Failover

The OneDB SQL Data store HA cluster supports automatic failover and manual failover. The default is set for automatic failover of the HA cluster. When the HDR primary server becomes non-responsive the HDR secondary needs to take over the primary responsibilities. This can happen automatically, or it can be configured to require manual intervention.

Auto failover functionality is designated by using the onedbcm.autofailover helm parameter in the OneDB SQL Data Store helm chart. By default, this value is set to true. If manual failover is preferred this can be set to false.

When failover is performed whether its automatic or manual the roles will toggle between pods onedb-server-0 and onedb-server-1. When a pod is restarted it will restart the OneDB database server as primary or secondary based on the peer pod and current state of that OneDB server.