When you enable OpenEdge Replication, you replicate a database on a primary machine to one or two secondary databases on a different machine or machines. In the event of a failure of either the primary database or the machine on which it runs, database activity can be failed over to one of the secondary databases. Activity continues on that second database and machine until the primary database or machine is again available.
Typically, a failure is indicated by a loss of communications between the Replication server on the primary database and the Replication agent on the secondary database. In this situation, the Replication agent enters a state known as pretransition, in which the agent waits until either the Replication server reconnects or transition is performed (automatically or manually). If communication is not resumed, transition begins.
Once you transition the secondary database to a source database, all database update activity is failed over to the secondary database. If you have configured a three database Replication Set, when one of the secondary (target) databases becomes the source, the other target continues its role as a replication target, and it receives updates from the new source. If you have a three database configuration, but no Replication Set, the second target database stops receiving replication updates.
Once you transition the secondary database to a source database, all database update activity is failed over to the secondary database. Once the primary machine is again available, all database update activity can be failed back to it. This process is known as failback. Failback is essentially a failover from the secondary to the primary machine.
Once the failback process completes, all database update operations are again performed on the primary database, and the secondary database returns to read-only access.
