In case the backup master fails, the master will still continue to replicate its updates to the slaves.
Applications still do their updates on the master. The replication cluster does not need reconfiguration.
ClusterControl™ detects the failure and will try to restart the backup master.
When the backup master comes up again, it will synchronize with the master.
When the backup master has caught up, it will be kept in sync via semi-synchronous replication.
At this point, the Replication cluster has been restored to its original topology.
Please sign in to leave a comment.