Consider this scenario:
Replication is all set up and running, but there are some network issues meaning you have to disable it for a short while until that's resolved. Now in order to start replication again you have to do a full backup and restore. Not exactly fun for a large database.
But at some point these 2 databases were in synch with each other. Can't they be brought into line more easily without having to do a full backup and restore?
When replication is interrupted due to network problems between Source and Target, and if Source still remains available to users, then all changes are accumulated in AI files. In this case, the Target database goes into the pre-transition state and waits for connection to the replication server.
Once the network problems are fixed, it is enough to restart the replication server on the Source database.
The agent on Target and the server on the Source will be connected and the synchronization process will begin.
Thus, if you have a properly configured OE Replication, then you do not need to restore the Target database from backup due to network problems.
The main condition is that you have enough space in AI extents to save all changes during the time when there is no connection with the agent.
Agreed Valeriy, but if you have to disable replication because AI space is running out then you have to start again.
In this case, instead of disabling replication, how about adding new AI extents and using prostrct reorder ai?