Transition of a Replication Set occurs in one of two scenarios:
Recovery transition — If the source fails, the two targets transition together. After the transition completes, the primary target is transitioned to a source database, and the other is its target.
Failover transition — If all three replicas are online, the source and two targets to transition together.
In both transition scenarios, replication continues after the transition completes.
The transition section of the server properties file must be configured for a Replication Set. For example:
Recovery transition of a Replication Set occurs when the source database is lost. The primary target transitions to the source, and the secondary target becomes a target for the new source.
During normal Replication Set processing, the three replicas are configured as shown:
If the source becomes unavailable and the Replication Set transitions, the resulting configuration is as shown:
Replication continues, but the Replication Set is suspended. Re-adding the third replica requires a rebase. After the database is rebased and added as a target, the Replication Set is reinstated.
Failover transition
Failover transition of a Replication Set is the planned switch of a target and source replica.
Failover transition is initiated by executing the following command on the source database:
dsrutil source-db -C transition failover
All three databases of the Replication Set should be online for failover transition to successfully execute.
Before failover transition, the Replication Set is configured as shown:
After executing failover transition on a Replication Set, all three members of the set remain, but the source and the first agent listed in the transition-to-agents property, have switched roles, as shown:
Recovery transition — If the source fails, the two targets transition together. After the transition completes, the primary target is transitioned to a source database, and the other is its target.
