Google Professional Cloud Database Engineer
Get started today
Ultimate access to all questions.
You launched a successful mobile game and utilize a 50 TB Cloud Spanner instance with PITR for storing game data in production. Upon reviewing game analytics, you discovered players exploiting a loophole to gain unfair points and climb the leaderboard. Additionally, a DBA mistakenly executed an emergency bugfix script that corrupted some production data. You must assess the scope of the corruption and restore the production environment. What steps should you take? (Select two.)
You launched a successful mobile game and utilize a 50 TB Cloud Spanner instance with PITR for storing game data in production. Upon reviewing game analytics, you discovered players exploiting a loophole to gain unfair points and climb the leaderboard. Additionally, a DBA mistakenly executed an emergency bugfix script that corrupted some production data. You must assess the scope of the corruption and restore the production environment. What steps should you take? (Select two.)
Explanation:
In a scenario where data corruption occurs in a PITR-enabled Cloud Spanner instance, the approach to recovery depends on the extent of the corruption. For significant corruption, using backup and restore with a specified recovery timestamp (Option A) is the most reliable method to restore the database to a state before the corruption occurred. This leverages Cloud Spanner's point-in-time recovery feature. For insignificant corruption, performing a stale read with a specified recovery timestamp and writing the results back (Option E) is more efficient as it allows for targeted recovery without the need for a full restore. Options B and C are less ideal; stale reads are not meant for significant data recovery, and import/export is not suitable for point-in-time recovery. Option D is incorrect because even for insignificant corruption, backup and restore is an overkill when a stale read can suffice.