The primary instance data is replicated frequently on each replica instance in the deployment. You can view the replication status of each replica instance in the Operations Console Replication Status Report page. When a replica instance shows an Out of Sync status, you must manually synchronize this replica instance. The synchronization process brings the replica instance data back into sync with the primary instance data.
When you synchronize a replica instance, only primary instance data is sent to the replica instance. The synchronization process does not send replica instance data to the primary instance.
A replica instance naturally goes out of sync when the following events take place:
A replica instance is promoted to become the new primary instance.
The primary instance is restored with a backup.
You must manually synchronize each replica instance with the primary instance after these events. When you manually synchronize a replica instance with the primary instance, the replica instance can contain data that is waiting to be sent to the primary instance. This pending data is lost during the synchronization process.
A replica instance can go out of sync when too much disk space is consumed as the result of the following events:
A network error prevents the replica instance from communicating with the primary instance.
An internal error prevents replica instance changes from being saved on the primary instance or primary instance changes from being saved on the replica instance.
You must manually synchronize each affected replica instance with the primary instance after the network or internal error is resolved.