Failover Testing

The size of a failover test operation for Oracle Database virtual machines is simply the size of the changes written to the test virtual machines during the test. All changes from the failover test virtual machines are deleted when the test is stopped.

Zerto best practice: Do not run a failover test for longer than the journal history is configured. If a longer test is required, then use the offsite-clone feature to create an entire VM-level clone of the virtual machines in the VPG onto any storage from any specified point-in-time available in the journal.

If recovering to a database consistent point-in-time, the failover test operation enables validation of the Oracle Database starting in backup mode to verify the consistency. It then enables validating removing the database from backup mode and mounting the database to ensure a successful recovery.

Zerto best practice: Perform test failovers to both crash and database consistent points-in-time to verify the recovery of both on a per Oracle database basis. The ability to recover from either should be recorded and included in a disaster recovery plan for future decision making in a failover scenario when deciding what point‑in-time to recover to and the amount of data loss to accept.

If IP address changes are required as part of a failover or move operation, they should also be performed in the failover test operation to ensure a successful recovery. This allows the IP change process to be verified and the full disaster recovery scenario to be simulated.