How to reset and re-establish database replication depends on the critical situations that you are experiencing and how many Address Manager servers you have in your database replication system.
Case 1—Network Communication Failure
If you have two Standby servers in replication and you have latency issues with one of the Standby servers, where the latency has exceeded the configured critical threshold or the disk space on the Primary server is being exhausted, remove the Standby server that can't be contacted. For more information on removing a Standby server from an active replication configuration, refer to Modifying database replication from the user interface. The Standby server can be added again after it is reset and server connectivity issues have been resolved.
If the connection between the Primary server and all Standby servers is lost, break replication on the Primary server so that the Primary server performs in a stand-alone state until the issue is solved. In addition, reset the Standby servers so that they can be added to a replication configuration again once connectivity issues have been resolved. For more information, refer to Breaking and resetting database replication.
Case 2—Disaster Recovery
If the Primary server goes offline due to unexpected circumstances, such as a result of physical damage or mechanical issues that cause it to become unusable, reset replication on the Standby servers to return the servers to standalone operation. If there was a single Standby server, the former Standby server will have retained the replicated copy of the database, and can be used as a standalone Address Manager server until issues with the former Primary are resolved and replication can be reconfigured. If there were two Standby servers, the former Standby servers will have retained the replicated copy of the database, and a new Primary/Standby replication relationship can be created using the former Standby servers, until issues with the former Primary are resolved. For more information, refer to Forcibly resetting database replication.