If the backend primary server fails, or if the whole backend node fails, the recovery must be handled according to rules and tools of the backend product in question. The solidDB^® server does not offer any means to correct the situation.

After the backend server is running as a new primary, an exact copy of the IBM InfoSphere CDC instance is restarted at the surviving node. You must reconfigure the subscription with the IBM InfoSphere CDC tools to reconnect the IBM InfoSphere CDC instances in question. The new subscription must continue from a full refresh (in both directions) before the mirroring can start.

In some cases, the state of the subscription replication might be lost, and a full refresh is needed.