When an error is detected in the container or POD of the master server, the cluster will perform an automatic failover by promoting one of the replicas to become the new master, and the connection destination of the database is switched. The database connection is broken, but you can reconnect by establishing a connection from the application again.

If an error occurs on the master server and an automatic failover occurs, the POD or container of the failed old master server is automatically restarted and reincorporated into the cluster as a replica server.

If a replica server fails, it automatically restarts and rejoins the cluster as a replica server.

You can manually switch any replica server to the master server. In this case, the original master server becomes the replica server.