PCDBE practice questions 11–20

1. Question 11 · 170 in bank

   You want to migrate your on-premises PostgreSQL database to Compute Engine. You need to migrate this database with the minimum downtime possible. What should you do?

   • APerform a full backup of your on-premises PostgreSQL, and then, in the migration window, perform an incremental backup.
   • BCreate a read replica on Cloud SQL, and then promote it to a read/write standalone instance.
   • CUse Database Migration Service to migrate your database.
   • DCreate a hot standby on Compute Engine, and use PgBouncer to switch over the connections.

   Show answer

   Answer: D

   Explanation

   D. A does not represent the minimum downtime. Read replicas in Cloud SQL depend upon a primary instance also in Cloud SQL. We’re talking about migrating to GCE, so B is wrong. The Database Migration Service migrates PostgreSQL to Cloud SQL or AlloyDB for PostgreSQL (Preview as of 3/16/23 which means it won’t be on the exam - yet). That makes C wrong. That leaves D. Pgbouncer is a connection pooler used to minimize application downtime. https://cloud.google.com/architecture/migrating-postgresql-to-gcp

   Source: Adapted from community discussion — dynamic_dba

   References

   • Google Cloud documentation

   Next question

2. Question 12 · 170 in bank

   You have an application that sends banking events to Bigtable cluster-a in us-east. You decide to add cluster-b in us-central1. Cluster-a replicates data to cluster-b. You need to ensure that Bigtable continues to accept read and write requests if one of the clusters becomes unavailable and that requests are routed automatically to the other cluster. What deployment strategy should you use?

   • AUse the default app profile with single-cluster routing.
   • BUse the default app profile with multi-cluster routing.
   • CCreate a custom app profile with multi-cluster routing.
   • DCreate a custom app profile with single-cluster routing.

   Show answer

   Answer: C

   Explanation

   C. To ensure you still get connected when a Bigtable instance has 2 or more clusters, you must use multi-cluster routing. The question states that a single cluster existed first, then a second cluster was added. Google’s documentation states, “if you created the instance with one cluster, the default app profile uses single-cluster routing. This ensures that adding additional clusters later does not change the behavior of your existing applications”. Simply adding a second cluster does not change the default profile from single-cluster routing to multi-cluster routing. Since you need multi-cluster routing, you’re going to need a custom app profile. So C is correct. https://cloud.google.com/bigtable/docs/app-profiles#default-app-profile

   Source: Adapted from community discussion — dynamic_dba

   References

   • Google Cloud documentation
   • Google Cloud documentation

   Next question

3. Question 13 · 170 in bank

   Your organization operates in a highly regulated industry. Separation of concerns (SoC) and security principle of least privilege (PoLP) are critical. The operations team consists of: Person A is a database administrator. Person B is an analyst who generates metric reports. Application C is responsible for automatic backups. You need to assign roles to team members for Cloud Spanner. Which roles should you assign?

   • Aroles/spanner.databaseAdmin for Person A roles/spanner.databaseReader for Person B roles/spanner.backupWriter for Application C
   • Broles/spanner.databaseAdmin for Person A roles/spanner.databaseReader for Person B roles/spanner.backupAdmin for Application C
   • Croles/spanner.databaseAdmin for Person A roles/spanner.databaseUser for Person B roles/spanner databaseReader for Application C
   • Droles/spanner.databaseAdmin for Person A roles/spanner.databaseUser for Person B roles/spanner.backupWriter for Application C

   Show answer

   Answer: A

   Explanation

   A. C is wrong because databaseUser (Person B) would allow database writes and the question says generate metric reports, which would be read access only. databaseReader (Application C) doesn't allow backups. D is wrong because databaseUser (Person B) would allow database writes. That leaves A and B. Based upon Google's own documentation, it must be A. B would work, but backupAdmin for Application C would allow backup deletion as well as creation. backupWriter is described in the docs as "is intended to be used by scripts that automate backup creation". https://cloud.google.com/spanner/docs/iam

   Source: Adapted from community discussion — dynamic_dba

   References

   • Google Cloud documentation

   Next question

4. Question 14 · 170 in bank

   Your organization works with sensitive data that requires you to manage your own encryption keys. You are working on a project that stores that data in a Cloud SQL database. You need to ensure that stored data is encrypted with your keys. What should you do?

   • AExport data periodically to a Cloud Storage bucket protected by Customer-Supplied Encryption Keys.
   • BUse Cloud SQL Auth proxy.
   • CConnect to Cloud SQL using a connection that has SSL encryption.
   • DUse customer-managed encryption keys with Cloud SQL.

   Show answer

   Answer: D

   Explanation

   D. By stored data, we assume it means stored in the database so there’s no point in exporting it to encrypt it. A is wrong. Cloud SQL Auth Proxy is about secure connections to Cloud SQL, not encrypted data. B is wrong and so is C for the same reason. That leaves D. Just use CMEK when you create the instance.

   Source: Adapted from community discussion — dynamic_dba

   References

   • Google Cloud documentation
   • Google Cloud documentation
   • Google Cloud documentation

   Next question

5. Question 15 · 170 in bank

   Your team is building an application that stores and analyzes streaming time series financial data. You need a database solution that can perform time series-based scans with sub-second latency. The solution must scale into the hundreds of terabytes and be able to write up to 10k records per second and read up to 200 MB per second. What should you do?

   • AUse Firestore.
   • BUse Bigtable
   • CUse BigQuery.
   • DUse Cloud Spanner.

   Show answer

   Answer: B

   Explanation

   Financial data, such as transaction histories, stock prices, and currency exchange rates. https://cloud.google.com/bigtable/docs/overview#what-its-good-for With SSD: Reads - up to 10,000 rows per second Writes - up to 10,000 rows per second Scans - up to 220 MB/s https://cloud.google.com/bigtable/docs/performance#typical-workloads

   Source: Adapted from community discussion — chelbsik

   References

   • Google Cloud documentation
   • Google Cloud documentation

   Next question

6. Question 16 · 170 in bank

   You are designing a new gaming application that uses a highly transactional relational database to store player authentication and inventory data in Google Cloud. You want to launch the game in multiple regions. What should you do?

   • AUse Cloud Spanner to deploy the database.
   • BUse Bigtable with clusters in multiple regions to deploy the database
   • CUse BigQuery to deploy the database
   • DUse Cloud SQL with a regional read replica to deploy the database.

   Show answer

   Answer: A

   Explanation

   A. Bigtable would not be a valid choice. BigQuery lends itself to OLAP not OLTP. Cloud SQL could not scale to multiple regions. It’s Spanner. A.

   Source: Adapted from community discussion — dynamic_dba

   Next question

7. Question 17 · 170 in bank

   You are designing a database strategy for a new web application in one region. You need to minimize write latency. What should you do?

   • AUse Cloud SQL with cross-region replicas.
   • BUse high availability (HA) Cloud SQL with multiple zones.
   • CUse zonal Cloud SQL without high availability (HA).
   • DUse Cloud Spanner in a regional configuration.

   Show answer

   Answer: D

   Explanation

   A and B add latency. C screams of an anti-pattern to me (no HA). Spanner satisfies the requirement without sacrificing availability, so D.

   Source: Adapted from community discussion — DBAgain

   Next question

8. Question 18 · 170 in bank

   You are running a large, highly transactional application on Oracle Real Application Cluster (RAC) that is multi-tenant and uses shared storage. You need a solution that ensures high-performance throughput and a low-latency connection between applications and databases. The solution must also support existing Oracle features and provide ease of migration to Google Cloud. What should you do?

   • AMigrate to Compute Engine.
   • BMigrate to Bare Metal Solution for Oracle.
   • CMigrate to Google Kubernetes Engine (GKE)
   • DMigrate to Google Cloud VMware Engine

   Show answer

   Answer: B

   Explanation

   B. Oracle is neither licensed nor supported in GCE. The only platform which supports RAC and all existing Oracle features is BMS.

   Source: Adapted from community discussion — dynamic_dba

   Next question

9. Question 19 · 170 in bank

   You are choosing a new database backend for an existing application. The current database is running PostgreSQL on an on-premises VM and is managed by a database administrator and operations team. The application data is relational and has light traffic. You want to minimize costs and the migration effort for this application. What should you do?

   • AMigrate the existing database to Firestore.
   • BMigrate the existing database to Cloud SQL for PostgreSQL.
   • CMigrate the existing database to Cloud Spanner.
   • DMigrate the existing database to PostgreSQL running on Compute Engine.

   Show answer

   Answer: B

   Explanation

   B. Firestore is a NoSQL database. Eliminate A. You could migrate to Spanner leveraging the PostgreSQL dialect, but costs need to be minimized so that wouldn’t be the cheapest option. Especially since the load doesn’t justify Spanner. Again, you could migrate like-for-like to a GCE VM, but that defeats minimizing the migration effort. The cheapest and easiest way to migrate would be Database Migration Service to Cloud SQL for PostgreSQL.

   Source: Adapted from community discussion — dynamic_dba

   Next question

10. Question 20 · 170 in bank

    Your organization is currently updating an existing corporate application that is running in another public cloud to access managed database services in Google Cloud. The application will remain in the other public cloud while the database is migrated to Google Cloud. You want to follow Google-recommended practices for authentication. You need to minimize user disruption during the migration. What should you do?

    • AUse workload identity federation to impersonate a service account.
    • BAsk existing users to set their Google password to match their corporate password.
    • CMigrate the application to Google Cloud, and use Identity and Access Management (IAM).
    • DUse Google Workspace Password Sync to replicate passwords into Google Cloud.

    Show answer

    Answer: A

    Explanation

    A. Updating passwords represents user disruption. Eliminate B. Eliminate C for the same reason. D doesn’t make sense, leaves A. From Google’s documentation, “Traditionally, applications running outside Google Cloud can use service account keys to access Google Cloud resources. However, service account keys are powerful credentials, and can present a security risk if they are not managed correctly. With identity federation, you can use Identity and Access Management (IAM) to grant external identities IAM roles, including the ability to impersonate service accounts. This approach eliminates the maintenance and security burden associated with service account keys.” https://cloud.google.com/iam/docs/workload-identity-federation

    Source: Adapted from community discussion — dynamic_dba

    References

    • Google Cloud documentation

    Next question

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

PCDBE practice, question 11. 170 questions available.