ARA-C01 Question 249
Single answerDisaster recoveryA global retailer runs a business-critical Snowflake production account in AWS us-east-1. The company must implement disaster recovery so that if the primary region becomes unavailable, reporting workloads can be resumed in AWS us-west-2 within 30 minutes, with minimal data loss. The architecture team wants to avoid maintaining duplicate ETL pipelines in the secondary region and wants a solution that can be regularly tested without disrupting production. Which approach best meets these requirements?
- A
Create a reader account in AWS us-west-2 and use database replication from the primary account. During a regional outage, direct reporting users to the reader account.
- B
Configure cross-region database replication and failover groups from the primary account to a secondary account in AWS us-west-2, include the required databases and account objects, and periodically promote the secondary during DR tests.
- C
Use Time Travel and Fail-safe on the production databases, then restore the databases into a new account in AWS us-west-2 when a disaster occurs.
- D
Unload critical tables to external stage files in S3 with Snowpipe enabled in AWS us-west-2, and rebuild the reporting environment from staged files during an outage.
Show answer and explanation
Correct answer: B
Explanation
For Snowflake disaster recovery across regions, the recommended pattern is to use account replication features, specifically database replication where appropriate and failover groups when account objects must also be protected. In real-world DR scenarios, simply copying data is often insufficient; workloads also depend on users, roles, warehouses, and related configuration. Failover groups are designed to support this broader recovery need and enable controlled failover/failback between accounts in different regions or cloud regions, subject to feature support and edition requirements. This approach aligns with the scenario's goals: low RPO/RTO relative to manual rebuild methods, no need to maintain duplicate ETL pipelines, and support for repeatable DR testing. By contrast, Time Travel and Fail-safe are for data recovery and historical restoration, not regional DR. Reader accounts support data sharing use cases rather than serving as a complete DR environment. Best practice is to define replication/failover scope carefully, validate object support, schedule replication to meet recovery objectives, and regularly test failover procedures.
- A. Incorrect.
Incorrect. Reader accounts are designed for data sharing scenarios and are not a full disaster recovery mechanism for an organization's own production environment. They do not address replication and failover of broader account-level objects needed to resume workloads, such as users, roles, warehouses, resource monitors, and integration-related configuration. This option also assumes database replication alone is sufficient, which can leave operational gaps during failover.
- B. Correct.
Correct. Cross-region replication combined with failover groups is the Snowflake-native DR approach for this scenario. Replication can keep databases synchronized to a secondary account in another region, and failover groups extend protection to supported account objects such as users, roles, warehouses, resource monitors, and other metadata required for rapid recovery. This minimizes data loss based on replication frequency and avoids building duplicate ETL pipelines because the replicated objects are maintained by Snowflake. Regular DR testing is supported by promoting/failing over to the secondary and then failing back after validation, assuming the required objects are included and the replication topology is properly designed.
- C. Incorrect.
Incorrect. Time Travel and Fail-safe protect against logical data issues and provide historical recovery within the same Snowflake account context; they are not a substitute for cross-region disaster recovery. They do not enable recovery into another region after a regional outage in the way required here. In particular, Fail-safe is a Snowflake-managed recovery mechanism and is not intended for customer-driven operational DR or to meet a 30-minute regional failover objective.
- D. Incorrect.
Incorrect. Unloading data to S3 and reconstructing the environment manually in another region creates significant operational overhead and generally does not meet a 30-minute recovery objective for a business-critical reporting platform. Snowpipe is for continuous data ingestion, not for preserving and failing over full Snowflake environments. This approach also does not replicate account objects or provide a straightforward, testable failover process.