Failback
Failback returns a service or workload from a failover destination to its restored original resource or a designated replacement.
Direction: Failback is a return of service responsibility, not merely restarting the old host. It may be unnecessary when the new primary can remain in service. The return destination can be a different receiving server, rather than the same physical hardware. [1] [4]
Data boundary: A reverse-resynchronization workflow transfers changes from the active destination back to the original source, followed by a final update and session checks before traffic returns. These steps belong to the cited replicated-volume workflow; they are not permission to overwrite either copy without resolving its history. [3]
Rejoining is different: PostgreSQL pg_rewind can prepare a former primary to follow the new one as a standby. Required WAL replay and recovery configuration still matter; adding that standby restores redundancy but does not itself return the primary role. [2]
Relations
Limitations
- Returning traffic without the required data synchronization can expose stale or conflicting state. [1, 3]
Citations
- Microsoft Azure reliability: failover and failback, planned/unplanned failovers and failback
- PostgreSQL 18: pg_rewind, Description and Failures While Rewinding
- NetApp Element 12.7: Perform a failback when source volume still exists, reverse resynchronization and final update
- AWS Elastic Disaster Recovery: Failback to on-premises environment, original or different target server