100% Latest & Updated Google Professional Cloud Architect Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 120
A regulated broker must prevent deletion of mandatory records until the documented hold period expires. Which operational action best maintains post-retention lifecycle handling after launch?
- Delete retained objects immediately when their business value falls
- Ignore separate legal holds because normal retention expired
- Review lifecycle transitions and legal holds before eligible object deletion
- Apply cheapest storage class without considering compliance access needs
Explanation
Correct Answer
C:Expiry of routine retention does not supersede a separate active legal hold
Incorrect Answers
A:Can violate mandatory hold. Expiry of routine retention does not supersede a separate active legal hold.
B:Different holds may still apply. Expiry of routine retention does not supersede a separate active legal hold.
D:May not be the right lifecycle decision. Expiry of routine retention does not supersede a separate active legal hold.
Question 121
A media production farm needs shared POSIX-style files accessed concurrently by Linux-based render workers. Which architecture most directly meets the need for managed shared filesystem?
- Use BigQuery for applications expecting POSIX file locks
- Write all simultaneous render outputs into each VM local disk
- Choose Filestore with an appropriate performance tier and client network placement
- Store every frame only as a new Cloud Storage object and assume mounted NFS behavior
Explanation
Correct Answer
C:Filestore supplies managed NFS semantics needed by shared file paths and file locking
Incorrect Answers
A:Warehouse tables do not provide NFS semantics. Filestore supplies managed NFS semantics needed by shared file paths and file locking.
B:Files will not be consistently shared. Filestore supplies managed NFS semantics needed by shared file paths and file locking.
D:Object APIs do not guarantee file-lock semantics. Filestore supplies managed NFS semantics needed by shared file paths and file locking.
Question 122
A media production farm needs shared POSIX-style files accessed concurrently by Linux-based render workers. Which implementation choice most directly delivers least-exposed NFS access?
- Expose the NFS export to all connected VPCs without review
- Limit clients through VPC network and export controls appropriate to Filestore
- Depend on public NFS connectivity from render workers
- Grant every worker root-level storage administration
Explanation
Correct Answer
B:NFS shares require explicit network access boundaries despite being private
Incorrect Answers
A:Broadens filesystem access. NFS shares require explicit network access boundaries despite being private.
C:Increases exposure and can break private routing. NFS shares require explicit network access boundaries despite being private.
D:Unnecessary for ordinary file reads and writes. NFS shares require explicit network access boundaries despite being private.
Question 123
A media production farm needs shared POSIX-style files accessed concurrently by Linux-based render workers. Which validation would best establish proof of required file semantics?
- Upload one small test file from one client
- Measure bucket listing performance instead of NFS IOPS
- Check only available file space
- Run concurrent client read/write and locking tests under representative load
Explanation
Correct Answer
D:A single file upload does not establish correct shared filesystem concurrency
Incorrect Answers
A:Does not exercise concurrent file locks. A single file upload does not establish correct shared filesystem concurrency.
B:Unrelated to the shared filesystem. A single file upload does not establish correct shared filesystem concurrency.
C:Cannot prove expected concurrency semantics. A single file upload does not establish correct shared filesystem concurrency.
Question 124
A media production farm needs shared POSIX-style files accessed concurrently by Linux-based render workers. Which operational action best maintains performance and capacity operations after launch?
- Monitor IOPS, throughput, and capacity growth and adjust service tier when needed
- Increase capacity every time a job slows without checking throughput
- Monitor only VM CPU for the render fleet
- Keep every scratch render forever
Explanation
Correct Answer
A:File-based workloads can bottleneck on throughput before reaching storage capacity
Incorrect Answers
B:May miss IOPS bottlenecks. File-based workloads can bottleneck on throughput before reaching storage capacity.
C:Does not reveal NFS service saturation. File-based workloads can bottleneck on throughput before reaching storage capacity.
D:Wastes storage without a retention policy. File-based workloads can bottleneck on throughput before reaching storage capacity.
Question 125
A data science group stages daily external files into Cloud Storage, then moves inactive data into less expensive classes. Which architecture most directly meets the need for appropriate bulk object ingestion?
- Run a custom one-time curl command as the full external ingestion program
- Treat cold storage as a substitute for transferring files
- Use Storage Transfer Service for supported repeatable external object transfers
- Require each external contributor to email large data batches
Explanation
Correct Answer
C:Managed transfer jobs provide scheduled movement and transfer-status visibility
Incorrect Answers
A:Lacks managed repeat schedules and failure reporting. Managed transfer jobs provide scheduled movement and transfer-status visibility.
B:Class selection does not ingest external objects. Managed transfer jobs provide scheduled movement and transfer-status visibility.
D:Does not scale or validate transfer integrity. Managed transfer jobs provide scheduled movement and transfer-status visibility.
Question 126
A data science group stages daily external files into Cloud Storage, then moves inactive data into less expensive classes. Which implementation choice most directly delivers retention-aware storage lifecycle?
- Move all newly written objects directly into an archive class regardless of reads
- Apply object lifecycle rules based on measured age and retrieval patterns
- Define lifecycle based only on the shortest storage price line
- Apply uniform expiration before checking legal holds
Explanation
Correct Answer
B:Moving frequently accessed objects to colder classes can increase access costs
Incorrect Answers
A:May increase retrieval expense. Moving frequently accessed objects to colder classes can increase access costs.
C:Ignores operation cost and access frequency. Moving frequently accessed objects to colder classes can increase access costs.
D:Could delete protected datasets. Moving frequently accessed objects to colder classes can increase access costs.
Question 127
A data science group stages daily external files into Cloud Storage, then moves inactive data into less expensive classes. Which validation would best establish proof of transfer correctness?
- Reconcile transferred object counts, sizes, and checksums against the manifest
- Check that the transfer job was created
- Compare only the top-level folder name
- Count total bytes billed but not per-object checksums
Explanation
Correct Answer
A:Successful scheduling is not evidence that every external object arrived intact
Incorrect Answers
B:Does not prove objects arrived. Successful scheduling is not evidence that every external object arrived intact.
C:Cannot detect missing files or byte corruption. Successful scheduling is not evidence that every external object arrived intact.
D:Can conceal individual transfer defects. Successful scheduling is not evidence that every external object arrived intact.
Question 128
A data science group stages daily external files into Cloud Storage, then moves inactive data into less expensive classes. Which operational action best maintains data growth optimization after launch?
- Track storage-class access charges and revisit thresholds as usage changes
- Switch all hot datasets to cold classes without monitoring usage
- Track only source API usage rather than destination classes
- Stop re-evaluating lifecycle policy after access changes
Explanation
Correct Answer
A:Real request frequency determines whether colder storage remains economical
Incorrect Answers
B:Can make regular access expensive. Real request frequency determines whether colder storage remains economical.
C:Does not reveal storage economy. Real request frequency determines whether colder storage remains economical.
D:Can lock in costly assumptions. Real request frequency determines whether colder storage remains economical.
Question 129
A SaaS product needs PostgreSQL high availability in one region while also retaining point-in-time restoration capabilities. Which architecture most directly meets the need for managed relational availability?
- Use a single-zone database with no HA replica
- Put database contents in Cloud Storage as unstructured objects
- Choose a self-managed fixed VM without a recovery plan
- Use Cloud SQL for PostgreSQL with a regional high-availability configuration
Explanation
Correct Answer
D:Regional HA provides managed failover of the primary service across zones
Incorrect Answers
A:Cannot tolerate the same zone loss. Regional HA provides managed failover of the primary service across zones.
B:Cannot provide PostgreSQL transaction behavior. Regional HA provides managed failover of the primary service across zones.
C:Adds administration and state failure exposure. Regional HA provides managed failover of the primary service across zones.
Question 130
A SaaS product needs PostgreSQL high availability in one region while also retaining point-in-time restoration capabilities. Which implementation choice most directly delivers recoverable database change history?
- Enable HA but disable historical backups
- Assume replicas are an immutable point-in-time archive
- Enable supported automated backups and point-in-time recovery settings
- Store the sole backup in the same application writable directory
Explanation
Correct Answer
C:Failover alone does not restore accidentally deleted or corrupted business records
Incorrect Answers
A:Cannot recover accidental logical changes. Failover alone does not restore accidentally deleted or corrupted business records.
B:Replicated deletes can propagate. Failover alone does not restore accidentally deleted or corrupted business records.
D:Not a protected separate recovery path. Failover alone does not restore accidentally deleted or corrupted business records.
Google Cloud Certified - Professional Cloud Architect (Professional Cloud Architect) Practice Tests