AZ-305 Question 143
Select 2A retail company hosts product images (unstructured data) in Azure Blob Storage and product catalog documents (semi-structured JSON data) in Azure Cosmos DB. They require continuous read/write availability even if an entire Azure region becomes unavailable. Which two strategies will fulfill the high availability requirement for both data types?
- A
A. Configure the Blob Storage account for Locally Redundant Storage (LRS) and enable a single write region in Azure Cosmos DB.
- B
B. Configure the Blob Storage account for Geo-Redundant Storage (GRS) and enable multi-region writes in Azure Cosmos DB.
- C
C. Configure the Blob Storage account for Read-Access Geo-Redundant Storage (RA-GZRS) and enable multi-region writes in Azure Cosmos DB.
- D
D. Configure the Blob Storage account for Zone-Redundant Storage (ZRS) paired with a manual failover region for Cosmos DB.
- E
E. Configure the Blob Storage account with Read-Access Geo-Redundant Storage (RA-GRS) and use a single read region for Cosmos DB.
Show answer and explanation
Correct answers: B, C
Explanation
For unstructured data, GRS or RA-GZRS ensures replication across geographically separated regions, enabling continued data access in the event a primary region goes offline. For semi-structured data, enabling multi-region writes in Azure Cosmos DB provides high availability for both reads and writes during a regional outage. Microsoft documentation recommends using geo-replicated storage configurations and multi-region write strategies in Cosmos DB for end-to-end resiliency (see Azure Storage redundancy documentation and the Cosmos DB global distribution best practices).
- A. Incorrect.
A. Incorrect. LRS provides high availability within a single data center, which does not protect against a region-wide outage. A single write region in Cosmos DB also does not provide continuous write availability if the primary region goes down.
- B. Correct.
B. Correct. GRS replicates data across two geographically separated regions. Combined with multi-region writes in Cosmos DB, this approach can ensure that both storage and database remain available for writes and reads even if the primary region fails.
- C. Correct.
C. Correct. RA-GZRS extends GRS by adding zone redundancy, and the read access feature lets you continue reading from the secondary endpoint in case of a primary region outage. Paired with multi-region writes in Cosmos DB, this ensures continuous read/write availability for both unstructured and semi-structured data.
- D. Incorrect.
D. Incorrect. ZRS only provides durability across availability zones within the same region. Manual failover for Cosmos DB could restore functionality, but there could be downtime and data might not remain instantly writable during an entire region outage.
- E. Incorrect.
E. Incorrect. RA-GRS provides read access to secondary replicas for Blob Storage, but using a single read region in Cosmos DB means you only have true read redundancy, not continuous write availability in the event of a region failure.