Examkin

Soru

Zorluk: OrtaDesign Storage and Non-Relational Data HA and DR

AquaShield Water Solutions is designing a disaster recovery solution for an application that processes real-time telemetry from water treatment facilities. The telemetry logs are written continuously to an Azure Blob Storage account in the East US region. The solution must meet the following requirements:
- During normal operations, the application must read and write data in the primary region.
- If a disaster disables the primary region, the application must be able to read existing telemetry logs immediately from a secondary region.
- The Recovery Point Objective (RPO) for data loss must be less than 15 minutes.
- If the primary region remains unavailable for an extended period, the application must be able to write new logs to the secondary region.
- Storage costs must be minimized.

Which storage configuration and disaster recovery strategy should you recommend?

  1. A
    Use Locally Redundant Storage (LRS) in the primary region. Configure Azure Backup to perform cross-region restores of the blobs to a separate storage account in the secondary region every 4 hours.
  2. B
    Use Read-Access Geo-Redundant Storage (RA-GRS). Configure the application to write directly to the secondary storage endpoint immediately if the primary region is unavailable.
  3. Use Read-Access Geo-Redundant Storage (RA-GRS). Configure the application to read from the secondary endpoint during a primary outage, and initiate a customer-managed failover to enable write operations in the secondary region if the outage is prolonged.Cevap
  4. D
    Use Geo-Redundant Storage (GRS). Configure a lifecycle management policy to archive all blobs to the Archive tier in the secondary region to minimize costs, and rehydrate them if a primary regional outage occurs.

Cevap

Use Read-Access Geo-Redundant Storage (RA-GRS). Configure the application to read from the secondary endpoint during a primary outage, and initiate a customer-managed failover to enable write operations in the secondary region if the outage is prolonged.
Read-Access Geo-Redundant Storage (RA-GRS) with a customer-managed failover strategy meets all requirements. RA-GRS provides a read-only secondary endpoint that is immediately readable during a primary regional outage, achieving a read RTO of 0. GRS/RA-GRS replication is asynchronous and typically has a replication lag of under 15 minutes, meeting the 15-minute RPO. Initiating a customer-managed failover promotes the secondary endpoint to primary, which enables write operations during a prolonged outage, all while maintaining lower storage costs than multi-region active-write configurations.

Adım Adım Çözüm

1
Identify the read and write requirements during a regional outage.
The solution requires immediate read access to existing data (RTO of 0 for reads) and the ability to write to the secondary region during a prolonged outage.
This filters out replication options that do not support read access to the secondary region prior to failover, and options that cannot accept write traffic post-failover.
2
Evaluate storage replication options against the RPO and read access requirements.
LRS is eliminated as it lacks regional redundancy. GRS alone does not allow read access prior to failover. RA-GRS provides a read-only endpoint in the secondary region with an asynchronous replication lag of under 15 minutes, satisfying the 15-minute RPO and the immediate read requirement.
Only RA-GRS supports immediate read access to the secondary region during a primary outage.
3
Determine the mechanism for enabling write operations in the secondary region during a prolonged outage.
Initiating a customer-managed failover on the RA-GRS account promotes the secondary region to the primary endpoint, enabling write operations.
Writing directly to the RA-GRS secondary endpoint is not permitted as it is read-only; a failover must be triggered to allow write access.

Anahtar Kavram

Azure Storage replication redundancy options (RA-GRS) and the customer-managed failover mechanism to achieve HA/DR for non-relational storage.
Bu soruyu puanla

Konular

Design Storage and Non-Relational Data HA and DR

Bu konudan daha fazla soru