COF-C03 Question 311
Select 2Data replication and failoverA company runs a business-critical application on Snowflake in AWS us-east-1 and needs a disaster recovery design that minimizes recovery time if the primary region becomes unavailable. The security team also requires that user authentication and role assignments continue to work after failover with minimal manual reconfiguration. Which TWO actions should the Snowflake architect implement?
- A
Configure account replication to a secondary account in another region and include failover of account objects such as users, roles, and grants.
- B
Create database replication only for the application database, because replicated databases automatically include account-level security objects needed for login after failover.
- C
Promote the secondary replicated account during a regional outage so it becomes writable and can serve production workloads.
- D
Use Time Travel on the primary account as the disaster recovery solution, because Time Travel can recover data even if the entire region is unavailable.
- E
Rely on reader accounts in the secondary region, because reader accounts can be failed over to become the primary production account.
Show answer and explanation
Correct answers: A, C
Explanation
The best solution is to use Snowflake replication and failover capabilities at the account level when the requirement includes both data availability and continuity of account-level security configuration. Replicating only databases may protect application data, but it does not fully address authentication and role-management needs after a regional outage. In practice, architects should replicate to a secondary account in another region or cloud-supported topology, keep replication current, and then promote the secondary during an outage. Snowflake documentation distinguishes database replication from account replication and explains that failover groups and account replication are used for broader disaster recovery scenarios, including supported account objects. Time Travel and Fail-safe are data recovery features, but they are not substitutes for cross-region disaster recovery.
- A. Correct.
Correct. For regional disaster recovery with minimal recovery time, Snowflake supports replication and failover/failback between accounts. Account replication is designed to replicate supported account objects such as users, roles, and grants in addition to other supported metadata, which helps preserve authentication and authorization configuration after failover. This is more appropriate than replicating only individual databases when the requirement includes account-level security continuity.
- B. Incorrect.
Incorrect. Database replication replicates the selected databases and supported objects within them, but it does not automatically replicate account-level objects such as users, roles, and related grants required for authentication and broader access control at the account level. This is a common misconception when teams focus only on data and overlook account object continuity.
- C. Correct.
Correct. In a disaster recovery scenario, the replicated secondary account can be promoted during failover so it becomes writable and takes over production activity. This is the key operational step that allows business continuity when the primary region is unavailable.
- D. Incorrect.
Incorrect. Time Travel helps recover historical versions of data within Snowflake, but it is not a regional disaster recovery mechanism. If the primary region is unavailable, Time Travel in that region does not provide access from another region. It addresses logical recovery scenarios, not cross-region outage failover.
- E. Incorrect.
Incorrect. Reader accounts are intended for data sharing to consumers who do not have their own Snowflake account. They are not a disaster recovery mechanism for promoting a standby production environment. They cannot be used as a replacement for account replication and failover.