Microsoft recently announced the public preview of a new failover type for Azure Storage with customer managed planned failover. This new failover type allows a storage account to failover while maintaining geo-redundancy, with no data loss or additional cost.
Earlier, the company provided unplanned customer managed failover as a disaster recovery solution for geo-redundant storage accounts. This has allowed their customers to fulfill their business needs for disaster recovery testing and compliance. Furthermore, another available failover type planned (Microsoft managed) is available. With the new customer managed planned failover type, users are no longer required to reconfigure geo-redundant storage (GRS) after their planned failover operation.
In a question on Stack Overflow, if Azure Storage failover and failback need two replications, the company replied hinting at the new failover type:
Yes, two replications would be required, and the process would work as documented. We currently have Planned Failover in private preview, allowing customers to test the Failover workflow while keeping geo-redundancy. As a result, there is no need to re-enable geo-redundancy after each failover and failback operation. This table highlights the key differences between Customer Managed Failover (currently GA) and Planned Failover (currently in the planned state).
If the storage service endpoints for the primary region become unavailable, users can fail over their entire geo-redundant storage account to the secondary region. During failover, the original secondary region becomes the new primary region, and all storage service endpoints are then redirected to the new primary region. After the storage service endpoint outage is resolved, users can perform another failover operation to fail back to the original primary region.
(Source: Microsoft Learn)
There are various scenarios where Planned Failover can be used according to the company:
- Planned disaster recovery testing drills to validate business continuity and disaster recovery.
- Recovering from a partial outage in the primary region where storage is unaffected. For example, suppose storage service endpoints are healthy in both regions, but another Microsoft or 3rd party service faces an outage in the primary region. In that case, users can fail over their storage services. In this scenario, once users fail over the storage account and all other services, their workload can continue to work.
- A proactive solution in preparation for large-scale disasters that may impact a region. To prepare for a disaster such as a hurricane, users can leverage Planned Failover to fail over to their secondary region and then fail back once things are resolved.
AWS and Google Cloud's competitive storage offerings offer similar disaster recovery options, such as multiple active regions (AWS) and dual/multiple buckets (GCP).
In a recent tweet emphasizing that backups are essential, Fletus Poston tweeted:
You likely know having a backup is essential—but it’s not enough. A backup that hasn't been validated could leave your business vulnerable when disaster strikes. Whether it's a ransomware attack, hardware failure, or accidental deletion, relying on untested backups can lead to incomplete or corrupted data recovery.
The new failover type is available in Southeast Asia, East Asia, France Central, France South, India Central, and India West Azure regions.