ADA-C01 Question 504
Select 2Domain 6.0: Disaster Recovery, Backup, and Data Replication (9%)A company runs a business-critical Snowflake deployment in AWS us-east-1 and must meet a disaster recovery objective that allows analytics workloads to be resumed in another region with minimal administrative effort. The source account contains several databases, including one shared from a different provider account and one database with an external stage that points to an S3 bucket in us-east-1. The Snowflake administrator configures account replication to a secondary account in AWS us-west-2 and performs regular refreshes. During a regional outage, the team plans to fail over to the secondary account. Which TWO statements correctly describe what the administrator should expect after failover?
- A
Primary databases that are included in the replication configuration can become writable in the secondary account after failover.
- B
The shared database that originated from another provider account will be fully replicated and available for failover in the secondary account.
- C
The external stage definition may exist after replication, but access depends on the target region having connectivity and appropriate access to the referenced S3 location.
- D
Virtual warehouses from the primary account will be replicated so compute capacity is immediately available after failover without additional setup.
- E
Historical query results cache from the primary account will be replicated so identical queries in the failover account can reuse the same cached results.
Show answer and explanation
Correct answers: A, C
Explanation
Snowflake disaster recovery relies on replication and failover capabilities for supported account and database objects, but administrators must distinguish between replicated metadata/data and non-replicated operational components. After failover, replicated databases become writable in the secondary account, enabling business continuity. However, inbound shared databases are not treated as fully owned primary databases for replication/failover, so they should not be assumed to appear as failover-ready copies. External stages require special attention because Snowflake can replicate object definitions, but the underlying external storage remains in the cloud provider account and region originally configured; access, network reachability, and IAM permissions must still work from the failover environment. Also, warehouses and result cache are not replicated, so DR planning should include provisioning compute in the target account and validating dependencies on external storage. These expectations align with Snowflake documentation on replication and failover, limitations for shared data, and the behavior of external stages and compute resources in cross-region disaster recovery designs.
- A. Correct.
Correct. With database or account replication and failover groups, replicated databases in the secondary account are read-only until failover occurs. After failover/promoting the secondary, the replicated objects become writable in the failover target, which is the core mechanism used to resume operations during a disaster recovery event.
- B. Incorrect.
Incorrect. A database that is itself obtained through data sharing from another account is not replicated as if it were a primary database owned by the source account. Replication covers eligible objects owned in the source account, but inbound shared databases are not fully replicated for failover in this manner. This is a common misconception when organizations mix native shares with replicated databases.
- C. Correct.
Correct. Replication can include metadata for supported objects, but external stages depend on the external cloud storage location and the related credentials/integration setup being usable from the failover environment. Because the stage points to an S3 bucket in us-east-1, the administrator must consider cross-region access, permissions, and whether the bucket remains reachable during the outage. Replication does not magically relocate the external data.
- D. Incorrect.
Incorrect. Virtual warehouses are compute resources and are not replicated as part of database/account replication for disaster recovery. In the failover account, warehouses must be created or otherwise made available separately. Candidates often confuse replicated data-plane metadata with compute infrastructure.
- E. Incorrect.
Incorrect. Persisted query results and cache state are not replicated for disaster recovery. Even if the same SQL is executed after failover, result reuse depends on conditions in the target account at that time, not on cache copied from the source account. Query history and execution context are not a substitute for replicated databases.