GCP Associate Cloud Engineer Practice Question

Your Cloud Bigtable instance has two clusters (us-central1 and europe-west1) using multi-cluster replication. Governance rules state that if a table is accidentally deleted you must be able to recreate it in less than 15 minutes, without exporting data outside Bigtable or performing manual copy jobs. Which strategy satisfies the requirement while keeping operational effort and storage costs low?

Take nightly persistent-disk snapshots of both clusters' underlying storage and restore the disks if the table is deleted.
Schedule a Cloud Dataflow job to export the table to Cloud Storage every night and re-import it if recovery is required.
Rely on multi-cluster replication alone and, after an accidental delete, immediately fail over to the europe-west1 cluster to recover the table.
Configure an hourly Cloud Bigtable backup on the us-central1 cluster and automatically copy each backup to the europe-west1 cluster; use the managed restore operation to recreate the table when needed.

Report Issue

Answer Description

Bigtable's managed backup feature lets you create point-in-time backups inside the same instance. A backup is stored in one cluster, but you can copy it asynchronously to a second cluster in the same instance. Copying keeps the data inside Bigtable, incurs only backup-class storage charges, and allows a fast, in-place restore that creates a new table without manual export-import steps. Disk snapshots or relying on replication will not protect against accidental deletes, because deletions replicate to all clusters. Cloud Storage exports work but require extra services and a much longer restore time, so they do not meet the 15-minute RTO.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.