Applies To:

Show Versions Show Versions

Manual Chapter: Basic Troubleshooting Tools
Manual Chapter
Table of Contents   |   << Previous Chapter   |   Next Chapter >>

38 
A log component is a source of syslog messages, typically an internal process or group of processes. The table below is an alphabetical list of all log components, with a brief description and an indication of the board(s) where the software runs.
You can set up auto-diagnostics to regularly gather system information and send it to F5 Support. F5 can use this information to monitor your ARX from off site, watching for gradual loss of service or resource degradation. The information sent to F5 is the output from a series of CLI show commands. You can use the additional-command operation to add another show command to this collection of diagnostics.
Use no additional-command to remove an extra show command from auto-diagnostics.
additional-command "show-command-string"
no additional-command "show-command-string"
show-command-string (1-128 characters) is a valid show command in the ARX CLI, such as show schedule. Surround the string with quotation marks (as shown) if it contains any spaces.
This command adds an additional show command to the following list, which is always included in the auto-diags collection:
show policy details
The no form of this command cannot remove any of these default show commands.
In a redundant pair of ARX machines (see redundancy), both peers collect and send auto-diagnostics information through E-mail. All of the commands in gbl mode are shared between the peers, so only the active peer collects gbl-mode information (such as show global-config and show namespace all). Both peers collect and send commands related to cfg mode (such as show running-config and show ron). This also applies to any commands you add to the collection with additional-command; if the command relates to gbl mode, it is only included in collections from the currently-active peer. If the command relates to cfg mode, it is included in auto-diagnostic collections from both peers.
prtlndA(gbl-auto-diag)# additional-command "show channel summary"
Use the no auto-diagnostics command to disable auto-diagnostics and remote monitoring.
This command places you in gbl-auto-diag mode, where you must use the schedule (gbl-auto-diag) command to assign a schedule for collecting and sending diagnostic information. You can optionally use the mail-to (gbl-auto-diag) command to add additional E-mail recipients; the attachment is only encrypted for the E-mail to F5 Support. You can also use additional-command to add more show commands to the E-mail. To test the configuration, you can use the auto-diagnostics test command in priv-exec mode. The show auto-diagnostics command shows the current configuration for this utility, along with the status of the most-recent auto-diagnostics collection.
In a redundant pair of ARX machines (see redundancy), both peers collect and send auto-diagnostics information through E-mail. All of the commands in gbl mode are shared between the peers, so only the active peer collects gbl-mode information (such as show global-config and show namespace all). Both peers collect and send commands related to cfg mode (such as show running-config and show ron). This also applies to any commands you add to the collection with additional-command; if the command relates to gbl mode, it is only included in collections from the currently-active peer. If the command relates to cfg mode, it is included in auto-diagnostic collections from both peers.
To address an immediate issue, use the collect command to gather a larger amount of diagnostic information and possibly send it to F5 Support.
The process always includes the following show commands in its E-mail attachment:
show policy details
bstnA(gbl)# auto-diagnostics
local (optional) sends the E-mail to local recipients only, defined by the mail-to (gbl-auto-diag) command. This does not send any E-mail to F5.
This command gathers auto diagnostics (as determined by the auto-diagnostics command and its sub-mode commands), compresses them into a Unix Tape Archive (tar) file, and sends the file through smtp as an E-mail attachment. This tests both the auto-diagnostics configuration and the smtp configuration.
By default, the E-mail goes to F5 Support and all configured mail-to (gbl-auto-diag) addresses. The E-mail attachment for F5 Support is encrypted, as it must travel outside your local network. The E-mail attachment is un encrypted for each of the local mail-to (gbl-auto-diag) addresses, for convenience. You can use the local option to send the test message only to the local addresses.
bstnA# auto-diagnostics test local
trap (1-64 characters) identifies the trap(s) to be cleared. Use show health to see a list of traps with open alarm conditions. Use the Event name from the output of show health.
The show health command shows all currently-raised traps on the system. For cases where a raised trap is unlikely to every be cleared by its corresponding clear trap, you can use this command to manually clear the trap. This clears all instances of the given trap. Then the alarm condition no longer appears in the show health output.
bstnA# clear health diskFail
Some machine problems require more diagnosis than is possible on-site. Use the collect command to collect diagnostic data on the switch and either upload it or store it locally.
all collects all diagnostic files on the switch.
diag-info collects all diagnostic files on the switch except reports, which can be excessively large.
config collects the configuration database, database logs, and the output from show running-config and show global-config.
cores gathers any core-dump files on the system, along with all shared-object libraries used by the system.
logs assembles the syslog, the procdat log (process data; a more-detailed log file for diagnosis by engineers), upgrade and installation logs, disk-usage logs, Kerberos state files, and logs from the systems service manager. For options to collect logs from specific times, see the documentation for collect logs.
stats-logs selects all stats-log files, .csv files with performance data relevant to the ARX, its filers and servers, and other external devices. You can use the show stats-logs command to list all of the stats-log files currently available.
reports chooses all reports.
state collects individual log files from every processor in the switch, three system log files (syslog, procdat, and traplog), a chronological list of all CLI commands used since the switch booted, the latest of each of the stats-log files, and the output from a series of useful show commands.
capture selects only the packet-capture files, if there are any. These files show the flow of IP traffic through the ARX during a capture session. You can use show capture to list all of the capture files currently available.
ftp://[user[:password]@] server/file.tgz (1-1024 characters) is the destination URL:
user[:password]@ (optional) is the username and password (for example, random:pw@). If you omit the password, the CLI prompts you for one.
server is the machine name (for example, mymachine.myco.com).
file.tgz is the desired path name for the file. Lead with an extra slash (/) if the path is absolute: for example, ...myserver//var/log/acopiaDiags.tgz specifies /var/log/acopiaDiags.tgz on myserver. Use only a single slash if the path is local to the users home directory.
async (optional) makes this command return immediately, rather than waiting for the collection process to finish. The CLI generates a report that you can use to chart the collection process; the report name appears after the process starts.
collect {all | diag-info | ... | capture}
scp://[
user@]server:file.tgz [accept-host-key] [async]
{all | ... | capture} define the collection process, as explained above.
scp://[user@]server:file.tgz (1-1024 characters) is the URL for the destination file:
user@ (optional if someone created an ip scp-user) is the username to present to the other end of the SCP connection. This user must be valid at the remote host. If you omit this and an ip scp-user is defined, it defaults to the username set by that command.
server: is the IP address or hostname for the SCP host. End with a colon (:).
file.tgz is the destination-file path. Lead with a slash (/) if the path is absolute (for example, scp://root@10.1.1.5:/var/diags/arx.tgz). Use no slash if the path is local to the home directory for user (for example, scp://root@10.1.1.5:diags.tgz).
accept-host-key (optional) indicates that if the other end of the connection has an unknown SSH host key (that is, if it is new, or if its key has changed since the last time the host was contacted), the ARX should accept the new host key and continue with the upload. Otherwise, the ARX stops the upload if the host presents an unknown key.
async (optional) was explained above.
{all | ... | capture} are explained above.
nfs | cifs is a required choice. This chooses the protocol for the file transfer.
namespace (1-30 characters) identifies the namespace to hold the collect file.
path (1-1024 characters) is the volume name.
file.tgz is the path to the collect file, starting at the root of the volume.
async (optional) was explained above.
{all | ... | capture} are explained above.
tftp://server/file.tgz (1-1024 characters) is the URL for the collect file:
server is the machine name (for example, mymachine.myco.com).
file.tgz is the desired path name for the file. Lead with an extra slash (/) if the path is absolute: for example, ...myserver//var/log/acopiaDiags.tgz specifies /var/log/acopiaDiags.tgz on myserver. Use only one slash if the path is local to the servers tftpboot directory. This conforms with the specification for FTP URLs in RFC 1738.
async (optional) was explained above.
{all | ... | capture} are explained above.
smtp://[e-mail-address/]file.tgz (1-1024 characters) is an E-mail destination for the collect file:
smtp:// declares that the destination is an E-mail address.
e-mail-address (optional) is the recipient of the E-mail in username@host format (for example, jsmith@myco.com). If you omit this, the CLI uses the default address set by the cfg-smtp to command.
file.tgz is the name of the copy. This is sent as an attachment to the outbound E-mail message. This is not a path, so do not use slashes (/). You must use a .tgz extension. For example, ...acopiaDiags.tgz is a valid file name.
async (optional) was explained above.
{all | ... | capture} are described above.
local-file.tgz (1-1024 characters) saves the file locally, to the diag-info directory. To preserve disk space, you can only save one diag-info file at a time. You must use a .tgz extension.
async (optional) was explained above.
(FTP only) If you omit the user:password from the FTP URL, the CLI uses the username and password set by the ip ftp-user command.
For SCP uploads, the CLI prompts for a password; enter the password for the user that you identified in the URL. The CLI also prompts for a password for FTP uploads where you do not supply one in the command.
The CLI prompts for confirmation before collecting the diagnostics. Answer yes to continue. The collection process can be time-consuming; prompts appear to show the progress of the operation as it occurs.
bstnA# collect diag-info ftp://jpublic@arxftp.f5.com/10-24-03.tgz
Password: jpassword
bstnA# collect diag-info cifs medarcv /rcrds admin/collect.tgz
bstnA# collect config ftp://arxftp.f5.com/10-24-03.tgz
bstnA# collect diag-info investigate.tgz
prtlndA# collect logs smtp://logsAndReports.tgz
collect logs [start-time start [utc]] [end-time end [utc]] destination [async]
start-time start [utc] (optional) sets a starting time for collected logs. If you omit this, the CLI collects all of its logs up to the end-time.
start is a time stamp in one of two formats:
mm/dd/yyyy:HH:MM:SS is designed for manual entry of a date and time.
yyyy-mm-ddTHH:MM:SS is designed for a copy-and-paste of a time stamp from a log file. Time stamps from log files are in UTC instead of local time.
utc (optional) declares that the start is in UTC instead of local time. This is typically used for time stamps copied from a log file (the second format from above).
end-time end [utc] (optional) sets an end time for collected logs. If you omit this, the CLI collects all of its logs after the start-time. The details of this time stamp are otherwise the same as above.
destination is a local-file name, an ARX volume, or a URL. The URL can use FTP, SCP, TFTP, or SMTP. The documentation for the collect command describes the exact syntax.
async (optional) makes this command return immediately, rather than waiting for the collection process to finish. The CLI generates a report that you can use to chart the collection process; the report name appears after the process starts.
This is an alternative syntax for the collect command. It is useful for collecting logs from a specific time frame; this can reduce the size of the collection as well as focus a diagnosis.
bstnA# collect logs start-time 06/05/2007:12:00:00 end-time 06/05/2007:13:00:00 ftp://jpublic@arxftp.f5.com/1hr
Password: jpassword
prtlndA# collect logs start-time 2007-01-07T05:00:34 utc sinceBoot.tgz
You can use the expect monitor command to repeat a show command at regular intervals.
expect monitor seconds show-command [search-string] [timeout timeout-seconds]
seconds is the number of seconds to wait between each invocation of the command.
show-command is a valid CLI show command, surrounded by quotation marks (). Use any of the show commands documented in this manual.
search-string (optional) also requires quotes. If you use this, the command stops repeating as soon as the search-string appears in the output. Otherwise, the command repeats until you use <Ctrl-C> to stop it.
timeout-seconds (optional, 1-2096) sets a time limit on this monitor process. The show commands stop repeating when this time expires. You can also use <Ctrl-C> to stop the process at any time.
timeout-seconds - 21,600 (6 hours)
You cannot issue other CLI commands in this session while the expect monitor command is running. Use <Ctrl-C> to stop the repetition at any time.
To run a full CLI script, copy the script onto the ARX (using copy ftp, copy scp, copy {nfs|cifs}, or copy tftp), and then use run to invoke it. To perform any command in the future (possibly on a schedule), use the at command.
bstnA# expect monitor 5 show namespace wwmed
repeats the show namespace wwmed command every 5 seconds, until you use <Ctrl-C> to stop it.
bstnA# expect monitor 10 show namespace medarcv /rcrds
repeats the show namespace medarcv command every 10 seconds until /rcrds appears in the output.
Use the logging destination command to define an external syslog server as an external logging destination.
Use the no form of the command to remove a syslog server.
logging destination server-name [udp | tcp] [port port]
server-name (1-80) is the IP address or hostname of the external syslog server.
udp | tcp (optional) is the transport protocol over which to send the syslog messages.
port port (optional, 1-65,535) is the UDP or TCP port number where the external server is listening.
Use the show logging destination command to list all configured external syslog servers.
bstnA(cfg)# logging destination 10.1.1.76 tcp
bstnA(cfg)# no logging destination 172.16.44.7
Each group of log messages, known as a log component, has a separately-tunable logging level. Critical level is the most terse, displaying critical log messages only; debug level is the most verbose, displaying all log messages from critical down to debug. The messages appear in the syslog file. Use the logging level command to set a logging level.
Use the no form of this command to revert a logging level back to the default.
logging level component {critical | error | warning | notice | info | debug | disable}
no logging level {component | all}
component (up to 255 characters) is the component to tune. See Log Components for a complete list of components.
all selects all components.
critical | error | warning | notice | info | debug | disable is a required choice. This is the logging level you choose for the component. The disable option stops all logging from the chosen component.
Use show logging levels to verify the log-level settings for each component.
From any mode, use show logs syslog or grep pattern logs syslog to view the log messages in the syslog file. See the manual, ARX Log Catalog, for a full list of log messages.
bstnA(cfg)# logging level NSCKRPT debug
bstnA(cfg)# logging level all error
bstnA(cfg)# no logging level all
bstnA(cfg)# logging level RON disable
bstnA(cfg)# no logging level POLICY_ACTION
Use the no mail-to command to remove an E-mail recipient.
mail-to recipient
no mail-to recipient
recipient (1-128 characters) is one E-mail recipient (for example, juser@nemed.com).
bstnA(gbl-auto-diag)# mail-to juser@wwmed.com
prtlndA(gbl-auto-diag)# no mail-to ex@nemed.com
Use the management source command to set a source address for sending syslog messages.
Use the no form to remove the management-source configuration.
mgmt is the switchs out-of-band management interface, the default. You specify the out-of-band management IP address at switch boot-up.
vlan-id (1-4096) identifies the VLAN to use as the source. Use the show vlan summary command for a list of all configured VLANs.
mgmt on platforms that support this command
vlan 1 on platforms that do not support this command.
any except ARX-VE
Use logging destination to send syslog messages to an external syslog server. This command sets the source address for those log messages.
Use the show logging destination command to view the current management source.
bstnA(cfg)# management source vlan 9
The ARX records the round-trip times between itself and external devices, such as CIFS clients and back-end filers. The stats-monitor process keeps moving averages of these round-trip times and other statistics, and records them in .csv files in the stats-log directory. The stats monitor also can log notification messages if the number of errors and/or response time from an external device increases dramatically. The stats monitor only logs a notification if the device has these increases multiple times in a row. Use the moving-average command to define the percentage of increase required, and the number of times that the errors or response times must increase that much to merit a notification.
Use no moving-average to return to the default thresholds.
response-time | errors is a required choice. This determines the type of threshold or interval you are setting, response times or errors.
threshold pct-threshold (optional, 1-65,535) is the percentage above the moving average that could trigger an alarm. For example, if you set moving-average response-time threshold 200, the stats monitor may trigger a notification if a devices response time is 200% above (or, 3 times) its current moving average. As another example, if you set moving-average errors threshold 150, stats monitor may trigger a notification if the devices number of errors is 150% above its current average. The stats monitor process only triggers the notification if this occurs for some number of samples, as set by the interval option.
interval num-samples (optional, 1-65,535) is the number of high samples in a row that are required to trigger a notification. If this many samples in a row are at least pct-threshold above the current average, the stats-monitor process logs a notification.
either threshold - 100%
either interval - 6
This is a beta-level command. Use the terminal beta command to unlock all beta-level commands, including this one.
The moving average is the average for a number (for example, 8) of the most-recent samples. The 8th time the software calculates a moving average, it takes the average of samples 1 to 8; the 9th time it runs, it averages samples 2 to 9; the 10th time, it takes the average of samples 3 to 10; and so on.
If a group of samples deviates from the current moving average, the stats monitor generates a notification. Use this command, moving-average, to decide how many samples in a row it takes to log the notification, and to decide how far over the current average a sample must be to qualify.
These notifications are recorded in the syslog; you can use show logs syslog to view them along with all other syslog messages. Look for the log component named SMON_ANALYSIS, or use grep to search for it. To create an additional SNMP trap for each notification, use the trap command.
The stats monitor collects data for a long period of time, called a moving-average window, before it starts using the moving average for notifications. This window is calculated from the moving average and sampling interval commands: it is 8 * moving-average interval * sampling interval. By default, that is 8 * 6 * 5 minutes, or four hours. The moving-average window restarts if you use the sampling interval command to change the sampling interval.
bstnA(gbl-stmon-nfy)# moving-average response-time threshold 10
bstnA(gbl-stmon-nfy)# moving-average response-time interval 3
sets the following tolerances for response times: if a devices response time is 10% (or more) slower than its current average for 3 samples in a row, stats monitor logs a notification message. It also sends an SNMP trap if trap is enabled.
bstnA(gbl-stmon-nfy)# moving-average errors threshold 5
bstnA(gbl-stmon-nfy)# moving-average errors interval 3
sets the following tolerances for errors: if a device returns 5% (or more) errors than its current average for 3 samples in a row, stats monitor logs a notification message. As above, stats monitor also sends an SNMP trap if trap is enabled.
The ARX records the round-trip times between itself and external devices, such as CIFS clients and back-end filers. You can use the notify command to trigger analysis for one group of external devices: CIFS-service clients, NFS-service clients, CIFS filers, and or NFS filers. Whenever an issue is discovered for one of the analyzed device groups, the stats-monitor process logs a notification in the syslog. Optionally, the stats monitor can also send an SNMP trap. This command also brings you into a submode, where you can alter the notification settings for the chosen group of devices.
You can use the negative form of the command, no notify, to deactivate this stats-monitor process for front-end or back-end services.
cifs-service | nfs-service is a required choice. The nfs-service option starts analyzing all nfs front-end services, and the cifs-service option starts analyzing all cifs front-end services.
cifs | nfs is a required choice if you choose to monitor filer shares. This chooses between analyzing CIFS or NFS on back-end filers.
This is a beta-level command. Use the terminal beta command to unlock all beta-level commands, including this one.
The notify command places you in gbl-stmon-nfy mode, where you can use other CLI commands to edit the notifications for the set of devices you have chosen. The notifications only occur when some number of samples deviates from the moving average by some percentage; for example, a notification occurs when 6 samples from a particular CIFS service have twice as many NetworkErrors as the moving average, or when 6 samples from a particular back-end NFS export have progressively-higher round-trip times for GETATTR requests. You can use the moving-average command to reset the number of samples to trigger a notification, and/or to change the threshold for each sample to qualify as too high. By default, a notification goes to the syslog only; the trap command enables SNMP traps in addition to those syslog messages.
The stats monitor records its notifications in the syslog; you can use show logs syslog to view them along with all other syslog messages. Look for the log component named SMON_ANALYSIS, or use grep to search for it. To create an additional SNMP trap for each notification, use the trap command.
bstnA(gbl-stmon)# notify cifs-service
Use no sampling interval to return to the default interval.
seconds (30-86,400) is the number of seconds between each sample. 86,400 seconds is one day.
This is a beta-level command. Use the terminal beta command to unlock all beta-level commands, including this one.
This command restarts the calculation of all moving averages. After you invoke this command, stats monitor must run for the full moving-average window before it can log any more notifications (or, possibly, send any more traps). This window is calculated from the sampling interval command as well as the moving-average command: it is 8 * moving-average interval * sampling interval. By default, that is 8 * 6 * 5 minutes, or four hours.
bstnA(gbl-stmon)# sampling interval 60
Use no schedule to remove the auto-diag schedule, thereby disabling auto diagnostics.
name (1-64 characters) identifies the schedule for auto diagnostics. Use show schedule for a list of configured schedules. Choose a schedule that has a frequency measured in days, weeks, or a longer time measure; for example, every day, every 3 days, every 1 week, or every 2 months. This command does not accept a schedule with a higher frequency, such as every 12 hours.
To create a schedule, use the gbl-mode schedule command. The schedule cannot run more frequently than once per day; you can define this with the every command.
minturnA(gbl-auto-diag)# schedule every2am
Schedule is the schedules configured name, chosen with the schedule (gbl-auto-diag) command. The fields in this table show the configuration of the schedule.
Description only appears if set by the gbl-schedule description (gbl-schedule) command.
Start Time can be reset by the gbl-schedule start command. This is the date and time of day used for all time-based calculations: for example, if a schedule runs every Saturday, this determines the time of day on Saturdays when the schedule fires.
Stop Time only appears if set by the gbl-schedule stop command. This is the expiration time for the schedule, after which the schedule never fires.
Interval is set by the gbl-schedule every command. This is the time between scheduled runs.
Duration appears only if set by the gbl-schedule duration command. This is not applicable to auto-diagnostics.
Status is Paused (the schedule is between run times), Running, or Waiting for start time (the schedule has never been invoked).
Previous is a sub-table showing the Run time (when the run started) and End time of the most-recent run.
Current appears if the rule is currently running. This has the same fields as the Previous table, Run time and End time.
Next shows when the schedules next run will begin. It also has the Run time and End time fields.
Mail-to is a list of local E-mail addresses, added with the mail-to (gbl-auto-diag) command. Every time the above schedule fires, the system sends a tar file with auto-diagnostic output to each of these addresses. It also sends an encrypted version of the same tar file to F5 Support.
Additional commands are extra CLI-show commands that are collected along with the default auto-diagnostics. You can add or remove these with the [no] additional-command command.
Status describes the most-recent auto-diagnostics collection and delivery operation since the last reboot. The following fields appear only if an auto-diagnostics collection has occurred:
Last time executed is the most-recent time that the schedule fired and the collection process started.
Status of last execution is either Success or Failure with a report name in square brackets. You can use show reports report-name to read the report and find the specific failure.
Date last sent is a time stamp for the most-recent E-mail delivery.
Number of times sent since last reboot shows how many times the system has collected and delivered auto-diagnostics data since the last ARX reboot.
bstnA(cfg)# show auto-diagnostics
bstnA(cfg)# show auto-diagnostics
show documentation msg-catalog-id
msg-catalog-id (1-255 characters) identifies the message. Every message-catalog message has an uppercase ID that appears with it in the syslog and in the CLI.
bstnA(cfg)# show documentation SQM_CANNOT_BE_PROMOTED_GRACE_PERIOD
show logs syslog
bstnA(cfg)# show documentation SQM_CANNOT_BE_PROMOTED_GRACE_PERIOD
Date is the date and time when the event occurred.
ID is the last three digits in the traps OID.
Event is the name of the trap.
Description explains the alarm condition.
Refer to the ARX SNMP Reference for a full list of all traps, along with explanations for each trap and procedures to address the issue.
Follow the Guidelines for the command, snmp-server traps, to set up a remote host to receive SNMP traps. For an overall view of the system hardware, use show chassis.
prtlndA(cfg)# show health
bstnA> show health
prtlndA(cfg)# show health
bstnA> show health
timeout seconds (optional, 1-3600) is the maximum number of seconds to wait for each server response. The default timeout is 3 seconds. The ARX makes two or more attempts in some cases, and this timeout applies to each attempt. Also, the timeout applies to each individual server; the overall time for all servers to respond may be significantly longer. You can use <Ctrl-C> to stop the command at any time.
IP Address identifies the network device.
Role shows the relationship of this device to the ARX. Possible values here are External Filer or Domain Controller. Use the show external-filer command for a full list of external filers used by this ARX. Use the show active-directory command for a full list of DCs and a general understanding of how the ARX is using them.
Observed Time Skew ... shows a time value in HH:MM:SS format, followed by ahead, behind, or (for a time value of 00:00:00) nothing.
bstnA> show health time-skew timeout 30
bstnA> show health time-skew
Namespace lists all namespaces on the switch, one per row. The ID column contains the numeric ID.
Volume lists all volumes on the switch, one per row. The Namespace column shows the volumes namespace, and the ID column contains the volumes numeric ID.
Share lists all shares on the switch, one per row. The Volume and Namespace columns show the shares context, and the ID column contains the shares numeric ID. Contact F5 Support if Not Found appears at the end of any of these rows; this indicates an issue in the configuration database.
Archive Name is a table of file-history archives on the ARX. A file-history archive is a repository for a managed volumes file-location records. You can use the file-history archive command to create one. The ID column in this table contains the archives numeric ID.
bstnA(cfg)# show id-mappings
show logs syslog
bstnA(cfg)# show id-mappings
Use the show logging destination command to view a list of configured external syslog servers.
Use the logging destination command to define a syslog server.
Use the management source command to define a source address for sending syslog messages and traps.
bstnA# show logging destination
component (optional) specifies that you only want the logging level for the given component.
The log components are listed in an earlier section; see Log Components. This command shows the logging levels in a graph: for each log component, an X appears in the column of its logging level. Use the logging level command to change any or all component-logging levels.
bstnA(cfg)# show logging levels
bstnA(cfg)# show logging levels POLICY
bstnA(cfg)# show logging levels
bstnA(cfg)# show logging levels POLICY
Statistics Monitor Configuration is a table with the high-level configuration for the stats monitor. You can edit this configuration with the stats-monitor command and its sub commands.
Status is always Enabled.
Sampling Interval is the time between the data samples (such as average RTTs and error rates) recorded in the stats-log. The stats monitor records these samples in a series of .csv files. Each .csv file represents one group of external devices (such as back-end CIFS filers or NFS clients for a particular NFS service), or an internal process (such as file migrations or an internal CIFS work queue). You can use show stats-logs to view these .csv files. The default sampling interval is adequate for most sites, but you can use the sampling interval command to change it if necessary.
Max Data Size is the storage limit for all statistics (.csv) files stored in the stats-logs partition.
Max Raw Data Size is the percentage of Max Data Size permitted for .csv files with raw (not hourly) in their names.
Used Raw Data Size are the amounts of the above storage that are already used. These are for all statistics (hourly and raw) and raw statistics, respectively.
Notification Parameters only appears if at least one notification is active; use the notify command to enable analysis and notification for front-end services or back-end filers. This shows each notification in one row of the table, with the following fields:
Applies To identifies the type of service, process, or device that is configured for notifications. You set this with the notify command. This is :
CIFS Services (related to a cifs front-end service and its clients; these statistics are similar to the CIFS output from show statistics global server),
NFS Services (for an nfs service and its clients, similar to the NFS output from show statistics global server),
CIFS Shares or NFS Shares (back-end shares behind an ARX volume, viewable with show statistics namespace ... summary on an individual share),
Statistics shows the type of notification. This is Errors for a notification of an unusual number of errors, or RespTime for a notification of an increase in response time. You can use the moving-average command to enable or disable notifications for this type of event.
Traps indicates whether or not the stats monitor sends an SNMP trap with this notification. This is Enabled or Disabled, and you can use the trap command to change the setting.
Metric is always Mov.Avg.
Parameters shows the threshold settings that trigger a notification. You can use the moving-average command to change these thresholds.
bstnA# show stats-monitor
bstnA# show stats-monitor
Use the show system tasks command to see a list of subsystem tasks running on the ARX.
Proc - is the module slot (number) and processor on which this task is running. This column only appears for platforms with more than one physical processor, such as the ARX-4000.
Subsystem - is the name of the subsystem task. If the task is a script, that fact is noted in parentheses: (script).
Instance - is how many instances of the same task are running on this processor. This column does not appear for the ARX-1500 or ARX-2500.
Memory (Kb) - is the amount of memory used by this task.
CPU% - measures the CPU cycles that the task is consuming. This percentage is measured since the last invocation of show system tasks. For a measurement of overall CPU usage, use show processors and/or show processors usage.
CPU-Time - only appears on an ARX-1500 or an ARX-2500. This shows the CPU time used by the current subsystem task since the process last started. The time is expressed in HH:MM:SS format.
bstnA> show system tasks
stoweA> show system tasks
bstnA> show system tasks
stoweA> show system tasks
The ARX records the round-trip times and errors between its own software and other external devices (such as back-end filers and front-end clients), as well as time required for migrations and internal operations. You can use the stats-monitor command to begin configuring an analysis process that keeps a running average of this data and monitors it for sudden changes. Whenever an issue is discovered, the stats-monitor process logs a notification of the issue. Optionally, it also sends an SNMP trap and raises and alarm that is visible with the show health command.
You can use the stats-monitor command to enter gbl-stmon mode and edit the analysis parameters.
This is a beta-level command. Use the terminal beta command to unlock all beta-level commands, including this one.
This command puts you into gbl-stmon mode, where you can use commands to change the stats-monitor configuration. The stats monitor takes periodic samples of errors, computation times, and round-trip times for analysis; you can use the sampling interval command to change the time between these samples. To change the set of external devices that are analyzed (such as front-end client machines and/or back-end filers), you can use the notify command. The notify command also leads to a submode, where you can enable traps for the group of devices and change the thresholds for the device sets notifications.
The stats monitor records its notifications in the syslog; you can use show logs syslog to view them along with all other syslog messages. Look for the log component named SMON_ANALYSIS, or use grep to search for it.
bstnA(gbl)# stats-monitor
Use no trap to stop sending an SNMP trap for these alarm conditions.
response-time | errors is a required choice. This determines the type of error that triggers an SNMP trap: slow response times or excessive errors.
This is a beta-level command. Use the terminal beta command to unlock all beta-level commands, including this one.
Refer to the ARX SNMP Reference for a full list of all traps, along with explanations for each trap and procedures to address the issue. Follow the Guidelines for the command, snmp-server traps, to set up a remote host to receive SNMP traps. To manually clear a raised trap, you can use the clear health command.
To send these traps in E-mail to a set of interested recipients, you can use the email-event command and its sub-commands to create an e-mail event group.
bstnA(gbl-stmon-nfy)# trap response-time
Table of Contents   |   << Previous Chapter   |   Next Chapter >>

Was this resource helpful in solving your issue?




NOTE: Please do not provide personal information.



Incorrect answer. Please try again: Please enter the words to the right: Please enter the numbers you hear:

Additional Comments (optional)