A BIG-IP® system provides high availability via packet mirroring across two chassis. When discussing redundancy, one should consider more than the initial failover. If the backup chassis also fails a fail-back will be required. The following tables provides a quick summary of the initial failover and the fail-back scenarios. Note, a BIG-IP system does not support Geo-Redundant failover. In other words, a BIG-IP system supports the concept of a local HA Pair. However, a BIG-IP system does not support a second HA Pair which will take over if the first HA Pair fails. This type of scenario is required where multiple redundant data centers are available to handle geographic failure scenarios.
|Failover (active failed)||Supported||Supported|
|Failback (new active failed after old active came back online)||Supported||Partial Support|
The above table shows that after the initial failover, if a second failover happens, the media streams, created after the failover, will be maintained. However, media streams created prior to the initial failover will be dropped.