Lab: Site Recovery drill and failover decision

Outcome

Replicate a supported disposable Azure VM to a secondary region, prove recovery readiness with a nonproduction test failover, and write the decision gates for actual failover, commit, reprotection, and failback.

Do

1. Check VM OS, compute, disk, network, target-region capacity, and support requirements. Create the Recovery Services vault, target resource group/network, replication policy, and replication configuration. Record source/target regions, target subnet, recovery points, and protected health.
2. Run a test failover to an isolated test network. Connect to the created VM, verify expected workload data and network isolation, record the selected recovery point and job output, then clean up the test failover. A test drill must not change production replication.
3. Write and review a two-person production runbook: outage declaration; target network/DNS; recovery-point choice; failover; workload validation; commit only after validation; reprotection; planned failback; and post-failback re-protection. Do not run unplanned/production failover for this lab.

Verify

Capture supported-resource checks, protected health, target configuration, test-failover job, isolated VM validation, cleanup, and the approved decision runbook. You should now understand: Backup restores data; Site Recovery replicates for continuity. Test failover is a no-disruption drill; actual failover and commit change the recovery state and require a return plan.

Break and fix

1. Configure an invalid target subnet/network or choose an unsupported resource. Diagnose prerequisites and correct target settings before replication.
2. Attempt test validation from a network that cannot reach the isolated test VM. Diagnose target NSG/DNS/subnet without opening production paths.
3. Simulate a request to commit before application validation. Stop, use the test checklist, and document why commit makes the selected recovery point final for that failover.

Scenario practice

1. Configure: Choose target region/network, replication policy, recovery point, test network, and RTO/RPO evidence for a multi-tier workload.
2. Diagnose: A replicated VM is unhealthy or test failover cannot be reached. Inspect replication health, target capacity/network, recovery point, NSG, DNS, and job details.
3. Distinguish: Contrast backup restore, test failover, planned failover, unplanned failover, commit, reprotection, and failback.

Clean up

Clean up the test failover, disable replication only for disposable test VMs, remove target resources and vault after confirming no recovery objective depends on them, and delete the training groups.

Sources and re-audit

• Azure Site Recovery overview, checked 2026-08-22.
• Fail over Azure VMs to a secondary region, checked 2026-08-22.

Re-audit when Site Recovery support, replication, or failover/failback behavior changes.

Related

azure-site-recovery · azure-backup · recovery-services-vault · workload · All labs