Google ACE Practice Question: Ensuring Successful Operation of a Cloud Solution

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Detach the disk, create the snapshot, then reattach.

  Why it's wrong here

  Detaching a persistent disk from a running instance is disruptive and risky: if the disk is in use by the OS or applications, the forced unmount can trigger I/O errors and unexpected application crashes. It also doesn't solve the consistency problem because the filesystem may still have unflushed buffers that are lost when the disk is detached. Google Cloud snapshots work on attached disks without detachment, so this extra step only adds downtime and danger without improving snapshot quality. Stop the instance instead to cleanly flush all writes.

• ✗

  Create the snapshot while the instance is running; snapshots are always consistent.

  Why it's wrong here

  The claim that snapshots are 'always consistent' misrepresents a crucial distinction: Google Cloud snapshots are crash-consistent at the storage level, meaning they represent a valid point-in-time recovery point, but they are not application-consistent. When the instance is running and the disk is being written, a snapshot may reflect a state where a database has committed a log record but has not yet flushed the data page, leaving the database in an inconsistent state. So creating a snapshot while the instance is running is safe in terms of storage integrity, but it does not guarantee that your application can cleanly recover from that snapshot.

• ✗

  Use gcloud compute disks snapshot without stopping; data will be consistent.

  Why it's wrong here

  Running `gcloud compute disks snapshot` on a running instance produces a crash-consistent snapshot, not a guaranteed application-consistent one. The snapshot is taken at a point in time, but it does not instruct the guest OS to flush filesystem caches, journal entries, or database transaction logs. If the instance is actively processing writes, the captured state may contain half-committed transactions or incomplete file operations that are not recoverable. Therefore, this command alone cannot be assumed to deliver 'consistent' data in the application sense.

• ✓

  Stop the instance, create the snapshot using gcloud compute disks snapshot, then restart the instance.

  Why this is correct

  Stopping the instance is the correct approach because it triggers a clean guest OS shutdown, allowing filesystem caches to be flushed and applications to close files gracefully. After the instance is in the `TERMINATED` state, the persistent disk is quiescent, and running `gcloud compute disks snapshot` captures a point-in-time image that is both crash-consistent and application-consistent for most setups. Once the snapshot completes, you can restart the instance with `gcloud compute instances start` and resume operations with confidence that the snapshot reflects a known-good state.

Go deeper

Related to this question