Sample questions
Gcp Databases Cloudsql Spanner FirestoreDifficulty 1
What is Cloud SQL?
- aA serverless NoSQL document database
- bA fully managed relational database service supporting MySQL, PostgreSQL, and SQL Server engines✓
- cA managed message queue for asynchronous processing
- dA distributed file storage system for unstructured data
Explanation:Cloud SQL is Google Cloud's fully managed relational database service — it runs standard MySQL, PostgreSQL, or SQL Server engines while Google handles patching, backups, and replication infrastructure. It is not a NoSQL document store (a), a queue (c), or a file storage system (d).
Gcp Databases Cloudsql Spanner FirestoreDifficulty 2
A team enables High Availability (HA) on a Cloud SQL instance. What does this configuration actually provide?
- aAn additional read-only replica in the same zone for load balancing reads
- bAutomatic sharding of the database across multiple regions
- cA standby instance with synchronously replicated data in a different zone of the same region✓
- dA backup copy of the database stored in Cloud Storage, restored manually on demand
Explanation:Cloud SQL HA (regional availability) provisions a standby instance in a different zone within the same region, keeping data synchronously replicated via the underlying regional persistent disk; Cloud SQL automatically fails over to the standby if the primary becomes unavailable. It is not a read replica (a, which is asynchronous and for read scaling), does not span multiple regions automatically (b), and is not a manually restored backup (d).
Gcp Databases Cloudsql Spanner FirestoreDifficulty 2
How does a Cloud SQL Read Replica relate to its primary instance?
- aIt asynchronously replicates data from the primary and can serve read traffic✓
- bIt synchronously mirrors every write before the primary acknowledges it, exactly like an HA standby
- cIt is promoted automatically and transparently whenever the primary instance restarts
- dIt only exists within Cloud Spanner instances, not Cloud SQL
Explanation:A Cloud SQL Read Replica replicates asynchronously from the primary, so it can lag slightly behind, and it is meant to offload read traffic rather than provide synchronous durability like an HA standby (b describes HA, not a read replica). Promotion to a standalone writable instance is a manual operation, not automatic on every restart (c). Read Replicas are a Cloud SQL feature, not exclusive to Spanner (d).
Gcp Databases Cloudsql Spanner FirestoreDifficulty 2
What must be enabled on a Cloud SQL instance for point-in-time recovery (restoring to an arbitrary moment, not just a daily backup snapshot) to be possible?
- aAutomated backups alone are sufficient; no other setting is required
- bA read replica must be attached to the instance
- cHigh Availability must be enabled
- dTransaction log retention (binary logging for MySQL, write-ahead log archiving for PostgreSQL) must be enabled alongside automated backups✓
Explanation:Point-in-time recovery replays transaction logs on top of a base backup to reach an arbitrary moment in time, so it requires transaction log retention (binlog for MySQL, WAL archiving for PostgreSQL) in addition to automated backups — daily backups alone (a) only let you restore to backup times, not arbitrary moments. HA (c) and read replicas (b) address availability and read scaling, not the ability to recover to a specific timestamp.
Gcp Databases Cloudsql Spanner FirestoreDifficulty 1
What is the defining characteristic of Cloud Spanner compared to a typical single-region relational database?
- aIt only supports NoSQL key-value access patterns, not SQL
- bIt is a horizontally scalable, globally distributed relational database with strong external consistency✓
- cIt is a caching layer placed in front of Cloud SQL to speed up reads
- dIt requires manual application-level sharding logic to scale beyond a single node
Explanation:Cloud Spanner's defining feature is horizontal scalability combined with strong, externally consistent transactions across a globally distributed deployment — a combination that is rare among relational databases, which typically trade off consistency or scale. It fully supports SQL (a is wrong), is not a caching layer (c), and scales automatically without requiring the application to implement its own sharding (d).
Gcp Databases Cloudsql Spanner FirestoreDifficulty 2
What is TrueTime's role in Cloud Spanner?
- aIt is a caching mechanism that stores recently read rows to reduce read latency
- bIt is the query optimizer that chooses execution plans for SQL statements
- cIt is a globally synchronized clock API with bounded uncertainty that Spanner uses to assign commit timestamps✓
- dIt is the name of Spanner's row-key based physical storage layout
Explanation:TrueTime is Google's globally synchronized clock API (backed by atomic clocks and GPS, exposing a bounded uncertainty interval) that Spanner uses to assign transaction commit timestamps such that transactions can be globally, externally ordered — this is the core mechanism behind Spanner's strong consistency guarantees across regions. It is not a cache (a), query optimizer (b), or storage layout (d).