A ticketing company's Cloud Spanner database must lose no committed transactions, even if an entire region becomes unavailable. Which configuration delivers a zero recovery point objective?
Show answer and explanation
Answer: A
Multi-region Spanner configurations replicate each write to replicas in several regions before it commits, using synchronous consensus. A regional outage therefore does not lose committed data, so the recovery point objective is zero, and failover is handled by the service. Nightly backups lose up to a day of commits, asynchronous replicas can lag and miss recent writes, and periodic exports leave gaps between runs.
Why the other options are wrong
- B. Nightly backups leave up to a full day of commits unprotected, so the recovery point is measured in hours rather than zero.
- C. Asynchronous replicas trail the primary, so writes made just before an outage can be missing from the replica. That produces a non-zero recovery point.
- D. Exports taken every six hours leave a gap in which committed transactions are lost during a disaster. Recovery point is bounded by the export interval, not zero.
Exam tip. A zero RPO across regions on Spanner means synchronous multi-region configuration, not backups or asynchronous replicas.