Switchover and Failover Processes

Switchover process

For administrative reasons, automatic switchovers happen in a controlled, planned manner during the scheduled maintenance. However, data is preserved during a switchover.

During a switchover, the IP address of the primary server is moved to the promoted replica server. Meanwhile, the client is signaled about the switchover by closing the TCP connection on the server. The client automatically closes the connection and reconnects to the database.

Failover process

IONOS DBaaS MariaDB has an automatic failover enabled by default. A failover can occur when the primary node fails. As a result, one of the replicas is chosen and promoted to be the new primary node. When the old primary node returns, it is added as a replica to the cluster.

Note: You may lose the data if it is not replicated but only committed to the primary node.