimes created when the system is under heavy load.

Component: Application Visibility and Reporting

Symptoms:
When system is under heavy load, aggregation of statistics tables in the database sometimes takes too much time and watchdog is triggered. When that happens, watchdog aborts the application and produces a core file.

Conditions:
-- System under heavy load.
-- Setting and resetting DoS profile on virtual servers.
-- Using AVR.
-- Displaying aggregated statistics.

Impact:
System produces monpd core file, when no real crash occurs.

Workaround:
None.

Fix:
Watchdog trigger no longer creates core by default under these conditions.


696544-1 : APM end users can not change/reset password when auth agents are included in per-req policy

Component: Access Policy Manager

Symptoms:
Users cannot change password when AD, Radius or LocalDb auth agents are included in per-req policy.

Conditions:
- Per-req policy is attached to Virtual Server.
- AD Auth, Radius Auth or LocalDB auth agents are included in per-req policy.
- End user is challenged to change/reset the password.

Impact:
Users can not change password.

Fix:
Users now can successfully change or reset the password.


696525-1 : B2250 blades experience degraded performance.

Component: Performance

Symptoms:
B2250 blades have degraded performance by up to 17%. This is caused by connections not being offloaded to hardware as often as expected.

Conditions:
This occurs when the FastL4 profile is configured to offload to hardware and the service provider DAG is configured and in use on B2250 blades.

Impact:
Performance will be degraded due to more connections being handled in software.

Workaround:
None.

Fix:
The performance issue for the B2250 blades has been fixed.


696383-1 : PEM Diameter incomplete flow crashes when sweeped

Component: Policy Enforcement Manager

Symptoms:
If a PEM Diameter flow is not fully created, for example suspended by an iRule, the sweeper may encounter a tmm crash.

Conditions:
-- PEM Diameter flow not fully created.
-- The flow is suspended by an iRule.
-- There is a CMP state change (likely) or a manual cluster-mirror change (less likely) while the flow is suspended.

Impact:
The tmm restarts and all flows reset. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
The PEM diameter flow is now created in such a way to prevent any crash by the sweeper.


696294-1 : TMM core may be seen when using Application reporting with flow filter in PEM

Component: Policy Enforcement Manager

Symptoms:
TMM core with flow filter when Application reporting action is enabled

Conditions:
If Application reporting is enabled along with flow filter

Impact:
TMM restart causing service interruption

Fix:
Initialize the application start buffer so as to prevent the TMM core


696265-5 : BD crash

Solution Article: K60985582

Component: Application Security Manager

Symptoms:
BD crash.

Conditions:
ecard_max_http_req_uri_len is set to a value greater than 8 KB.

Impact:
Potential traffic disturbance and failover.

Workaround:
Change the value of ecard_max_http_req_uri_len to a size lower than 8 KB.

Fix:
Fixed a BD crash scenario.


696260-1 : GUI Network Map as Start Screen presents database error

Solution Article: K53103420

Component: TMOS

Symptoms:
If the Network Map is set as the Preferences Start Screen, the GUI will display a database error page.

Conditions:
Set System :: Preferences : Start Screen to Network Map.

Impact:
Error page is displayed.

Workaround:
Navigate to the Network Map via the left navigation menu: Local Traffic :: Network Map.

Fix:
The Screen Start now launches successfully into the Network Map page.


696212-1 : monpd does not return data for multi-dimension query

Component: Application Visibility and Reporting

Symptoms:
When querying 'time-series' data for multiple-dimensions, most multi-dimension queries receive an empty response.

Conditions:
This occurs because the order of entities in the query is not sorted by priority.

Impact:
The corresponding dashboard displays incorrect statistics.

Workaround:
There is no workaround at this time.

Fix:
The monpd process now performs two queries in order to get the 'time-series' data for multi-dimensions:
-- The first query gets the top entities.
-- The second query gets data that is 'drilled down' by the top entities, the ones received from the first query.


696201-1 : Anomaly threshold floor calculation (for bins with very low learnt threshold) during signature generation

Component: Advanced Firewall Manager

Symptoms:
AFM might generate a dynamic signature for those bins that have a very low learnt threshold during the learning phase, if the current traffic rate spikes and increases above the anomaly threshold floor db variable value as specified by l4bdos.anomaly.threshold.floor

Conditions:
AFM dynamic signature feature is enabled.

Impact:
This might cause AFM to generate signatures with higher false positives.

This is specifically due to incorrect application of db variable setting 'l4bdos.anomaly.threshold.floor' that should be interpreted as the 'floor' value of learnt thresholds for any bin. So, if the learnt threshold of a bin is lower than this db variable, the baseline threshold of the bin should be set to the db variable for anomaly detection phase.

Workaround:
There is no workaround at this time.

Fix:
This issue is fixed by making sure that db variable 'l4bdos.anomaly.threshold.floor' is used as the 'floor' value of baseline thresholds for those bins that have a learnt threshold lower than this db variable.


696113-3 : Extra IPsec reference added per crypto operation overflows connflow refcount

Component: TMOS

Symptoms:
The size of the refcount field in connflow became smaller, making the length of some crypto queues in IPsec able to reach and exceed the maximum refcount value.

Conditions:
When a large data transfer under an IPsec SA creates a queue of crypto operations longer than the connflow's refcount can handle, the refcount can overflow.

Impact:
Unexpected tmm failover after refcount overflow.

Workaround:
There is no workaround at this time.

Fix:
An object tracking crypto operations now adds a sole reference to the connflow as long as the count of crypto operation pending is nonzero.


696073-2 : BD core on a specific scenario

Component: Application Security Manager

Symptoms:
bd process crashes, and core file created in the /shared/core/ directory.

Conditions:
Specific request and response characteristics that relates to CSP headers sent by the server.

Impact:
Failover in high availability units.

Workaround:
Disable CSP headers handling in ASM by running the following commands:

/usr/share/ts/bin/add_del_internal add csp_enabled 0
bigstart restart asm

Fix:
The system now reinitializes the CSP headers before each response headers event, so this issue no longer occurs.


696049-1 : High CPU load on generic message if multiple responses arrive while asynchronous Tcl command is running

Component: Service Provider

Symptoms:
High CPU load on generic message if multiple responses arrive while asynchronous Tcl command is running.

Conditions:
Multiple response messages arrive on a connection while an asynchronous Tcl command is running on that connection.

Impact:
High CPU load might occur as multiple responses will be assigned the same request_sequence_number.

Workaround:
None.

Fix:
Request_sequence_numbers are not assigned to response messages until the Tcl event is executed for that message. This avoids assigning the same number to multiple events.


695968-1 : Memory leak in case of a PEM Diameter session going down due to remote end point connectivity issues.

Component: Policy Enforcement Manager

Symptoms:
Memory leak resulting in a potential OOM scenario.

Conditions:
1. PEM configured with Gx
2. Flaky Diameter connection
3. Subscriber creation via PEM

Impact:
Potential loss of service.

Workaround:
There is no workaround at this time.

Fix:
Freed Diameter messages appropriately.


695953-1 : Custom URL Filter object is missing after load sys config TMSH command

Component: Access Policy Manager

Symptoms:
Cannot see the custom URL Filter object that is created either through TMSH/GUI. If the filter object is referred in an Access Policy, the policy fails to load when running the command: load sys config. The system logs errors similar to the following:

01070712:3: Values (/Common/custurlfilter) specified for URL Filter Lookup Agent (/Common/prp_act_url_filter_lookup_ag): foreign key index (name_FK) do not point at an item that exists in the database.
Unexpected Error: Loading configuration process failed.

Conditions:
-- Configure custom URL Filter object
-- SWG is not provisioned.

Impact:
The access policy fails to load if it refers the URL Filter object. Running the 'load sys config' command in TMSH removes the filter.

Workaround:
You can use either of the following workarounds:
-- Provision SWG, and recreate the URL Filter.
-- Edit bigip.conf to include the URL Filter object.

Fix:
Now during 'load sys config', custom URL filters get saved properly.


695925-1 : tmm crash when showing connections for a CMP disabled virtual server

Component: Local Traffic Manager

Symptoms:
tmm crashes when performing 'tmsh show sys connection' and there is a connection from a secondary blade to a CMP-disabled virtual server.

Conditions:
This occurs when all of the following conditions are met:

-- There is a CMP-disabled virtual server.

-- There is a connection to that server from the control plane of a secondary blade (this can include monitoring traffic).

-- Connections are displayed that include the connection from the secondary blade ('tmsh show sys connection').

Impact:
tmm crashes and restarts impacting traffic.

Workaround:
Do not use 'cmp-enabled no' virtual servers when there will be connections from the BIG-IP control plane to the virtual server.

Avoid using tmsh show sys connection


695901-1 : TMM may crash when processing ProxySSL data

Solution Article: K46940010


695707-5 : BIG-IP does not retransmit DATA_FIN when closing an MPTCP connection

Component: Local Traffic Manager

Symptoms:
BIG-IP does not retransmit DATA_FIN when closing an MPTCP connection.

Conditions:
Close an MPTCP connection.

Impact:
If a DATA_ACK is not received for the DATA_FIN, the connection will stall until it times out.

Workaround:
There is no workaround at this time.

Fix:
Keep the retransmission timer running if an MPTCP connection can retransmit a DATA_FIN.


695563-1 : Improve speed of ASM initialization on first startup

Component: Application Security Manager

Symptoms:
ASM initialization on first startup takes a long time.

Conditions:
Provision ASM.

Impact:
ASM initialization takes a long time.

Workaround:
There is no workaround at this time.

Fix:
ASM initialization on first startup is faster.


695072-2 : CVE-2016-8399 CVE-2017-1000111 CVE-2017-1000112 CVE-2017-11176 CVE-2017-14106 CVE-2017-7184 CVE-2017-7541 CVE-2017-7542 CVE-2017-7558

Solution Article: K23030550


694922-5 : ASM Auto-Sync Device Group Does Not Sync

Component: Application Security Manager

Symptoms:
In rare circumstances a device may enter an untrusted state and confuse the device group.

Conditions:
1) ASM sync is enabled on an autosync device group
2) A new ASM entity is created on a device

Impact:
ASM configuration is not correctly synchronized between devices

Workaround:
1) Remove ASM sync from the device group (Under Security ›› Options : Application Security : Synchronization : Application Security Synchronization)
2) Restart asm_config_server.pl on both devices and wait until they come back up
3) Change the device group to a manual sync group
4) On the device with the good configuration re-enable ASM sync for the device group
5) Make a spurious ASM change, and push the configuration.
6) Change the sync type back to automatic

Fix:
Devices no longer spuriously enter an untrusted state


694897-2 : Unsupported Copper SFP can trigger a crash on i4x00 platforms.

Component: TMOS

Symptoms:
PFMAND can crash when an unsupported Proline Copper SFP is inserted in the 1G interfaces.

Conditions:
-- Using Proline CuSFP, Part number FCLF8521P2BTLTAA.
-- Inserted into 1 GB interfaces.
-- On i4x00 platforms.

Impact:
PFMAND cores.

Workaround:
Use only F5 branded Copper SFPs

Fix:
This release updates SFP string parsing in PFMAND to account for NULL terminated vendor information.


694849-1 : TMM crash when packet sampling is turned for DNS BDOS signatures.

Component: Advanced Firewall Manager

Symptoms:
TMM crashes upon traffic matching a DNS BDOS signature if packet sampling is turned on by enabling db variable (l4bdos.signature.sample.packet.frequency).

Conditions:
DB variable l4bdos.signature.sample.packet.frequency is modified to a non-zero value (to collect DNS packet info upon matching a DNS dynamic signature).

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Disable the packet sampling feature for BDOS signatures by setting the db variable l4bdos.signature.sample.packet.frequency to default value (0).

Fix:
TMM no longer crashes when packet sampling is turned on and traffic matches DNS BDOS signature.


694778-1 : Certain Intel Crypto HW fails to decrypt data if the given output buffer size differs from RSA private key size

Component: Local Traffic Manager

Symptoms:
SSO-enabled Native RDP resources cannot be accessed via hardware (HW) BIG-IP systems with 'Intel Cave Creek' coprocessor (i.e., SSL connection cannot be established with the db variable 'crypto.hwacceleration' enabled, and RSA key used).

Conditions:
The failure might occur in the following scenario:
-- Running on Intel Cave Creek Engine (e.g., BIG-IP 2000 (C112) or 4000 (C113)).
-- Client OS is Mac, iOS, or Android.
-- HW crypto is enabled
-- Using a virtual server with a client SSL profile and 2048 bit RSA key on.
-- Native RDP resource with enabled SSO is used on hardware BIG-IP with 'Intel Cave Creek' coprocessor.
-- Output buffer size differs from RSA private key size.

Impact:
-- SSL connection fails.
-- RDP client cannot launch the requested resource (desktop/application).

Workaround:
There is no workaround other than to disable crypto HW acceleration with following command:
tmsh modify sys db crypto.hwacceleration value disable

Fix:
SSL connection can now be established as expected. SSO-enabled Native RDP resources now can now be accessed via hardware BIG-IP systems with 'Intel Cave Creek' coprocessor (e.g., BIG-IP 2000 (C112) or 4000 (C113) platforms) from Mac, iOS, and Android clients.


694740-3 : BIG-IP reboot during a TMM core results in an incomplete core dump

Component: TMOS

Symptoms:
If an HSB lockup occurs on i10600 and i10800 platforms, then HSB panics and generates a core file for post-analysis. HSB also triggers a nic_failsafe reboot. On this platform, the reboot occurs before the core file is fully written, resulting in a truncated core.

Conditions:
An HSB lockup occurs, triggering a core dump and a nic_failsafe reboot.

Impact:
The reboot happens before the core dump completes, resulting in an incomplete core dump which is not useful for analyzing why the HSB lockup occurred. Traffic disrupted while tmm restarts.

Workaround:
No workaround; does not hinder device operation, but does prevent post-crash analysis.

Fix:
Reboot is delayed until TMM core file is completed.


694717-1 : Potential memory leak and TMM crash due to a PEM iRule command resulting in a remote lookup.

Component: Policy Enforcement Manager

Symptoms:
TMM crashes

Conditions:
PEM iRule command that would result in an inter-TMM lookup on a long lived flow that would result in the PEM iRule command being hit several times. For example, a long lived flow with multiple HTTP transactions.

Impact:
Traffic disrupted while tmm restarts.

Fix:
Always release the connFlow reference associated with the TCL command to avoid a memory leak and potential crash.


694696-5 : On multiblade Viprion, creating a new traffic-group causes the device to go Offline

Component: TMOS

Symptoms:
All devices in the failover device group will go offline, resulting in traffic disruption and possible failovers.

Conditions:
When a new traffic-group is created on a multiblade Viprion system that is a member of a sync-failover device group.

Impact:
Traffic to all other traffic-groups is disrupted for several seconds.

Workaround:
There is no workaround at this time.

Fix:
Creating a new traffic-group does not disrupt existing traffic-groups.


694656-1 : Routing changes may cause TMM to restart

Solution Article: K05186205

Component: Local Traffic Manager

Symptoms:
When routing changes are made while traffic is passing through the BIG-IP system, TMM might restart. This isn't a general issue, but can occur when other functionality is in use (such as Firewall NAT).

Conditions:
-- Routing changes occurring while BIG-IP is active and passing traffic.

-- Dynamic routing, due to its nature, is likely increase the potential for the issue to occur.

-- Usage of Firewall NAT functionality is one specific case known to contribute to the issue (there may be others).

Impact:
TMM restarts, resulting in a failover and/or traffic outage.

Workaround:
If dynamic routing is not in use, avoiding static route changes may avoid the issue.

If dynamic routing is in use, there is no workaround.

Fix:
TMM now properly manages routing information for active connections.


694624-1 : SSO enabled Native RDP resources can't be accessed via hardware BIG-IP with "Intel Cave Creek" coprocessor

Component: Access Policy Manager

Symptoms:
APM Webtop's SSO enabled Native RDP resources can't be accessed via hardware BIG-IP with "Intel Cave Creek" coprocessor (e.g. BIG-IP 2000 (C112) or 4000 (C113)) from Mac, iOS and Android clients. Each launch attempt generates following errors in /var/log/apm:
... err vdi[123] ... {45.C} RsaDecryptData error: AsyncError:5: InvalidData
... err vdi[123] ... {45.C} An exception is thrown: handshake: decryption failed or bad record mac

Conditions:
Native RDP resource with enabled SSO is used on hardware BIG-IP with "Intel Cave Creek" coprocessor (e.g. BIG-IP 2000 (C112) or 4000 (C113)).
The client OS is Mac, iOS or Android.

Impact:
RDP client can't launch requested resource (desktop/application).

Workaround:
Disable crypto HW acceleration with following command:
tmsh modify sys db crypto.hwacceleration value disable

Fix:
SSO enabled Native RDP resources now can be accessed via hardware BIG-IP with "Intel Cave Creek" coprocessor (e.g. BIG-IP 2000 (C112) or 4000 (C113) platforms) from Mac, iOS and Android clients.


694547-2 : TMSH save sys config creates unneeded generate_config processes.

Solution Article: K74203532

Component: TMOS

Symptoms:
When saving a configuration through TMSH or iControl REST, the system creates an unneeded process named generate_config.

Conditions:
Run tmsh save sys config, or the same command through iControl REST.

Impact:
One generate_config process is generated per save operation. If config save occurs often, these extraneous processes can slowly fill up the process table.

Workaround:
There is no real workaround except to not save the config often enough to fill up process table with these extraneous processes.

If this issue has already occurred, you can recover by locating the parent process that has an associated zombie process, and restart the parent process to purge the zombie processes. icrd_child and/or scriptd are the parent processes known to cause this issue. To find out which daemon to restart and how to restart it, perform the following procedure:

Impact of workaround: Restarting any daemon on the BIG-IP system may cause service disruption, and F5 recommends performing this procedure only during a scheduled maintenance period. For more information about daemons' functions, refer to K05645522: BIG-IP daemons (13.x) (https://support.f5.com/csp/article/K05645522).

1. If you are still logged on to the tmsh command-line utility that was performing the configuration-save operation, exit from it first.
2. Login to the BIG-IP system's advanced shell using an account with Administrator credentials.
3. Locate the zombie process and its parent process using the following command:
ps --forest | grep -B1 generate_config.*defunct
 
4. With the parent process name discovered in the previous step, restart the associated daemon using the following commands that apply:
-- For the icrd_child process: tmsh restart /sys service restjavad
-- For the scriptd process: tmsh restart /sys service scriptd

Fix:
tmsh save sys config no longer generates generate_config processes.


694319-1 : CCA without a request type AVP cannot be tracked in PEM.

Component: Policy Enforcement Manager

Symptoms:
May cause diagnostic issues as not all CCA messages cannot be tracked.

Conditions:
1. PEM with Gx/Gy configured
2. PCRF sends CCA's without a request type AVP

Impact:
May hamper effective diagnostics.

Workaround:
Mitigation:
Configure the PCRF to always include a request type in its CCAs.

Fix:
Add a statistics counter to track CCA's that do not request type AVPs.
Name of new counter:cca_unknown_type


694318-1 : PEM subscriber sessions will not be deleted if a CCA-t contains a DIAMETER_TOO_BUSY return code and no request type AVP.

Component: Policy Enforcement Manager

Symptoms:
Subscriber sessions in PEM will be stuck in a "Marked for Delete" state.

Conditions:
1. PEM provisioned with Gx.
2. PCRF responds to a CCR-t with a DIAMETER_TOO_BUSY return code and no Request type AVP.

Impact:
Subscriber sessions stuck in delete pending resulting in a potential increase in memry consumption over a period of time.

Workaround:
Mitigation:
PCRF (remote Diameter end point) must send a CCA-t with a request type AVP in case a DIAMETER_TOO_BUSY return code is present.

Fix:
Handle the DIAMETER_TOO_BUSY return code on a CCA-t regardless of the request type AVP.


694274-1 : [RHSA-2017:3195-01] Important: httpd security update - EL6.7

Solution Article: K23565223


694078-1 : In rare cases, TMM may crash with high APM traffic

Component: Access Policy Manager

Symptoms:
Intermittent tmm core under load.

Conditions:
-- Provision APM (at least).

-- Additional required conditions are not well understood.

-- Seems more likely to occur when APM is provisioned with other modules, especially ASM or AVR.

Impact:
The BIG-IP system stops processing traffic while the TMM restarts.

Workaround:
None.

Fix:
Tmm core no longer occurs with high APM traffic.


694073-3 : All signature update details are shown in 'View update history from previous BIG-IP versions' popup

Component: Application Security Manager

Symptoms:
If you are running a BIG-IP release named with 4 digits (e.g., 12.1.3.1), all signature update details are shown only in 'View update history from previous BIG-IP versions' popup. The 'Latest update details' section is missing.

Conditions:
Running a BIG-IP software release named with 4 digits (e.g., 12.1.3.1).

Impact:
Low and incorrect visibility of signature update details.

Workaround:
Signature update details can be viewed in 'View update history from previous BIG-IP versions' popup.

Fix:
Signature updates are now shown correctly for all versions.


693996-5 : MCPD sync errors and restart after multiple modifications to file object in chassis

Solution Article: K42285625

Component: TMOS

Symptoms:
Upon modifying file objects on a VIPRION chassis and synchronizing those changes to another VIPRION chassis in a device sync group, the following symptoms may occur:

1. Errors are logged to /var/log/ltm similar to the following:

-- err mcpd[<#>]: 0107134b:3: (rsync: link_stat "/config/filestore/.snapshots_d/<_additional_path_to/_affected_file_object_>" (in csync) failed: No such file or directory (2) ) errno(0) errstr().
-- err mcpd[<#>]: 0107134b:3: (rsync error: some files could not be transferred (code 23) at main.c(1298) [receiver=2.6.8] syncer /usr/bin/rsync failed! (5888) () Couldn't rsync files for mcpd. ) errno(0) errstr().
-- err mcpd[<#>]: 0107134b:3: (rsync process failed.) errno(255) errstr().
-- err mcpd[<#>]: 01070712:3: Caught configuration exception (0), Failed to sync files..

2. MCPD may restart on a secondary blade in a VIPRION chassis that is receiving the configuration sync from the chassis where the file object changes were made.

Conditions:
Making multiple changes to the same file objects on a VIPRION chassis and synchronizing those changes to another VIPRION chassis in a device sync group.

Impact:
Temporary loss of functionality, including interruption in traffic, on one or more secondary blades in one or more VIPRION chassis that are receiving the configuration sync.

Workaround:
After performing one set of file-object modifications and synchronizing those changes to the high availability (HA) group members, wait for one or more minutes to allow all changes to be synchronized to all blades in all member chassis before making and synchronizing changes to the same file-objects.


693979 : Autoscale is not functional because of the change in file permission of the /shared/vadc/aws/iid-document document

Component: TMOS

Symptoms:
The /shared/vadc/aws/iid-document's file permission changed and as a result the autoscale feature was failing.

Conditions:
Whenever autoscale is triggered

Impact:
The autoscale feature does not work

Workaround:
The permission of /shared/vadc/aws/iid-document was never set explicitly. It inherited file permission flags from /shared/vadc/. We set the file permission explicitly.

Fix:
The autoscale feature is functional after changing file permissions of /shared/vadc/aws/iid-document.


693966-1 : TCP sndpack not reset along with other tcp profile stats

Component: Local Traffic Manager

Symptoms:
TCP sndpack stat added is not being properly reset when a tmsh reset-stats command is issued.

Conditions:
When tmsh reset-stats command is issued.
-- tmsh reset-stats /ltm profile tcp <profile-name>

Impact:
TCP sndpack stat doesn't reset when tmsh reset-stats command is issued.

Workaround:
There is no workaround.

Fix:
With this fix, TCP sndpack stat will reset when tmsh reset-stats command is issued.


693964-1 : Qkview utility may generate invalid XML in files contained in Qkview

Component: TMOS

Symptoms:
When Qkview runs, it may gather XML files that are not well-formed, and contain ASCII control characters. This is most commonly seen with mcp_module.xml.

An XML validator may report an error such as:

    mcp_module.xml:536081: parser error : PCDATA invalid Char value 29
      <msgs></msgs>
            ^

Conditions:
-- Running Qkview.
-- An ASCII control character exists within a certain string field.

Impact:
The control character will be written verbatim into XML without encoding. Automated tools (e.g., iHealth) that attempt to process these files may fail.

Workaround:
iHealth automatically detects and corrects this issue in uploaded Qkviews.

You can analyze the XML files with some other tool, a tar.gz, so it can be unpacked, the XML files edited to correct the formatting, and then repacked. The xmllint command-line tool (present on the BIG-IP system) can also recover valid XML by removing the invalid characters.

To do so, you can run a command similar to the following:

    xmllint --recover mcp_module.xml --output mcp_module.xml

Fix:
Qkview no longer writes control characters in XML text, but instead processes them as expected.


693910-4 : Traffic Interruption for MAC Addresses Learned on Interfaces that Enter Blocked STP State (2000/4000/i2800/i4800 series)

Component: Local Traffic Manager

Symptoms:
Some MAC addresses might experience delayed forwarding after the interface on which they are learned transition to a STP blocked state. This is because the FDB entries are not being flushed on these interfaces after they change to a block state. Traffic destined to these MAC addresses gets dropped until their associated entries in the FDB age out.

Conditions:
MAC addresses are learned on a STP forwarding interface that transitions to a blocked interface following a topology change.

Impact:
Traffic interruption for up to a few minutes for MAC addresses learned on the affected interface.

Workaround:
None.

Fix:
FDB entries are now flushed by interface whenever an interface transitions to a STP block state.


693844-1 : APMD may restart continuously and cannot come up

Solution Article: K58335157

Component: Access Policy Manager

Symptoms:
apmd process cannot start correctly and restarts in an infinite loop. When apmd initializes the Allow Ending agent, the process tries to load all resources (including ACLs, the webtop, and all resources for every webtop, app tunnels, rdp, etc.). The most likely configuration to encounter this issue is with ACLs. For example, if you have thousands of ACL records, the Allow agent pulls them all at once. If the mcpd process is consumed with other operations, it might be that apmd cannot initialize the Allow agent in 30 secs, so it restarts, at which point the process tries to load all resources, cannot complete within the 30 seconds, and restarts in a loop.

Conditions:
Too many resources assigned in an Access Policy profile
for example thousands of ACLs configured.

apmd cannot initialize the Allow Ending agent in 30 seconds and decides it has stopped responding. Then it restarts by it's own, but problem is not solved, so it restarts in a loop

Impact:
APM end users cannot authenticate.

Workaround:
Reduce amount of resources so every agent can initialize within the 30-second timeframe.


693780-1 : Proactive Bot Defense sends CAPTCHA to the UC browser on iOS devices

Component: Application Security Manager

Symptoms:
When a request arrives from UCBrowser running on iOS and without TSPD_101 (proactive bot defense cookie). The big-ip respond with captcha challenge.

Conditions:
Dos profile attached to a virtual.
Dos profile has application security enabled.
Dos profile has proactive bot defense enabled.

Impact:
UC browser end-user presented with captcha challenge.

Workaround:
Increase proactive bot defense score.
list sys db dosl7.browser_legit_min_score_captcha
sys db dosl7.browser_legit_min_score_captcha {
    value "60"
}

Fix:
User agent parser has been changed (adjusted) for the UC browser. The UC browser is detected as safari ios.


693744-4 : CVE-2018-5531: vCMP vulnerability

Solution Article: K64721111


693694-1 : tmsh::load within IApp template results in unpredicted behavior

Component: iApp Technology

Symptoms:
tmsh::load command within IApp template triggers transaction within transaction and it is not supported by the MCP. One of the unexpected behavior seen is with the template having ASM policy and LTM policy. IApp framework doesn't let user to reconfigure the application service without turning off strict updates and also on rerunning, breaks association of LTM Policy with ASM Policy

Conditions:
tmsh::load command need to be used in in template to create ASM policy. With this tmsh::create there is no issue seen.

Impact:
Association b/w LTM Policy and ASM Policy broken

Workaround:
Use tmsh::create or tmsh::modify to create/update ASM policy through IApp template


693663-1 : Proactive Bot Defense sends CAPTCHA to the Firefox browser on iOS only in desktop mode

Component: Application Security Manager

Symptoms:
When a request arrives from Firefox running on iOS in desktop mode and without TSPD_101 (proactive bot defense cookie). The big-ip respond with captcha challenge.

Conditions:
Dos profile attached to a virtual.
Dos profile has application security enabled.
Dos profile has proactive bot defense enabled.

Impact:
Firefox (iOS desktop mode only) end-user presented with captcha challenge.

Workaround:
Increase proactive bot defense score.
list sys db dosl7.browser_legit_min_score_captcha
sys db dosl7.browser_legit_min_score_captcha {
    value "60"
}

Fix:
User agent parser has been changed (adjusted) for the Firefox browser running in desktop mode. The browser is detected as safari pc and the browser version is taken from Mac version number.


693611-3 : IKEv2 ike-peer might crash on stats object during peer modification update

Solution Article: K76313256

Component: TMOS

Symptoms:
A crash occurs upon passing traffic through the IPsec interface.

Conditions:
When an ike-peer is updated, or first defined at startup.

Impact:
Tmm restarts on crash.

Workaround:
No workaround is known at this time.

Fix:
IKEv2 ike-peer no longer crashes on stats object during peer modification update.


693578-2 : switch/color_policer tmctl table is not available in 12.1.0 - 13.1.0

Component: TMOS

Symptoms:
switch/color_policer tmctl table is not available in 12.1.0 - 13.1.0

Conditions:
None

Impact:
switch/color_policer tmctl table is not available in 12.1.0 - 13.1.0

Workaround:
None

Fix:
No fix.


693359-1 : AWS M5 and C5 instance families are supported

Component: TMOS

Symptoms:
AWS has announced support for M5 and C5 instance families. No previous BIG-IP Virtual Edition (VE) software supports these new AWS instance families.

Conditions:
Attempting to install pre-v14.0.0.1 VE software on M5 and C5 instance families.

Impact:
The system experiences a kernel panic and might crash.

Workaround:
None.

Fix:
All necessary components are added to support AWS M5 and C5 instance families.

Behavior Change:
No previous BIG-IP Virtual Edition (VE) software supports the newly announced AWS M5 and C5 instance families. Version 14.0.0.1 officially supports the AWS M5 and C5 instance families.


693312-1 : vCMPd may crash when processing bridged network traffic

Solution Article: K03165684


693308-1 : SSL Session Persistence hangs upon receipt of fragmented Client Certificate Chain

Component: Local Traffic Manager

Symptoms:
When a very large Client Certificate Chain, typically exceeding 16,384 bytes, is received by BIG-IP on a virtual service, and Session Persistence is enabled, the handshake hangs.

Conditions:
[1] SSL client authentication is enabled on the backend server
[2] No SSL profile is specified on the BIG-IP device for the virtual service, on both, client and server side
[3] An SSL connection is initiated from the front-end client, via the BIG-IP's virtual service, to the backend server.
[4] The client certificate chain is passed to the BIG-IP device as part of initiating the connection.

Impact:
The backend server will not be securely accessible via SSL because the connection hangs

Workaround:
Disable SSL Session Persistence.

Fix:
Whenever a fragmented message is received by a BIG-IP virtual service, subsequent messages contain a 5-byte header, each, which should be accounted for. Upon taking this into consideration, no more multiple-of-5 bytes are found missing while the message is being parsed by the Session Persistence parser, and the parser no longer hangs.


693244-2 : BIG-IP not sending RST for SYN,ACK packets when ASM is provisioned

Component: Local Traffic Manager

Symptoms:
BIG-IP silently drop the serverside TCP flow, when it receives a client-side reset and the server-side flow is still in the SYN-SENT state.

Conditions:
BIG-IP receives a client-side reset when client-side TCP flow is in ESTABLISHED state and server-side TCP flow is in SYN-SENT state, serverside flow is silently dropped.

Impact:
Since serverside pool member does not receive the RST, it remains in SYN-RECEIVED state until it runs out of syn retransmissions and eventually, due to timeout, it returns to LISTEN state.

Fix:
BIG-IP resets serverside TCP flow with RST when it receives a client-side reset and the server-side flow is still in the SYN-SENT state.


693007-1 : Modify b.root-servers.net IPv4 address 192.228.79.201 to 199.9.14.201 according to InterNIC

Component: Global Traffic Manager (DNS)

Symptoms:
The current IPv4 address for b.root-servers.net is 192.228.79.201. The IPv4 address for b.root-servers.net will be renumbered to 199.9.14.201, effective 2017-10-24. The older number will be invalid after that date.

Conditions:
Several profiles contain the b.root-servers.net IPv4 address as 192.228.79.201.

Impact:
The impact is likely minimal, at most a single timeout for pending TLD queries when they happen to round-robin onto an old IP address, probably not more often than the hint's TTL, which is more than a month, and even this should cause a timeout only when the old IP actually stops responding.

Workaround:
Update the root hints for all affected profiles manually except the hardwired ones.

Fix:
The IPv4 address for b.root-servers.net has been renumbered to 199.9.14.201, effective 2017-10-24.

Behavior Change:
The IPv4 address for b.root-servers.net has been renumbered to 199.9.14.201, effective 2017-10-24, according to InterNIC. The old IPv4 address (192.228.79.201) will continue to answer queries for at least 6 months.


692970-2 : Using UDP port 67 for purposes other than DHCP might cause TMM to crash

Component: Local Traffic Manager

Symptoms:
DHCP relay presumes that a flow found via lookup is always a server-side flow of type DHCP relay. Hence TMM can crash when DHCP relay makes a server connection if UDP port 67 is used for another purpose, in which case a wrong DHCP server flow could be selected.

Conditions:
A UDP port 67 is configured for a purpose other than DHCP relay.

Impact:
TMM restart causes traffic interruption or failover.

Workaround:
Do not use UDP port 67 for other virtual servers, or configure a drop listener on certain VLANs that cannot avoid using UDP port 67.

Fix:
TMM no longer crashes with DHCP flow validation.


692941-1 : GTMD and TMM SIGSEGV when changing wide IP pool in GTMD

Component: Global Traffic Manager (DNS)

Symptoms:
Changing wide IP causes gtmd and tmm core under certain conditions.

Conditions:
-- GTM pool is removed when it is referenced by a persist record.
-- That record is accessed before it is purged.

Impact:
gtmd and tmm core. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Changing wide IP no longer causes gtmd and tmm core when GTM pool is removed when it is referenced by a persist record, and that record is accessed before it is purged.


692890-3 : Adding support for BIG-IP 800 in 13.1.x

Component: TMOS

Symptoms:
Installing software version 13.1.0 fails on BIG-IP 800.

# tmsh show sys soft


---------------------------------------------------------Sys::Software Status
Volume Product Version Build Active Status
---------------------------------------------------------
HD1.1 BIG-IP 13.1.0 0.0.1868 no failed (Failed to install.)
HD1.2 BIG-IP 13.0.0 0.0.1645 yes complete
HD1.3 BIG-IP 11.6.0 0.0.401 no complete

---------------------------
Sys::Software Update Check
---------------------------
  Check Enabled true
  Phonehome Enabled true
  Frequency weekly
  Status none
  Errors 0

The system logs the following messages in /var/log/liveinstall.log:

info: Hardware is lm capable
info: System is lm capable
info: Adding application-package ltm7-application/noarch to transaction.
info: Adding application-package ros7-application/noarch to transaction.
info: Adding application-package sam-main/noarch to transaction.
info: Adding application-package sum-application/noarch to transaction.
info: Adding application-package ts-application/noarch to transaction.
info: Adding application-package wa-master/noarch to transaction.
info: Adding application-package (lm) woc-application-lm/noarch to transaction.
error: Product has no root package for Mercury
error: couldn't get package list file for LTM.ROS.SAM.SUM.TS.WA.WOC group Terminal error: Failed to install.
*** Live install end at 2018/01/02 13:29:45: failed (return code 255) ***

Conditions:
-- Installing/upgrading to v13.1.x.
-- Using the BIG-IP 800 platform.

Impact:
Install/upgrade will fail.

Workaround:
None.

Fix:
Installation now completes successfully on the BIG-IP 800 platform.


692753-1 : shutting down trap not sent when shutdown -r or shutdown -h issued from shell

Component: TMOS

Symptoms:
Shutting down trap not sent when shutdown -r or shutdown -h issued from shell.

Conditions:
When user access the linux shell and issues "shutdown -h" or "shutdown -r", the BIG-IP does not send shutting down trap.

Impact:
None.
Since this is user triggered command, the user is aware of the shutdown event, so the lack of trap is not critical.

Workaround:
None

Fix:
The shutdown trap is sent when user issues "shutdown -r" or "shutdown -h" from the linux shell.


692683-1 : Core with /usr/bin/tmm.debug at qa_device_mgr_uninit

Component: TMOS

Symptoms:
Running a debug version of tmm (/usr/bin/tmm.debug) on BIG-IP 2xxx and 4xxx platforms, crashes at qa_device_mgr_uninit when issuing either of the following commands:
-- bigstart stop tmm
-- bigstart restart tmm

Conditions:
Running a debug version of tmm.
-- BIG-IP 2xxx and 4xxx platforms.
-- Running either of the following commands:
   + bigstart stop tmm
   + bigstart restart tmm

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround other than not using a debug version of tmm on BIG-IP 2xxx and 4xxx platforms.

Fix:
tmm no longer halts and restarts under these conditions.


692557-1 : When BIG-IP as SAML IdP processes signed authentication requests from external SAML SP, a block of memory may become corrupted.

Component: Access Policy Manager

Symptoms:
After processing signed authentication requests from external SAML SP, SAML IdP may corrupt a block of tmm memory.

Conditions:
- BIG-IP system is configured as SAML IdP.
- IdP is configured with external SP's signing certificate.
- External SP sends signed authentication request.

Impact:
Various possible negative effects, including TMM core. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
BIG-IP as SAML IdP no longer causes memory corruption when handling certain traffic.


692328-1 : Tmm core due to incorrect memory allocation

Component: Advanced Firewall Manager

Symptoms:
In a rare condition after providing afm, we get a tmm core.
You will see the following line in avrd.log
/usr/bin/avrinstall -c20 -t10 -s2401000 --provisionAVR=0 --provisionASM=0 --provisionAFM=0 --provisionPBD=0 --provisionAPM=0 --provisionFPS=0 --provisionPEM=0 --provisionVCMP=0

Conditions:
AFM provisioned.
Attack started.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
We check that the shared memory was allocated correctly before reporting on an attack.


692310-2 : ICAP modified HTTP 1.1 request/response without content-length gets "chunked" even with no body

Solution Article: K69250459

Component: Service Provider

Symptoms:
When a modified HTTP v1.1 request or response is returned from an ICAP server and has no body (and also no content-length), a 'Transfer-Encoding: chunked' header is added and a zero-length chunk is appended.

Conditions:
-- HTTP profile.
-- Either request-adapt or response-adapt profile.
-- Modified HTTP v1.1 request/response has no body (and no content-length header).

Impact:
The HTTP server or client to which the modified request or response is destined, receives invalid HTTP and might behave in an unexpected manner.

Workaround:
Use an iRule to detect the lack of a Content-Length header in the modified request or response, and insert 'Content-Length: 0'.

For example with modified request:

when ADAPT_REQUEST_HEADERS {
    if {([HTTP::method] eq "GET") and (not [HTTP::header exists 'Content-Length'])} {
        HTTP::header insert Content-Length 0
    }
}

Similarly when ADAPT_RESPONSE_HEADERS {} for a response.

Fix:
A modified HTTP v1.1 request or response with no body is never 'chunked'.


692307-3 : User with 'operator' role may not be able to view some session variables

Component: Access Policy Manager

Symptoms:
When a user with 'operator' role tries to view the session variables, the GUI may show the following error: An error has occurred while trying to process you request.

Conditions:
This occurs when there is a huge blob of data associated with the user whose session variable is being viewed. For example Active Directory (AD) user accounts with thumbnailphoto and userCertificate user attributes containing binary data.

Impact:
User cannot view the session variables for those particular sessions. This data is available, however, via clicking on the Session ID.

Workaround:
Find this data via clicking on the session ID.

Fix:
User with 'operator' role can now view all expected session variables


692189-1 : errdefsd fails to generate a core file on request.

Component: TMOS

Symptoms:
Should errdefsd crash or be manually cored for the purpose of gathering diagnostic information, no core file will be generated.

Conditions:
Forcing errdefsd to core for diagnostic purposes.

Impact:
Increased communication required with F5 Support when attempting to diagnose problems with errdefsd.

Workaround:
1. Add the following lines to the official errdefsd script, /etc/bigstart/scripts/errdefsd:
16a17,19
> # set resource limits, affinity, etc
> setproperties ${service}
>
2. Run the following command: bigstart restart errdefsd

Fix:
errdefsd now generates a core file when forced to core.


692179-1 : Potential high memory usage from errdefsd.

Component: TMOS

Symptoms:
errdefsd memory usage grows with each config-sync or config update.

Conditions:
Ever increasing errdefsd memory usage is visible when the following conditions are met:
-- Config updates are frequent.
-- management-port logging is not being used.

Impact:
In extreme cases, errdefsd memory usage might increase userland memory-pressure. Note, however, that it isn't actually using the extra memory, so while Virtual Memory Size (VSZ) might be high, Resident Set Size (RSS) need not be.

Workaround:
A workaround for this problem is to periodically send logs to a management-port destination. The destination may be as simple as a dummy UDP port on 127.0.0.1. Adding such a destination to the publisher for an existing, active log source will work. errdefsd just needs a reason to process and flush accumulating configurations.

Fix:
The changes in this bug force errdefsd to check for config changes once every second in addition to checking whenever management-port destination logs come in.


692158-1 : iCall and CLI script memory leak when saving configuration

Component: TMOS

Symptoms:
Whenever an iCall script or CLI script saves the configuration, the scriptd process on the device leaks memory.

Conditions:
Use of iCall or CLI scripts to save the configuration.

Impact:
Repeated invocation might cause the system to run out of memory eventually, causing tmm to restart and disrupting traffic.

Workaround:
There is no workaround other than not saving the configuration from iCall or CLI scripts.

Fix:
scriptd process on the device no longer leaks memory when iCall and CLI scripts are used to save the configuration.


692123 : GET parameter is grayed out if MobileSafe is not licensed

Component: Fraud Protection Services

Symptoms:
GET parameter is grayed out if MobileSafe is not licensed.

Conditions:
-- Provision FPS on a system whose license has at least one active feature.
-- Do not license MobileSafe.

Impact:
In FPS Parameter's list, the GET method is always grayed out.

Workaround:
Use either of the following workarounds:
-- License MobileSafe.
-- Use TMSH or REST.

Fix:
The GET method is not grayed out if MobileSafe is not licensed.


692095-1 : bigd logs monitor status unknown for FQDN Node/Pool Member

Solution Article: K65311501

Component: Local Traffic Manager

Symptoms:
While monitoring FQDN nodes or pool members, bigd may log the current or previous monitor status of the node or pool member as 'unknown' in messages where that state internally could have been logged as 'checking' or 'no address' for FQDN template nodes. Other states for FQDN configured nodes or pool members log monitor status as expected. Messages are similar to the following:

notice bigd[####]: 01060141:5: Node /Common/node_name monitor status unknown [ ip.address: unknown ] [ was up for ##hrs:##mins:##sec ]
notice bigd[####]: 01060141:5: Node /Common/node_name monitor status up [ ip.address: unknown ] [ was unknown for ##hrs:##mins:##sec ]
notice bigd[####]: 01060141:5: Node /Common/node_name monitor status up [ ip.address: up ] [ was unknown for ##hrs:##mins:##sec ]
notice bigd[####]: 01060145:5: Pool /Common/pool_name member /Common/node_name monitor status unknown. [ ] [ was unchecked for ##hrs:##mins:##sec ]
notice bigd[####]: 01060145:5: Pool /Common/pool_name member /Common/node_name monitor status up. [ ] [ was unknown for ##hrs:##mins:##sec ]

Conditions:
This may occur of the FQDN template node or pool member is in a 'checking' or 'no address' state.
The 'checking' state may occur if the DNS resolution of the FQDN node or pool member name is in progress.
The 'no address' state may occur if no IP addresses were returned by the DNS server for the configured FQDN node or pool member name.

Impact:
Unable to triage state of FQDN nodes or pool members identified in these log messages, to determine whether further troubleshooting is required, or what specific problem condition might require further investigation.

Workaround:
None.

Fix:
An FQDN-configured node or pool member logs each internal monitor status, including for scenarios of 'checking' and 'no address' for FQDN template nodes which were previously logged as 'unknown'.


691897-3 : Names of the modified cookies do not appear in the event log

Component: Application Security Manager

Symptoms:
A modified domain cookies violation happened. When trying to see additional details by clicking the violation link, the name and value of the modified cookie is empty.

Conditions:
A modified domain cookies violation happens.

Note: This can happen only if there are also non-modified or staged cookies.

Impact:
Expected violation details are not displayed.

Workaround:
There is no workaround at this time.

Fix:
Issue with modified domain cookie violation details is now fixed.


691806-1 : RFC 793 - behavior receiving FIN/ACK in SYN-RECEIVED state

Solution Article: K61815412

Component: Local Traffic Manager

Symptoms:
The BIG-IP system resets connection with RST if it receives FIN/ACK in SYN-RECEIVED state.

Conditions:
The BIG-IP system receives FIN/ACK when it is in SYN-RECEIVED state.

Impact:
The BIG-IP system resets connection with RST.

Workaround:
None.

Fix:
The BIG-IP system now responds with FIN/ACK to early FIN/ACK.


691785-1 : The bcm570x driver can cause TMM to core when transmitting packets larger than 6144 bytes

Component: Local Traffic Manager

Symptoms:
The bcm570x driver will cause TMM to core with the log message:

panic: ifoutput: packet_data_compact failed to reduce pkt size below 4.

Conditions:
-- Running on a platform that uses the bcm570x driver (BIG-IP 800, 1600, 3600).
-- A packet larger than 6144 bytes is transmitted from the BIG-IP system.

Impact:
TMM core. Failover or outage. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Removed the panic statement that caused TMM to core. TMM will now log an error and drop the packet instead.


691670-5 : Rare BD crash in a specific scenario

Component: Application Security Manager

Symptoms:
BD crash or False reporting of signature ID 200023003.

Conditions:
JSON/XML/parameters traffic (should not happen with the enforce value signature).

Impact:
Failover, traffic disturbance in the core case. False positive violation or blocking in the other scenario.

Workaround:
Removing attack signature 200023003 from the security policy stops the issue.

Fix:
Fix a bug in the signatures engine that causes a false positive reporting of a signature. In some rare cases, this false reporting may cause a crash.

A newly released attack signature update changes the signature in a way that it no longer causes the issue to happen.


691609-1 : 1NIC: Configuration error: Clearing IP on interface mgmt failed - Cannot assign requested address

Component: TMOS

Symptoms:
The error:

Configuration error: Clearing IP on interface mgmt failed - Cannot assign requested add

Conditions:
Starting VE in 1NIC mode without a DHCP server and configuring the management interface.

Impact:
No management IP or Self IP.

Workaround:
There is no workaround at this time.

Fix:
Configuring the management IP in 1NIC mode now works.


691589-4 : When using LDAP client auth, tamd may become stuck

Component: TMOS

Symptoms:
When a virtual server uses an LDAP auth profile for client authentication, the system can get into a state where all authentications time out. This condition persists until tamd restarts. Traffic analysis between the BIG-IP system and the LDAP server shows that the BIG-IP system makes a TCP handshake with the server, then immediately sends FIN. Tamd cores may be seen.

Conditions:
-- Virtual server using LDAP auth profile.
-- BIG-IP system makes a TCP handshake with the server, then immediately sends FIN.

Impact:
Authentication to the virtual server fails until tamd is restarted.

Workaround:
To recover, restart tamd by running the following command: bigstart restart tamd

Fix:
tamd no longer becomes stuck when using LDAP client auth.


691504-1 : PEM content insertion in a compressed response may cause a crash.

Solution Article: K54562183


691498-3 : Connection failure during iRule DNS lookup can crash TMM

Component: Global Traffic Manager (DNS)

Symptoms:
The TMM crashes in the DNS response cache periodic sweep.

Conditions:
The DNS resolver connection fails after a successful lookup response is cached. This has been reproduced with a failure due to a lost route. Other failures such as down ports do not cause a crash.

Impact:
The TMM cores and automatically restarts, Traffic disrupted while tmm restarts.

Workaround:
No known workaround.

Fix:
The reference counting of the resolver connection was fixed.


691497-2 : tmsh save sys ucs <file> fails due to missing patch file in /config/.diffVersions

Component: TMOS

Symptoms:
The .diffVersions directory has config change-diffs saved into patch files every time tmsh saves the configs. This is used for diagnostic purposes only.

Conditions:
When doing a 'tmsh save sys ucs <file>', it is possible that a background 'tmsh save sys config' gets run at the same time, causing a patch file to be deleted that the ucs-save feature now expected to exist.

Impact:
The ucs-save feature complains about the missing patch file and exits.

Workaround:
Re-run the 'tmsh save sys ucs <file>' ignoring the missing patch file message. That missing file doesn't need to be restored.

Fix:
With this defect fixed, patch files that end up missing once 'tmsh load sys ucs <file>' is started will not be reported as an error, and the tmsh command will complete normally.


691491-5 : 2000/4000, 10000, i2000/i4000, i5000/i7000/i10000, i15000, B4000 platforms may return incorrect SNMP sysIfxStatHighSpeed values for 10G/40G/100G interfaces

Solution Article: K13841403

Component: TMOS

Symptoms:
2000/4000, 10000, i2000/i4000, i5000/i7000/i10000, i15000, B4000 platforms may return incorrect SNMP sysIfxStatHighSpeed values for 10G/40G/100G interfaces

Conditions:
-- 2000/4000, 10000, i2000/i4000, i5000/i7000/i10000, i15000, B4000 platforms.
-- SNMP query of network interfaces via OID sysIfxStatHighSpeed.

Impact:
Value returned for 10G/40G/100G interfaces may be incorrect.

Workaround:
Use OID sysInterfaceMediaActiveSpeed.

Fix:
The BIG-IP system now correctly returns SNMP sysIfxStatHighSpeed values for 10G/40G/100G interfaces.


691477-2 : ASM standby unit showing future date and high version count for ASM Device Group

Component: Application Security Manager

Symptoms:
Policy builder is changing configuration of standby unit.

Conditions:
The system state changes from active to standby (also when blade is changed from master to non-master).

Impact:
Unexpected changes are made to the policy on standby device (CID increment).

Workaround:
Restart pabnagd when switching device from active to standby (also when blade is changed from master no non-master):

killall -s SIGHUP pabnagd

Fix:
Policy builder now updates its state correctly and doesn't make changes to a policy on a standby device.


691462-1 : Bad actors detection might not work when signature mitigation blocks bad traffic

Component: Anomaly Detection Services

Symptoms:
When signature detected and mitigating no bad actors detection

Conditions:
1. Signatures detected and mitigating
2. Attack traffic is not significantly higher than the good traffic

Impact:
No bad actors detected.
Only signatures provides DoS protection.
BIG-IP CPU utilization is higher than necessary

Workaround:
No workaround at this time.

Fix:
The fix takes in account also SIGNATURES DROPS to decide when bad actors detection should be more agressive.


691287-1 : tmm crashes on iRule with GTM pool command

Component: Global Traffic Manager (DNS)

Symptoms:
tmm crashes with a SIGSEGV when a GTM iRule executes a 'pool' command against Tcl variables that have internal string representations, which can occur when a value is a result of (some) string commands (e.g., 'string tolower') or if the value comes from a built-in iRules command (such as 'class').

For example:

when DNS_REQUEST {
    pool [string tolower "Test.com"]
}

or:

when DNS_REQUEST {
    pool [class lookup pool-dg key-value]
}

Conditions:
GTM iRule executes a 'pool' command against Tcl variables that have internal string representations.

Impact:
Traffic disrupted while TMM restarts.

Workaround:
Pass the 'pool' argument through 'string trim'. For instance:

when DNS_REQUEST {
    pool [string trim [class lookup pool-dg key-value]]
}

Fix:
tmm no longer crashes on GTM iRules that use the 'pool' command.


691224-3 : Fragmented SSL Client Hello message do not get properly reassembled when SSL Persistence is enabled

Solution Article: K59327001

Component: Local Traffic Manager

Symptoms:
Node Server rejects received-and-incomplete ClientHello message and connection terminates.

Conditions:
This occurs when the following conditions are met:
-- SSL Persistence is enabled.
-- There is no ClientSSL and ServerSSL profile.
-- The BIG-IP device receives fragments of a ClientHello message (typically, 11 bytes each) from an SSL front-end client.

Impact:
With Session Persistence enabled
-- The parser fails to reassemble fragmented ClientHello messages prior to passing it on to the backend server.
-- As a result, the backend server responds as if it has received an incomplete ClientHello message, rejects the handshake, and terminates the connection.

Workaround:
The issue disappears when SSL Persistence is disabled.


691210-1 : Traffic stops on tmm restart when guest VLAN tagging is used for BIG-IP VE.

Component: TMOS

Symptoms:
Traffic stops after tmm restart. BIG-IP Virtual Edition (VE) becomes unresponsive and requires power cycle.

Conditions:
This occurs when the following conditions are met:
-- Using VE.
-- Data plane interfaces are SR-IOV VF.
-- Guest VLAN tagging is used.
-- tmm restart.

Impact:
BIG-IP system stops working, and management connection may be lost, requiring power cycle.

Workaround:
Use VLAN tagging from host.

Fix:
The BIG-IP system now continues to work after tmm restart when guest VLAN tagging is used with SR-IOV interfaces for BIG-IP VE.


691095-1 : CA bundle manager loses certificates in the CA bundle if the serial number is longer than 4 bytes

Component: Local Traffic Manager

Symptoms:
CA certificates with long but different serial numbers are treated identical and duplicate, thus get lost in the CA certificate merge operation. Only one would be left.

Conditions:
- The CA bundle file is managed by the CA bundle manager.

- The file contains certificates with large serial numbers.

Impact:
Certificates with large serial numbers are treated as duplicate, and removed.

Workaround:
There is no workaround at this time.

Fix:
Large serial numbers are treated correctly.


691048-1 : Support DIAMETER Experimental-Result AVP response

Solution Article: K34553736

Component: Service Provider

Symptoms:
When the Diameter server returns an answer message with an Experimental-Result AVP, but doesn't have Result-Code AVP inside, then the server side flow is aborted.

Conditions:
Diameter answer message doesn't have Result-Code AVP available, but with Experimental-Result AVP.

Impact:
The server side flow is aborted.

Workaround:
Use iRule to insert Result-Code AVP in all Diameter answer messages.

Fix:
This release supports DIAMETER Experimental-Result AVP response.


690890-1 : Running sod manually can cause issues/failover

Component: TMOS

Symptoms:
If multiple instances of the system failover daemon are executed, improper behavior results. The system failover daemon is run internally by the system service manager (bigstart). When accidentally or intentionally executing the command 'sod', the second running instance will disrupt the failover system.

Conditions:
Accidentally or intentionally executing the command 'sod'.

Impact:
System might failover, reboot, or perform other undesirable actions that result in traffic interruption.

Workaround:
Do not attempt to invoke the 'sod' daemon directly. There is no use case for executing 'sod' directly, it is managed by 'bigstart'.

Fix:
The failover daemon detects that an instance is already running, and exits without disrupting the system.


690883-1 : BIG-IQ: Changing learning mode for elements does not always take effect

Component: Application Security Manager

Symptoms:
When changing learning mode for an element type (e.g., WebSocket URLs), if no other changes are made to the default '*' entity, then suggestions are not created correctly.

Conditions:
Changes are deployed from a BIG-IQ device, where the learning mode for an element type (e.g., WebSocket URLs) is changed (e.g., from Never to Always), and no other changes are made to the default '*' entity.

Impact:
Suggestions are not created correctly.

Workaround:
Modify the '*' entity as well (change description).

Fix:
Learning mode changes are correctly handled from BIG-IQ.


690819-1 : Using an iRule module after a 'session lookup' may result in crash

Component: TMOS

Symptoms:
'session lookup' does not clean up an internal structure after the call finishes. If another iRule module uses the values in this internal structure after a 'session lookup', it may result in a core or other undesired behavior.

Conditions:
Calling 'session lookup' in an iRule where a result is successfully retrieved, and then calling another module.

Impact:
The system may core, or result in undefined and/or undesired behavior.

Workaround:
Check the return value of 'session lookup' before using another iRule module.

If 'session lookup' says that the entry exists, call 'session lookup' again for a key that is known to not exist.


690793-1 : TMM may crash and dump core due to improper connflow tracking

Solution Article: K25263287

Component: TMOS

Symptoms:
In rare circumstances, it is possible for the embedded Packet Velocity Acceleration (ePVA) chip to try to process non-ePVA connflows. Due to this improper internal connflow tracking, TMM can crash and dump core.

Conditions:
This issue can occur on any system equipped with an ePVA and configured with virtual servers that make use of it to accelerate flows.

While no other conditions are required, it is known that modifying a FastL4 virtual server to Standard while the virtual server is processing traffic is very likely to cause the issue.

Impact:
TMM crashes and dumps core. A redundant unit will fail over. Traffic may be impacted while TMM restarts.

Workaround:
It is not recommended to switch virtual server profiles while running traffic. To change virtual server profiles, it is recommended to halt traffic to the system first.

However, this does not eliminate entirely the chances of running into this issue.

Fix:
The system now checks for HSB flow status update data and prevents false positive matches to virtual servers with non-FastL4 profiles.


690778-1 : Memory can leak if the STREAM::replace command is called more than once in the STREAM_MATCHED event in an iRule

Solution Article: K53531153

Component: Local Traffic Manager

Symptoms:
Memory leak; the memory_usage_stat cur_allocs will increase each time the iRule is invoked.

Conditions:
This occurs when using an iRule that calls the STREAM::replace command more than once in the iRule's STREAM_MATCHED event.

Impact:
Memory leak; eventually the system will run out of memory and TMM will restart (causing a failover or outage). Traffic disrupted while tmm restarts.

Workaround:
Change the way the iRule is written so that the STREAM::replace command is not called more than once in the iRule's STREAM_MATCHED event.

Fix:
Prevented memory leak in stream code.


690756-1 : APM depends on undocumented internal behavior of HTTP iRule commands after a retry is initiated

Component: Local Traffic Manager

Symptoms:
Using the ACCESS::restrict_irule_events disable command to allow iRule events triggered by APM-generated responses to be visible to the iRule no longer works.

Conditions:
-- ACCESS::restrict_irule_events disable.
-- HTTP iRules commands used in HTTP_RESPONSE_RELEASE after a retry has been triggered by APM.

Impact:
iRule execution is aborted.

Workaround:
The only possible workaround is to abandon the iRule, and implement the functionality using a VIP-targeting-VIP configuration.

Note: This might not be acceptable in many cases either because of functionality loss (e.g., client certificate auth), or because there are complicated issues specifically solved by iRules.

Fix:
APM triggers a new iRule event when it retries a request. This new event allows iRules to be notified when this occurs.

The HTTP_RESPONSE_RELEASE event is no longer triggered on an internal retry as no response will be sent.

A BigDB variable has been added to disable run-time validation of HTTP iRule commands. This is intended to ease the roll-forward of old APM iRules.


690166-1 : ZoneRunner create new stub zone when creating a SRV WIP with more subdomains

Component: Global Traffic Manager (DNS)

Symptoms:
Creating SRV wideip will result in stub zone creation even there are already matching zones.

Conditions:
Creating SRV wideip with three more layers than existing zone.

Impact:
Unnecessary stub zones created.


690116-1 : websso daemon might crash when logging set to debug

Component: Access Policy Manager

Symptoms:
If the authentication type is HTTP headers and the log level is set to debug, an incorrect parameter gets printed, and if it happens to be NULL the websso daemon crashes.

Conditions:
-- Authentication type is HTTP headers.
-- Log level is debug for WebSSO (the single-sign-on (SSO) functionality for Web access through the BIG-IP APM system).

Impact:
websso daemon might crash.

Workaround:
Set log level to Informational.

Note: The data logged specifically for debug level is targeted toward developers, and is rarely useful in a production environment.

Fix:
The websso daemon no longer crashes when running in debug logging mode and handling certain traffic.


690042-1 : Potential Tcl leak during iRule suspend operation

Solution Article: K43412307

Component: Local Traffic Manager

Symptoms:
TMM's Tcl memory usage increases over time, and does not decrease. Memory leak of Tcl objects might cause TMM to core.

Conditions:
-- iRules are in use.
-- Some combination of nested proc calls and/or loops must go at least five levels deep.
-- Inside the nested calls, an iRule executes a suspend operation.

Impact:
Degraded performance. TMM out-of-memory crash. A failover or temporary outage might occur. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
TMM no longer leaks memory.


689730-3 : Software installations from v13.1.0 might fail

Component: TMOS

Symptoms:
Installation terminates with the following final log messages:

info: updating shared filesystem directories...
progress: 10/100
error: mkdir /mnt/tm_install/3543.JENFeQ/core failed - File exists
Terminal error: Failed to install.

Conditions:
-- BIG-IP Virtual Editions, or the following appliances:
   + i2600
   + i2800
   + i4600
   + i4800
   + i5600
   + i5800
   + i5820
   + i7600
   + i7800
   + i7820
   + i10600
   + i10800
   + i11600
   + i11800

-- Running BIG-IP software v13.1.0 or earlier.
-- Installing BIG-IP software with --instslot option.

Impact:
Installation of new software cannot proceed.

Workaround:
Remove the '/shared/core' symlink, the restart the installation.

Fix:
The installer now properly detects the symlink and proceeds without error.


689691-2 : iStats line length greater than 4032 bytes results in corrupted statistics or merge errors

Component: TMOS

Symptoms:
You can create dynamic statistics using the istats command and iStats directive in iRules. The maximum length of the line (the sum of all columns) is 4032 bytes. If the user attempts to create an iStat whose column sizes when summed exceed this value then there will be errors in the ltm and logs, and the statistic will not be incremented or merged. Log messages appear similar to the following:
-- notice 4: tmstat_row_alloc(1520): tmstat segment blade/tmm0 appears to be damaged at 0x42e2d50.
-- err tmm[21822]: 01220001:3: TCL error: /Common/istat_it <HTTP_REQUEST> - Error: tmstat_row_alloc(1520): tmstat segment blade/tmm0 appears to be damaged (line 1) invoked from within "ISTATS::incr "ltm.virtual [virtual name] counter $host-$path" 1".

Conditions:
An iStat is created or modified such that the sum of the column widths is greater than 4032 bytes.

Impact:
Statistics corruption or merge errors occur. The statistic is not maintained. This is a system limit. An iStat should not be created such that its record length exceeds the 4032-byte limit.

Workaround:
This is a system limit. An istat should not be created such that it's record length exceeds the limit.

Fix:
Line length enforcement was added and an error log is output when the length is exceeded. Now, when the limit is reached, there are no corruption or merge errors. The system posts messages similar to the following in the tmm log file:

-- notice iStat for table 'ltm_virtual' column 'www_qqwabc3584' cannot be added as row size '4040' is too long at 0x46dcd90

To avoid errors like this, do not add columns to iStats in iRule directives.


689591-2 : When pingaccess SDK processes certain POST requests from the client, the TMM may restart

Component: Access Policy Manager

Symptoms:
BIG-IP's tmm may restart when processing certain client's POST requests body on which need to be inspected by the PingAccess policy server.

Conditions:
- BIG-IP virtual server is configured as policy decision point with PingAccess policy server.
- User sends a POST request to BIG-IP.
- Policy configured on PingAccess server requires inspection of the body of the POST request sent by the user.

Impact:
Traffic will be temporarily disrupted while tmm restarts.

Fix:
TMM will no longer restart when processing client's POST requests that need to be inspected by the PingAccess policy server.


689577-3 : ospf6d may crash when processing specific LSAs

Solution Article: K45800333

Component: TMOS

Symptoms:
When OSPFv3 is configured and another router in the network performs a graceful restart, ospf6d may crash.

Conditions:
-- OSPFv3 in use.
-- Graceful restart initiated by another system.

Impact:
OSPFv3 routing will interrupted while the daemon restarts and the protocol re-converges.

Workaround:
Disabling graceful restart on other network systems will prevent ospf6d from crashing on the BIG-IP system. However, it will cause a routing interruption on a system that restarts.

Fix:
The ospf6d daemon no longer crashes when a graceful restart occurs in the network.


689561-1 : HTTPS request hangs when multiple virtual https servers shares the same ip address

Component: Local Traffic Manager

Symptoms:
SSL forward proxy reuses the server ssl session when client ip, server ip and server port matches the ssl session. when multiple virtual https servers share the same ip address, it could happen server ssl reuse a session previously from other virtual server. in such a situation, client cannot forge certificate and hangs the ssh handshake.

Conditions:
multiple virtual https servers share the same ip address, and they internally share the ssl sessions. we saw it happens in several google domain.

Impact:
client cannot access some https web server.

Workaround:
A workaround is disabling the "Session Ticket" in the server ssl profile, since we do not support session id resumption in the server ssl, this will cause it do full handshake to web server every time, so server_certchain will not be NULL.

Fix:
it matches the client ip, server ip and port as well as the server name in the SNI to the server ssl session cache. it will not reuse the sessions does not match virtual server name after the fix.


689540-1 : The same DOS attack generates new signatures even if there are signatures generated during previous attacks.

Component: Anomaly Detection Services

Symptoms:
The same DOS attack generates new signatures even if there are signatures generated during previous attacks.

Conditions:
Repeated DOS attack with the same attacking traffic

Impact:
Generated redundant useless signatures.

Workaround:
There is no workaround at this time.

Fix:
Prevent generation of new signatures handles requests which are already covered by the old ones.


689449-1 : Some flows may remain indefinitely in memory with spdy/http2 and http fallback-host configured

Component: Local Traffic Manager

Symptoms:
As a result of a known issue, in some circumstances the system may experience an unconstrained TMM memory growth when a virtual server is configured for spdy/http2 and http with fallback-host.

Conditions:
- VIP configured with spdy/http2 and http with fallback-host.

Impact:
TMM may eventually enter aggressive sweeper mode where this memory will be released. In the process it is possible that some legitimate connections will killed.

Workaround:
No workaround at this time.

Fix:
BIG-IP no longer attempts to send a response with a configured fallback host in HTTP profile, when a connection is aborted by a client or due to an internal error. It prevents the internal flow to stay in memory after the connection has died.


689437-1 : icrd_child cores due to infinite recursion caused by incorrect group name handling

Solution Article: K49554067

Component: TMOS

Symptoms:
Every time the virtual server stats are requested via REST, icrd_child consumes high CPU, grows rapidly toward the 4 GB max process size (32-bit process), and might eventually core.

Conditions:
Virtual server stats are requested via iControl REST with a special string that includes the dotted group names.

Impact:
icrd_child consumes high CPU, grows rapidly, and might eventually core.

Workaround:
Clear the virtual server stats via reset-stats and icrd_child no longer cores.

Fix:
icrd_child parsing logic update is needed to not enter recursion.


689375-1 : Unable configure 'Generic Alert' setting on SSL client/server profile through TMUI when 'Proxy SSL' is enabled

Solution Article: K01512833

Component: TMOS

Symptoms:
TMUI hides several SSL client/server settings when 'Proxy SSL' is enabled, including the 'Generic Alert' setting, so you cannot set it.

Conditions:
Enable 'Proxy SSL' on SSL client/server profile in TMUI.

Impact:
TMUI hides many config settings, including 'Generic Alert'. You cannot modify 'Generic Alert' setting on SSL client/server profile using the GUI when 'Proxy SSL' is enabled.

Workaround:
Modify the same setting through TMSH modify client/server SSL profile command, as follows:

tmsh modify ltm profile client-ssl <profile_name> generic-alert disabled

tmsh modify ltm profile server-ssl <profile_name> generic-alert disabled

Fix:
You can now modify 'Generic Alert' setting on SSL client/server profile through TMUI when 'Proxy SSL' is enabled.


689343-2 : Diameter persistence entries with bi-directional flag created with 10 sec timeout

Component: Service Provider

Symptoms:
Diameter persistence entries have timeout value less than 10 seconds, although the configured persistence timeout is 120 seconds

Conditions:
When Diameter custom persistence "DIAMETER::persist key 1" bi-directional iRule is used.

Impact:
The persist timeout is less than 10 seconds, thus subsequent requests will be dispatched to other pool member.

Workaround:
Don't use bi-directional persistence iRule. Use AVP persistence type with session-id as the persist key.

Fix:
When the Diameter custom persistence iRule "DIAMETER::persist key 1" is used, the persist timeout value will be set correctly as configured.


689089-1 : VIPRION cluster IP reverted to 'default' (192.168.1.246) following unexpected reboot

Component: Local Traffic Manager

Symptoms:
The cluster configuration file can be lost or corrupted, resulting in the out-of-band cluster management IP reverting to the default value.

Conditions:
Unexpected system restart while the configuration file is being updated may cause the file to become corrupted. If this occurs, the following error will be logged during blade startup:

"err clusterd[8171]: 013a0027:3: Chassis has N slots, config file has 0, ignoring config file"

Where "N" is the number of physical slots in the chassis (2, 4, or 8).

Impact:
Management IP reverts to 192.168.1.246, resulting in loss of access to the chassis through the out-of-band management network.

Workaround:
If this occurs, the management IP can be restored using TMSH or the UI through an in-band self IP, or with TMSH through the management console port.

Fix:
The configuration file update logic has been changed to prevent file corruption during update.


689002-3 : Stackoverflow when JSON is deeply nested

Component: TMOS

Symptoms:
When the returned JSON payload from iControl REST is very large and deeply nested, the JSON destruction could trigger stack overflow due to deep recursion. This will crash icrd_child.

Conditions:
Deeply nested JSON returned from iControl-REST.

Impact:
icrd_child process coredumps.

Workaround:
None.

Fix:
The fix changes the destruction mechanism into an iterative solution, to completely avoid the stack overflow.


688942-5 : ICAP: Chunk parser performs poorly with very large chunk

Component: Service Provider

Symptoms:
When an ICAP response contains a very large chunk (MB-GB), the BIG-IP system buffers the entire chunk and reevaluates the amount of data received as each new packet arrives.

Conditions:
ICAP server returns large payload entirely in a single chunk, or otherwise generates very large chunks (MB to GB range).

Impact:
The BIG-IP system uses memory to buffer the entire chunk. In extreme cases the parser can peg the CPU utilization at 100% as long as packets are arriving.

Workaround:
If possible, configure the ICAP server to chunk the response payload in mulitple normal sized chunks (up to a few tens of KB).

Fix:
When an ICAP response contains a very large chunk (MB-GB), the BIG-IP system streams content back to the HTTP client or server as it arrives, without undue memory use or performance impact.


688911-1 : LTM Policy GUI incorrectly shows conditions with datagroups

Solution Article: K94296004

Component: TMOS

Symptoms:
When editing an LTM policy rule, the GUI defaults to using the datagroup value, overriding previous rule values, because the policy rule introduced the datagroups.

Conditions:
Editing a policy rule.

Impact:
The previous rule values are overridden by the datagroup's values.

Workaround:
Use TMSH to modify the rule.

Fix:
The GUI was updated to default to using the policy rule's values and not the datagroup values.


688813-2 : Some ASM tables can massively grow in size.

Solution Article: K23345645

Component: Application Visibility and Reporting

Symptoms:
/var/lib/mysql mount point gets full.

Conditions:
Many combinations of IP addresses, Device IDs (and other ASM dimensions) in traffic over very long period of time (potentially weeks/months).

Impact:
While other dimensions are being collapsed correctly, the Device ID field is not being collapsed causing the growth in size.
-- High disk usage.
-- Frequent need to clear AVR data.

Workaround:
Manually delete AVR mysql partitions located in /var/lib/mysql/AVR.

Fix:
Over time, no of the AVR_STAT_ASM_HTTP_CLIENT_IP_X#...MYD file exceeds 300 MB, so this problem no longer occurs.


688744-1 : LTM Policy does not correctly handle multiple datagroups

Solution Article: K11793920

Component: Local Traffic Manager

Symptoms:
Policy in use where the conditions reference two or more datagroups, but only the conditions that refer to the first datagroup have any effect.

Conditions:
LTM Policy where the conditions reference two or more datagroups.

Impact:
LTM Policy conditions do not check against second and subsequent datagroups, resulting in policy not working as intended.

Workaround:
Only mitigation is to refactor the policy so that it does not refer to more than one datagroup-based condition.

Fix:
LTM Policy correctly handles policies referencing multiple datagroups


688629-1 : Deleting data-group in use by iRule does not trigger validation error

Solution Article: K52334096

Component: Local Traffic Manager

Symptoms:
iRule aborts due to failed commands, causing connflow aborts.

Conditions:
-- Delete a data group.
-- iRule uses that data-group on a virtual server

Impact:
In use data-group can be deleted without error. iRule aborts leading to connflow aborts.

Workaround:
Don't delete data-groups in use by an iRule.

Fix:
An attempt to delete a data-group in use by an iRule now triggers a validation error.


688625-5 : PHP Vulnerability CVE-2017-11628

Solution Article: K75543432


688571-2 : Untrusted cert might be accepted by the server-ssl even though when 'untrusted-cert-response-control drop' is configured in the server-ssl profile.

Solution Article: K40332712

Component: Local Traffic Manager

Symptoms:
If the server-ssl profile is configured with 'untrusted-cert-response-control drop', and the system receives the certificate from the backend server which is not trusted by current BIG-IP system, the expected behavior is that the system should end the connection.

But the current server-side behavior is that the system still accepts the untrusted certificate and establishes the SSL connection with backend server.

Conditions:
-- The BIG-IP system receives a certificate from the backend server that is not trusted by the BIG-IP system.

-- Configure the 'untrusted-cert-response-control drop' in the server-ssl profile.

-- The corresponding server-ssl is configured at the virtual server.

Impact:
Virtual server might communicate with the backend server that sends the untrusted certificate to the BIG-IP system. Untrusted cert could still be accepted by the server-ssl virtual server even though 'untrusted-cert-response-control drop' is configured in the server-ssl profile.

Workaround:
None.

Fix:
When the system receives the untrusted certificate from backend server and the server-ssl profile is configured with 'untrusted-cert-response-control drop', the system will end the current SSL handshake procedure instead of continuing to proceed to finish it.


688570-5 : BIG-IP occasionally sends MP_FASTCLOSE after an MPTCP connection close completes

Component: Local Traffic Manager

Symptoms:
After an MPTCP connection closes properly, the BIG-IP will occasionally start sending MP_FASTCLOSE.

Conditions:
An MPTCP connection is closed.

Impact:
The MPTCP connection on the remote device is closed, but the connection on the BIG-IP remains open until the fastclose retransmission times out.

Workaround:
There is no workaround at this time.

Fix:
Fixed event processing at the end of a connection.


688557-1 : Tmsh help for ltm sasp monitor incorrectly lists default mode as 'pull'

Solution Article: K50462482

Component: Local Traffic Manager

Symptoms:
The description of the 'mode' parameter shown by tmsh help for the ltm sasp monitor indicates that the default value for the 'mode' parameter is 'pull' (where the load balancer sends Get Weight Requests to the GWM).
As of BIG-IP v13.0.0 and v12.1.2-hf1, the default value for the 'mode' parameter is actually 'push' (where the load balancer receives Send Weights from the GWM).
This is a change from prior versions of BIG-IP, where the default value for the 'mode' parameter was 'pull'.

Conditions:
The incorrect description appears when issuing the 'tmsh help ltm monitor sasp' command on BIG-IP v13.0.0, v12.1.2-hf1, and later.

Impact:
Incorrect information about the default value for the 'mode' parameter for the ltm sasp monitor.

Workaround:
The default value for the 'mode' parameter is actually 'push' (where the load balancer receives Send Weights from the GWM).

Fix:
The 'tmsh help ltm monitor sasp' command now lists the correct default value for the 'mode' parameter.


688516-1 : vCMPd may crash when processing bridged network traffic

Solution Article: K03165684


688406-1 : HA-Group Score showing 0

Solution Article: K14513346

Component: TMOS

Symptoms:
The 'show sys ha-group' command incorrectly displays a "0" for the total score even if the pools/trunks/clusters components have non-zero scores.

Conditions:
If the sys ha-group object is not currently assigned to any traffic-group.

Impact:
The total score is not calculated. An incorrect score value is displayed.

Workaround:
Refer to the component Score Contributions displayed using the following command: show sys ha-group detail.

Fix:
The total HA-Group Score is now displayed correctly.


688246-1 : An invalid mode in the LSN::persistence command causes TMM crash

Component: Carrier-Grade NAT

Symptoms:
When an iRule is triggered and the LSN::persistence command is passed an invalid persistence mode, TMM will crash.

Conditions:
An iRule using the LSN::persistence command with an invalid persistence mode that is attached to a Virtual Server and is triggered by traffic.

Impact:
TMM restarts. Traffic is interrupted. It is likely that the iRule will be triggered again causing repeated crashes.

Workaround:
The persistence mode must be set to one of "none", "address", "address-port" or "strict-address-port".

Fix:
TMM no longer crashes when an invalid persistence mode is used. Instead the LSN::persistence command returns an error.


688011-7 : Dig utility does not apply best practices

Solution Article: K02043709


688009-7 : Appliance Mode TMSH hardening

Solution Article: K46121888


687987 : Presentation of signatures in human-readable format

Component: Anomaly Detection Services

Symptoms:
When publishing signature with predicates such as http.referer and http.uri, the system presents the result of the hash operation as follows: http.uri_file_hashes-to 42

Conditions:
Always when publishing signature with predicates such as http.referer and http.uri.

Impact:
It is not clear what '42' means.

Workaround:
None.

Fix:
When publishing signatures, the system now presents the hashes as follows:

http.referer_hashes-like '/zzz'
http.uri_file_hashes-like '/123'


687986 : High CPU consumption during signature generation, not limited number of signatures per virtual server

Component: Anomaly Detection Services

Symptoms:
The number of the signatures per virtual server is not limited. This can result in a very large number of generated signatures during sophisticated attacks that use changing patterns. After a time, when a system experiences a number of attacks, the list of generated signatures can be too long.

Conditions:
-- Sophisticated attacks that use changing patterns.
-- System experiences a large number of attacks.

Impact:
High CPU utilization when mitigating. Overloaded GUI signatures screen.

Workaround:
Manually remove old / not-often-used signatures.

Fix:
The system now limits the number of signatures per virtual servers, and optimizes per-signatures operation.


687984 : Attacks with randomization of HTTP headers parameters generates too many signatures

Component: Anomaly Detection Services

Symptoms:
When attackers randomize HTTP headers parameters, Behavioral DoS (BADoS) might generate too many signatures.

Conditions:
Attacks with randomization of HTTP headers parameters.

Impact:
The list of generated signatures is too long. It produces unnecessary CPU utilization for attack mitigation.

Workaround:
None.

Fix:
Improved algorithm that detects a randomization.


687937-1 : RDP URIs generated by APM Webtop are not properly encoded

Component: Access Policy Manager

Symptoms:
RDP URIs used to launch Native RDP resources form APM Webtop on Android/iOS/Mac are not properly encoded. As a result, RDP client might misinterpret the URI, in cases in which some RDP parameter contains the ampersand ( & ) symbol.

Conditions:
Native RDP resource is launched via RDP URI from APM Webtop.

One of RDP parameters contains symbol that should be URI encoded, e.g., '&'.

Impact:
RDP client misinterprets URI, which may result in failure to open RDP resource.

Workaround:
None.

Fix:
RDP URIs used to launch Native RDP resources from APM Webtop on Android/iOS/Mac are now properly encoded.


687905-2 : OneConnect profile causes CMP redirected connections on the HA standby

Solution Article: K72040312

Component: TMOS

Symptoms:
When virtual server uses OneConnect profile in HA setup, it can cause Clustered Multiprocessing (CMP) redirected connections and memory leak on high availability (HA) standby systems, including high memory usage on standby units.

Conditions:
-- Virtual server uses OneConnect profile in HA configuration.
-- Mirroring is enabled.
-- BIG-IP platform supports CMP.

Impact:
Redirected connections and memory leak on a standby device.

Workaround:
Remove OneConnect profile from the virtual server.


687759-1 : bd crash

Component: Application Security Manager

Symptoms:
A bd crash.

Conditions:
-- A config change follows a bd crash.
-- There is a policy that is misconfigured, for example, a form-data parsing is applied on a non-form-data payload (such as XML or JSON).

Impact:
bd crashes; system fails over; traffic disturbance occurs.

Workaround:
Set the following internal parameter to 0: max_converted_length_to_cache


687658 : Monitor operations in transaction will cause it to stay unchecked

Component: TMOS

Symptoms:
If a monitored object is deleted and created or modified in the same transaction, and any of its monitor configuration is changed (either the monitor, or the state user-down), the monitor state will become unchecked.

Conditions:
This only happens within transactions.

Note: Using the command 'modify ltm pool <name> members replace-all-with' is considered a transaction containing a delete and create of pool members.

Impact:
Monitor state never returns to its correct value.

Workaround:
Do not do these operations in transactions. For pool members, use 'modify ltm pool <name> members modify' instead of replace-all-with.


687635-1 : Tmm becomes unresponsive and might restart

Solution Article: K58002142

Component: Local Traffic Manager

Symptoms:
Under certain conditions, the tmm process can become unresponsive, and eventually be restarted by the watchdog process.

Conditions:
Problem occurs when there is an unexpected interaction between HTTP and SSL handlers during an abnormal connection shutdown.

Impact:
Tmm becomes unresponsive and restarts. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Tmm correctly shuts down HTTPS connection.


687534-1 : If a Pool contains ".." in the name, it is impossible to add a Member to this pool using the GUI Local Traffic > Pools : Member List page

Component: TMOS

Symptoms:
- Create a pool with name containing two dots (that is, the string '..')
- Go to the GUI Local Traffic :: Pools : Member List page and click the Add button to add a member.
- There is a No Access error preventing you from adding a member to the pool

Conditions:
This issue occurs when a pool name contains .. in the name.

Impact:
Cannot add a Member to the pool using the GUI.

Workaround:
Use tmsh to add pool members to an existing pool using a command similar to the following.
 tmsh modify ltm pool <pool name> members add { <member info> }

Fix:
For pools with '..' in the name, it is now possible to add members after pool creation using the GUI Local Traffic :: Pools : Member List page.


687353-1 : Qkview truncates tmstat snapshot files

Solution Article: K35595105

Component: TMOS

Symptoms:
Qkview truncates the snapshot files it collects in /shared/tmstat/snapshots/.

Conditions:
Files are larger than 5 MiB, or the 'max file size' limit specified when running Qkview (the -s argument).

Note: 5 MiB is qkview utility's default maximum file size value.

Impact:
Snapshot data may not be collected in qkview. This may result in data being lost if the issue is only identified once important data has rotated out of history.

Workaround:
To specify no file size limit when collecting qkviews, use the following tmsh command:
qkview -s0


687205-2 : Delivery of HUDEVT_SENT messages at shutdown by SSL may cause tmm restart

Component: Local Traffic Manager

Symptoms:
During flow shutdown, SSL may deliver HUDEVT_SENT messages, causing additional messages to be queued by higher filters, which may result in a tmm crash and restart.

Conditions:
This happens in response to a relatively rare condition that occurs during shutdown, such as HUDEVT_SENT queued after HUDCTL_SHUTDOWN.

Impact:
Possible tmm restart. Traffic disrupted while tmm restarts.

Workaround:
None.


687128-1 : gtm::host iRule validation for ipv4 and ipv6 addresses

Component: Global Traffic Manager (DNS)

Symptoms:
gtm::host iRule isn't validating that the host given matches the type of WideIP it is attached to.

Conditions:
An AAAA-type wideip with a ipv4 gtm::host iRule, or A-type wideip with an ipv6 gtm::host iRUle.

Impact:
Incorrect host information was being returned.

Workaround:
Only attach gtm::host of IPv4 type to A-type WideIPs, and gtm::host rules of IPv6 to AAAA-type WideIPs.

Fix:
FIxed issue allowing incorrect gtm::host iRule commands to trigger on incorrect wideip types.


686972-4 : The change of APM log settings will reset the SSL session cache.

Component: Local Traffic Manager

Symptoms:
If you change the configuration of APM log settings, it might cause the SSL session cache to be reset. Also, subsequent resumption of SSL sessions may fail after such change causing a situation where full ssl handshakes may occur more frequently.

Conditions:
-- Change the configuration of APM log settings.
-- SSL session cache is not empty.

Impact:
The change of APM log settings resets the SSL session cache, which causes the SSL session to initiate full-handshake instead of abbreviated re-negotiation.

Workaround:
Follow this procedure:
1. Change access policy.
2. The status of that access policy changes to 'Apply Access Policy'.
3. Re-apply that.

Fix:
The change of APM log settings now limits its effect on APM module instead of affecting other (SSL) module's data.


686906-2 : Fragmented IPv6 packets not handled correctly on Virtual Edition

Component: TMOS

Symptoms:
Use of IP fragmentation with IPv6 packets might not be handled correctly by BIG-IP Virtual Edition (VE) platforms. The initial fragmented are received, but subsequent fragments are discarded.

Conditions:
VE with IPv6 packets and IP fragmentation.

Impact:
Traffic which depends upon fragmented IPv6 packets will not be successfully processed.

Workaround:
There is no workaround at this time.

Fix:
These fragments are now handled correctly in the same manner as IPv4.


686890-1 : X509_EXTENSION memory blocks leak when C3D forges the certificate.

Component: Local Traffic Manager

Symptoms:
One X509_EXTENSION memory block leaks when C3D forges the certificate.

Conditions:
When C3D forges the certificate.

Impact:
X509_EXTENSION memory blocks leak when forged certificate is successful.

Workaround:
None.

Fix:
The system now frees the leaked X509_EXTENSION when C3D forges the certificate.


686765-2 : Database cleaning failure may allow MySQL space to fill the disk entirely

Component: Application Security Manager

Symptoms:
Database cleaning failure may allow MySQL space to fill the disk entirely, which prevents any modification of ASM policy configuration.

In /var/log/ts/asm_config_server.log you might see these errors repeatedly:

Jun 9 09:38:45 10gal-f5-5000-2 crit g_server_rpc_handler.pl[16654]: 01310027:2: ASM subsystem error (asm_config_server.pl,F5::ASMConfig::Handler::handle_error): Code: 999, Error message = DBD::mysql::db do failed: The table 'NEGSIG_SIGNATURES' is full

Conditions:
This occurs if database cleaning failures occur.

Impact:
Disk will fill up, and you will be unable to modify ASM policies.


686631-2 : Deselect a compression provider at the end of a job and reselect a provider for a new job

Component: Local Traffic Manager

Symptoms:
The system might potentially retain a compression context, even though there is no data to be compressed or decompressed. This can affect the calculation of the load of the compression provider.

Conditions:
-- A connection is up.
-- Compression context is active.
-- There is no data for the compression provider.

Impact:
It affects the compression provider selection.

Workaround:
None.

Fix:
The system now deselects a provider at the end of a compression/decompression operation, and reselects a provider at the beginning of another compression/decompression operation.


686517-2 : Changes to a parent policy that has no active children are not synced to the secondary chassis slots.

Component: Application Security Manager

Symptoms:
Changes to a parent policy that has no active children are not synced to the secondary chassis slots.

Conditions:
-- ASM provisioned.
-- Having a parent policy that has no active children.

Impact:
On a chassis failover, the new Primary slot will have an outdated version of the parent policy.

Workaround:
None.

Fix:
Changes to a parent policy that has no active children are now synced to the secondary chassis slots.


686510-1 : If tmm was restarted during an attack, the attack might appear ongoing in GUI

Component: Application Visibility and Reporting

Symptoms:
Attack appears ongoing, even though it ended.

Conditions:
Rare condition of tmm restart during an attack.

Impact:
The GUI falsely shows the attack as ongoing, even though it ended.

Workaround:
No workaround.

Fix:
Now, when tmm is restarted during an attack, this specific attack is shown as ended in DoS overview page after 15 minutes.


686470-1 : Enable AJAX Response Page or Single Page Application support causes the part of the web page failed to load.

Component: Application Security Manager

Symptoms:
AJAX requests are not sent, JavaScript errors, AJAX-based web-app malfunctions.

Conditions:
1a. Single page application enabled either via a DoS application profile or in an ASM policy.
1b. AJAX Response Page enabled via ASM policy.

2. Web Application client side code uses jQuery or any other AJAX clientside framework.

Impact:
AJAX request might not be sent, and the overall website's clientside functionality related to the AJAX requests might not work as expected.

Workaround:
Disable Single Page Application support.

Fix:
Fixed Single Page Application AJAX hook to support the AJAX response onload callback re-assignment.


686452-1 : File Content Detection Formats are not exported in Policy XML

Component: Application Security Manager

Symptoms:
If a policy is configured with Data Guard enabled with File Content Detection, the selected File Content Detection Formats are not correctly exported in the Policy XML.

When the policy is then imported, the Data Guard settings will be invalid and cannot be changed until the File Content Detection Formats are configured again.

Conditions:
A policy is configured with Data Guard enabled with File Content Detection, and then exported in XML format.

Impact:
When the policy is then imported, the Data Guard settings will be invalid and cannot be changed until the File Content Detection Formats are configured again.

The formerly selected file content formats will not be correctly identified.

Workaround:
Use Binary Policy import/export.

Fix:
File Content Detection Formats are correctly exported.


686389-1 : APM does not honor per-farm HTML5 client disabling at the View Connection Server

Component: Access Policy Manager

Symptoms:
Current logic for determining whether to offer HTML5 client option works for Horizon 6.x (and earlier) but it does not work for Horizon 7.x.

With Horizon 7.x, VMware has enhanced the XML so that each resource includes a flag to indicate whether HTML5 client is enabled (absence of <html-access-disabled/> tag). APM does not honor this flag to show HTML5 client option to APM end user only if it has been enabled for that resource.

Conditions:
-- APM webtop with a VMware View resource assigned.
-- HTML5 Access disabled for some of the RDS farms managed by the broker.

Impact:
APM offers HTML5 client launch option and actually runs it if requested, although it is disabled at the backend.

Workaround:
There is no workaround at this time.

Fix:
For Horizon 7.x, the system now honors the <html5-access-disabled> flag in broker responses to disable HTML5 client for those RDS desktops and apps that have this flag set.

Behavior Change:
Before this fix, all the RDS desktops and apps were available for HTML5 client if it was installed on VCS.
Now, for those desktops and apps where HTML5 access has been deliberately disabled at the broker, only the native client option will be available.


686376-2 : Scheduled blob and current blob no longer work after restarting BIG-IP or PCCD daemon

Component: Advanced Firewall Manager

Symptoms:
When there are scheduled firewall rules, and the BIG-IP system is restarted or PCCD daemon is restarted, new blob compilation succeeds, but TMM fails to activate the new blob. Both GUI and TMSH show error status: Firewall rules deployment failed. After the system gets in this state it cannot be fixed except by removing or disabling all scheduled firewall rules.

Conditions:
-- There are scheduled firewall rules.
-- The BIG-IP system is restarted or the PCCD daemon is restarted.

Impact:
After this failure, firewall rules are not applied on data traffic.

Workaround:
Remove or disable all scheduled firewall rules.

Fix:
New blob deployed and new firewall rules applied successfully.


686307-3 : Monitor Escaping is not changed when upgrading from 11.6.x to 12.x and later

Solution Article: K10665315

Component: Local Traffic Manager

Symptoms:
When upgrading, monitor attributes such as receive string and send string might contain escape sequences that must be processed after the upgrade. However, due to a problem introduced by the LTM policy upgrade script, this processing is not performed, resulting in monitors not functioning correctly after the upgrade.

Note: Without LTM policies in the configuration, monitors upgrade without problem.

Conditions:
-- Upgrading and rolling forward monitor configuration data.
-- LTM policy data present.

Impact:
Monitors may not work after upgrade.

Workaround:
No workaround at this time.

Fix:
This release addresses the underlying problem so the issue no longer occurs.


686305-1 : TMM may crash while processing SSL forward proxy traffic

Solution Article: K64552448


686228-1 : TMM may crash in some circumstances with VLAN failsafe

Solution Article: K23243525

Component: Local Traffic Manager

Symptoms:
TMM may crash when managing traffic in response to the VLAN failsafe traffic generating mechanisms

Conditions:
- VLAN failsafe is configured with low timers.
- VLAN failsafe is triggered and multiple responses are received for traffic generating in fast succession.

Impact:
A TMM may core file may be produced. Traffic disrupted while tmm restarts.

Workaround:
Relax the timer to the default VLAN failsafe timer setting.

Fix:
TMM no longer crashes in some circumstances with VLAN failsafe.


686190-1 : LRO performance impact with BWC and FastL4 virtual server

Component: TMOS

Symptoms:
Using Bandwidth controller (BWC) might result in a very large drop in performance of up to 75%. In this release, Large receive offload (LRO) is enabled by default.

Conditions:
-- BWC is configured.
-- Virtual server has a FastL4 profile assigned.
-- LRO is enabled (enabled by default in 13.1.0).

Impact:
Very large performance impact to the BWC policy (up to 75%). For example, if the BWC policy rate limit is set to 100Mb, the actual rate limit could be 25Mb.

Workaround:
Disabling LRO recaptures most of the performance degradation related to using FastL4. To disable LRO (this is a system-wide setting), run the following command:
 tmsh modify sys db tm.largereceiveoffload value disable

Important note: Although you can disable LRO to recapture much of the 13.0.0-level performance, you will likely still experience some impact: 2-5% for small files, 17-22% degradation for the '10 requests per connection' benchmark. The only guaranteed way to avoid performance degradation is to remain on version 13.0.0.


686108-1 : User gets blocking page instead of captcha during brute force attack

Component: Application Security Manager

Symptoms:
Unexpected blocking page while captcha is configured.

Conditions:
-- Brute force configured with alarm and captcha mitigation.
-- The only source configured is username.
-- These are the first failed login requests after a system start up or after a login page configuration change.

Impact:
Unexpected blocking page mitigation where captcha mitigation was expected.

Workaround:
There are two workarounds:

-- Access the login page at least 10 times within 5 minutes.

-- Run the following command: tmsh modify sys db asm.cs_qualified_urls value <YOUR_LOGIN_URL>

Fix:
Fixed an issue with unexpected blocking page while captcha is configured.


686065-2 : RESOLV::lookup iRule command can trigger crash with slow resolver

Component: Local Traffic Manager

Symptoms:
If thousands of connections are serviced by an iRule that performs a lookup for the same FQDN before the FQDN can be resolved, tmm may crash.

Conditions:
iRule with RESOLV::lookup.
Slow DNS resolver.
Thousand of connections triggering resolution of the same name.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Remove RESOLV::lookup from the workflow if it is not required.

Fix:
The scenario now works as expected and no longer results in a crash.


686029-2 : A VLAN delete can result in unrelated VLAN FDB entries being flushed on shared VLAN member interfaces

Component: TMOS

Symptoms:
FDB flushing on VLAN deletes is performed by VLAN member interface reference only, without regard to VLAN tags. This can result in unrelated VLAN FDB entries also being flushed on shared VLAN member interfaces.

Conditions:
Issuing a VLAN delete with other VLANs using shared tagged member interfaces with the VLAN being deleted.

Impact:
FDB entries for unrelated VLANs will be flushed if they share the same tagged VLAN member interfaces as the VLAN being deleted.

Workaround:
None.

Fix:
Correct FDB flushing on VLAN deletes, by limiting the scope to be VLAN specific.


685964-1 : cs_qualified_urls bigdb does not cause configured URLs to be qualified.

Component: Application Security Manager

Symptoms:
cs_qualified_urls is configured but is not functional.

Conditions:
-- cs_qualified_urls is configured.
-- A request to the URL listed in the bigdb arrives.
-- The URL is seen as non-qualified although configured.

Impact:
URLs that are not supposed to getting through configuration.

Workaround:
None.

Fix:
Fixed a bigdb issue with cs_qualified_urls variable.


685862-1 : BIG-IP as SAML IdP/SP may include last x509 certificate found in the configured bundle in signed SAML Response or single logout message

Component: Access Policy Manager

Symptoms:
When BIG-IP is used as SAML IdP, and signing is configured, BIG-IP will sign the message by configured signing key, and include last certificate from the configured signing certificate chain in the SAML protocol message. Expected behavior is to include first certificate from the configured signing certificate chain.
The same applies to SAML SP generating SLO request/response messages.

Conditions:
All of the following:
- BIG-IP is used as SAML IdP or SAML as SP with SLO configured.
- BIG-IP generates signed SAML response containing assertion or SLO request/response
- Configured on BIG-IP signing certificate is a security chain and not a single certificate

Impact:
Impact is based on the SAML implementation on the receiving end of message sent by BIG-IP.
Some implementation may drop signed SAML message if last certificate from the bundle is included in the message. Other implementations will accept such signed messages. Note that signing operation itself is performed correctly by BIG-IP using configured signing certificate, and digital signature will contain correct value.

Workaround:
Instead of using signing certificate chain, change BIG-IP SAML IdP/SP configured signing certificate to refer to a standalone signing certificate (single X509 object) by extracting first certificate from the chain.

Fix:
After the fix, BIG-IP will include first certificate found within configured signing certificate (chain).


685771-1 : Policies cannot be created with SAP, OWA, or SharePoint templates

Component: Application Security Manager

Symptoms:
Attempting to create a security policy using "OWA Exchange", "SAP NetWeaver", or "SharePoint" Policy Template fails.

Conditions:
Attempt to create a security policy using "OWA Exchange", "SAP NetWeaver", or "SharePoint" Policy Templates

Impact:
Policy creation fails.

Workaround:
None.

Fix:
Policies can be created using these factory templates.


685708-4 : Routing via iRule to a host without providing a transport from a transport-config created connection cores

Component: Service Provider

Symptoms:
Using MR::message route command without specifying a transport to use (virtual or config) will core if the connection receiving the request was created using a transport-config.

Conditions:
Using MR::message route command without specifying a transport to use (virtual or config) will core if the connection receiving the request was created using a transport-config.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Specify a transport to use for creating a new outgoing connection in the MR::message route command.

Fix:
The system will no longer core.


685628-1 : Performance regression on B4450 blade

Component: Performance

Symptoms:
Performance degradation may occur for certain types of traffic when the system is under heavy traffic load. L4 and L7 performance may be degraded by up to 5% compared to previous BIG-IP releases.

Conditions:
- L4 and L7 traffic when system is under heavy traffic load.
- VIPRION B4450 blades.

Impact:
You may encounter a performance degradation for certain types of traffic upon upgrading.

Workaround:
None.

Fix:
Performance regression on B4450 blade has been eliminated.


685615-4 : Incorrect source mac for TCP Reset with vlangroup for host traffic

Solution Article: K24447043

Component: Local Traffic Manager

Symptoms:
BIG-IP outbound host TCP RST packets have incorrect source-mac-address.

Conditions:
BIG-IP host traffic is exiting via VLANs in a VLAN group.

Impact:
TCP Reset for traffic exiting the BIG-IP system with incorrect source-mac-address, which could include monitor traffic.

Workaround:
Use transparent mode on the VLAN group.

Fix:
source-mac-address for host traffic is correctly set.


685582-7 : Incorrect output of b64 unit key hash by command f5mku -f

Component: TMOS

Symptoms:
The output of the b64 unit key hash is inconsistent upon each 'f5mku -f' command, whereas the hex version of the unit key hash was always correct/consistent.

Conditions:
Viewing output of 'f5mku -f' command.

Impact:
Inconsistent output of the b64 unit key.

Workaround:
Adding the verbose option (v) to the f5mku command will print additional information. The following command prints the hex version of the unit key header hash, which will be stable and can be used to detect changes to the unit key:

 f5mku -vf

For example:

# f5mku -vf
...
-- hdr.hash = c9:0d:13:2a:74:d4:7e:31:a4:78:5e:c8:3e:9c:b5:3d:7b:65:9c:7d
...

Fix:
The unit key hash is now the correct length and is consistent upon each 'f5mku -f' command.


685475-1 : Unexpected error when applying hotfix

Solution Article: K93145012

Component: TMOS

Symptoms:
Software 'install hotfix' commands fail with a message similar to the following: Image (BIG-IP-11.6.1.0.0.317.iso) has a software image entry in MCP database but does not exist on the filesystem.

Conditions:
Before installing the hotfix, the necessary full (base) software image prerequisite was added to the images directory, but then was removed before issuing the hotfix command.

For example, to apply 'Hotfix-BIG-IP-11.6.1.2.0.338-HF2.iso', the system must have access to the base image; 'BIG-IP-11.6.1.0.0.317.iso'.

Here is another example: on multi-bladed VIPRION systems, where it is resolved by running 12.1.3.6.

1) Install and boot into 12.0.0 on the VIPRION system:
-- install /sys software image 12.0.0.iso create-volume volume HD1.test
-- reboot volume HD1.test
2) Install and boot into 12.1.2.0.402.249:
-- install /sys software hotfix Hotfix-BIG-IP-12.1.2.0.402.249-ENG.iso create-volume volume HD1.test2
-- reboot volume HD1.test2
3) Delete 12.0.0.iso and volume HD1.test:
-- delete sys software image 12.0.0.iso
-- delete sys software volume HD1.test
4) Copy over Hotfix-BIG-IP-13.1.0.7.0.17.1-ENG.iso without the 13.1.0.7 base image.
5) Check the /var/log/ltm logs for the following message:
-- lind[6288]: 013c0006:5: Image (BIG-IP-12.0.0.0.0.606.iso) has a software image entry in MCP database but does not exist on the filesystem.

Impact:
Cannot apply hotfix until the full base image is present.

Workaround:
Perform the following procedure:
1. Copy the base image to the images directory on the BIG-IP system.
2. Restart the 'lind' daemon.
3. Try the hotfix installation operation again.

Fix:
Issuing a 'install hotfix' command when the base image is not available sends the system into a 'wait' state. The process status is 'waiting for base image', which should make clear what needs to be done. When the base image becomes available (in the images directory), the hotfix installation proceeds.


685467-1 : Certain header manipulations in HTTP profile may result in losing connection.

Solution Article: K12933087

Component: Local Traffic Manager

Symptoms:
HTTP profile has an option 'Insert X-Forwarded-For' which adds a header when a request is forwarding to a pool member. When a virtual server has iRule with a collect command like SSL::collect, it is affected by the HTTP profile processing. Same issue has an option 'Request Header Erase' available in HTTP profile.

Conditions:
A virtual server meets the following conditions:
-- ClientSSL profile.
-- HTTP profile with option 'Insert X-Forwarded-For' enabled and/or configured option 'Request Header Erase'.
-- iRule that has an 'SSL::collect' command in at least two events (e.g., CLIENTSSL_HANDSHAKE and CLIENTSSL_DATA).

Impact:
TCP connection is reset, and no response is provided to a client.

Workaround:
Use iRule to insert X-Forwarded-For with an appropriate IP address and/or remove headers configured in 'Request Header Erase' option of HTTP profile.

Fix:
An issue of a resetting connections due to configuration options 'Insert X-Forwarded-For' and 'Request Header Erase' in HTTP profile no longer happens.


685458-7 : merged fails merging a table when a table row has incomplete keys defined.

Solution Article: K44738140

Component: TMOS

Symptoms:
There is as timing issue in merged where it will fail processing a table row with incomplete keys defined.

Conditions:
There are no specific conditions required, only that merged is running, which is true on every BIG-IP system, when the BIG-IP system is processing a table row with incomplete keys defined.

Although this issue is not dependent on configuration or traffic, it appears that it is more prevalent on vCMP hosts.

Impact:
There will be a few second gap in available statistics during the time when a core is being created and merged restarts.

Workaround:
None.

Fix:
merged now detect this scenario, a table row with incomplete keys defined, and does not fail.


685230-3 : memory leak on a specific server scenario

Component: Application Security Manager

Symptoms:
The bd process memory increases.

Conditions:
A specific server scenario of handling the traffic.

Impact:
Swap may be used. The kernel OOM killer may be invoked. Possible traffic disturbance.

Workaround:
There is no workaround at this time.

Fix:
A memory leaked related to a specific server scenario was fixed.


685207-1 : DoS client side challenge does not encode the Referer header.

Component: Application Security Manager

Symptoms:
XSS reflection when DoS client side is enabled as a mitigation, or a proactive bot defense is enabled.

Conditions:
1. Login to the client IP address and send the ab request.
2. Once the DoS attack starts, sends the curl request
hl=en&q=drpdrp'-alert(1)-'drpdrp".
3. Unencoded Referer header is visible.

Impact:
The XSS reflection occurs after triggering the DoS attack.

Workaround:
None.

Fix:
DoS client side challenge now encodes the Referer header.


685193-1 : If Inheritance is None in the Parent Policy and there are at least 1 child policy, the number of Comments shown in Inheritance Settings is equal to number of child policies

Component: Application Security Manager

Symptoms:
If Inheritance is None in the Parent Policy and there is at least one child policy, the number of Comments shown in Inheritance Settings is equal to number of child policies.

Conditions:
1) Create Parent policy and set some section's Inheritance to None.
2) Create child policy and assign it to the parent created above.
3) Go to the Parent Policy Inheritance Setting tab, you will see that number of comments for sections with None will be equal to number of child policies.

Impact:
There is an incorrect number of Comments shown in Inheritance Settings

Workaround:
None.

Fix:
The correct number of comments will be shown for each section in Inheritance Setting tab for Parent Policy. In case of None inheritance nothing will be shown.


685164-1 : In partitions with default route domain != 0 request log is not showing requests

Component: Application Security Manager

Symptoms:
No requests in request log when partition selected, while they can be seen when 'All [Read Only]' is selected.

Conditions:
Select a partition whose default route domain is not 0 (zero).

Impact:
No requests in request log.

Workaround:
As a partial workaround, you can use [All], but it's read only.

Fix:
Fixed filter by Source IP, which worked incorrectly in partitions whose default route domain was not 0 (zero).


685110-1 : With a non-LTM license (ASM, APM, etc.), ephemeral nodes will not be created for FQDN nodes/pool members.

Solution Article: K05430133

Component: Local Traffic Manager

Symptoms:
1. FQDN Node/pools fails to populate with members.

2. An error similar to the following is logged in /var/log/ltm when an FQDN node or pool member is created:

err mcpd[####]: 01070356:3: Ratio load balancing feature not licensed.

Conditions:
1. License a BIG-IP system with non-LTM license lacking the ltm_lb_ratio feature.
Such licenses include APM and/or ASM licenses for certain newer platforms which do not support the AAM module.
Affected platforms include certain iSeries and Virtual Edition (VE) releases.
2. Configure an FQDN node/pool member. Do not specify a 'ratio' value.

Impact:
Unable to use FDQN nodes/pool members with non-LTM license.

Workaround:
None.

Fix:
Non-LTM license (ASM, APM, etc.), ephemeral nodes are now created for FQDN nodes/pool members.


685056 : VE OVAs is not the supported platform to run VMware guest OS customization

Component: TMOS

Symptoms:
VMware vCenter fails to create customization specification wizard because the BIG-IP Virtual Edition (VE) OVA's OSType is set to 'Other 64-bit'.

Conditions:
When applying VMware guest OS customization on VMware BIG-IP VE.

Impact:
VMware guest OS customization fails (cannot create customization specification wizard).

Workaround:
You can use either of the following workarounds:
 - Apply VMware guest OS customization with 'ovftool'.
 - Manually set OSType to 'Other 3.x Linux 64-bit'.

Fix:
OS type embedded in .ovf file in VE OVAs has been changed from 'Other 64-bit' to 'Other 3.x Linux 64-bit' to enable VMware guest OS customization.

Behavior Change:
In this release, the OS type set in .ovf file in the BIG-IP VE SCSI OVA images for VMware has been changed from 'Other 64bit' to 'Other 3.x Linux 64bit'. This enables 'VMware Guest Customization' via VMware vCenter.


685020-3 : Enhancement to SessionDB provides timeout

Component: TMOS

Symptoms:
In some cases, calls made to SessionDB never return from the remote TMM.

Conditions:
-- Using add, update, delete, and lookup commands to remote TMM.
-- SessionDB request is not returned.

Impact:
Calls made to SessionDB never return from the remote TMM.

Workaround:
None.

Fix:
The system initiates a timeout after 2 seconds. If a timeout occurs, the calling command receives a result of err==ERR_TIMEOUT.

A new sys db variable was added to enable the timeout for the iRule table command. It defaults to false. Below is the new definition followed by the tmsh command to set the value.

# Enable or disable iRule table cmd timeout override to cause all
# requests to be 'remembered' so that we do not leak them
# if subsystems fail.
[Tmm.SessionDB.table_cmd_timeout_override]
default=false
type=enum
realm=common
enum=|true|false|

Behavior Change:
A new sys db variable was added to enable the timeout for the iRule table command. It defaults to false. Below is the new definition followed by the tmsh command to set the value.

# Enable or disable iRule table cmd timeout override to cause all
# requests to be 'remembered' so that we do not leak them
# if subsystems fail.
[Tmm.SessionDB.table_cmd_timeout_override]
default=false
type=enum
realm=common
enum=|true|false|


684937-3 : [KERBEROS SSO] Performance of LRU cache for Kerberos tickets drops gradually with the number of users

Solution Article: K26451305

Component: Access Policy Manager

Symptoms:
APM performance of handling HTTP request drops gradually when Kerberos SSO is being used over period of time.
Websso process CPU usage is very high during this time. The latency can vary between APM end users.

Conditions:
-- A large number of APM end users have logged on and are using Kerberos SSO.
-- Running APM.

Impact:
Increased latency of HTTP request processing.

Workaround:
Reduce the number of cached Kerberos user tickets by lowering the cache lifetime.

Fix:
LRU cache performance no longer drops linearly with the number of caches Kerberos tickets, the latency of HTTP request processing has been significantly improved.


684852-1 : Obfuscator not producing deterministic output

Component: Fraud Protection Services

Symptoms:
Proactive defense challenge is not passed.

Conditions:
The obfuscator does not produce the same output for the same pair of key and seed. Therefore, on multi-blade devices, or on active-active deployments, when the request to the page (url=/) and the request to the javascript (/TSPD/*?type=10) each go to a different blade or a different device.

More frequently, it happens when the page and javascript are loaded from the same blade, but the javascript is stored in the cache.

Then another refresh, and the request goes to the second blade. Because the javascript in the cache was received from the first blade, it does not match the page.

Impact:
Proactive defense challenge is not passed; challenge remains on blank page on chassis.

Workaround:
None.

Fix:
Obfuscator now uses common Random object.


684583-1 : Buitin Okta Scopes Request object uses client -id and client-secret

Component: Access Policy Manager

Symptoms:
Buitin Okta Scopes Request object uses client credentials instead of resource server credentials.

Conditions:
Buitin Okta Scopes Request object

Impact:
Scope request with Buitin Okta Scopes Request object fails.

Workaround:
Use modified Request object.

Fix:
Buitin Okta Scopes Request object is fixed to use resource server credentials.


684391-3 : Existing IPsec tunnels reload. tmipsecd creates a core file.

Component: TMOS

Symptoms:
Existing IPsec tunnels reload. tmipsecd creates a core file.

Conditions:
Although an exact replication of the issue has not been reliable, the one instance occurred after the following conditions:
-- Remote peer repeatedly tries to establish a tunnel.
-- The IPsec IKEv1 daemon 'racoon' is dead.

Impact:
Existing IPsec tunnels reload when tmipsecd aborts. tmipsecd creates a core file.

Workaround:
None.

Fix:
Exception handling in tmipsecd has been improved so that tmipsecd will not reload when encountering some unusual conditions.


684370-1 : APM now supports VMware Workspace ONE integration with VIDM as ID Provider

Component: Access Policy Manager

Symptoms:
When VMware Horizon resources are behind APM, you can see available desktops and application on VMware Workspace One (WS1) portal, but you cannot launch them.

Conditions:
-- VMware virtual desktops and applications are hosted on VMware Horizon behind APM.
-- Authenticate with VMware Identity Manager (VIDM) and see available virtual desktops and applications on WS1 portal.
-- Attempt to launch a virtual desktop or application with VMware HTML5 client.

Impact:
BIG-IP users get authenticated with VIDM and can see available desktops and applications on the WS1 portal, but cannot launch a desktop or application with View HTML5 client.

Workaround:
Not applicable.

Fix:
APM now supports VMware Workspace One (WS1) with VMware Identity Manager (VIDM) as the Identity Provider and APM as a service provider, protecting VMware Horizon desktops and applications.


684333-1 : PEM session created by Gx may get deleted across HA multiple switchover with CLI command

Component: Policy Enforcement Manager

Symptoms:
PEM sessions may get cleaned up with terminate cause of FATAL GRACE TIMEOUT, if multiple high availability (HA) failover is being performed using the following command: tmsh run sys failover standby.

Conditions:
Multiple HA failover performed using the following command: tmsh run sys failover standby.

Impact:
PEM session created using Gx may get deleted.

Workaround:
Initiate failover using alternate commands, such as the following:
 tmm big start restart.


684325-1 : APMD Memory leak when applying a specific access profile

Component: Access Policy Manager

Symptoms:
Access profile having CheckMachineCert agent, while updating profile using 'Apply access policy', each time it leaks 12096 bytes of memory.

Conditions:
-- Access profile configured with agent 'CheckMachineCert'.
-- Repeatedly update the profile using 'Apply access policy'.

Impact:
APMD process stops after repeated application of the script.

Workaround:
None.

Fix:
APMD no longer leaks memory when applying Access profile configured with agent 'CheckMachineCert'.


684312-1 : During Apply Policy action, bd agent crashes, causing the machine to go Offline

Solution Article: K54140729

Component: Application Security Manager

Symptoms:
During Apply Policy action, bd agent crashes, causing with this error:
--------------------
crit perl[21745]: 01310027:2: ASM subsystem error (bd_agent,): bd_agent exiting, error:[Bit::Vector::new_Dec(): input string syntax error at /usr/local/share/perl5/F5/CfgConvert.pm line 66, <$inf> line 1. ]
--------------------

Causing bd and bd_agent processes restart, and causing the machine to go Offline.

Conditions:
-- ASM provisioned.
-- Applying policy.
-- Corrupted data was attempted to be loaded during an Apply Policy action.

Impact:
bd and bd_agent processes restart, causing the machine to go Offline while the processes restart..

Workaround:
None.

Fix:
During Apply Policy action, bd agent no longer crashes when attempting to load corrupted data.


684218-1 : vADC 'live-install' Downgrade from v13.1.0 is not possible

Component: TMOS

Symptoms:
vADC v13.1.0 has a new storage format that is incompatible with earlier versions. It should be possible to use the '--format' option of 'image2disk' to downgrade a v13.1.0 vADC system, but it is not.

Conditions:
Running v13.1.0, attempt software downgrade to 11.5.4, for example:

image2disk --format=volumes --nosaveconfig 11.5.4

Impact:
request is not allowed. no changes are made.

Workaround:
deploy a new 11.5.4 software image via the hypervisor environment


684033-3 : CVE-2017-9798 : Apache Vulnerability (OptionsBleed)

Solution Article: K70084351


683741-1 : APM now supports VMware Workspace ONE integration with vIDM as ID Provider

Component: Access Policy Manager

Symptoms:
When VMware Horizon resources are behind APM, APM end user is able to see available desktops and application on VMware Workspace ONE portal but is not able to launch them.

Conditions:
-- VMware virtual desktops and applications are hosted on VMware Horizon behind APM.
-- APM end user authenticates with VMware Identity Manager (IDM) and sees available virtual desktops and applications on Workspace ONE portal.
-- APM end user attempts to launch a virtual desktop or application with VMware native client.

Impact:
Users authenticates but is not able to launch a desktop or application with View native client.

Workaround:
None.

Fix:
APM now supports VMware Workspace ONE with VMware IDM as Identity Provider and APM as service provider, protecting VMware Horizon desktops and applications.


683508-1 : WebSockets: umu memory leak of binary frames when remote logger is configured

Solution Article: K00152663

Component: Application Security Manager

Symptoms:
ASM out of memory error messages in /var/log/asm.

Conditions:
-- Virtual server configured with WebSocket profile.
-- ASM remote logger configured and assigned to the virtual server.

Impact:
ASM out of memory, memory leak.

Workaround:
Remove ASM remote logging profile from a virtual server.

Fix:
This release correctly releases unused memory after WebSocket message is sent to the logging destination.


683474 : The case-sensitive problem during comparison of 2 Virtual Servers

Component: Application Visibility and Reporting

Symptoms:
Failed to load "incident types volume graph" if incident was filtered by Virtual Server

Impact:
Chart of incident data will not be displayed.

Workaround:
Avoid to create virtual servers that have the same letters, differing only by capital letters verses small letters.

Fix:
monpd process uses a case-sensitive comparison of virtual servers


683389-3 : Error #2134 when attempting to create local flash.net::SharedObject in rewritten ActionScript 3 file

Component: Access Policy Manager

Symptoms:
Flash ActionScript3 application shows Error #2134 when trying to call flash.net::SharedObject.getLocal with localPath specified.

Conditions:
Attempt to create local SharedObject.

Impact:
Affected Flash applications are not working when accessed through Portal Access.

Workaround:
None.

Fix:
Addressed an issue in Portal Access which caused rewritten Flash files to show Error #2134 on attempt to create local SharedObject.


683297-2 : Portal Access may use incorrect back-end for resources referenced by CSS

Component: Access Policy Manager

Symptoms:
If HTML page contains reference to external CSS file with URL beginning with '//', then host-less references in this CSS file are handled incorrectly by Portal Access.

Conditions:
- HTML page at http://example.host/page.html:

    <link rel=stylesheet href=//another.host/some/path/my.css>

- and this CSS contains reference with absolute path like this:

    html { background-image: url(/misc/image/some.png); }

Portal Access uses 'http://example.host' as back-end for this image instead of correct 'http://another.host'.

Impact:
Web application may not work correctly.

Workaround:
Use iRule to correct back-end host.

Fix:
Portal Access uses correct back-end host for references in CSS files included with scheme-less URL.


683241-1 : Improve CSRF token handling

Solution Article: K70517410

Component: Application Security Manager

Symptoms:
Under certain conditions, CSRF token handling does not follow current best practices.

Conditions:
CSRF is configured.

Impact:
CSRF token handling does not follow current best practices.

Workaround:
None.

Fix:
CSRF token handling now follows current best practices.


683131-1 : Hotfix install fails on vCMP guest when both guest and host have the same base version ISO present

Component: TMOS

Symptoms:
BIG-IP software installations will fail and report a status of:

    waiting for cleanup; multiple base builds found (BIG-IP 13.0.0)

Conditions:
- Perform software installation in a vCMP guest when guest is running v13.0.0 or newer.
- Ask the guest to install a hotfix (e.g. "tmsh install /sys software hotfix 13.0.0-hf2.iso volume HD1.2 create-volume") to a boot location that does not have a base software version installed.
- Hypervisor has a valid copy of the the correct base image (e.g. v13.0.0 build 0.0.1645)
- Guest has a valid copy of the correct base image (e.g. v13.0.0 build 0.0.1645)

Impact:
Software installation fails, and will not complete/continue.

Workaround:
Delete the base software image from either the hypervisor or guest's file system

Fix:
The condition no longer causes an error; the installation request successfully runs to completion.


683114-2 : Need support for 4th element version in Update Check

Component: TMOS

Symptoms:
Previously, there was no 4th element version Update Check functionality.

Conditions:
Using Update Check.

Impact:
No 4th element version support provided.

Workaround:
None.

Fix:
There is now 4th element version support in Update Check.


683113-3 : [KERBEROS SSO][KRB5] The performance of memory type Kerberos ticket cache in krb5 library drops gradually with the number of users

Solution Article: K22904904

Component: Access Policy Manager

Symptoms:
APM performance of handling HTTP request drops gradually when Kerberos SSO is being used over a period of time.

Websso CPU usage is very high.

The BIG-IP system response can rate drop to the point that the clients disconnect after waiting for a response. The system logs error messages similar to the following: Failure occurred when processing the work item.

Conditions:
-- Running APM.
-- A large number of APM end users (~20 KB) have logged on and are using Kerberos SSO.

Impact:
Increased latency of HTTP request processing.

Workaround:
Reduce the number of cached Kerberos user tickets by lowering the cache lifetime.

Fix:
Improvements to the krb5 library have been implemented for better scalability, so the latency of HTTP request processing has been significantly improved.


683029-1 : Sync of virtual address and self IP traffic groups only happens in one direction

Component: TMOS

Symptoms:
If you have a virtual address and a self IP that both listen on the same IP address, changing the traffic group of the self IP will make an equivalent change to the traffic group of the virtual address. However, this does not work in reverse. Changing the traffic group of the virtual address will not cause a change of the traffic group of the self IP.

Conditions:
You have a virtual address and a self IP that both listen on the same IP address. (The subnet mask need not be the same.)

Impact:
This is by design, but is counterintuitive, and there is no warning message that this is the case. Setting the traffic group on both objects will always work properly.

Workaround:
Care should be taken to ensure that the desired traffic group is set on both objects.


682944-1 : key-id missing for installed netHSM key for standby BIG-IP system in HA setup

Component: Local Traffic Manager

Symptoms:
In a BIG-IP high availability (HA) configuration, the nethsm key installed has empty key-id string for the standby BIG-IP system. That is, the BIG-IP system that actually gets the key installed has the key-id string properly displayed. But its peer BIG-IP system does not display a key-id string associated with the installed key.

Conditions:
-- nethsm key installed.
-- Standby BIG-IP system in an HA configuration.

Impact:
The peer BIG-IP system has no key-id string properly displayed.

Workaround:
Even though key-id does not display, the key is present on the peer BIG-IP system and can be used there.

Fix:
The netHSM key for standby BIG-IP system in HA configurations now shows up after a successful configsync.


682500-2 : VDI Profile and Storefront Portal Access resource do not work together

Component: Access Policy Manager

Symptoms:
Accessing a Citrix Storefront portal access resource and clicking on the application does not work since VDI returns HTTP status 404.

Conditions:
-- VDI profile is attached to the Virtual server.
-- Access policy has Citrix Storefront portal access resource.
-- Citrix remote-desktop resource is attached.

Impact:
Citrix Storefront portal access resource cannot be used to launch applications.

Workaround:
None.

Fix:
Citrix Storefront portal access resources can now be used with Citrix Remote desktop resources.


682335-1 : TMM can establish multiple connections to the same gtmd

Component: Global Traffic Manager (DNS)

Symptoms:
TMM can establish multiple connections to the same gtmd, and tmm may core.

Conditions:
This timing-related issue involves coordination between license blob arrival, gtmd connection teardown/establishment, and gtmd restart.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
Fixed, if there is an existing connflow, don't start another connection.


682283-2 : Malformed HTTP/2 request with invalid Content-Length value is served against RFC

Component: Local Traffic Manager

Symptoms:
HTTP/2 request can include Content-Length header. When the value of a Content-Length header does not match the sum of lengths of all DATA frames from the stream, RFC requires that the system reset the stream.

Conditions:
-- A virtual server is configured with HTTP/2 profile.
-- The value of Content-Length header does not match the sum of lengths of all DATA frames from the stream.

Impact:
The BIG-IP system sends a request to a server and serves a provided response, which is not in conformance with the RFC.

Workaround:
None.

Fix:
Now, when a client sends a request over an HTTP/2 connection with a malformed HEADERS frame in which Content-Length does not match the payload size in DATA frames, the BIG-IP system correctly resets the stream with RST_STREAM frame.


682213-1 : TLS v1.2 support in IP reputation daemon

Solution Article: K31623549

Component: TMOS

Symptoms:
The IP reputation daemon opens SSL connections to the Webroot BrightCloud server using TLS 1.0 protocol.

Conditions:
This occurs when using IP reputation.

Impact:
Because IP reputation services are used to accept/deny connections to critical business applications, there might be concerns about the service. Also some configurations might require that all transactions exfiltrating a PCI-controlled environment leverage secure protocols and ciphers, which won't be the case for IP reputation services.

Workaround:
None.

Fix:
Webroot updated BrightCloud servers to support TLS 1.2. This is additional support. To preserve backward compatiblity, the servers support TLS 1.0, TLS 1.1, TLS 1.2, SSL 2.0 and SSL 3.0.

In addition, this software version supports TLS 1.2 on the client side by customizing the SDK used by the IP reputation daemon.


682104-3 : HTTP PSM leaks memory when looking up evasion descriptions

Component: Local Traffic Manager

Symptoms:
http_psm_description_lookup leaks xfrags containing PSM evasion descriptions.

Conditions:
When PSM looks up evasion descriptions.

Impact:
Memory leaked each time might eventually cause out of memory to the TMM.

Workaround:
None.

Fix:
This fix will stop the memory leakage.


681955-1 : Apache CVE-2017-9788

Solution Article: K23565223


681782-6 : Unicast IP address can be configured in a failover multicast configuration

Component: TMOS

Symptoms:
Failover multicast configuration does not work when configured with a unicast IP address. Although this is an invalid configuration, the system does not prevent it.

Conditions:
Specify a unicast IP address under 'Failover Multicast Configuration' for network failover in high availability configurations.

Impact:
Failover multicast configuration does not work.

Workaround:
Specify a multicast IP address under 'Failover Multicast Configuration' for network failover.

Fix:
The system now prevents specifying a unicast IP address when configuring multicast failover.


681757-3 : Upgraded volume may fail to load if a Local Traffic Policy uses the forward parameter 'member'

Solution Article: K32521651

Component: Local Traffic Manager

Symptoms:
In response to this issue, you might see the following symptoms:
-- System remains inoperative
-- Screen alerts similar to the following are posted: 'Configuration has not yet loaded'.
-- Configuration fails to load after upgrade to v12.1.0 or higher.

The system records an error message similar to the following in the ltm log file:

 emerg load_config_files: "/usr/bin/tmsh -n -g load sys config partitions all " - failed. -- 010716de:3: Policy '/Common/Policy', rule 'Policy-Rule'; target 'forward' action 'select' does not support parameter of type 'member'. Unexpected Error: Loading configuration process failed.

Conditions:
Local Traffic Policy with a forward action that selects a target of type 'member'.

Impact:
Configuration fails to load on upgrade.

Workaround:
Modify the local traffic policies to use a target type of 'node' before upgrading, or, after upgrading, edit the config file and modify 'member' to 'node' in the local traffic policy, and then reload the configuration.

Fix:
Upon upgrade to v12.1.0 or later, policies that perform the action 'forward - select - member' will be automatically changed to 'forward - select - node', and configuration will load successfully.


681673-4 : tmsh modify FDB command permits multicast MAC addresses, which produces unexpected results

Component: Local Traffic Manager

Symptoms:
TMSH does not block modify FDB commands that add a multicast MAC addresses.

Conditions:
This occurs when the following is configured using tmsh commands when the mac-address is multicast:
 fdb vlan <vlan> records add {<mac-address> {interface <slot>/<port>}}.

Impact:
There is not enough information to map the outgoing MAC address to a multi-cast group, and therefore it gets a default entry added that has no ports mapped. The result is that the frame will not go out the interface indicated in the tmsh command yet no warning is provided.

Workaround:
None.

Fix:
TMSH modify FDB command is no longer permitted to add multicast MAC addresses, so this issue no longer occurs.


681415-3 : Copying of profile with advanced customization or images might fail

Component: Access Policy Manager

Symptoms:
Copying a profile with advanced customization or images produces an error message similar to the following: 'Please specify language code' or similar

Conditions:
Access Profile has advanced customization group or customization image assigned to any of object connected to profile.

Impact:
Unable to copy policy.

Workaround:
None.

Fix:
Copying of profile with advanced customization or images now succeeds as expected.


681385-2 : Forward proxy forged cert lifespan can be configured from days into hours.

Component: Local Traffic Manager

Symptoms:
Once support for OCSP in place, you may need to forge certificates in lifespan shorter than one day. Previously, there was no way to configure that.

Conditions:
Configure forward proxy forged cert lifespan shorter than a day.

Impact:
None. This is a request for enhancement.

Workaround:
None.

Fix:
A new DB variable (tmm.ssl.certlifespaninhours) is added to support specifying hours instead of days:

[root@localhost:Active:Standalone] config # tmsh list sys db tmm.ssl.certlifespaninhours
sys db tmm.ssl.certlifespaninhours {
    value "disable"
}
[root@localhost:Active:Standalone] config # tmsh modify sys db tmm.ssl.certlifespaninhours value enable
[root@localhost:Active:Standalone] config # tmsh list sys db tmm.ssl.certlifespaninhours
sys db tmm.ssl.certlifespaninhours {
    value "enable"
}

When this variable is enabled, the configured lifespan is treated as hours. When this variable is disabled, the configured lifespan is treated as days.

Behavior Change:
Configured Forward proxy forged cert lifespan allows changing
from days to hours using a new DB variable: tmm.ssl.certlifespaninhours.


681175-3 : TMM may crash during routing updates

Solution Article: K32153360

Component: Local Traffic Manager

Symptoms:
When dynamic routing is configured and ECMP routes are received, certain routing updates may lead to a TMM crash.

Conditions:
-- Dynamic routing.
-- ECMP routes.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
Disable ECMP routes by configuring "max-paths 1" in ZebOS.

Fix:
TMM no longer crashes on routing updates when ECMP is in use.


680850-2 : Setting zxfrd log level to debug can cause AXFR and/or IXFR failures due to high CPU and disk usage.

Solution Article: K48342409

Component: Global Traffic Manager (DNS)

Symptoms:
Enabling debug logging on zxfrd (DNSX) can result in excessive CPU and disk usage, as well as errors during DNS AXFR or IXFR processing.

Conditions:
log.zxfrd.level is set to debug by running the following command: tmsh modify sys db log.zxfrd.level value debug

Impact:
IXFRs or AXFRs may fail and be rescheduled due to high CPU usage by zxfrd, which causes it to fail to process data packets during a transfer.

Workaround:
To avoid this issue, do not set log.zxfrd.level to debug.

Fix:
With this change, the dump to /var/tmp/zxfrd.out occurs only when a new db variable, dnsexpress.dumpastext, is set to true. This enables turning on logging for debug without consuming all the CPU and disk necessary to dump packets and zone contents.

Note: Setting this value to true will cause the information to be dumped to stderr, which reveals the original issue, potentially causing transfer failures and high system resource usage. F5 recommends that you enable it only when directed to do so by F5 support staff.

Behavior Change:
Previously, setting the zxfrd log level to debug caused all AXFR and IXFR requests and responses to be logged to /var/tmp/zxfrd.out. Doing so also caused the contents of all zones to be dumped, as text, to /var/tmp/zxfrd.out when the database was saved.

This was extremely CPU-, memory-, and disk-intensive. The CPU load could cause zxfrd to fail to process transfer data packets in a timely fashion, which could cause the master DNS server to close the connection.

With this fix, setting log.zxfrd.level debug no longer outputs this information.

Although it is not generally useful to output the contents of the transfer packets or the contents of the database, if this information is required for troubleshooting or information-verification purposes, you can set the new db variable: dnsexpress.dumpastext.

Note: Setting this value to true will cause the information to be dumped to stderr, which reveals the original issue, potentially causing transfer failures and high system resource usage. F5 recommends that you enable it only when directed to do so by F5 support staff.


680729-1 : DHCP Trace log incorrectly marked as an Error log.

Solution Article: K64307999

Component: Policy Enforcement Manager

Symptoms:
The following sample DHCP debug log may be found repeatedly in the TMM logs.

<#> <date> <slot#> notice DHCP:dhcpv4_xh_timer_callback/1053: Entering: <mac-addr>

Conditions:
Send a DHCP request through a DHCP virtual and wait for 30 seconds for the DHCP callback to trigger.

Impact:
Possible clutter in the TMM logs.

Workaround:
Set the db variable to critical. To do so, run the following command: setdb tmm.dhcp.log.level critical

Fix:
The following log can be seen only when DHCP debug logs are set to enabled.
<#> <date> <slot#> notice DHCP:dhcpv4_xh_timer_callback/1053: Entering: <mac-addr>


680388-1 : f5optics should not show function name in non-debug log messages

Component: TMOS

Symptoms:
For logging thresholds other than debug, the function name appears in log messages created by f5optics.

Conditions:
-- BIG-IP is running.
-- Logging thresholds is set to a value other than debug.

Impact:
Log files contain unexpected data.

Workaround:
There is no workaround at this time.

Fix:
With the fix, f5optics is not displaying function names in non-debug logging messages.


680353-1 : Brute force sourced based mitigation is not working as expected

Component: Application Security Manager

Symptoms:
Brute force mitigations are not working by the configured order under some conditions - for example a captcha is arriving instead of a drop.

Conditions:
-- Brute force is configured.
-- There is more than one source (for example, User and IP address).

Impact:
The incorrect mitigation is received.

Workaround:
None.

Fix:
Fixed an issue with brute force mitigations.


680264-2 : HTTP2 headers frame decoding may fail when the frame delivered in multiple xfrags

Solution Article: K18653445

Component: Local Traffic Manager

Symptoms:
Intermittently, HTTP2 experiences protocol resets.

Conditions:
-- xfrag is 2 bytes in length.
-- The header length is greater than 128.
-- xfrag starts.

For example, the following returns the incorrect header length:
 (0xFF BYTE1) next byte, http2_arbint_read.

Impact:
Unexpected loss of HTTP2 frames due to protocol resets.

Workaround:
No effective workaround.

Fix:
HTTP2 now parses the request, regardless of its xfrags distribution.


680086 : md5sum check on BMC firmware fails

Component: TMOS

Symptoms:
Checking the md5 sum of BMC firmware fails when issuing the command 'md5sum -c /usr/firmware/shuttle_x.x.xx.ima_enc.md5'. The command fails with the following message: (...) listed file could not be read".

Conditions:
iSeries appliances:
- i2000
- i4000
- i5000
- i7000
- i10000
- i15000

Impact:
'md5sum -c' does not work for BMC firmware checksums.

Workaround:
Indirectly check the md5sum by calculating it with 'md5sum /usr/firmware/shuttle*.ima_enc' and comparing it to 'cat /usr/firmware/shuttle*.ima_enc.md5'. Or use this command:

diff -sy <(md5sum < /usr/firmware/shuttle*.ima_enc | awk '{ print $1 }') <(cat /usr/firmware/shuttle*.ima_enc.md5 | awk '{ print $1 }')

Fix:
Fixed 'md5sum -c' not working for BMC firmware checksums.


680074-2 : TMM crashes when serverssl cannot provide certificate to backend server.

Component: Local Traffic Manager

Symptoms:
TMM halts and restarts when server SSL cannot provide a certificate to the backend server.

Conditions:
-- The backend server is configured to require a client certificate to complete the SSL handshake.
-- The server SSL profile is not configured with a client certificate.

Impact:
TMM halts and restarts. Traffic disrupted while tmm restarts.

Workaround:
No workaround at this time.

Fix:
TMM no longer halts and restarts when server SSL cannot provide a certificate to the backend server.


680069-1 : zxfrd core during transfer while network failure and DNS server removed from DNS zone config

Solution Article: K81834254

Component: Global Traffic Manager (DNS)

Symptoms:
zxfrd cores and restarts.

Conditions:
While zone transfer is in progress, the network fails and the DNS server is removed from the DNS zone configuration.

Impact:
zxfrd cores.

Workaround:
None.

Fix:
zxfrd no longer cores during transfer while network failure and DNS server removed from DNS zone config.


679861 : Weak Access Restrictions on the AVR Reporting Interface

Component: Application Visibility and Reporting

Symptoms:
The AVR reporting interface does not follow best practices for access restrictions.

Conditions:
AVR provisioned

Impact:
If accessed the AVR reporting interface may disclose:
 - Client and server IP addresses
 - URIs from client requests
 - Metadata about attacks detected by BIG-IP

Workaround:
Ensure that network access to the management port is restricted and that Port Lockdown setting for Self-IPs is not set to "Allow All". The default port lockdown of "Allow Default" provides mitigation against access via Self-IP.

Fix:
Stronger access restrictions enforced on the AVR reporting interface.


679613-1 : i2000/i4000 Platforms Improperly Handle VLANs Created with a Value of '1'

Solution Article: K23531420

Component: Local Traffic Manager

Symptoms:
When an interface is associated with a VLAN whose tag value is '1', traffic is incorrectly sent out as untagged.

Conditions:
1. Create a VLAN with a tag value of '1'.
2. Associate an interface with the VLAN whose value is '1'.
3. Send traffic out that interface.

Impact:
Incorrect routing/switching of traffic.

Workaround:
Use VLANs with a tag value different from '1'.


679496-2 : Add 'comp_req' to the output of 'tmctl compress'

Component: Local Traffic Manager

Symptoms:
The output of 'tmctl compress' displays the total numbers of requests (tot_req), but does not distinguish between deflate (compression) requests and inflate (decompression) requests.

Conditions:
Viewing the output of the 'tmctl compress' command.

Impact:
Cannot determine the different types of requests.

Workaround:
There is no workaround at this time.

Fix:
This release now distinguishes between deflate (compression) requests and inflate (decompression) requests, as follows: there is an indicator, 'comp_req', for compression requests. The number of decompression request is tot_req - comp_req.


679494-1 : Change the default compression strategy to speed

Component: Local Traffic Manager

Symptoms:
The current default compression.strategy is 'latency', which does not perform properly, i.e., the provider selection algorithm does not react to load change fast enough.

Conditions:
Using compression.strategy to distribute workload among hardware and software compression providers.

Impact:
The work load may not be distributed evenly among hardware and software compression providers when compression.strategy is 'latency'.

Workaround:
Modify the tmsh sys db variable compression.strategy to 'speed'.

Fix:
The default compression strategy is now set to 'speed'.


679384-3 : The policy builder is not getting updates about the newly added signatures.

Solution Article: K85153939

Component: Application Security Manager

Symptoms:
The policy builder is not getting updates about the newly added signatures.

Conditions:
When ASU is installed or user-defined signatures are added/updated.

Impact:
No learning suggestions for some of the newly added signatures.

Workaround:
Use either of the following workarounds:
-- One workaround is restarting the policy builder. This will revert the learning progress made in the last 24 hours:
 killall -s SIGHUP pabnagd

-- Manually change some Policy Attack Signature Set in Learning and Blocking Settings (e.g., disabling and re-enabling Learn checkbox).

Fix:
After the fix, Policy Builder will be aware of all newly added signatures.


679347-2 : ECP does not work for PFS in IKEv2 child SAs

Solution Article: K44117473

Component: TMOS

Symptoms:
The original racoon2 code has no support for DH generate or compute using elliptic curve algorithms for (perfect forward security).

Additionally, the original interfaces are synchronous, but the only ECP support present uses API with async organization and callbacks, so adding ECP does not work.

Conditions:
Changing an ike-peer definition from the default phase1-perfect-forward-secrecy value of modp1024 to something using ECP: ecp256, ecp384, or ecp512.

Note: The first child SA is negotiated successfully.

Impact:
Once the first child SA expires (or is deleted), the IKEv2 tunnel goes down when another SA cannot be negotiated.

Workaround:
Use MODP for perfect-forward-secrecy instead of ECP.

Fix:
Full support for ECP as PFS has now been added, so a new child-SA negotiated in a IKEV2EXCH_CREATE_CHILD_SA exchange works as expected for ecp256, ecp384, and ecp512.


679221-2 : APMD may generate core file or appears locked up after APM configuration changed

Component: Access Policy Manager

Symptoms:
Right after clicking the 'Apply Access Policy' link, APMD may generate a core file and restart; or appears to be locked up.

Conditions:
-- Changing APM configuration, especially Per-Session and Per-Request Policy.
-- Configured for AD or LDAP Auth Agent.

Impact:
If APMD restarts, then AAA service will be interrupted for a brief period of time. If APMD appears to be locked up, then AAA service will be stopped until manually restarted.

Workaround:
None.

Fix:
APMD now processes the configuration changes correctly during 'modify apm profile access <profile name> generation-action increment' (TMSH) or 'Apply Access Policy' (GUI), and no service interruption occurs.


679149-1 : TMM may crash or LB::server returns unexpected result due to reused lb_result->pmbr[0]

Component: Global Traffic Manager (DNS)

Symptoms:
TMM may crash or LB::server returns unexpected result.

Conditions:
GTM rule command LB::server is executed before a load balance decision is made or the decision is not to a real pool member.

Impact:
GTM rule command LB::server returns unexpected result. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
GTM rule command LB::server is now executed at the correct time, so TMM does not crash and LB::server returns expected results.


679135-2 : IKEv1 and IKEv2 cannot share common local address in tunnels

Component: TMOS

Symptoms:
When IKEv1 and IKEv2 IPsec tunnels are configured to use the same local IP address, either all the IKEv1 or all the IKEv2 tunnels will not establish.

Note: This is as designed: the system does not support using the same local self IP to establish both IKEv1 and IKEv2 tunnels. However, the system does not prevent it, and there is no indication of the reason for the failure.

Conditions:
-- Use the same self IP as the local address of an IPsec tunnel for IKEv1, as well as the local address of a tunnel for IKEv2.
-- Try to create competing listeners.

Impact:
Either the IKEv1 or IKEv2 tunnel will not work, because the listener for that tunnel fails to establish. Usually the IKEv1 tunnel will not work after tmm restart or BIG-IP reboot.

Workaround:
Use another self IP for the tunnel local address to keep IKEv1 and IKEv2 local tunnel addresses separate.

Note: If IKEv1 tunnels use one local address, while IKEv2 tunnels use another, everything works as expected.

Fix:
Logging in /var/log/ltm now reveals failure to establish listener, along with a suggestion to avoid sharing one local address across IKEv1 and IKEv2 tunnels.


679114-4 : Persistence record expires early if an error is returned for a BYE command

Component: Service Provider

Symptoms:
When an error is returned for a SIP command, the persistence timeout is set to the transaction timeout.

Conditions:
An error is returned for a any SIP command.

Impact:
The persistence record will expire early when the call has not been ended.

Workaround:
None.

Fix:
For BYE commands, the timeout is not set to transaction timeout on failure.


679088-1 : Avr reporting and analytics does not display statistics of many source regions

Component: Application Visibility and Reporting

Symptoms:
1. The network reporting does not show the statistics related to some Source Regions.
2. In the Security=>Reporting=>Network=>Enforced Rules dashboard are impossible to select or find some Source Region using filtering .
For example, there are list of some missing Source Regions:
France, Ile-de-France, Ukraine, Kyyiv,Russian Federation, Tambovskaya oblast, South Africa, Western Cape and Spain,Madrid

Conditions:
This occurs when attempting to filter on the affected source regions.

Impact:
The network reporting does not show the statistics related to some Source Regions.


678872-3 : Inconsistent behavior for virtual-address and selfip on the same ip-address

Component: Local Traffic Manager

Symptoms:
Inconsistent ICMP/ARP behavior for self IP address or virtual-address when virtual-address and self IP address have the same IP address.

Conditions:
Virtual-address and self IP address have the same IP address.
-- Virtual-address ICMP and ARP disabled.

Impact:
ICMP echo reply and ARP for the IP address might be inconsistent. Self IP address might override the ARP setting of a virtual address.

Workaround:
No workaround.

Fix:
This implements the initialization-order-independent set of rules of whether particular IP address should have ARP/ICMP enabled for multiple maching vaddrs. The lookup is performed from the most fine netmask to the most coarse netmask. If for particular netmask there is no maching vaddr then more coarse netmask is lookedup. Otherwise if any machnig vaddr for particular netmask have ARP/ICMP enabled then IP address will have ARP/ICMP enabled. If none of matching vaddrs for particular netmask have ARP/ICMP enabled the then IP address will have ARP/ICMP disabled.

The rule above have one exception, due to the performance optimizations. If the vaddr have both ARP and ICMP disabled then the vaddr is considered deleted.


678861-1 : DNS:: namespace commands in procs cause upgrade failure when change from Link Controller license to other

Component: Global Traffic Manager (DNS)

Symptoms:
Upgrade fails with a message similar to the following.

emerg load_config_files: "/usr/bin/tmsh -n -g load sys config partitions all " - failed. -- 01070356:3: Link Controller feature not licensed. Unexpected Error: Loading configuration process failed.

Conditions:
Previously had Link Controller with DNS:: commands in an iRule proc.

Impact:
Upgrade fails.

Workaround:
Remove DNS:: commands from procs before upgrade.

Or use AFM instead of iRules.


678851-3 : Portal Access produces incorrect Java bytecode when rewriting java.applet.AppletStub.getDocumentBase()

Component: Access Policy Manager

Symptoms:
Java applets containing call of getDocumentBase() through a reference to java.applet.AppletStub are incorrectly rewritten.

Attempt to call incorrectly patched method causes following exception:
java.lang.VerifyError: (...) Illegal type in constant pool

Conditions:
This occurs when using rewrite on Java applets that call getDocumentBase().

Impact:
Affected Java applets cannot be started through Portal Access.

Workaround:
None.

Fix:
Rewritten applets with calls of java.applet.AppletStub interface methods are no longer causing java.lang.VerifyError exception during execution.


678820-1 : Potential memory leak if PEM Diameter sessions are not created successfully.

Component: Policy Enforcement Manager

Symptoms:
Memory leak resulting in reduction in available memory.

Conditions:
1. PEM configured with Gx.
2. PCRF Gx end point operationally DOWN
3. Subscriber creation attempt.

Impact:
Loss of service

Workaround:
There is no workaround at this time.

Fix:
Diameter context is freed in case of a failed Diameter session creation.


678801-4 : WS::enabled returned empty string

Component: Local Traffic Manager

Symptoms:
WS::enabled command returned empty string instead of 0 or 1 for status.

Conditions:
-- WS::enabled command is used to query the status of WebSocket processing.
-- WebSocket and HTTP profiles are configured on the virtual server.

Impact:
Unable to determine the status of WebSocket processing using iRule commands.

Workaround:
There is no workaround at this time.

Fix:
Invoke appropriate method via WebSocket Tcl code.


678524-1 : Join FF02::2 multicast group when router-advertisement is configured

Component: Local Traffic Manager

Symptoms:
MLD snooping switches may not deliver router solicitation packets to BIG-IP, which breaks BIG-IP's router advertisement functionality. MLD snooping switches may not deliver the packets because BIG-IP has not joined the FF02::2 multicast group.

Conditions:
router-advertisement configured, MLD snooping switches.

Impact:
IPv6 hosts never receive router advertisements from BIG-IP in response to their router solicitations.

Workaround:
Disable MLD snooping on switches.

Fix:
BIG-IP now joins the FF02::2 multicast group when router-advertisements are configured.

Behavior Change:
BIG-IP now joins the FF02::2 multicast group when router-advertisement is configured.


678488-1 : BGP default-originate not announced to peers if several are peering over different VLANs

Solution Article: K59332320

Component: TMOS

Symptoms:
BGP default-originate is not announced to peers if several are configured as peers over different VLANs.

Conditions:
-- default-originate is configured on four similar neighbors.
-- The neighbors are reachable over different interfaces/subnets.

Impact:
Only some of the peered neighbors get the default route.

Workaround:
Add the following to the the BGP configuration:
 network 0.0.0.0/0

Fix:
All peered neighbors now get the default route.


678427-1 : Safari 11 displays F5 EPI and F5 VPN launch confirmation dialogs twice

Solution Article: K03138339

Component: Access Policy Manager

Symptoms:
Safari 11 displays confirmation dialogs to launch F5 EPI or F5 VPN app twice. Although functionality is not affected, the user experience might be confusing.

Conditions:
-- Safari 11, F5 EPI, or F5 VPN app installed.
-- Endpoint check or VPN configured in access policy.

Impact:
None. The extra dialog box does not affect system functionality.

Workaround:
None.

Fix:
Confirmation dialog is now displayed only once during VPN establishment with Safari browser.


678388-1 : IKEv1 racoon daemon is not restarted when killed multiple times

Solution Article: K00050055

Component: TMOS

Symptoms:
IPsec IKEv1 tunnels will fail and stay down indefinitely if the IKEv1 racoon daemon crashes. racoon does not get restarted by tmipsecd. This can occur if racoon has crashed more than once beforehand.

Conditions:
The IPsec IKEv1 racoon daemon crashes, or is killed manually, multiple times.

Impact:
IPsec IKEv1 tunnels cannot be established because the racoon daemon is dead. The user will receive no CLI or web UI clues to indicate that racoon is dead. Attempts to reconfigure IPsec while racoon is dead will not resolve the problem.

Workaround:
Run the following command to restart the IPsec IKEv1 racoon and tmipsecd daemons at the same time:
tmsh restart sys service tmipsecd

Fix:
Fixed tmipsecd so it correctly tracks whether the IKEv1 racoon daemon is still running or needs a restart. This also covers odd timing, such as killing racoon right after it starts.


678293-2 : Uncleaned policy history files cause /var disk exhaustion

Solution Article: K25066531

Component: Application Security Manager

Symptoms:
There are hundreds of policy history files for non-existent policies stored under /ts/dms/policy/policy_versions, which might cause /var disk exhaustion.

Conditions:
There are hundreds of policy history files for non-existent policies stored under /ts/dms/policy/policy_versions.

Two possible causes might explain what caused the history files to be copied:
-- The device was synchronized from itself.
-- There was a UCS loaded on the device.

Impact:
/var disk usage is high.

Workaround:
Use the following one-liner to find unreferenced policy history files that can be deleted:

----------------------------------------------------------------------
perl -MData::Dumper -MF5::DbUtils -MFile::Find -e '$dbh = F5::DbUtils::get_dbh(); $sql = ($dbh->selectrow_array(q{show tables in PLC like ?}, undef, q{PL_POLICY_VERSIONS})) ? q{select policy_version, policy_id from PLC.PL_POLICY_VERSIONS} : q{select revision, policy_id from PLC.PL_POLICY_HISTORY}; %known_history_files = map { (qq{$_->[1]/$_->[0].plc} => 1) } @{$dbh->selectall_arrayref($sql)}; find({ wanted => sub { next unless -f $_; $_ =~ m|/policy_versions/(.*)$|; if (! $known_history_files{$1}) { print qq{$_\n} } }, no_chdir => 1, }, q{/ts/dms/policy/policy_versions});'
----------------------------------------------------------------------

Manually verity the file list output. If it seems correct, you can then delete the files by piping the output into 'xargs rm'.

In addition, you can delete the following file: /var/ts/var/install/recovery_db/conf.tar.gz.


678254-1 : Error logged when restarting Tomcat

Component: TMOS

Symptoms:
An error is logged after restarting Tomcat and using the web UI.

Conditions:
Using the web UI to restart tomcat.

Impact:
An error is logged after restarting Tomcat and using the web UI.

Workaround:
There is no workaround.

Fix:
When restarting Tomcat and using the web UI, and error will be logged only if the debug flag is enabled.


677958-4 : WS::frame prepend and WS::frame append do not insert string in the right place.

Component: Local Traffic Manager

Symptoms:
When WS::frame prepend and WS::frame append are used together in the same event, the strings are not inserted in the right place.

Conditions:
-- Both WS::frame prepend and WS::frame append commands are present in the same iRule event.
-- WebSocket and HTTP profile are configured on the virtual.
-- Client/server send and receive WebSocket frames.

Impact:
The user-supplied string is not inserted in the right place when sent to the end-point.

Workaround:
None.

Fix:
Separate buffers were now used for append and prepend, instead of reusing the same buffer.


677919-4 : Enhanced Data Manipulation AJAX Support

Component: Fraud Protection Services

Symptoms:
Need enhanced data manipulation detection to protect against modifying parameters in real-time (malware script in the browser) that are sent by JSON.

Conditions:
There is a malware script in the browser performing real-time modification of parameters that are sent by JSON.

Impact:
End-users already under attack could send manipulated JSON data to backend servers.

Workaround:
None.

Fix:
The Enhanced Data Manipulation Check has been improved so that it can now detect JSON data manipulation in the browser.


677666-2 : /var/tmstat/blades/scripts segment grows in size.

Solution Article: K60909141

Component: Local Traffic Manager

Symptoms:
Over time the /var/tmstat/blade/scripts file size grows.

Conditions:
-- Using istats.
-- Deleting configurations.
-- Using ASM.

Impact:
Virtual size of merged process grows linearly with /var/tmstat/blades/scripts segment. This could lead to out of memory condition.

Workaround:
No known workarounds.

Fix:
Condition corrected.


677525-2 : Translucent VLAN group may use unexpected source MAC address

Component: Local Traffic Manager

Symptoms:
When a VLAN group is configured in translucent mode, IPv6 neighbor discovery packets sent from the BIG-IP system may have the locally unique bit flipped in the source MAC address.

Conditions:
VLAN group in translucent mode.

Impact:
In an HA configuration, switches in the network may have FDB entries for the standby system assigned to the port of the active system.

Workaround:
No workaround at this time.

Fix:
Translucent VLAN group no longer send neighbor discovery packets whose source MAC has the locally unique bit flipped.


677494-1 : Flow filter with Periodic content insertion action could leak insert content record

Component: Policy Enforcement Manager

Symptoms:
Subscriber using flow filter and periodic insert content could create multiple records for same insert content action.

Conditions:
If two flows belonging to the same subscriber matching 2 different rules of the same policy and alternates and in the meanwhile policy rule action is updated.

Impact:
More than one record being created for the same insert content action.

Workaround:
There is no workaround at this time.

Fix:
Update the insert content array as soon as the pemdb record is updated.


677485-1 : Discovery of DSC clustered BIG-IP systems fails due to secure value decryption error

Component: TMOS

Symptoms:
After initially configuring a DSC cluster, iControl-REST on BIG-IP systems might fail to decrypt the secure values due to a stale BIG-IP master key in its cache, and returns the secure values encrypted by the BIG-IP master key. BIG-IQ is unable to decrypt these secure values and fails to discover the BIG-IP system.

Conditions:
-- DSC cluster.
-- iControl REST.
-- BIG-IP system with stale BIG-IP master key in its cache.
-- BIG-IQ attempts to decrypt the secure values.

Impact:
Discovery fails due to secure value decryption error.

Workaround:
Restart iControl-REST server on the BIG-IP system.

On BIG-IP v12.0.0 and later:
-- In TMSH, run the following command:
restart sys service restjavad
-- On the console, run the following command:
bigstart restart restjavad

On BIG-IP v11.x.x:
-- In TMSH, run the following command:
restart sys service icrd
-- On the console, run the following command:
bigstart restart icrd

Fix:
The system now enforces obtaining the BIG-IP master key if the first decryption fails to proceed properly.


677473-3 : MCPD core is generated on multiple add/remove of Mgmt-Rules

Component: Advanced Firewall Manager

Symptoms:
MCP crashes with core. MCP automatically restarts causing other dependent daemons, including tmm, to restart. Corresponding messages (restarting mcp, restarting tmm, and so on) are broadcast in all command line connections (terminals). Core files are written into /shared/core directory. The BIG-IP system might become unusable while daemons restart, so both control-plane (tmsh/GUI) and data traffic processing are stopped until all daemons restart.

Conditions:
-- AFM is licensed and provisioned.
-- There are firewall policies/rules defined in configuration.
-- Remove firewall rules/policies, especially rules attached to management-IP (might have to repeat this).

Impact:
The BIG-IP system might become unusable for few minutes while daemons restart, so both control-plane (tmsh/GUI) and data traffic processing are stopped. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
MCP no longer crashes, and other dependent daemons no longer restart. The BIG-IP system remains operational in both control-plane (tmsh/GUI) and data traffic processing.


677368-2 : Websso crash due to uninitialized member in websso context object while processing a log message

Component: Access Policy Manager

Symptoms:
Websso crashes occasionally on processing a log message on TMEVT_CLOSE event. This happens when a TMEVT_CLOSE event is received without receiving a request.

Conditions:
TMEVT_CLOSE event is received without receiving a request.

Impact:
Websso process crash.

Workaround:
None.

Fix:
A rare condition that caused the websso module to core is fixed by removing the webssocontext object from the logging function.


677148-1 : Periodic content insertion creates duplicate records if same policy is added to Global high and subscriber specific

Component: Policy Enforcement Manager

Symptoms:
If same pem policy with insert content is added to global high and subscriber specific, insert content could add duplicate records. This result in a case where if the periodic content tag is absent, the periodic content insertion will not scheduled immediately, but will add only after the expiry of the current interval.

Conditions:
If same pem policy with insert content is added to global high and subscriber specific.

Impact:
if the periodic content tag is absent, the periodic content insertion will not scheduled immediately.

Workaround:
This is a wrong configuration, a pem policy should be included either in Global High, or subscriber specific, not both.

Fix:
Re-use the already created record in case of same policy attached to Global high and subscriber specific


677088-2 : Qkview does not follow current best practices

Solution Article: K01067037


676457-5 : TMM may consume excessive resource when processing compressed data

Solution Article: K52167636


676346-2 : PEM displays incorrect policy action counters when the gate status is disabled.

Component: Policy Enforcement Manager

Symptoms:
Action counters are incorrect.

Conditions:
PEM policy actions enabled with gate status of disabled.

Impact:
May provide an inconsistent view of PEM actions.

Workaround:
There is no workaround.

Fix:
Counters are managed correctly regardless of the gate status.


676203-3 : Inter-blade mpi connection fails, does not recover, and eventually all memory consumed.

Component: TMOS

Symptoms:
TMM memory usage suddenly increases rapidly.

Conditions:
The inter-blade mpi connection fails and does not recover.

Impact:
Inter-blade mpi requests do not complete and the system eventually exhausts memory.

Workaround:
None.

Fix:
Inter-blade mpi connection now continues as expected, without memory issues.


675921-2 : Creating 5th vCMP 'ssl-mode dedicated' guest results in an error, but is running

Component: TMOS

Symptoms:
Creating 'ssl-mode dedicated' guests on the BIG-IP i5800, the 5th guest and beyond get an error, however they do become deployed with Status of 'running'.

Conditions:
-- Creating 5 (or more) 'ssl-mode dedicated' vCMP guests.
-- Running on the BIG-IP i5800 platform.

Impact:
5th guest and beyond result in an error.

Workaround:
There is no workaround other than not creating more than 4 'ssl-mode dedicated' vCMP guests when provisioning vCMP guests on the i5800 platform.

Fix:
The system now limits the maximum number of 'ssl-mode dedicated' vCMP guests to the number that the BIG-IP i5800 can physically support.


675866-4 : WebSSO: Kerberos rejects tickets with 2 minutes left in their ticket lifetime, causing APM to disable SSO

Component: Access Policy Manager

Symptoms:
Kerberos rejects tickets with 2 minutes left in their ticket lifetime. This causes tickets to be rejected by KDC, causing APM to disable SSO.

Conditions:
This occurs with Kerberos-protected resources using Windows Server 2012-based DC due to issue described in the Microsoft KB: Kerberos authentication fails when the computer tries to request a service ticket from a Windows Server 2012-based DC, https://support.microsoft.com/en-us/help/2877460/kerberos-authentication-fails-when-the-computer-tries-to-request-a-ser.

Impact:
Cannot access the Kerberos-protected resources.

Workaround:
None.

Fix:
Kerberos SSO (S4U) tickets are not used when the remaining lifetime is less than 5 minutes. Existing tickets with more than half the configured lifetime or at least 1 hour of lifetime remaining are used. If there are no such tickets, then new tickets are acquired and used.


675775-4 : TMM crashes inside dynamic ACL building session db callback

Component: Access Policy Manager

Symptoms:
Race condition between PPP tunnel close and Session expired happening on different TMM almost at the same time may cause TMM restart in dynamic ACL building session db callback.

Conditions:
Race condition between PPP tunnel close and Session expired happening on different TMM almost at the same time.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
None

Fix:
Guard against NULL pointer dereference for dynamic ACL build.


675232-6 : Cannot modify a newly created ASM policy within an iApp template implementation or TMSH CLI transaction

Component: Application Security Manager

Symptoms:
Errors encountered -

In TMSH CLI transaction:
----------------
transaction failed: 01020036:3: The requested ASM policy (/Common/<some_policy>) was not found.
----------------

In iApp template implementation:
----------------
script did not successfully complete: (01020036:3: The requested ASM policy (/Common/<some_policy>) was not found.
----------------

Conditions:
In an iApp template implementation or TMSH CLI transaction, create a new ASM policy and then try to modify it's active state.

Impact:
The policy is created but the modify action cannot find the policy.

Workaround:
iApps are built to work with ASM Policy Templates.

A new ASM Policy Template can be created from the desired ASM Policy.

That can be done via GUI and starting from from v13.0 via REST as well.

Then, the newly created ASM Policy Template can be referenced in the iApp template implementation or TMSH CLI transaction as follows:
-----------------
tmsh::create asm policy <some_policy> active policy-template NEWLY_CREATED_POLICY_TEMPLATE
-----------------

Fix:
iApp template implementation and TMSH CLI transaction can now modify a newly created ASM policy.


674747-4 : sipdb cannot delete custom bidirectional persistence entries.

Solution Article: K30837366

Component: Service Provider

Symptoms:
Custom bidirectional SIP persistence entries cannot be deleted using the sipdb tool.

Conditions:
Rules and SIP messages created custom bidirectional SIP persistence entries.

Impact:
Custom bidirectional SIP persistence entries exist and can be viewed with the sipdb utility. They cannot be deleted,
however.

Workaround:
None.

Fix:
The sipdb tool now supports deletion of bidirectional SIP persistence entries.


674576-3 : Outage may occur with VIP-VIP configurations

Component: Local Traffic Manager

Symptoms:
In some VIP-VIP configurations, TMM may produce a core with a 'no trailing data' assert.

Conditions:
VIP-VIP configuration.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
No workaround at this time.

Fix:
TMM no longer produces a core with a 'no trailing data' assert.


674494-4 : BD memory leak on specific configuration and specific traffic

Solution Article: K77993010

Component: Application Security Manager

Symptoms:
RSS memory of the bd grows.

Conditions:
-- Remote logger is configured.
-- IP has ignore logging configured.
-- Traffic is coming from the ignored logging IP.

Impact:
Potential memory exhaustion. The kernel might run out of memory and may kill bd, causing traffic disruption.

Workaround:
None.

Fix:
Freeing up the remote loggers data when deciding not to log remotly.


674455-5 : Serial console baud rate setting lost after running 'tmidiag -r' in Maintenance OS

Component: TMOS

Symptoms:
When booted into the Maintenance OS image from the grub boot menu, running tmidiag -r drops the serial console from the grub kernel line, which causes a loss of communication on the serial console after rebooting.

Conditions:
-- Booted into Maintenance OS.
-- Running the command: tmidiag -r

Impact:
Serial console baud rate settings are incorrect. Uses the bios baud rate on the console.

Workaround:
When booting, edit the grub kernel line to include console=ttyS0.

Note: The value is "tty", an uppercase "S" character, and zero, so ttyS0.

Fix:
tmidiag has been fixed to not strip out console=ttyS0.


674256-2 : False positive cookie hijacking violation

Solution Article: K60745057

Component: Application Security Manager

Symptoms:
A false positive cookie hijacking violation.

Conditions:
-- Several sites are configured on the policy, without subdomain.
-- TS cookies are sent with the higher domain level then the configured.
-- A single cookie from another host (that belongs to the same policy) arrives and is mistaken as the other site cookie.

Impact:
False positive violation / blocking.

Workaround:
Cookie hijacking violation when the device ID feature is turned off is almost never relevant, as it should be able to detect only cases where some of the TS cookies were taken. The suggestion is to turn off this violation.


673996-2 : Unable to set 'media-fixed' on management port on BIG-IP i15000 platforms

Component: TMOS

Symptoms:
Changing 'media-fixed' on management port on BIG-IP i15000 platforms using tmsh command 'tmsh modify net interface mgmt media-fixed <speed>' does not take effect.

Conditions:
-- Connecting two BIG-IP i15000 units via management port.
-- Changing the 'media-fixed' value.

Impact:
Changing the 'media-fixed' value does not work.

Workaround:
Pull the management cable out and plug it back in to get the link up at the respective speeds.

Fix:
Users can now change the 'media-fixed' value using tmsh commands.


673832-1 : Performance impact for certain platforms after upgrading to 13.1.0.

Component: Performance

Symptoms:
Performance impact for certain platforms after upgrading to 13.1.0.

Conditions:
The following platforms, with Fast HTTP/OneConnect/Full Proxy configured.

-- i2800
-- i4800
-- i5800
-- i7800
-- i10800
-- i11800
-- B2250
-- B4450

Impact:
The performance impacts occur on the following platforms under the associated conditions:

-- i2800 2%-3% Full Proxy traffic.
-- i4800 2%-3% Full Proxy traffic.
-- i5800 3%-8% Fast HTTP/Full Proxy traffic.
-- i7800 3%-7% Fast HTTP/Full Proxy traffic.
-- i10800 3%-7% Fast HTTP/Full Proxy traffic.
-- i11800 2%-3% Fast HTTP traffic.
-- B2250 3%-6% OneConnect/Full Proxy traffic.
-- B4450 4%-10% Fast HTTP/OneConnect/Full Proxy traffic.

Workaround:
None.

Fix:
Performance impact for certain platforms has been eliminated.


673717-3 : VPE loading times can be very long

Component: Access Policy Manager

Symptoms:
When a configuration contains 2000-3000 Access Policy objects, the Visual Policy Editor (VPE) loading operations can take tens of seconds to load.

Conditions:
-- Access Policy configured with 2000-3000 Access Policy objects.
-- Open in VPE.

Impact:
Policies with thousands of entries can take tens of seconds or more to load.

Workaround:
None.

Fix:
Configuration load has been optimized, so VPE loading time is significantly shorter for these types of configurations.


673664-1 : TMM crashes when sys db Crypto.HwAcceleration is disabled.

Component: Local Traffic Manager

Symptoms:
TMM crashes when sys db Crypto.HwAcceleration is disabled.

Conditions:
This occurs when sys db Crypto.HwAcceleration is disabled.

Impact:
TMM crash. Traffic disrupted while tmm restarts.

Workaround:
Enable crypto hardware acceleration using the following command:
tmsh modify sys db crypto.hwacceleration value enable


673607-9 : Apache CVE-2017-3169

Solution Article: K83043359


673595-9 : Apache CVE-2017-3167

Solution Article: K34125394


673399-3 : HTTP request dropped after a 401 exchange when a Websockets profile is attached to virtual server.

Component: Local Traffic Manager

Symptoms:
The client sends a GET request with switching protocols and the server responds with a 401. Subsequent GET request from client is dropped.

Conditions:
HTTP and Websocket profiles are attached to the virtual server. The client sends a GET request with switching protocols indicating upgrade to Websockets, and the server responds with a 401. Subsequent GET request from client is dropped.

Impact:
Connection is reset.

Workaround:
Disable Websockets profile on the virtual server.

Fix:
We now check whether the Websockets filter is on the virtual server before attempting an insert.


672667-6 : CVE-2017-7679: Apache vulnerability

Solution Article: K75429050


672514-1 : Local Traffic/Virtual Server/Security page crashed

Component: Advanced Firewall Manager

Symptoms:
Local Traffic/Virtual Server/Security page crashes when AFM policy with 20k rules is attached to a virtual.

Conditions:
1. AFM provisioned.
2. AFM policy attached to Virtual with 20k rules with UUID.
3. Traffic hitting Virtual.

Impact:
Unable to manage the AFM policy using the management utility.

Workaround:
Use tmsh to attach AFM policy and other security items

Fix:
The issue is fixed. Policy Rules are displayed on Virtual security page only for Management context.


672504-2 : Deleting zones from large databases can take excessive amounts of time.

Solution Article: K52325625

Component: Global Traffic Manager (DNS)

Symptoms:
When deleting a zone or large number of Resource Records, zxfrd can reach 100% CPU for large amounts of time.

Conditions:
With a significantly sized database, deletes might be very time-intensive.

Impact:
Because zxfrd takes an excessive amount of time deleting records, it can delay transfer requests

Workaround:
None.

Fix:
Dramatically improved algorithm, to remove significant delay in deletions.


672312-3 : IP ToS may not be forwarded to serverside with syncookie activated

Component: Local Traffic Manager

Symptoms:
IP ToS field may not be preserved on forwarding in a FastL4 virtual server when syncookie is activated.

Conditions:
-- FastL4 ip-forwarding virtual server.
-- ToS in passthrough mode.
-- Syncookie is activated in hardware mode.
-- On a hardware platform with ePVA.

Impact:
IP ToS header is not forwarded to the serverside.

Workaround:
None.

Fix:
The BIG-IP system now forwards IP ToS in syncookie mode.


672124-6 : Excessive resource usage when BD is processing requests

Solution Article: K12403422


671716-1 : UCS version check was too strict for IPS hitless upgrade

Component: Protocol Inspection

Symptoms:
When we upgrade from one minor release to another, e.g. from 13.1 to 13.2, then UCS upgrade of IPS IM packages fail.

Conditions:
During upgrade from one minor release to another.

Impact:
The default library will be used instead of the last updated IM/IPS library in last build.

Workaround:
Install the IM package available for that new release.


671712-2 : The values returned for the ltmUserStatProfileStat table are incorrect.

Component: TMOS

Symptoms:
An SNMP table entry that intermittently comes back incorrectly. Specifically, an LTM profile user statistic value is returned with an OID that is off by one in the table.

Conditions:
A statistics profile is attached to a virtual server with an iRule that increments statistics on the profile. The SNMP walk of the table created by the statistics profile intermittently returns the values shifted up one OID such that the last row is zero and the first row is lost. A subsequent request is correct.

Impact:
Incorrect data returned in SNMP walk of LTM profile table.

Workaround:
The statistics themselves are kept correctly and can be accessed without using SNMP.

Fix:
The values in the ltmUserStatProfileStat table are always correct.


671627-3 : HTTP responces without body may contain chunked body with empty payload being processed by Portal Access.

Solution Article: K06424790

Component: Access Policy Manager

Symptoms:
Some HTTP responses do not contain any body. For instance, responses with status codes 1xx, 204, or 304 must not include body. Portal Access adds 'Transfer-Encoding: chunked' header and may add chunked body with empty payload to such responses.

Conditions:
HTTP response without body processed by Portal Access

Impact:
Most browsers ignore invalid 'Transfer-Encoding' header and/or body for responses which must not include body at all. And yet, some traffic validators may refuse to pass invalid responses.

Workaround:
Use an iRule to remove 'Transfer-Encoding' header and/or body from HTTP responses with status codes 1xx, 204, and 304.

Fix:
Now Portal Access does not add invalid 'Transfer-Encoding' header and/or body to responses which have no body.


671597-3 : Import, export, copy and delete is taking too long on 1000 entries policy

Component: Access Policy Manager

Symptoms:
Huge policies with 10^3 items are impossible to import, export and copy.

Conditions:
When access policy has 1000+ entires.

Impact:
Import, export and copy are abandoned or fail due to out of memory condition.

Workaround:
Use ng_export, ng_import and ng_profile rather than UI to import/export.

Fix:
ng_export speed has been improved 5 times
ng_import and ng_profile are working 50 times faster because of avoiding denormalisation and optimisation

ng_export is still should be used from the console.


671323-1 : Reset PIN Fail if Token input field is not 'password' field

Component: Access Policy Manager

Symptoms:
User is not able to reset the PIN when the password source field in RSA SecurID or RADIUS Auth agent is not set to default value(%{session.logon.last.password})

Conditions:
- APM is licensed and provisioned.
- RSA SecurID or RADIUS Auth agent is included in an access policy.
- Password source field in this agent is changed to a custom value.
- APM end user is challenged to reset the PIN or reenter the PIN/token.

Impact:
APM end users cannot reset the PIN or do not get authenticated.

Workaround:
There is no workaround other than not changing the default value in password source fields for RADIUS or RSA SecureID auth agent.

Fix:
APM end users can now successfully reset the PIN or reenter the token. They can also use custom password session variables for authentication.


670528-4 : Warnings during vCMP host upgrade.

Solution Article: K20251354

Component: TMOS

Symptoms:
- Log message repeats every 5 seconds in /var/log/ltm
     slot<#>/<host> warning vcmpd[<pid>]: 01510005:4: Failed to find value for enum::cli_id (ha_feature_t::provisioning-failed).

Conditions:
- Configure vCMP host in 12.1.x or 11.6.x.
 - Deploy 13.x guest.
 - Monitor /var/log/ltm.

Impact:
Warning message displayed every 5 seconds.

Workaround:
Run the following command:
 tmsh create sys log-config filter stop_vcmpd_log message-id 01510005 publisher none


670103-1 : No way to query logins to BIG-IP in TMUI

Component: TMOS

Symptoms:
Cannot use the GUI to query logins to the BIG-IP system based on a time range or a specific user.

Conditions:
-- Using the GUI.
-- Gather login information.

Impact:
No support for queries.

Workaround:
None.

Fix:
Added support for using using the GUI to query logins to the BIG-IP system.

Behavior Change:
The ability to query logins on the BIG-IP, using the GUI, was added at System >> Logins : History. Users can query all available login data that is present on the BIG-IP. This information can be filtered by time, username, status, access method, and host.


669585-1 : The tmsh sys log filter is unable to display information in uncompressed log files.

Component: TMOS

Symptoms:
You notice missing log information when reviewing system logs using the tmsh show sys log command.

Conditions:
One or more of the BIG-IP sytem backup log files, designated with .1, .2, etc are not compressed.
Note: Backup log files should end with the .gz extension. For example, ltm.1.gz.
You use the tmsh show sys log command to view log information for one or more days in the past.

Impact:
Unable to view the full range of backup log information.

Workaround:
To log in to the Advanced shell (bash).
To ensure all backup logs for a particular log type are compressed, use the following command syntax:

gzip /var/log/<log>.*

For example, to compress the full set of backup logs for the ltm log type, type the following command:

Note: The following message is expected if the log file is already compressed: gzip: /var/log/<log>.gz already has .gz suffix -- unchanged'

gzip /var/log/ltm.*

Fix:
Increased flexibility of log reading mechanism, to look for both compressed (ending in .gz) and uncompressed (ending in .#) log files.


669462-2 : Error adding /Common/WideIPs as members to GTM Pool in non-Common partition

Component: TMOS

Symptoms:
Unable to use Pool Members from /Common/ when outside of /Common/

Conditions:
Adding /Common/WideIPs as members in non-Common GTM Pool

Impact:
Unable to use pool-members from /Common/ when outside of /Common/

Workaround:
No workaround at this time.

Fix:
Fixed issue preventing users from using GTM pool-members within /Common/ on GTM Pools outside of /Common/


668826-1 : File named /root/.ssh/bigip.a.k.bak is present but should not be

Component: TMOS

Symptoms:
In AWS instances, a file /root/.ssh/bigip.a.k.bak is present which should not be. It is harmless to users other than that it is confusing.

Conditions:
After the first boot, this file should be deleted, but it is not.

Impact:
No real impact other than possibly confusion as this file isn't used in this environment. The file does not contain any sensitive data as it's a dangling symlink.

Workaround:
No need to workaround as the presence of the file is harmless. Users could manually remove this file if desired.

Fix:
This file is no longer present which is the correct state.


668276-1 : BIG-IP does not display failed login attempts since last login in GUI

Component: TMOS

Symptoms:
The BIG-IP does not have a mechanism in the GUI to display information about login attempts.

Conditions:
n/a

Impact:
Administrators cannot use the GUI to evaluate login attempts to the BIG-IP.

Workaround:
Administrators can view the logs at /var/log/secure.

Fix:
New GUI pages were create to allow administrators, resource admins, and auditors to view information about login attempts to the BIG-IP. These pages are available at System >> Logins in the GUI.
The user logins summary, available at System >> Logins : Summary can be set as the default start screen for BIG-IP users. However, this process is not as straightforward as other pages, as these pages are available only to users with a role of admin, resource admin, or auditor. Because of these restrictions, setting this page as default is accomplished by setting a DB variable, UI.Users.RedirectSuperUsersToAuthSummary, to true.
When this DB variable is set to true, users with roles of admin, resource admin, or auditor will be redirected to the System >> Logins : Summary page. Users with other roles will be redirected to the Start Screen that is set in System >> Preferences.


668273-1 : Logout button not available in Configuration Utility when using Client Cert LDAP

Solution Article: K12541531

Component: TMOS

Symptoms:
When the BIG-IP system is configured to use the Client Cert LDAP for Remote Authorization, the Logout button is not available.

Conditions:
A BIG-IP system is configured to use Client Cert LDAP for Remote Authorization.

Impact:
BIG-IP system users cannot end the session on the BIG-IP system.

Workaround:
Close all windows to end the session.

Fix:
Now, when the BIG-IP system is configured to use Client Cert LDAP as the Remote Auth method, there is a Logout button in the window, and when the Logout button is clicked, the system displays a modal window to instruct the user on how to end the session.


668184-2 : Huge values are shown in the AVR statistics for ASM violations

Component: Application Security Manager

Symptoms:
Huge values are shown in the AVR statistics for ASM violations.

Conditions:
Out-of memory-condition in the ASM. Some other extreme conditions might also cause this behavior.

Impact:
ASM violation numbers are incorrectly reported.

Workaround:
None.

Fix:
An issue with bd sending wrong numbers to AVR was fixed.


667770-1 : SIGSEGV to tmm when performing multiple, repeated SSL profile updates, or during UCS restore

Solution Article: K12472293

Component: Local Traffic Manager

Symptoms:
The BIG-IP system sends a SIGSEGV to the TMM process when performing multiple change operations on configurations that contain a combination of SSL profiles and AVR (analytics).

Conditions:
-- Configuration contains a combination of SSL profiles and AVR.

-- Performing multiple, repeated SSL profile updates, or during UCS restore.

Impact:
The BIG-IP system sends a SIGSEGV to the TMM process. Traffic disrupted while tmm restarts.

Workaround:
There is no workaround.

Fix:
TMM no longer halts and restarts when performing multiple change operations on configurations that contain a combination of SSL profiles and AVR (analytics).


667542-6 : DNS Express does not correctly process multi-message DNS IXFR updates.

Component: Global Traffic Manager (DNS)

Symptoms:
If DNS Express sends an IXFR query to a DNS server, and that server responds with an IXFR update that is larger than one DNS message, DNS Express processes only the first message.

DNS Express then updates the SOA serial number to match that of the IXFR, marks the IXFR as successful and the Zone as 'Green'.

There is no indication that the IXFR was incomplete.

DNS Express might then have, and might serve, incorrect data for that Zone.

Conditions:
An IXFR response from a DNS server spans multiple DNS messages.

Note: This is not a common condition, but it is possible.

Impact:
This might result in incomplete or otherwise inaccurate Zone data, which DNS Express will serve.

Workaround:
Note: Although this does have a workaround, there is no way for you to determine that the Zone is complete other than to query the entire zone and compare it to the zone from the master DNS server.

To workaround this issue:
1. Stop zxfrd.
2. Remove the database /var/db/zxfrd.bin.
3. Restart zxfrd.

This triggers a full transfer (AXFR) of the zone, as well as all the other zones.

Fix:
The system now continues the processing of DNS messages until the closing SOA RR is encountered.


667469-3 : Higher than expected CPU usage when using DNS Cache

Solution Article: K35324588

Component: Global Traffic Manager (DNS)

Symptoms:
CPU usage shows higher than expected usage when using the DNS Cache.

Conditions:
Usage of any of the 3 DNS Cache types, particularly on chassis with multiple blades.

Impact:
Higher than expected CPU usage.

Workaround:
No workaround at this time.

Fix:
Improvements have been made to the efficiency of the DNS Cache inter-tmm mirroring. These efficiencies may result in better CPU utilization and/or higher responses per second.


667353 : Intermittent TMM crash when AFM dynamic (behavioral) signature is enabled and past attack signatures incorrectly exist in the correlation stats table

Component: Advanced Firewall Manager

Symptoms:
Intermittent TMM crash when AFM dynamic (behavioral) signature is enabled and past attack signatures incorrectly exist in the correlation stats table - issue is due to TMM (self) abort due to memory corruption in one of the TMSTAT tables AFM uses for correlating dynamic signatures.

Conditions:
Following conditions suffice to trigger the TMM crash due to self abort in one of the TMSTAT tables:

a) Generate a set of N dynamic signatures (few context).

b) When attack stops, the current set of signatures are moved to 'past' attack state.

c) If in between, TMM restarts (or receives MCP config again e.g via load), these past attack signatures are incorrectly created in tmstat table which is used only for the current attack signatures - this is the *cause* of the issue!

d) New attack appears that somewhat overlap with the 'past' signatures and this causes the following TMSTAT table to be corrupted over period of time.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
There is no workaround at this time.

Fix:
This issue is fixed, the past attack signatures are never created in the correlation stats table (even for conditions explicitly described above)


667173-1 : 13.1.0 cannot join a device group with 13.1.0.1

Component: TMOS

Symptoms:
13.1.0.1 cannot form device trust with a 13.1.0 device.

Conditions:
A device running 13.1.0 wanting to establish device trust with a device running 13.1.0.1 or vice versa.

Impact:
Cannot form Device Trust.

Workaround:
13.1.0 cannot initially form device trust with a 13.1.0.1 device. However, if you establish trust from the 13.1.0.1 device and then bring in the 13.1.0 device from 13.1.0.1, you can mitigate this issue. Once trust is formed, there should be no issue.

Fix:
13.1.0.1 now can form device trust with a 13.1.0 device.


667148-3 : Config load or upgrade can fail when loading GTM objects from a non-/Common partition

Solution Article: K02500042

Component: TMOS

Symptoms:
GTM configuration fails to load.

Conditions:
GTM config referencing non-/Common partition objects from /Common.

Impact:
GTM configuration fails to load, which may keep a system from becoming active

Workaround:
No workaround.

Fix:
Fixed issue preventing GTM configurations from loading when non-Common partitioned items present.


665992-2 : Live Update via Proxy No Longer Works

Solution Article: K40510140

Component: Application Security Manager

Symptoms:
BIG-IP devices that need to use a proxy server to communicate with callhome.f5.com, no longer receive, or check for, automatic updates.

Conditions:
The BIG-IP device is behind a network firewall and outbound communication must be through a proxy.

Impact:
The BIG-IP will not be able to contact the callhome server to check for, or receive, updates.

Workaround:
Updates can be downloaded manually from the F5 Downloads server and installed directly on the BIG-IP.

Fix:
Proxy settings are correctly used when contacting the F5 callhome server.


665362-2 : MCPD might crash if the AOM restarts

Component: TMOS

Symptoms:
In very rare circumstances, mcpd might crash when the AOM restarts.

Conditions:
This can occur while AOM is restarting.

Impact:
System goes offline for a few minutes.

Workaround:
None.

Fix:
Added error handling to prevent crash. If this error occurs in the future it will not crash, but a restart of mcpd is required.


665354-1 : Silent reboot, identified with bad_tlp_status and completion_time_out in the sel log

Solution Article: K31190471

Component: TMOS

Symptoms:
The most common symptom is a reboot of the unit without much detail in the normal tmm or ltm logs. From there, inspect the SEL logs. In the SEL logs, you will see a message about a bad_tlp_status, followed shortly by a message about completion_time_out_status.

Those two messages together indicate this known issue.

Conditions:
-- There are empty 10 GB ports or 10 GB ports that have optics but are not connected to a proper link.
-- Running on one of the following platforms: i2600, i2800, i4600, i4800.

Impact:
The unit intermittently reboots.

Workaround:
To prevent the issue from occurring, you must populate all 10 Gb ports with optic cables and ensure they are connecting to a working peer link. A single 10 Gb empty or improperly connected port can cause a system reboot.

If that is not possible, however, there is no workaround, and you must contact F5 Technical Support to request a software update or engineering hotfix.

Important: A device Return Materials Authorization (RMA) will not prevent this issue.

Fix:
There is a BIG-IP system software update to disable the 10 Gb FPGA mac receiver until a valid link is detected. This eliminates the issue and prevents the ultra jumbo packet from being sent to the FPGA datapath.


664528-2 : SSL record can be larger than maximum fragment size (16384 bytes)

Solution Article: K53282793

Component: Local Traffic Manager

Symptoms:
SSL record containing handshake data can exceed maximum fragment size of 16384 bytes because handshake data is not fragmented.

Conditions:
This usually happen when a large certificate or certificate chain is configured for server or client authentication.

Impact:
SSL handshake will fail with client or server that properly checks the record size.

Workaround:
Use a certificate that is smaller in size.

Fix:
Properly fragment handshake data.


663874-2 : Off-box HSL logging does not work with PEM in SPAN mode.

Solution Article: K77173309

Component: Policy Enforcement Manager

Symptoms:
While on-box HSL logging works, off-box HSL logging does not work with PEM in SPAN mode.

Conditions:
-- PEM in SPAN mode.
-- Off-box HSL logging is configured.

Impact:
Cannot use off-box HSL logging with PEM in SPAN mode; must use on-box HSL logging instead.

Workaround:
There is no workaround at this time.

Fix:
Off-box HSL logging now works with PEM in SPAN mode.


663821-1 : SNAT Stats may not include port FTP traffic

Solution Article: K41344010

Component: Local Traffic Manager

Symptoms:
Using 'tmsh show ltm snat' command or the GUI, SNAT stats are not updated for port 21 traffic (FTP).

Conditions:
-- SNAT is configured.
-- FTP connections transverse the SNAT.

Impact:
Stats are not incremented in tmsh or GUI

Workaround:
None.

Fix:
SNAT Stats now include port FTP traffic, and values are incremented as expected.


662850-6 : Expat XML library vulnerability CVE-2015-2716

Solution Article: K50459349


662311-1 : CS alerts should contain actual client IP address in XFF header

Component: Fraud Protection Services

Symptoms:
When no XFF header exists, the alert server will use the sender IP address as the client IP address. Doing so is incorrect behavior because the sender IP address is always the BIG-IP system's IP address. Even if XFF headers exist, the client IP address as known to the BIG-IP system may be missing in the XFF header.

Conditions:
This occurs under either of the following conditions:
-- There is no XFF header in the original request.
-- An XFF header exists, but it does not contain the actual client IP address (as seen by the BIG-IP system).

Impact:
Alert server/BIG-IQ does not show the actual client IP address.

Workaround:
None.

Fix:
FPS now always appends the client IP address to the end of the last XFF header in the alert request. If there is no XFF header, FPS inserts one.


661939-2 : Linux kernel vulnerability CVE-2017-2647

Solution Article: K32115847


660239-6 : When accessing the dashboard, invalid HTTP headers may be present

Component: TMOS

Symptoms:
When accessing parts of the BIG-IP dashboard via the GUI, there might be invalid HTTP headers in the responses.

Conditions:
Access the dashboard via Statistics :: Dashboard.

Impact:
The invalid HTTP headers might cause issues with the dashboard if there are intervening proxies between the browser and the BIG-IP.

You may see such errors in the http error logs

Feb 20 08:47:58 myBIG-IP err httpd[13777]: [error] [client 10.20.30.40] Response header name '<PostData><![CDATA[table=log%5Fstat]]></PostData>Cache-Control' contains invalid characters, aborting request, referer: https://mybigip.com/tmui/dashboard/MonitorDashboardModule.swf

Workaround:
There is no workaround at this time.

Fix:
Eliminated invalid header data.


658410-2 : icrd_child generates a core when calling PUT on ltm/data-group/internal/

Component: TMOS

Symptoms:
icrd_child generates a core file when calling PUT on ltm/data-group/internal/.

Conditions:
Calling PUT on ltm/data-group/internal/.

Impact:
The iControl REST API is temporarily not available for configuration queries or modifications.

Workaround:
There is no workaround at this time.

Fix:
icrd_child no longer cores when calling PUT on ltm/data-group/internal/.


658278-1 : Network Access configuration with Layered-VS does not work with Edge Client

Component: Access Policy Manager

Symptoms:
When Network Access is configured in virtual server-to-virtual server targeting, the Edge client cannot connect.

Conditions:
Network Access is configured as follows :
-- The external-client-facing virtual server has the SSL profile attached.
-- The internal virtual server has the Access profile and connectivity profile attached.
-- The external-client-facing virtual server has an iRule that forwards the HTTP requests to the internal virtual server.

Impact:
When Edge client connects, the external-client-facing virtual server issues a request for '/pre/config.php?version=2.0', and the Edge client hangs.

Workaround:
None.

Fix:
Network Access configuration with Layered-VS now works with Edge Client.


656901-3 : MRF add 'existing_connection_only' and 'outgoing_connection_instance_seed' two iRule commands

Component: Service Provider

Symptoms:
If the MRF 'existing_connection_only' is not there, then MRF will forward the new message to either the existing connection or creating a new connection.

If the MRF 'outgoing_connection_instance_seed' is not there, then the generation of the connection's instance number will use some internal originating connection id. Same client IP with different src_port may end up to different outgoing connection.

Conditions:
If these two new iRule commands were not there.

Impact:
1. Won't always reuse the existing connection.
2. The requests from same client IP with different src_port, the outgoing connection may be different.

Workaround:
There is no workaround at this time.

Fix:
MR::message existing_connections_only <boolean> Gets or sets a flag that instructs the MRF to only forward the message using existing connections,
and if a connection to the selected host does not exist then the route will fail.

MR::message outgoing_connection_instance_seed <integer>Gets or if been set by this iRule then this seed will be used to generate the connection instance number instead of this generated by some internal originating connection id. (See MR::connection_instance iRule command).

If the number received is larger than 32 bit then the 64 bit number will be hashed to 32 bit number.


655233-2 : DNS Express using wrong TTL for SOA RRSIG record in NoData response

Solution Article: K93338593

Component: Global Traffic Manager (DNS)

Symptoms:
DNS Express returns an incorrect TTL for the SOA RRSIG record in a NoData response.

Conditions:
-- DNS Express configured.
-- A query that results in a NoData response and DNSSEC signing requested.

Impact:
This brings the behavior in line with RFC2308. There is no known functional impact.

Workaround:
There is no workaround.

Fix:
The TTL of the RRSIG record now matches the TTL of the covered SOA record.


653976-4 : SSL handshake fails if server certificate contains multiple CommonNames

Solution Article: K00610259

Component: Local Traffic Manager

Symptoms:
SSL server side handshake fails when the external server certificate's Subject field contains multiple CommonNames.

Conditions:
This issue occurs when both of the following conditions are met:
-- The external server certificate's Subject field contains multiple CommonNames.
-- The certificate does not contain subjAltName extension (or if it does, the same names are not included in the subjAltName's dNSName list).

Impact:
Connection with external server cannot be established. In case of forward proxy, bypass or intercept will fail.

Workaround:
In case of forward proxy bypass, configure IP address bypass instead of hostname bypass since IP address bypass check happens before SSL handshake.

The second option is to update the external server's certificate to include the list of CommonNames in subjAltName extension as dNSName.

Fix:
The system now checks all CommonNames in a certificate's Subject field instead of checking only the longest one in length.


653759-1 : Chassis Variant number is not specified in C2200/C2400 chassis after a firmware update

Component: TMOS

Symptoms:
Chassis Variant number is not specified when checking the log file /var/log/ltm, for example:

#grep queryFDD /var/log/ltm
...debug chmand[12982]: 012a0007:7: queryFDD returned 1 items for: update|F100|||NONE|NONE|NONE|0x0

This should contain the Variant number 400-0028-04, as follows:
...debug chmand[32663]: 012a0007:7: queryFDD returned 1 items for: update|F100|400-0028-04||NONE|NONE|NONE|0x0

Conditions:
-- B2100/B2150/B2200 blade in C2200/C2400 chassis.
-- Checking for the Chassis Variant number.

Impact:
This has no impact, since there are no Variants currently defined for the C2200/C2400 chassis.

Workaround:
There is no workaround at this time.

Fix:
Chassis Variant number is printed out as expected in the log file.


653201-2 : Update the default CA certificate bundle file to the latest version and remove expiring certificates from it

Component: Local Traffic Manager

Symptoms:
The default CA certificate bundle file used by the system contains some older certificates, e.g., expired or soon-to-be expired.

Conditions:
If the default CA certificate bundle file is configured in SSL profiles, it is used as a set of built-in trusted certificates when verifying peer's certificate during SSL handshake.

Impact:
When the built-in trusted certificates are obsolete, i.e., containing a certain number of expired certificates, the systems might fail to verify peers certificate correctly.

Workaround:
You can either directly update the default CA certificate bundle file /config/ssl/ssl.crt/ca-bundle.crt with proper certificates and then 'bigstart restart tmm'

Alternatively, you can use a separate certificate, for example:
tmsh install sys crypto cert better_ca_bundle from-local-file /shared/better_ca_bundle.pem
tmsh modify ltm profile client-ssl cssl ca-file better_ca_bundle.crt

Fix:
This release updates the default CA certificate bundle file by adding the latest certificates and removing the expired certificates.


652877-5 : Reactivating the license on a VIPRION system may cause MCPD process restart on all secondary blades

Component: TMOS

Symptoms:
All services on one or all secondary blades in a VIPRION chassis restart, and MCPD logs errors similar to the following:

-- err mcpd[9063]: 01070734:3: Configuration error: DB validation exception, unique constraint violation on table (sflow_vlan_data_source) object ID (1168). A duplicate value was received for a non-primary key unique index field. DB exception text (Cannot update_indexes/checkpoint DB object, class:sflow_vlan_data_source status:13)
-- err mcpd[9063]: 01070734:3: Configuration error: Configuration from primary failed validation: 01070734:3:Configuration error: DB validation exception, unique constraint violation on table (sflow_vlan_data_source) object ID (1168). A duplicate value was received for a non-primary key unique index field. DB exception text (Cannot update_indexes/checkpoint DB object, class:sflow_vlan_data_source status:13)... failed validation with error 17237812.

In versions prior to v11.6.0, the error is: 'Can't save/checkpoint DB object,' rather than 'Can't update_indexes/checkpoint DB object'.

Conditions:
Multi-bladed VIPRION system, where the 'if-index' value for VLANs differs between blades.

You can check the 'if-index' value by running the following command on each blade: tmsh list net vlan all if-index.

Impact:
MCPD restart on all secondary blades results in partial service outage.

Workaround:
Reactivate the license only on a system that is standby/offline.

Fix:
Reactivating the license on a VIPRION system no longer causes MCPD process restart on one or all secondary blades.


651741-2 : CVE-2017-5970 kernel.el7: ipv4: Invalid IP options could cause skb->dst drop

Solution Article: K60104355


651413-4 : tmsh list ltm node does not return an error when node does not exist

Solution Article: K34042229

Component: TMOS

Symptoms:
TMSH does not post an error message in response to the tmsh command to list a specific, non-existent LTM node, or when listing a set of non-existent nodes using regular expressions.

Conditions:
-- Running the command: tmsh list ltm node.
-- Running a regular expression to list a set of nodes.
-- The specified node does not exist.

Impact:
The command produces no output or error message. No indication of why there is no output, nor is there a description of the possible error condition.

Workaround:
None.

Fix:
TMSH now posts the appropriate, node-not-found error message when LTM nodes do not exist when running the command: tmsh list ltm node.


649161-2 : AVR caching mechanism not working properly

Solution Article: K42340304

Component: Application Visibility and Reporting

Symptoms:
The AVR caching mechanism fails to store dimension-based queries properly, which leads to incorrect reports.

Conditions:
Using AVR caching mechanism (turned-on by default).

Impact:
Reports will be incorrect.

Workaround:
Using the following TMSH command should solve the problem:
tmsh modify sys db avr.requestcache value disable

* NOTE: the above might cause AVR to perform a bit slower.

Fix:
The system no longer stores the dimension-based queries in the AVR cache.


648802-1 : Required custom AVPs are not included in an RAA when reporting an error.

Component: Policy Enforcement Manager

Symptoms:
Custom Attribute-Value-Pairs (AVPs) defined on the system will not be visible in a Re-Auth-Answer (RAA).

Conditions:
This occurs when the following conditions are met:
-- PEM enabled with Gx.
-- Custom AVPs configured.
-- PEM responds with an error code in an RAA against a Policy and Charging Rules Function (PCRF) triggered action.

Impact:
Lack of custom AVPs in such an RAA may result in loss of AVP dependent service.

Workaround:
There is no workaround at this time.

Fix:
Custom AVPs included regardless of an error code in an RAA.


648766-2 : DNS Express responses missing SOA record in NoData responses if CNAMEs present

Solution Article: K57853542

Component: Global Traffic Manager (DNS)

Symptoms:
A valid NoData response can contain CNAMEs if a partial chase occurred without final resolution. DNS Express is not including the expected SOA record in this scenario.

Conditions:
-- DNS Express configured.
-- Partial CNAME chase resulting in incomplete resolution.

Impact:
A valid DNS response with a a partial chase but missing the SOA record may not be considered authoritative due to the missing record.

Workaround:
None.

Fix:
The SOA record is now included as appropriate.


648320-5 : Downloading via APM tunnels could experience performance downgrade.

Solution Article: K38159538

Component: Local Traffic Manager

Symptoms:
Multiple DTLS records can be packed into one UDP packet. When packet size is too large, packet fragmentation is possible at IP layer. This causes high number of packet drops and therefore performance downgrade.

Conditions:
When downloading using APM tunnels.

Impact:
High number of packet drops and inferior performance.

Workaround:
None.

Fix:
One DTLS record is now contained in each UDP packet to avoid packet fragmentation.


648242-2 : Administrator users unable to access all partition via TMSH for AVR reports

Solution Article: K73521040

Component: Application Visibility and Reporting

Symptoms:
Using the TMSH for AVR reports can fail if it contains partition based entities, even with an administrator user (which should have permissions to all partitions).

Conditions:
Using the TMSH for querying partitioned based stats with an administrator user.

Impact:
AVR reports via TMSH will fail when using partition based entities.

Workaround:
None.

Fix:
Allowing for administrator users to get all partitions available on query.


646615-2 : Improved default storage size for DNS Express database

Component: Global Traffic Manager (DNS)

Symptoms:
A tweak has been made to the DNS Express database to improve the initial database size.

Conditions:
DNS Express with configured zones.

Impact:
Possibly reduced database size.

Workaround:
N/A as this is an improvement.

Fix:
A tweak has been made to the DNS Express database to improve the initial database size.


645615-6 : zxfrd may fail and restart after multiple failovers between blades in a chassis.

Solution Article: K70543226

Component: Global Traffic Manager (DNS)

Symptoms:
zxfrd may fail and restart after multiple failovers between blades in a single chassis.

Conditions:
DNS Express must be configured in a multi-blade chassis. If a blade transitions from active to backup to active states and the DNS Express (tmmdns.bin) database has been re-created while the blade was in backup status, zxfrd may fail when attempting to reference old data.

Impact:
zxfrd will create a core file and restart, picking up where it left off.

Workaround:
None.

Fix:
The cause of the failure is now addressed.


644822 : FastL4 virtual server with enabled loose-init option works differently with/without AFM provisioned

Solution Article: K19245372

Component: Advanced Firewall Manager

Symptoms:
If AFM provisioned, a FastL4 virtual server with enabled loose-init option drops all RST packets that do not relate to any existing flows.

This behavior does not match the BIG-IP behavior when AFM is not provisioned.

Conditions:
AFM provisioned.
-- FastL4 virtual server.
-- Loose-init option enabled.

Impact:
RST packets that do not relate to any existing flows are dropped, while they should not be dropped if the loose-init option enabled.

Workaround:
No workaround.

Fix:
Fixed, so FastL4 virtual servers with enabled loose-init option will forward any RST packets.


642923-6 : MCP misses its heartbeat (and is killed by sod) if there are a large number of file objects on the system

Solution Article: K01951295

Component: TMOS

Symptoms:
MCP may timeout and be killed by the sod watchdog, causing mcpd to restart.

Conditions:
Certain operations, under certain conditions, on certain platforms, may take longer to complete than the mcpd heartbeat timeout (300 seconds). When that happens, the system considers mcpd unresponsive, and will kill mcpd before it has finished its task, resulting in this issue.

There are a number of ways that this issue may manifest.

For example, the default mcpd heartbeat timeout might be reached when loading a configuration file with a large number* of file objects configured (e.g., SSL certificates and keys, data-groups, APM customizations, EPSEC file updates, external monitors, or other data present in the filestore (/config/filestore)).

*Note: Depending the operations mcpd is performing, the performance of the hardware, the speed of disk access, and other potential factors, 3,000 is a relative estimate of the number of filestore objects that might cause this issue to occur.

Impact:
mcpd restarts, which causes a system to go offline and restart services.

Workaround:
To prevent the issue from occurring, you can temporarily disable the heartbeat timeout using the following command:

   modify sys daemon-ha mcpd heartbeat disable

Important: Disabling the heartbeat timer means that, should the mcpd process legitimately become unresponsive, the system will not automatically restart mcpd to recover.

Note: If you have a large number of objects (more than 3,000) in the filestore, and are able to reduce this by deleting their related configuration objects, you may be able to work around the issue.

To determine the specific cause of the issue, you can open a support case with F5, to inspect the resulting mcpd core file.

Fix:
A possible case where mcpd goes too long without updating the heartbeat has been fixed by replacing one algorithm with a more efficient one.


642068-4 : PEM: Gx sessions will stay in marked_for_delete state if CCR-T timeout happens

Component: Policy Enforcement Manager

Symptoms:
PEM sessions stay in the marked-for-delete state if CCR-T times out.

Conditions:
This occurs if PCRF does not respond to CCR-T packets from the BIG-IP system during session termination.

Impact:
PEM sessions remain in the marked-for-delete state.

Workaround:
Configure the required timeout value in the sys db variable tmm.pem.session.timeout.endpointdeleteresponse.

Note: The value must be greater than 0 (zero).

Fix:
PEM sessions no longer stay in the marked-for-delete state if CCR-T times out.


641101-7 : httpd security and bug fix update CVE-2016-8743

Solution Article: K00373024


640766-2 : Linux kernel vulnerability: CVE-2016-10088 CVE-2016-9576

Solution Article: K05513373


638091-6 : Config sync after changing named pool members can cause mcpd on secondary blades to restart

Component: TMOS

Symptoms:
After performing a ConfigSync, mcpd restarts and the following error is seen in /var/log/ltm:

     01070734:3: Configuration error: Invalid mcpd context, folder not found <foldername>

Conditions:
- Chassis cluster with at least two blades
- sync-failover device group set to full-sync and auto-sync disabled
- Changing a named pool-member in non-default partition without syncing between delete and create

Impact:
Secondary blades do not process traffic as they restart

Workaround:
To prevent blade restart, follow the workaround in K16592: ConfigSync may fail when deleting and recreating a pool member with a node name set (https://support.f5.com/csp/article/K16592).

To work around this issue, you can synchronize the configuration just after deleting the pool member and node, before re-creating the pool member. To do so, perform the following procedure:

Impact of workaround: Performing the following procedure may impact client connectivity to the node. You should perform this procedure only during a maintenance window.

1. Log in to the BIG-IP system Configuration utility.
2. Navigate to Local Traffic :: Pools, and select the Pool with the member you want to delete.
3. From the top of the menu, click Members.
4. Select the checkbox next to the pool member you want to delete, and click Remove.
5. Navigate to Local Traffic :: Nodes.
6. Select the checkbox next to the node with the same name, and click Delete.
7. Navigate to Device Management :: Overview.
8. Select the local device by hostname (self).
9. Click the Sync option.
10. If the ConfigSync was successful, you may now re-create the pool member.

Fix:
Config sync after changing named pool members no longer causes mcpd on secondary blades to restart.


636997-1 : big3d may crash

Component: Global Traffic Manager (DNS)

Symptoms:
Under certain conditions, big3d may crash or malfunction while processing GTM data

Conditions:
GTM enabled

Impact:
iQuery connection may be reset

Workaround:
None

Fix:
big3d processes GTM data as expected


636994-1 : big3d may crash

Component: Global Traffic Manager (DNS)

Symptoms:
Under certain conditions, big3d may crash or malfunction while processing GTM data

Conditions:
GTM enabled

Impact:
iQuery connection may be reset

Workaround:
None

Fix:
big3d processes GTM data as expected


636992-1 : big3d may crash

Component: Global Traffic Manager (DNS)

Symptoms:
Under certain conditions, big3d may crash or malfunction while processing GTM data

Conditions:
GTM enabled

Impact:
iQuery connection may be reset

Workaround:
None

Fix:
big3d processes GTM data as expected


636986-1 : big3d may crash

Component: Global Traffic Manager (DNS)

Symptoms:
Under certain conditions, big3d may crash or malfunction while processing GTM data

Conditions:
GTM enabled

Impact:
iQuery connection may be reset

Workaround:
None

Fix:
big3d processes GTM data as expected


636982-1 : big3d may crash

Component: Global Traffic Manager (DNS)

Symptoms:
Under certain conditions, big3d may crash or malfunction while processing GTM data

Conditions:
GTM enabled

Impact:
iQuery connection may be reset

Workaround:
None

Fix:
big3d processes GTM data as expected


635509-1 : APM does not support Vmware'e Blast UDP

Component: Access Policy Manager

Symptoms:
APM does not support Blast Extreme Adaptive Transport (BEAT) protocol which is required for Blast UDP

Conditions:
1. Vmware View Connection Server is configured for Blast UDP
2. Client attempts Blast UDP

Impact:
Since APM does not support Blast UDP, Vmware Horizon Client always uses TCP transport even when the network conditions dictate that UDP transport would be more efficient

Workaround:
None

Fix:
APM now adds support for Blast Extreme Adaptive Transport protocol, which in turn enables Blast UDP.


635191-2 : Under rare circumstances TMM may crash

Component: Local Traffic Manager

Symptoms:
tmm crash and BIG-IP failover.

Conditions:
There are no known, reproducible conditions under which this occurs. However, the tmm restart happens once, and then does not recur. The only way to determine that the issue exists is through a review of the core stack, which must be completed by F5 Support.

Impact:
tmm restart and failover. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
The tmm restart and failover no longer occur.


633441-1 : Datasync Background Tasks running even without features requiring it

Component: TMOS

Symptoms:
The Datasync Background Tasks are running daily for several hours and consuming CPU. This is expected and required to generated dynamic versions of obfuscated JavaScript. However, this is running even if there are no features enabled which require JavaScript.

Conditions:
ASM is provisioned.

Impact:
Spikes of daily CPU usage during several hours even if there are no features requiring JavaScript.

Workaround:
If there are no features requiring JavaScript, then this command limits to a single version of obfuscated JavaScript, causing this CPU spike to remain a short one, and only once daily.

tmsh modify security datasync local-profile cs-asm-dosl7 max-gen-rows 1

Important: It is not recommended to keep this configuration if any of the JavaScript features are enabled in either ASM Policy or DoS profile, because it will significantly reduce the JavaScript security.

To re-enable full JavaScript obfuscation, run this command:

tmsh modify security datasync local-profile cs-asm-dosl7 max-gen-rows infinite

The log /var/log/datasync/datasyncd.log can be used to monitor the Background Tasks.

Fix:
The Datasync Background Tasks are now running only if there are features requiring JavaScript.


632875-5 : Non-Administrator TMSH users no longer allowed to run dig

Solution Article: K37442533


632646-1 : APM - OAM login with ObSSOCookie results in error page instead of redirecting to login page, when session cookie (ObSSOCookie) is deleted from OAM server.

Component: Access Policy Manager

Symptoms:
APM - OAM login with invalid ObSSOCookie results in error page instead of redirecting to login page.

Conditions:
This happens occasionally if a session cookie (ObSSOCookie) is deleted from OAM server, or an OAM session is deleted from server.

Impact:
OAM login with invalid ObSSOCookie results in error page. However, expected behavior is that user is redirected to login page if login with ObSSOCokkie fails.

Workaround:
No Workaround

Fix:
Issue is fixed - On authenticate with ObSSOCookie, read getStatus() API call to check the ObSSOCookie status and redirect to IDP if it is not 1 (LOGGEDIN, AWAITINGLOGIN). With this fix user will be redirected to IDP on logging with cookie that is deleted manually from the OAM server.


631418-1 : Packets dropped by HW grey list may not be counted toward AVR.

Component: Advanced Firewall Manager

Symptoms:
If the system supports hardware grey list, packets dropped by HW grey list may not be counted toward AVR.

Conditions:
AFM license, HW grey list support.

Impact:
User visibility.

Workaround:
There is no workaround at this time.

Fix:
The issue is fixed.


631316-2 : Unable to load config with client-SSL profile error

Solution Article: K62532020

Component: TMOS

Symptoms:
Config loading fails with an error similar to the following: 'Client SSL profile cannot contain more than one set of same certificate/key type.'

Conditions:
This occurs when both of the following conditions are met:
 -- The system is loading config.
 -- The config contains a client SSL profile which has an RSA cert-key-chain whose key is default (/Common/default.key), but whose chain is non-empty, or the cert is different from /Common/default.crt. For example:

    cert-key-chain {
        cert /Common/default.crt <==== default cert
        chain /Common/chainCA.crt <==== non-empty
        key /Common/default.key <==== default key
        rsa {
            cert /Common/default.crt <==== default cert
            chain /Common/chainCA.crt <==== non-empty
            key /Common/default.key <==== default key
        }
    }

Impact:
Configuration can not be loaded.

Workaround:
Remove or adjust the problematic client SSL profile by editing the appropriate bigip.conf file (/config/bigip.conf or /config/partitions/<name>/bigip.conf, depending on the partition the profile resides in), and then load the configuration again.

Steps:
1. Open the configuration file in a text editor.
2. Load the file /config/bigip.conf (or /config/partitions/<name>/bigip.conf, if the client SSL profile is in a partition).
   
3. Update the client SSL profile by setting .crt and .key to non-default, as shown in the following example:

    cert-key-chain {
        cert /Common/kc.crt <==== changed to non-default
        chain /Common/chainCA.crt
        key /Common/kc.key <==== changed to non-default
        rsa {
            cert /Common/kc.crt <==== changed to non-default
            chain /Common/chainCA.crt
            key /Common/kc.key <==== changed to non-default
        }
    }

4. Save your changes, and then run the following command:
 tmsh load sys conf


631286-3 : TMM Memory leak caused by APM URI cache entries

Component: Access Policy Manager

Symptoms:
Tmctl stats for "access_uri_info" gradually grows and can lead to TMM memory exhaustion.

Conditions:
APM or SWG in use.

Impact:
TMM memory exhaustion.

Workaround:
Restart tmm.

Fix:
This release implements a limit of how many entries the system stores in the URI cache. The default is 2048 entries. The DB variable allows a range of 2048 - 8192. You can the following DB variable to control the max limit:

access.max.euie_uri.cache.entries


630137-2 : Dynamic Signatures feature can fill up /config partition impacting system stability

Component: Advanced Firewall Manager

Symptoms:
When the AFM DoS Dynamic Signatures feature is enabled, inadequate file housekeeping results in the /config/filestore partition filling up. mcpd halts the other running daemons and the system becomes unresponsive.

Conditions:
AFM DoS Dynamic Signatures feature enabled

Configuration changes made but not saved

Device receives traffic.

Impact:
Traffic disrupted while TMM restarts.

Workaround:
Make all configuration changes via Configuration Tool (UI) or issue a 'save sys config partitions all' command.

If rolling back the configuration is a requirement, before making changes to the configuration, save a configuration snapshot to a file with the 'save sys config file <filename>' command. You can then load the previous configuration with a 'load sys config file <filename>' command.

Fix:
AFM DoS Dynamic Signatures file housekeeping improved, /config filestore no longer fills up.


629334-1 : Portal Access: JavaScript expressions in parentheses may be rewritten incorrectly

Component: Access Policy Manager

Symptoms:
In some cases Portal Access rewrites incorrectly JavaScript expressions enclosed into parentheses.

Conditions:
JavaScript code with the following constructions:
- (a.b) (...)
- (a[b]) (...)
- (b) = ...
Assuming 'b' is an element to be rewritten.
Some examples:
- (window.open) ("", "_blank");
- (form["submit"])();
- (location) = "http://some.org/";

Impact:
JavaScript code may not work correctly. In some cases, JavaScript code becomes syntactically incorrect.

Workaround:
Use iRule to remove parentheses around JavaScript expressions where necessary.

Fix:
Now JavaScript expressions in parentheses are rewritten correctly.


624231-4 : No flow control when using content-insertion with compression

Component: Policy Enforcement Manager

Symptoms:
Packets can get queued in PEM and cause performance impact.
It could cause memory corruptions in some cases

Conditions:
This issue can happen when system there is are a lot of connections with compression enabled, hardware offload is not enabled, and content insertion is enabled

Impact:
Performance impact to flows and possible system crash.

Workaround:
Enable hardware offload and use the pem throttle feature for content insertion


617643-2 : iControl.ForceSessions enabled results in GUI error on certain pages

Component: TMOS

Symptoms:
GUI pages display 'An error has occurred while trying to process your request.'

Conditions:
Visiting pages related to PKI (cert/key), SNMP, AFM or licensing tasks when iControl.ForceSessions is enabled.

Impact:
Unable to use GUI for certain tasks when iControl.ForceSessions is enabled.

Workaround:
Use shell for related administrative tasks or if feature is not used, disable with the following command:

tmsh# modify sys db icontrol.forcesessions value disable

Fix:
Enabled GUI to adapt when the iControl.ForceSessions is set to 'enable'.


616008-1 : TMM core may be seen when using an HSL format script for HSL reporting in PEM

Solution Article: K23164003

Component: Policy Enforcement Manager

Symptoms:
TMM core resulting in potential loss of service.

Conditions:
Requires a PEM HSL reporting action with an HSL format script against a virtual server.

Impact:
Traffic disrupted while tmm restarts.

Workaround:
In an iRule against the virtual server, set a Tcl variable in the first line after an iRule event. Unset the same Tcl variable in the last line of the iRule event.

Fix:
TMM core no longer occurs when using an HSL format script for HSL reporting in PEM.


612792-1 : Support RDP redirection for connections launched from APM Webtop on iOS

Component: Access Policy Manager

Symptoms:
Launching Native RDP resource from APM Webtop might fail on iOS.

Conditions:
1. Native RDP resource is launched from APM Webtop on iOS.
2. The RDP connection is redirected from one RDP server to another. This typically happens in RDP farm (multiple RDP servers) deployments.

Impact:
Native RDP resource can't be launched.

Workaround:
iOS RDP client version 8.1.35 allows workaround with following “Variable Assign” agent in Access Policy:
  Custom Variable:
    session.client.platform
  Custom Expression:
    set client_os [mcget {session.client.platform}];
    return [expr {$client_os == "iOS" ? "Android" : $client_os}];

Fix:
RDP redirection is now supported for connections launched from APM Webtop on iOS. Launching RDP resources from APM Webtop now requires at least version 8.1.35 of iOS RDP client.


612118-2 : Nexthop explicit proxy is not used for the very first connection to communicate with the backend.

Component: Access Policy Manager

Symptoms:
In SWG / forward proxy, nexthop explicit proxy is not used for the very first connection to communicate with the backend.

Conditions:
SWG per-request policy with proxy select agent.

Impact:
The BIG-IP system directly communicates with the backend to fetch server certificates.

Workaround:
None.

Fix:
Next-hop proxy gets used for all the connections that use proxy-select agent even for fetching the backend cert. In earlier version it would use the default route to fetch the certificate.

In transparent mode for https traffic, the proxy select agent is able to use the host & port information gathered from the backend certificate as the per-request policy can run before the cert fetching process. Therefore there is no longer a requirement for the per-request policy to have a category lookup agent before the proxy select agent.


608988-1 : Error when deleting multiple ASM Policies

Component: Application Security Manager

Symptoms:
Error when attempting to delete multiple ASM policies at once.

Conditions:
Multiple ASM policies are selected for deletion that have multiple XML profiles configured on their URLs.

Impact:
Operation fails with ASM subsystem error messages in asm log.

Workaround:
Delete policies one at a time.

Fix:
Multiple ASM policy delete finishes successfully.


606983-2 : ASM errors during policy import

Component: Application Security Manager

Symptoms:
Import failure when importing ASM policy with many Session Awareness Data Points.

ASM logs errors similar to the following:
-- crit g_server_rpc_handler_async.pl[10933]: 01310027:2: ASM subsystem error (asm_config_server.pl,F5::ASMConfig::Handler::log_error_and_rollback): Could not update the Export Policy Task 'Export Policy Task (1469519765.114479)'. DBD::mysql::db do failed: Got a packet bigger than 'max_allowed_packet' bytes.

Conditions:
-- ASM provisioned.
-- Session Awareness enabled.
-- Many (more than 1000) active Session Awareness 'Block All' data points are present.
-- Export a policy.
-- Import the same policy.

Impact:
asm_config_server crash occurs. asm_config_server recovers automatically, within ~15-30 seconds.

Workaround:
Either release all Session Awareness Data Points before export, or remove them from the exported policy before importing it back.

Fix:
Import failure no longer occurs when importing ASM policy with more than 1000 Session Awareness Data Points. Now, there is a maximum of 1000 Session Awareness Data Points exported into an XML policy export.


602708-4 : Traffic may not passthrough CoS by default

Component: Local Traffic Manager

Symptoms:
Traffic being forwarded by TMM may not passthrough the Class of Service (CoS) received.

Conditions:
-- IP forwarding Virtual server.
-- Traffic received with priority other than 3.

Impact:
Traffic is set to L2 QoS priority 3 and may cause issues on other networking devices.

Workaround:
Create a default CoS configuration or apply L2 QoS settings in the FastL4 profile.

Fix:
TMM now correctly passes through CoS by default.


594751-1 : LLDP VLAN Information not Transmitted to Neighbors When Interfaces are Added to a Trunk after the Trunk has Already Been Assigned to a VLAN

Solution Article: K90535529

Component: Local Traffic Manager

Symptoms:
Vlan Name and Vlan Tag values are not seen by the LLDP neighbors of the BIG-IP system.

Conditions:
1. LLDP is enabled globally and per interface.

2. Interfaces are added to a trunk after it has already been assigned to a VLAN.

For instance, assume the following protocol were followed for creating an LLDP trunk:

tmsh modify net lldp-globals enabled
tmsh modify net interface 1.1 lldp-admin txrx lldp-tlvmap 114680
tmsh modify net interface 1.2 lldp-admin txrx lldp-tlvmap 114680
tmsh create net trunk myTrunk
tmsh create net vlan myVlan
tmsh modify net trunk myTrunk interfaces add { 1.1 1.2 }
tmsh modify net vlan myVlan interfaces add { myTrunk }

The neighbor to this BIG-IP unit would not see the Vlan Name and Vlan Tag information of this trunk because the interfaces were added after the trunk was already assigned to a VLAN.

Impact:
LLDP Neighbors are unable to see VLAN information for the LLDP interfaces of the BIG-IP.

Workaround:
If the configuration has not yet been performed, the issue can be prevented by assigning all the desired interfaces to a trunk before assigning the trunk to a VLAN.

If the problem already exists, it can be remedied by performing a restart of the LLDP service. This should not impact dataplane services outside of LLDP. To do so, run the following command:
 bigstart restart lldpd

Fix:
VLANs are now properly applied to any interfaces added to a trunk if the trunk already belongs to any VLANs.


589233-2 : vCMPd may crash when processing bridged network traffic

Solution Article: K03165684


589083-6 : TMSH and iControl REST: When logged in as a remote user who has the admin role, cannot save config because of permission errors.

Component: TMOS

Symptoms:
When a remotely authenticated user who has the admin role uses TMSH or iControl to save the configuration, the operation fails because of permission errors.

Using iControl, the system posts an error similar to the following: Error processing request for URI:http://localhost:8110/mgmt/tm/sys/config
{code:400,message: Can't create tmsh temp directory \"/config/.config.backup\" Permission denied, errorStack:[]}.

Using TMSH (e.g., running the command: tmsh save sys config), the system posts an error similar to the following:

Can't create tmsh temp directory "/config/.config.backup" Permission denied

Conditions:
This occurs when the following conditions are met:
-- Remote Authentication is configured.
-- User is logged in as a remote user who has the admin role.
-- Using TMSH or iControl for remotely authenticated user operations.

Impact:
Cannot save the configuration.

Workaround:
Use one of the following workarounds:
-- Use the GUI to save the configuration.
-- Have a locally authenticated user with admin role save the configuration.

Fix:
When a remotely authenticated user who has the admin role uses TMSH or iControl to save the configuration, the operation now completes as expected, without permission errors.


581851-6 : mcpd process on secondary blades unexpectedly restarts when the system processes multiple tmsh commands

Solution Article: K16234725

Component: TMOS

Symptoms:
The Master Control Program Daemon (MCPD) on secondary blades may unexpectedly restart when the BIG-IP system processes multiple, concurrent TMOS Shell (tmsh) commands.

Under these circumstances, a race condition may occur and cause the mcpd process on the secondary blades to fail to correctly process concurrent updates from the primary blade.

As a result of this issue, you may encounter one or more of the following symptoms:

-- The mcpd process on secondary blades unexpectedly restarts.
-- You notice error messages in the /var/log/ltm file on the BIG-IP system that appears similar to the following example:
 + err mcpd[<PID>]: 01070823:3: Read Access Denied: The current update partition ([None]) does not match the object's partition (Common), stats not reset

 + err mcpd[<PID>]: 01070734:3: Configuration error: Configuration from primary failed validation: 01070823:3: Read Access Denied: The current update partition ([None]) does not match the object's partition (Common), stats not reset

-- Depending on your high availability (HA) configuration, the device may unexpectedly fail over to another system in the device group.

Conditions:
This issue occurs when all of the following conditions are met:

-- You have a VIPRION platform or Virtual Clustered Multiprocessing (vCMP) guest configuration that uses two or more blades.
-- You attempt to run multiple, concurrent tmsh commands on the BIG-IP system. For example, you run a tmsh command to continually reset persistence records and at the same time run another tmsh command to continually reset the TCP statistics.

Impact:
The BIG-IP system may experience performance degradation when the secondary blades become unavailable while the mcpd process restarts. Depending on your HA configuration, the device may fail over.

Workaround:
None.

Fix:
This issue no longer occurs.


580537-3 : The GeoIP update script geoip_update_data cannot be used to install City2 GeoIP data

Component: Global Traffic Manager (DNS)

Symptoms:
The geoip_update_data script will not install a City2 GeoIP data.

Conditions:
Attempting to install the City2 GeoIP data.

Impact:
The City2 GeoIP data must be installed manually.

Workaround:
The City2 GeoIP data can be installed manually by extracting the contents of the RPM and updating the associated files. The commands are:

rpm -U <full path to the City2 GeoIP RPM file>
rm /shared/GeoIP/F5GeoIP.dat
ln -s /shared/GeoIP/F5GeoIPCity2.dat /shared/GeoIP/F5GeoIP.dat
rm /shared/GeoIP/v2/F5GeoIP.dat

Fix:
The geoip_update_data script was updated to support installing City2 GeoIP data.


571651-4 : Reset Nitrox3 crypto accelerator queue if it becomes stuck.

Solution Article: K66544028

Component: Local Traffic Manager

Symptoms:
Certain configuration parameters used during an SSL handshake can elicit a 'queue stuck' message from the accelerator. When this happens, the /var/log/ltm log file will contain a message similar to the following:

    'n3-cryptoX request queue stuck'.

Conditions:
The BIG-IP system uses Nitrox 3 encryption hardware to perform SSL encryption.

An SSL handshake sent to the BIG-IP system is incorrectly configured or contains bad information.

Impact:
In-flight and queued contexts for the specific accelerator are dropped. After device recovery, new requests will be accelerated.

Workaround:
Disable crypto acceleration.

Fix:
The crypto accelerator is gracefully reset when the accelerator stalls due to a misconfigured request.

Note: This issue resolution is limited to SSL handshake issues, and is not a resolution for all possible causes of a 'queue stuck' event.


563661-1 : Datastor may crash

Component: TMOS

Symptoms:
In rare cases, datastor may crash to protect the system from corruption. After datastor restarts, tmm may crash when attempting to communicate with datastor.

Conditions:
WAM provisioned and enabled.

Impact:
Datastor and TMM crashes. Traffic disrupted while tmm restarts.

Workaround:
None.

Fix:
This issue has been fixed.


562921-5 : Cipher 3DES and iQuery encrypting traffic between BIG-IP systems

Component: Global Traffic Manager (DNS)

Symptoms:
BIG-IP systems use the iQuery protocol to securely communicate with other BIG-IP systems. The BIG-IP system supports the AES/3DES ciphers for encrypting iQuery traffic. Some of these ciphers are now considered unsecure.

Conditions:
The value is hardcoded into the product.

Note: This is completely independent of the TMM profiles or the httpd cipher values.

Impact:
There is no way to configure this; the value is hardcoded. Scanner operations performed on your configuration will report this as an unsecure cipher.

Workaround:
If you do not need iQuery at all, you can block port 4353 completely. For those who do need it, there is no workaround.

Fix:
The cipher list in use is now
"AESGCM:AES:!ADH:!AECDH:!PSK:!aECDH:!DSS:!ECDSA:!AES128:-SHA1:AES256-SHA"


551925-4 : Misdirected UDP traffic with hardware acceleration

Component: TMOS

Symptoms:
UDP traffic might be forwarded to the wrong destination when using hardware acceleration.

Conditions:
If the UDP timeout is lower than the embedded Packet Velocity Acceleration (ePVA) aging timeout.

This occurs because UDP connections are accelerated until the ePVA aging timeout expires for the connection. If the ePVA aging timeout is greater than the UDP timeout, then TMM removes the connection from software, but the connection is still accelerated in the ePVA. Subsequent traffic then matches to the original connection, causing it to be sent to the wrong destination.

Impact:
Traffic can be sent to the wrong destination.

Workaround:
You can use either or both of the following workarounds:
-- Increase the UDP timeout (60s or more).
-- Disable UDP hardware acceleration.


536831-1 : APM PAM module does not handle local-only users list correctly

Component: Access Policy Manager

Symptoms:
The following log messages are shown in /var/log/secure, when remote-auth (APM based) is configured and when trying to authenticate local users:

-- notice httpd[8281]: pam_apm: module returning Failure, ClientHandler auth failed!(admin)
-- notice httpd[8281]: 01070417:5: AUDIT - user admin - RAW: httpd(mod_auth_pam): user=admin(admin) partition=[All] level=Administrator tty=/sbin/nologin host=X.X.X.X attempts=1 start="Wed Jan 17 14:49:21 2018"

This failure log shows that the system first attempts to authenticate local users (like admin, root, etc.) remotely.

Conditions:
This occurs when following conditions are met:
- APM is provisioned on a BIG-IP system.
- APM-based remote-auth is configured.
- Local users (like admin, root, etc.) attempt to log into the management interface of that BIG-IP system.

Impact:
Local users credentials are sent to remote authentication servers which will return auth failure. However, in the second attempt, the system attempts to authenticate a user locally, and it will succeed, as expected. Check below logs:

-- notice httpd[8281]: pam_apm: module returning Failure, ClientHandler auth failed!(admin)
-- notice httpd[8281]: 01070417:5: AUDIT - user admin - RAW: httpd(mod_auth_pam): user=admin(admin) partition=[All] level=Administrator tty=/sbin/nologin host=X.X.X.X attempts=1 start="Wed Jan 17 14:49:21 2018"

Workaround:
None.

Fix:
Local users are authenticated locally. The system no longer sends request to remote servers for local users.


514703-3 : gtm listener cannot be listed across partitions

Component: TMOS

Symptoms:
Unable to reference (perform operations: list, create, modify ...) gtm listeners across partitions.

Conditions:
-- In one partition.
-- Listener in another partition.
-- Attempt to perform operations on the listener in the other partition.

For example, the current partition is /Common, and a listener exists in /DifferentPartition, and you try to perform operations on the listener under /DifferentPartition.

Impact:
Cannot perform any operations on that listener. The listener will be listed as non-existent.

Workaround:
Change to the partition where the listener exists before performing any operations on it.

Fix:
The system can now reference GTM listeners across partitions.


495443-9 : ECDH negotiation failures logged as critical errors.

Solution Article: K16621

Component: Local Traffic Manager

Symptoms:
When a failure occurs in an SSL negotiation involving Elliptic Curve Diffie-Hellman (ECDH) key agreement, a critical error may be logged. However, an SSL negotiation failure is not a critical issue.

Conditions:
An SSL negotiation failure involving ECDH key agreement.

Impact:
Spurious critical error logs.

Workaround:
Treat SSL ECDH negotiation failures as non-critical errors.

Fix:
These ECDH failures are now logged as non-critical errors.


495242-4 : mcpd log messages: Failed to unpublish LOIPC object

Component: Local Traffic Manager

Symptoms:
The system posts the following error: err mcpd[7143]: 010716d6:3: Failed to unpublish LOIPC object for (loipc_name.1417443578.297505208). Call to (shm_unlink) failed with errno (2) errstr (No such file or directory).

Conditions:
This is an intermittent issue that occurs on standby systems in High Availability (HA) configurations. In this case, the system is attempting to remove a file/directory that does not exist. Either the file has already been removed or it was not created.

Impact:
This is a benign error that can be safely ignored.

Workaround:
None.

Fix:
The system now suppresses logging when attempting to delete non-existent file.


471237-4 : BIG-IP VE instances do not work with an encrypted disk in AWS.

Solution Article: K12155235

Component: TMOS

Symptoms:
BIG-IP VE instances cannot use encrypted disks in AWS as encryption-decryption introduces some data corruption in the disk that causes failure in some of the TMOS daemons at run-time.

Conditions:
Deploy a BIG-IP VE guests in AWS with an encrypted disk.

Impact:
TMM cores at startup, and does not start.

Workaround:
Do not use encrypted disks in AWS for BIG-IP VE instances.

Fix:
BIG-IP VE instances can now work with an encrypted disk in AWS.


464650-6 : Failure of mcpd with invalid authentication context.

Component: TMOS

Symptoms:
MCPd cores.

Conditions:
It is not known what triggers this core.

Impact:
Mcpd restarts

Workaround:
None.

Fix:
Failure of mcpd with invalid authentication context no longer occurs.


463097-5 : Clock advanced messages with large amount of data maintained in DNS Express zones

Component: Local Traffic Manager

Symptoms:
Clock advanced messages with a large amount of data maintained in DNS Express (DNSX) zones, the TMM can suffer from clock advances when performing the DB reload.

Conditions:
Large enough zones into DNSX (several hundred thousand to several million records depending on hardware).

Impact:
Clock advance messages in log. No traffic can be passed for this duration. When DNS Express zones are updated, you may see messages similar to the following in the /var/log/ltm log: notice tmm[25454]: 01010029:5: Clock advanced by 121 ticks.

Workaround:
Prevent all updates to DNSX zones.

Fix:
AXFR and IXFR to DNS Express (DNSX) with large zones has been significantly improved. DNSX DB now reside in /shared to resolve DB size issues.


452283-5 : An MPTCP connection that receives an MP_FASTCLOSE might not clean up its flows

Component: Local Traffic Manager

Symptoms:
An MPTCP connection that never expires can be seen using the command "tmsh show sys conn". Its idle time periodically resets to 0.

Conditions:
A virtual server is configured with a TCP profile with "Multipath TCP" enabled.
BIG-IP receives an MP_FASTCLOSE while the BIG-IP is advertising a zero window.

Impact:
A connection remains that never expires; its idle time periodically resets to 0.

Workaround:
There is no workaround at this time.

Fix:
Fixed MP_FASTCLOSE handling.


440620-1 : New connections may be reset when a client reuses the same port as it used for a recently closed connection

Component: Local Traffic Manager

Symptoms:
If a client reuses the same port that it used for a recently closed connection, the new connection may receive a RST in response to the client's SYN.

Conditions:
A client reuses the same port that it used for a recently closed connection. The 4-tuple of local address, local port, remote address, and remote port must be the same to trigger this issue.

Impact:
New connections reusing a 4-tuple may be reset for a brief period following a connection close.

Workaround:
Lowering the "Close Wait" and "Fin Wait 1" timeouts in the TCP profile will shorten the amount of time that a particular 4-tuple remains unusable.

Fix:
Improved abort handling to better clean up hanging connections.


251162-1 : The error message 'HTTP header exceeded maximum allowed size' may list the wrong profile name

Solution Article: K11564

Component: Local Traffic Manager

Symptoms:
If you apply a custom HTTP profile to a virtual server, and the maximum header size defined in the profile is exceeded, the BIG-IP system lists the wrong profile name in the corresponding log message. Instead of logging the profile name associated with the virtual server, the BIG-IP system logs the profile name as http.

For example:

tmm1[5133]: 011f0005:3: HTTP header (34083) exceeded maximum allowed size of 32768 (Client side: vip=http_10.1.0.30 profile=http pool=apache2)

Conditions:
-- You apply a custom HTTP profile to a virtual server.
-- The maximum header size defined in the profile is exceeded.

Impact:
The BIG-IP system lists the wrong profile name in the corresponding log message. This is a cosmetic error, as the correct profile is affected. Only its name is incorrectly reported.

Workaround:
None.



Known Issues in BIG-IP v13.1.x


TMOS Issues

ID Number Severity Solution Article(s) Description
698085 1-Blocking   Transparent mode VLAN group may not work on vCMP guests
682837-2 1-Blocking   Compression watchdog period too brief.
750586-1 2-Critical   HSL may incorrectly handle pending TCP connections with elongated handshake time.
749388 2-Critical   'table delete' iRule command can cause TMM to crash
748205-1 2-Critical   SSD bay identification incorrect for RAID drive replacement
746464-3 2-Critical   MCPD sync errors and restart after multiple modifications to file object in chassis
743790-3 2-Critical   BIG-IP system should trigger a HA failover event when expected HSB device is missing from PCI bus
743271-3 2-Critical   Querying vCMP Health Status May Show Stale Statistics
742419-1 2-Critical   BIG-IP NIC teaming of SR-IOV interfaces does not work under VMware ESXi
741423-2 2-Critical   Secondary blade goes offline when provisioning ASM/FPS on already established config-sync
737692-1 2-Critical   Handle x520 PF DOWN/UP sequence automatically by VE
737055-2 2-Critical   Unable to access BIG-IP Configuration Utility when BIG-IP system is behind a Reverse proxy
734539-3 2-Critical   The IKEv1 racoon daemon can crash on multiple INVALID-SPI payloads
726487-2 2-Critical   VIPRION secondary MCPD, Node Name encodes IP address which differs from supplied
724556-2 2-Critical   icrd_child spawns more than maximum allowed times (zombie processes)
714281-2 2-Critical   NSH tunnel reject inner packet from other vendor
713380 2-Critical K23331143 Multiple B4450 blades in the same chassis run into inconsistent DAG state
710277-1 2-Critical   IKEv2 further child_sa validity checks
708968-2 2-Critical   OSPFv3 failure to create a route entry for IPv4-Mapped IPv6 Address
707013 2-Critical   vCMP host secondary member's cluster.conf file may replaced by that of vCMP guest
705730-1 2-Critical   Config fails to load due to invalid SSL cipher after upgrade from v13.1.0
699515-1 2-Critical   nsm cores during update of nexthop for ECMP recursive route
697424-1 2-Critical   iControl-REST crashes on /example for firewall address-lists
693206 2-Critical   iSeries LCD screen is frozen on a red spinning 'please wait' indicator
688148-3 2-Critical   IKEv1 racoon daemon SEGV during phase-two SA list iteration
680556-1 2-Critical   Under unknown circumstances sometimes a sessionDB subkey entry becomes corrupted
677937-3 2-Critical   APM tunnel and IPsec over IPsec tunnel rejects isession-SYN connect packets
671741-3 2-Critical   LCD on iSeries devices can lock at red 'loading' screen.
668041-2 2-Critical K27535157 Config load fails when an iRule comment ends with backslash in a config where there is also a policy.
621260-4 2-Critical   mcpd core on iControl REST reference to non-existing pool
608511-1 2-Critical   Message router profile is not inheriting the traffic-group from the parent folder
593536-6 2-Critical   Device Group with incremental ConfigSync enabled can report "In Sync" when devices have differing configurations
751409-3 3-Major   MCP Validation does not detect when virtual servers differ only by overlapping VLANs
751024-2 3-Major   i5000/i7000/i10000 platforms: SFP/QSFP I2C problems may not be cleared by bcm56xxd
751021-3 3-Major   One or more TMM instances may be left without dynamic routes.
751011-1 3-Major   ihealth.sh script and qkview locking mechanism not working
751009-1 3-Major   Generating Qkviews or tcpdumps via GUI or running the 'ihealth' command removes /var/tmp/mcpd.out
750447-1 3-Major   GUI VLAN list page loading slowly with 50 records per screen
750318-1 3-Major   HTTPS monitor does not appear to be using cert from server SSL profile
749785-1 3-Major   nsm can become unresponsive when processing recursive routes
748206 3-Major   Browser becomes unresponsive when loading the network map with a virtual server that contains a forwarding rule policy in the second position
748187-2 3-Major   'Transaction Not Found' Error on PATCH after Transaction has been Created
747799-2 3-Major   'Unable to load the certificate file' error and configuration-load failure upgrading from 11.5.4-HF2 to a later version, due to empty cert/key in client SSL profile
747676-1 3-Major   Remote logging needs 'localip' to set source IP properly
746758 3-Major   Qkview produces core file if interrupted while exiting
746657-3 3-Major   tmsh help for FQDN node or pool member shows incorrect default for fqdn interval
746650 3-Major   Stale packets in HSB transmit queue causes HSB DMA lockup
746333 3-Major   Setting the hostname to non-FQDN value prevents upgrade
746266-1 3-Major   Vcmp guest vlan mac mismatch across blades.
745825-3 3-Major   The "audit_forwarder is disabled as the configuration is incomplete" message can be confusing if logged when the configuration is loading
744936 3-Major   Adding a default tmm gateway in AWS breaks failover between two instances if the default tmm gateway can't provide route to the ec2 metadata service at 169.254.169.254.
744520-3 3-Major   virtual server with perm profile drops traffic received from Vxlan-GRE tunnel interface
744252-3 3-Major   BGP route map community value: either component cannot be set to 65535
743803-2 3-Major   IKEv2 potential double free of object when async request queueing fails
743132-4 3-Major   mcpd might restart on secondary blades after modify sys httpd ssl-certchainfile
742753-2 3-Major   Accessing the BIG-IP system's WebUI via special proxy solutions may fail
741902-3 3-Major   sod does not validate message length vs. received packet length
740589-3 3-Major   mcpd crash with core after 'tmsh edit /sys syslog-all-properties'
740517-3 3-Major   Application Editor users are unable to edit HTTPS Monitors via the Web UI
740413-3 3-Major   sod not logging Failover Condition messages
740345-1 3-Major   TMM core files seen on standby device after failover, when connection mirroring, session mirroring and OCSP stapling are enabled.
740280-1 3-Major   Configuration Utility and tmsh may not validate Certificate Authority profile names
740135-3 3-Major   Traffic Group ha-order list does not load correctly after reset to default configuration
739872-2 3-Major   The 'load sys config verify' command can cause HA Group scores to be updated, possibly triggering a failover
739533-4 3-Major   In rare circumstances, config sync may fail to delete files in /config/filestore/.snapshots_d/, filling up /config
739400 3-Major   iControl REST fails to list virtual servers
739118-3 3-Major   Manually modifying a self IP address in bigip_base.conf file and reloading the configuration results in routing misconfiguration
738445-2 3-Major   IKEv1 handles INVALID-SPI incorrectly in parsing SPI and in phase 2 lookup
737901-2 3-Major   Management MAC address and host VLAN MAC address is the same on iSeries platforms when in vCMP mode
737536-1 3-Major   Enabling 'default-information originate' on one of the several OSPF processes does not inject a default route into others.
737437-2 3-Major   IKEv1: The 4-byte 'Non-ESP Marker' may be missing in some IKE messages
737397-3 3-Major   User with Certificate Manager role is unable to archive certificates using GUI and iControlSOAP
737346-3 3-Major   After entering username and before password, the logging on user's failure count is incremented.
735565-1 3-Major   BGP neighbor peer-group config element not persisting
734846-3 3-Major   Redirection to logon summary page does not occur after session timeout
734836-3 3-Major   Network Map summary counts pool members more than once if they are shared across pools
725985-1 3-Major   REST API takes more than 20 seconds when 1000+ virtual servers with SNAT-Pool configured
725791-4 3-Major   Potential HW/HSB issue detected
724889 3-Major   BIG-IP VE in AWS does not failover NATs in same availability zone
724706-1 3-Major   iControl REST statistics request causes CPU spike
724143-1 3-Major   IKEv2 connflow expiration upon ike-peer change
723794-3 3-Major   PTI (Meltdown) mitigation should be disabled on AMD-based platforms
723579-4 3-Major   OSPF routes missing
722691 3-Major   Available datagroup list does not contain datagroups with the correct type.
722380-2 3-Major   The BIG-IP system reboots while TMM is still writing a core file, thus producing a truncated core.
721806 3-Major   Traffic Policy edit to datagroup errors on adding ASM disable action
721805 3-Major   Traffic Policy edit to datagroup errors on adding ASM disable action
721020-3 3-Major   Changes to the master key are reverted after full sync
721016 3-Major   vcmpd fails updating VLAN information on vcmp guest
720819-2 3-Major   Certain platforms may take longer than expected to detect and recover from HSB lock-ups
720610 3-Major   Updatecheck logs bogus 'Update Server unavailable' on every run
720269-2 3-Major   TACACS audit logging may append garbage characters to the end of log strings
720110-2 3-Major   0.0.0.0/0 NLRI not sent by BIG-IP when BGP peer resets the session.
718817-2 3-Major   Software installation on secondary blade of VIPRION or VIPRION-based vCMP guest can fail.
718800-2 3-Major   Cannot set a password to the current value of its encrypted password
718405-1 3-Major   RSA signature PAYLOAD_AUTH mismatch with certificates
718397-1 3-Major   IKEv2: racoon2 appends spurious trailing null byte to ID payloads
716166-4 3-Major   Dynamic routing not added when conflicting self IPs exist
715379-1 3-Major   IKEv2 accepts asn1dn for peers-id only as file path of certificate file
715115 3-Major   Application Security roles are not showing all accessible objects in GUI
715061-2 3-Major   vCMP: tmm core in guest when stopping vCMP guest from host
714986-3 3-Major   Serial Console Baud Rate Loses Modified Values with New Logins on iSeries Platforms Without Reboot
714626-2 3-Major   When licensing through a proxy, setting the db variables for proxy.host, proxy.port, etc., has no effect.
714216-2 3-Major   Folder in a partition may result in load sys config error
713708 3-Major   Update Check for EPSEC shows OPSWAT description without EPSEC version on GUI
712033-2 3-Major   When making a REST request to an object in /stats that is an association list, the selfLink has a duplicate name
710666-1 3-Major   VE with interface(s) marked down may report high cpu usage
710173 3-Major   TMSH dns-resolver allows route-domain from another partition
709559-2 3-Major   LTM v12.1.2 Upgrade fails if config contains "/Common/ssh" object name
709544-2 3-Major   VCMP guests in HA configuration become Active/Active during upgrade
707320-2 3-Major   Upgrades from pre-12.0.0 BIG-IPs to 12.0.0 with WideIPs with ipv6-no-error-response enabled will no longer delete AAAA-type WideIPs
706104-3 3-Major   Dynamically advertised route may flap
705651-1 3-Major   Async transaction may ignore polling requests
705442-1 3-Major   GUI Network Map objects search on Virtual Server IP Address and Port does not work
705037-2 3-Major K32332000 System may exhibit duplicate if_index, which in some cases lead to nsm daemon restart
704546 3-Major   Symlinks may be corrupted by upgrade
704449-2 3-Major   Orphaned tmsh processes might eventually lead to an out-of-memory condition
703090 3-Major   With many iApps configured, scriptd may fail to start
702310-1 3-Major   The ':l' and ':h' options are not available on the tmm interface in tcpdump
701898-1 3-Major   Certain virtual address route-advertisement settings break upgrades from 13.0.0 hotfix rollups
701722-1 3-Major   Potential mcpd memory leak for signed iRules
701529-1 3-Major   Configuration may not load or not accept vlan or tunnel names as "default" or "all"
701341-1 3-Major K52941103 If /config/BigDB.dat is empty, mcpd continuously restarts
700897-1 3-Major   sod is unable to handle the maximum (127) allowable traffic groups if there are 8 devices in the DG
700827-4 3-Major   B2250 blades may lose efficiency when source ports form an arithmetic sequence.
698947-2 3-Major   BIG-IP may incorrectly drop packets from a GRE tunnel with auto-lasthop disabled.
698933-1 3-Major   Setting metr