Restore Verification Checklist

A restore is considered valid only if all these checks pass:

• Schema Exists: Verify that database objects and expected migrations are present.
• Read/Write Check: Execute a sample read query and write query to confirm access.
• App Smoke Check: Verify that the application can successfully connect and function.
• Permissions: Ensure all required database roles and grants are intact.
• Row Count: Confirm that the record counts match the state of the backup.

Core Exercises (Required)

1. Verify Scheduled Backups: Use kubectl get backup -n develop to identify the most recent successful backup.
2. Execute Manual Backup: Manually trigger a backup for the develop Postgres cluster and verify its completion.
3. Simulate a Restore: Restore the database into a separate namespace. Use a temporary pod to query the data and verify its integrity.
4. Identify Gaps: Intentionally delete a database role before taking a backup. Restore the backup and document how this affects the application.

Challenge Exercise (Optional)

Corrupted Backup Restore: Intentionally corrupt a CNPG backup artifact, attempt a restore, and verify that the restore verification checklist correctly identifies the failure. Document which check caught the corruption.

Done When

You have completed this chapter when:

• You can verify that scheduled backups are running and completing successfully.
• You have successfully executed a manual backup.
• You have performed a restore simulation and validated the recovered data.
• You understand why a “successful” backup status is not recovery proof.

Knowledge Check

Before finishing this chapter, complete the Quiz to verify your understanding of the guardrail principles.