Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team manages a mission-critical Cloud SQL for PostgreSQL (Enterprise Plus) database in us-central1 with a designated cross-region disaster recovery (DR) replica in us-east4 and Point-in-Time Recovery (PITR) enabled. You must design a quarterly DR drill to validate the following requirements:
Which DR testing approach should you implement?
In Cloud SQL Enterprise Plus, advanced disaster recovery (DR) provides a switchover operation designed for planned role reversals and routine DR drills between a primary instance and its designated cross-region DR replica. Combining this with Point-in-Time Recovery (PITR) restoration drills in an isolated non-production environment allows teams to safely test both infrastructure failover and logical corruption recovery.
This approach directly exercises every failure mode—both regional outages and logical data corruption—while protecting production workloads and generating accurate, repeatable RTO and RPO validation metrics.
Replica failover is an emergency operation in Cloud SQL advanced DR that immediately promotes a designated cross-region DR replica to become the new primary instance when the primary region becomes unavailable. Cloud Monitoring backup metrics track whether scheduled backup jobs succeeded or failed.
You were responding to an actual, unplanned regional outage where the primary region is completely unreachable and immediate recovery takes priority over zero data loss.
For routine DR drills, switchover is specifically recommended over replica failover because it synchronizes all transactions first, ensuring zero data loss while still exercising the failover path.
An in-place Point-in-Time Recovery (PITR) restoration overwrites or replaces the live production database state using historical backups and transaction logs. Confirming that the database accepts SQL connections checks only basic database engine availability.
You had already experienced a confirmed production data corruption event, declared an outage, and needed to perform an actual emergency PITR recovery in production.
Even during an emergency PITR recovery, engineers typically restore to a new or cloned instance first to inspect data consistency before cutting over traffic, avoiding destructive overwrites of the original instance.
Manually shutting down a production Cloud SQL instance and severing network connectivity simulates a hard failure directly in the live environment. Recording only the database promotion time measures an isolated infrastructure step rather than total service recovery.
You were conducting a tightly scoped chaos engineering experiment in a dedicated, production-equivalent sandbox specifically designed to measure isolated database engine promotion speed.
Comprehensive DR drills track multiple statistics—including failure detection time, recovery initiation time, RTA, and RPA—to ensure the complete business workflow meets SLA and SLO commitments.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated