How would you perform a test failover for a MetroCluster?

• A. Run a controlled failover/failback using the defined runbook, verify data integrity, and validate client connectivity.
• B. Power down the primary site and switch to the secondary.
• C. Reimage all nodes and restore from tape.
• D. Disable all network interfaces and observe.

Answer: (A)

Testing a MetroCluster failover is about validating the disaster recovery process in a controlled and repeatable way. The best approach is to perform a controlled failover and failback using the predefined runbook, verify data integrity after the switch, and confirm that clients can reconnect and access services from the surviving site. This method ensures DR readiness without causing an unplanned outage, and it provides a clear rollback path to bring everything back to the original site once the test is complete. It also allows you to confirm that data remains consistent across sites and that the network and client paths function correctly during and after the failover.

Powering down the primary site and switching over is an actual outage and not a controlled test, risking data loss and service disruption. Reimaging all nodes and restoring from tape is destructive and not how MetroCluster testing is conducted. Disabling network interfaces and observing does not simulate a real failover and does not validate data integrity or client access.