Handling of master server failure
Updates are done on the master server only.
If the master fails, replication will stop. Applications doing updates reconnect to the backup master and then continue to operate.
Effectively, the backup master has become the new master. The slaves are failed over to the new master and continue to get updates.
When the old master comes up again, it will synchronize with the new master (which is handling application updates).
This is orchestrated by ClusterControl™.
Once the old master is up-to-date with the new master, ClusterControl™ turns it into a backup master.
Semi-sync replication between the master and the backup master ensures that both instances have the same data,
The Replication cluster has been restored to its original topology.
Please sign in to leave a comment.