Google Associate Cloud Engineer Practice Test 19
Question 180
A team has a successful Persistent Disk snapshot named data-before-migration and wants to use its contents as a read/write data volume attached to a replacement VM. What step is required?
- Mount the snapshot resource directly using the normal Linux block-device mount command; no disk must exist.
- Rename the snapshot to the replacement VM name; Google Cloud automatically converts its snapshot into that VM's writable disk.
- Create a compatible new Persistent Disk with that snapshot as the source, attach it to the replacement VM, and mount the restored filesystem.
- Change the snapshot retention policy to zero; making it expire converts the snapshot into an active writable Persistent Disk.
Explanation
Correct Answer
C:Create a compatible new Persistent Disk with that snapshot as the source, attach it to the replacement VM, and mount the restored filesystem. The documented restore flow creates a new disk from the snapshot; only that disk can be attached and mounted as writable storage.
Incorrect Answers
A:Mount the snapshot resource directly using the normal Linux block-device mount command; no disk must exist. A snapshot is a backup resource, not a block device available for direct mounting by a running guest OS.
B:Rename the snapshot to the replacement VM name; Google Cloud automatically converts its snapshot into that VM's writable disk. Changing a name does not restore disk data or attach a storage volume to an instance.
D:Change the snapshot retention policy to zero; making it expire converts the snapshot into an active writable Persistent Disk. Retention controls backup lifecycle, not disk restoration or attachment. Deleting the only recovery point would make the situation worse.
Question 181
An application needs the same base operating system across newly created VM instances in an instance group. Which Compute Engine artifact should define it?
- Create an approved instance template referencing the required boot image and use it for the managed instance group.
- Take one VM snapshot and expect the group’s template to switch images on every reboot.
- Rely on the group’s autoscaling policy to install an operating system on blank disks.
- Create a firewall policy and use it as the operating-system source for replacement VMs.
Explanation
Correct Answer
A:Managed instance groups create replacements from their instance template. Reference a tested public or custom image for consistent boot configuration, then roll out a new template version when the base OS changes.
Incorrect Answers
B:A snapshot is a backup artifact; it does not automatically become a new managed instance group image template.
C:Autoscaling determines capacity, not which base operating-system image starts on new instances.
D:Firewall rules control traffic, not the guest OS that Compute Engine provisions.
Question 182
A VM needs an attached filesystem shared via NFS by multiple instances, but engineers are proposing one read-write zonal Persistent Disk attached everywhere. What service should be considered?
- Use one zonal block disk in exclusive read-write mode and mount it from every VM in different zones.
- Provision a compatible managed NFS service such as Filestore and mount its share on authorized clients.
- Use a VM image as a writable shared file server filesystem.
- Create an external IP for each VM, which enables native NFS sharing on its boot disk.
Explanation
Correct Answer
B:Filestore provides a managed file system accessed over NFS from eligible clients. Plan the share’s network reachability, access controls, performance and geographic support for the group of VMs.
Incorrect Answers
A:Ordinary block disks have attachment and write-mode limits; they are not automatically a multi-client NFS filesystem.
C:An image supplies a creation template and is not a live shared POSIX filesystem.
D:External IPs do not turn separate local boot disks into an NFS server.
Question 183
A stateless web application needs multiple identically configured Compute Engine VMs that automatically scale with load and recover failed members. Which architecture best fits?
- Place each VM in a manually created unmanaged instance group with no capacity policy and restart failed VMs by hand.
- Use an autoscaled managed instance group built from an instance template, with appropriate health checking and a load balancer.
- Run all web traffic on one large VM and increase its boot disk size when CPU demand grows.
- Create a separate Cloud Billing account per VM; billing isolation automatically implements autoscaling and repairs.
Explanation
Correct Answer
B:Use an autoscaled managed instance group built from an instance template, with appropriate health checking and a load balancer. A managed instance group controls replicated instance configuration and can autoscale, repair unhealthy instances and serve through a load balancer.
Incorrect Answers
A:Place each VM in a manually created unmanaged instance group with no capacity policy and restart failed VMs by hand. An unmanaged group does not provide the requested template-governed automated creation, scale-out and autohealing behavior.
C:Run all web traffic on one large VM and increase its boot disk size when CPU demand grows. A single VM is a single serving failure domain and disk capacity is not a response to traffic-based instance demand.
D:Create a separate Cloud Billing account per VM; billing isolation automatically implements autoscaling and repairs. Separate billing accounts track charges but have no mechanism to add or repair a fleet of VM instances.
Question 184
A managed instance group repeatedly recreates healthy application VMs. Its autohealing probe checks /health-old, which the new app version no longer serves. How should the rollout be repaired?
- Disable every VPC firewall rule to eliminate all possible health-probe network restrictions.
- Raise the CPU quota only, because an unsupported probe URL becomes healthy when quota increases.
- Update autohealing to use a valid application health endpoint with suitable initial delay and conservative failure thresholds.
- Switch the group to Spot VMs to keep failed probes from triggering autohealing.
Explanation
Correct Answer
C:Update autohealing to use a valid application health endpoint with suitable initial delay and conservative failure thresholds. MIG autohealing recreates VMs when its health check fails. A correct endpoint and realistic thresholds prevent false unhealthy classifications.
Incorrect Answers
A:Disable every VPC firewall rule to eliminate all possible health-probe network restrictions. Removing firewall controls broadly introduces unrelated exposure; the stated problem is a stale probe path, not a denied network path.
B:Raise the CPU quota only, because an unsupported probe URL becomes healthy when quota increases. The health checker evaluates an application endpoint. Additional CPU quota does not create a missing /health-old route.
D:Switch the group to Spot VMs to keep failed probes from triggering autohealing. Spot changes preemption risk and cost, not whether the configured health check correctly evaluates the application.
Question 185
A regional managed instance group must continue serving after one zone is impaired. Which configuration is essential?
- Place the entire group in one preferred zone so its load balancer uses fewer backends.
- Turn off autohealing so unhealthy zonal VMs remain registered during an incident.
- Configure a regional MIG spanning supported zones, with sufficient regional capacity and health-checked traffic distribution.
- Use a zonal MIG and rely on an unused second subnet to recover the same VMs.
Explanation
Correct Answer
C:Regional MIGs can distribute instances across zones. Combine multiple-zone placement, available capacity, autohealing and an appropriate load balancer so one impaired zone need not eliminate service.
Incorrect Answers
A:Concentrating instances in one zone preserves a single-zone failure domain.
B:Without health recovery, failed instances can leave the group below target capacity.
D:A subnet is not a second zone deployment and does not replicate a zonal instance group.
Question 186
A managed instance group must move its stateless web tier to a new instance template without replacing every VM simultaneously. Which update method allows gradual rollout?
- Configure the MIG updater for a controlled rolling or canary template update, including the required disruption limits.
- Delete the MIG and recreate it without a health check, so every instance changes at once.
- Edit the project display name because the MIG detects name changes as an instruction to update its VM image.
- Increase the Cloud Billing budget, which automatically updates instances to the latest disk image.
Explanation
Correct Answer
A:Configure the MIG updater for a controlled rolling or canary template update, including the required disruption limits. Google's MIG updater supports controlled rollouts, including configurable disruption and canary versions for instance templates.
Incorrect Answers
B:Delete the MIG and recreate it without a health check, so every instance changes at once. Deleting the group is disruptive and lacks the requested gradual replacement behavior and rollout controls.
C:Edit the project display name because the MIG detects name changes as an instruction to update its VM image. A project display-name change does not modify the managed instance group's desired template version.
D:Increase the Cloud Billing budget, which automatically updates instances to the latest disk image. Budget increases authorize spending but have no effect on the VM template version selected by a managed instance group.
Question 187
A developer creates a new instance template with an updated boot image, but existing VMs in the managed instance group still run the old image. What is the next deployment action?
- Restart the Cloud Billing service because VM image changes are picked up only after invoice recalculation.
- Change the old template's display name, which mutates the already-running VM disks immediately.
- Increase the VPC subnet secondary IP range; extra addresses cause Compute Engine to update each VM template.
- Set the MIG's desired version to the new instance template and initiate or allow the controlled update rollout.
Explanation
Correct Answer
D:Set the MIG's desired version to the new instance template and initiate or allow the controlled update rollout. The MIG controller must be told which template version to apply. Its updater then reconciles existing instances under the approved update policy.
Incorrect Answers
A:Restart the Cloud Billing service because VM image changes are picked up only after invoice recalculation. Billing service state does not control the desired software image of running managed instance group VMs.
B:Change the old template's display name, which mutates the already-running VM disks immediately. Running VMs do not automatically replace their boot images when an unrelated template label changes.
C:Increase the VPC subnet secondary IP range; extra addresses cause Compute Engine to update each VM template. Subnet capacity affects IP allocation, not which image or template version the MIG is configured to deploy.
Question 188
A web tier has no load growth at night, but its managed instance group remains fixed at the daytime peak. Which feature reduces idle capacity safely?
- Keep a fixed daytime instance count and reduce Cloud Logging retention at night.
- Set the instance template’s memory lower so the group automatically deletes idle VMs.
- Disable autoscaling and manually delete all backends at midnight.
- Enable a MIG autoscaler with an appropriate workload signal and min/max capacity bounds.
Explanation
Correct Answer
D:An autoscaler can lower the group’s target size during quiet periods while respecting a minimum replica floor. Choose a useful CPU, load-balancing or custom metric and test scaling behavior before production.
Incorrect Answers
A:Reducing log retention could lower storage costs but does not release unneeded compute replicas at night.
B:Machine sizing controls resources per instance, not how many replicas run as demand changes.
C:Deleting all backends ignores availability requirements and requires manual intervention to restore service.
Question 189
A regional group update sets maxSurge to one VM while the group spans three zones, triggering validation errors. What constraint should the update design account for?
- Leave fixed maxSurge=1 and set maxUnavailable=0; validation ignores the number of zones.
- Convert maxSurge to an unsupported negative value to avoid creating extra instances.
- Set the group name to include all zone IDs; a name change removes updater constraints.
- Use a valid regional MIG maxSurge setting: zero or a fixed number at least the number of zones, subject to update rules.
Explanation
Correct Answer
D:Regional MIG updates distribute changes across zones. A fixed maxSurge must be zero or at least the number of selected zones; verify maxUnavailable and quota together before launching the rollout.
Incorrect Answers
A:For fixed surge values on a regional MIG, one extra VM can be invalid when the group spans three zones.
B:Surge counts cannot be negative and that would not create a valid update plan.
C:Resource names are descriptive identifiers; they do not change the rolling updater’s multi-zone fixed-surge validation.
Question 190
An operator groups web VMs in an unmanaged instance group but expects it to recreate failed VMs automatically. What change supplies managed autohealing behavior?
- Add a Cloud Billing alert for network spend; the alert automatically replaces failed unmanaged VMs.
- Set the group's display name to managed; an unmanaged group acquires autohealing from a naming convention.
- Migrate the fleet to an appropriately configured managed instance group with an instance template and autohealing health check.
- Enable serial console access and manually reboot each failed VM without changing the unmanaged group.
Explanation
Correct Answer
C:Migrate the fleet to an appropriately configured managed instance group with an instance template and autohealing health check. Managed instance groups can monitor application health and recreate members. Unmanaged instance groups do not provide that automated repair behavior.
Incorrect Answers
A:Add a Cloud Billing alert for network spend; the alert automatically replaces failed unmanaged VMs. Billing alerts provide financial notifications, not an instance-group repair controller.
B:Set the group's display name to managed; an unmanaged group acquires autohealing from a naming convention. Management type is a group architecture and resource control feature, not a string in the group name.
D:Enable serial console access and manually reboot each failed VM without changing the unmanaged group. Manual reboot may recover one VM but does not fulfill the requirement for automatic replacement when members become unhealthy.