Examkin

Question

Difficulty: Very hardDesign Storage and Non-Relational Data HA and DR

VeloMed Solutions is designing a storage architecture for patient medical images and session recordings. The primary storage region is West US 3, and the secondary region is East US. The solution must meet the following requirements:
- The primary storage must tolerate the failure of a single Availability Zone in West US 3 with zero data loss (RPO=0RPO = 0) and immediate availability (RTO=0RTO = 0).
- If a catastrophic regional disaster occurs in West US 3, read access to existing images must be available in the secondary region with an RTO of less than 15 minutes.
- The maximum acceptable data loss (RPO) for files replicated to the secondary region is 15 minutes.
- Write operations must be restored in the secondary region with an RTO of less than 1 hour during a regional disaster.
- Storage costs must be minimized while meeting all recovery objectives.

Which of the following configurations and disaster recovery procedures should the cloud architect recommend to satisfy these requirements? (Select TWO.)

  1. Deploy the storage account using Read-Access Geo-Zone-Redundant Storage (RA-GZRS).Answer
  2. Initiate a customer-managed failover of the storage account to the secondary region if write capabilities must be restored during a regional outage.Answer
  3. C
    Configure the application to perform write operations directly to the secondary storage endpoint if the primary endpoint is unavailable.
  4. D
    Deploy the storage account using Read-Access Geo-Redundant Storage (RA-GRS).
  5. E
    Deploy the storage account using Geo-Zone-Redundant Storage (GZRS) and rely on Microsoft-managed failover to restore read access to the secondary region within 15 minutes.

Answer

The storage account must be configured with Read-Access Geo-Zone-Redundant Storage (RA-GZRS), and a customer-managed failover must be initiated to the secondary region to restore write capabilities during a regional disaster.
The solution requires high availability across availability zones in the primary region (ruling out RA-GRS) and immediate read availability in the secondary region in under 15 minutes (ruling out standard GZRS). Deploying the storage account with Read-Access Geo-Zone-Redundant Storage (RA-GZRS) satisfies these parameters. Furthermore, because write operations cannot be routed to a secondary geo-redundant endpoint under normal or outage states without changing the account configuration, initiating a customer-managed failover is the only valid way to restore write capabilities within the 1-hour RTO constraint.

Step-by-Step Solution

1
Analyze the primary region high availability requirements.
The requirement for zero data loss and immediate availability during a zone outage in West US 3 necessitates zone redundancy in the primary region (ZRS or GZRS/RA-GZRS). Locally redundant options (LRS/GRS/RA-GRS) are ruled out.
LRS does not replicate data across availability zones.
2
Analyze the disaster recovery read requirements.
The 15-minute RTO for reads during a regional outage requires immediate read availability in the secondary region. Standard GZRS does not expose a readable secondary endpoint unless a failover is completed, which would exceed the time limit. Therefore, Read-Access Geo-Zone-Redundant Storage (RA-GZRS) is required.
RA-GZRS provides a secondary read-only endpoint that is immediately available.
3
Analyze the write restoration requirements.
Writing directly to a secondary geo-redundant endpoint is impossible. To restore writes in the secondary region within 1 hour, a customer-managed failover must be triggered, which changes the DNS to point writes to the secondary region (converting the storage account to LRS in the process).
Only a failover can grant write access to the replicated data in the secondary region.

Key Concept

Azure Storage Replication and Customer-Managed Failover Mechanics
Rate this question

Topics

Design Storage and Non-Relational Data HA and DR

More questions from this topic