ADA-C01 Question 519
Single answerDifferentiate between replication Groups and failover groupsA global company runs a Snowflake application in AWS us-east-1 and must improve disaster recovery for a business-critical workload. The DR region will be AWS us-west-2. The platform team must meet these requirements: (1) customer-facing applications must be able to continue using the same account-level objects such as users, roles, and warehouses after a regional outage, and (2) several read-only reference databases must also be copied to the DR region for reporting, but those databases do not need account failover. Which Snowflake design best satisfies these requirements?
- A
Create a replication group for users, roles, warehouses, and the reference databases, because replication groups can replicate both account objects and databases and can be promoted during failover.
- B
Create a failover group for the account-level objects that must be available after account failover, and use a separate replication group for the read-only reference databases that only need cross-region replication.
- C
Create only a failover group that includes all required account objects and all databases, because failover groups are the standard mechanism for every cross-region database replication requirement.
- D
Create database replication for the reference databases and manually recreate users, roles, and warehouses in the DR account, because failover groups cannot replicate account objects across regions.
Show answer and explanation
Correct answer: B
Explanation
The key distinction is purpose: replication groups are for replicating supported objects, commonly when you want secondary copies for availability, reporting, or staged recovery workflows, while failover groups are for business continuity/disaster recovery scenarios where you need to fail over supported account objects and databases to a secondary account so workloads can continue running. In this scenario, the requirement to preserve and continue using account-level objects such as users, roles, and warehouses after a regional outage points directly to a failover group. At the same time, databases that only need to be copied to another region for read-only reporting do not necessarily require failover semantics, so a separate replication group is appropriate. This aligns with Snowflake documentation on replication and failover, which distinguishes replication groups from failover groups based on whether account failover capability is required.
- A. Incorrect.
Incorrect. Replication groups are used to replicate certain account-level objects and some related objects for business continuity scenarios, but they are not the mechanism that provides account failover semantics for continuing operations with replicated account objects after a regional outage. The option also incorrectly implies replication groups are the right choice when the requirement is to fail over account-level objects such as users, roles, and warehouses for continued application use.
- B. Correct.
Correct. Failover groups are designed for business continuity and disaster recovery when account objects must be failed over to another account/region so workloads can continue operating. They can include supported account objects such as users, roles, warehouses, resource monitors, and databases. If some databases only need replication for read-only or reporting use cases and do not need account failover behavior, a separate replication group is appropriate for those databases. This design cleanly separates failover needs from simple replication needs.
- C. Incorrect.
Incorrect. This overgeneralizes failover groups. Although failover groups are the correct choice for objects that must participate in failover, they are not the required mechanism for every database replication use case. Snowflake also supports replication groups for database replication scenarios where failover of account-level objects is not needed. Using only a failover group for everything may not reflect the stated requirement to differentiate simple database replication from account failover.
- D. Incorrect.
Incorrect. This reflects a common misconception. Failover groups are specifically intended to replicate supported account objects across regions/clouds for disaster recovery. Manually recreating users, roles, and warehouses would increase operational risk and would not meet the requirement to continue using the same account-level objects after an outage.