Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
A financial services company operates a write-heavy Cloud SQL for MySQL primary instance in us-central1. The team needs to deploy five read replicas in us-east1 to support regional analytics and disaster recovery (DR). The database engineers have the following requirements:
us-central1 experiences a regional outage.Which replication architecture and DR strategy should the team implement?
us-east1, and disable binary logging and point-in-time recovery (PITR) on the primary instance to eliminate disk fsync write overheadfsync overhead areIn Cloud SQL for MySQL, binary logging (which is coupled with point-in-time recovery, or PITR) records data modifications so they can be replayed for recovery or sent to replicas. Cloud SQL enforces row-based replication with sync_binlog=1 and innodb_support_xa=true, which requires an additional disk fsync per write operation and introduces measurable write performance overhead on the primary instance.
sync_binlog cannot be relaxed on the primary, it can be adjusted on a read replica if binary logging is enabled there—though it automatically resets to 1 if that replica is promoted to a standalone instance.The workload did not require read replicas or PITR at all, and you were deleting all replicas specifically to reclaim write performance on a standalone primary instance.
us-east1 with four child replicas underneath it, and manually promote the top-level us-east1 replica during a regional outageCascading read replicas allow you to build a multi-level replication hierarchy in which a read replica serves as the replication source for additional downstream read replicas. In Cloud SQL for MySQL, you can create up to four levels of replicas (including the primary instance), and a single parent replica can support up to eight sibling replicas underneath it.
us-central1 streaming binary log changes independently to five separate cross-region targets, it replicates data to only a single top-level cascading replica in us-east1. That parent replica offloads the replication work by forwarding changes to the four child replicas.us-central1 and the parent replica in us-east1. Replication from the parent replica to the four child replicas stays within us-east1, where in-region network transfer is free.us-central1 fails and you manually promote the top-level replica in us-east1 to a standalone writable primary instance, its entire downstream hierarchy is preserved. All four child replicas continue replicating from the newly promoted primary automatically without needing to be deleted and recreated.Compared to attaching five direct cross-region replicas to the primary, a cascading hierarchy cuts cross-region egress and replication fan-out by 80% while eliminating the lengthy process of deleting and recreating sibling replicas after a failover.
us-east1 attached to the primary instance, and promote one replica during an outage so the remaining four automatically replicate from itDirect read replicas connect straight to the primary instance and asynchronously apply changes from the primary's binary logs. While they effectively offload read queries and analytics traffic, each direct replica maintains its own replication stream from the primary.
us-central1 to service five separate replication streams and send five identical copies of the transaction stream across regions, increasing primary overhead and cross-region network egress costs.us-east1, administrators must manually delete the four orphaned replicas (as well as the failed primary when accessible) and create four brand-new read replicas with different names from the newly promoted primary.You only needed a single cross-region replica in us-east1 for DR and had no requirement to preserve multiple downstream read replicas immediately upon promotion.
us-east1, and rely on HA to automatically promote one replica during a regional outageCloud SQL High Availability (HA) configures a regional deployment consisting of a primary instance in one zone and a passive standby instance in a secondary zone within the same region, backed by synchronous disk replication. If the primary zone fails, Cloud SQL automatically fails over to the standby instance within roughly 60 seconds.
us-central1, not a full regional outage. Read replicas are completely separate from the HA standby instance and never fail over automatically during an outage.us-central1) becomes unavailable, both the primary and its in-region standby instance are affected.The requirement was to provide automatic failover for a zonal outage within us-central1 rather than cross-region DR for a full regional outage.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated