The automatic failover feature uses the synchronization service, which uses Postgres Logical Replication to synchronize the designated-active and designated-standby repositories.
If the underlying synchronization service is broken in a high-availability environment, use dmcluster switchback to re-configure and re-synchronize the repositories.
- Perform the following, based on the TVI alert received when server management logging is enabled:
TVI Alert Action 4604001 or 4604003 These alerts are indicative of a broken sync configuration. - Re-configure and re-synchronize the repositories using the the steps in Synchronizing the Active and Standby Repositories.
4604002 This alert is triggered when the replication process has had significant delays and the designated-standby repository can no longer keep up with the designated-active repository. - Refer to Understanding Replication Lag to troubleshoot networking and performance issues causing the performance delay.
- Reconfigure and resynchronize the repositories using the the steps in Synchronizing the Active and Standby Repositories.