I’m trying to design a multi-region architecture in Azure, with automatic failover in the event of a region-wide outage. I’m using Azure Kubernetes Service. One of the key applications uses a file system back-end (not something I can change). Since Azure Files is listed as one of the ReadWriteMany volume types, this is what I’m using for my AKS volumes to allow multiple instances of this container to read/write simultaneously.

What I need to figure out is an architecture that allows failing to a secondary region with a hot deployment. Ideally, it would be an active/active architecture with load balancing across the two regions so there’s immediate and automatic failover, but I don’t see a way to achieve that with Azure Files. While Azure Files supports GZRS, the secondary region is only for disaster recovery and can’t be used for writing. I cannot find any documentation on how the secondary region would be promoted to the “primary” region to allow writing.

I’m coming from the AWS world, so Azure in general is fairly new to me. How would one design this to allow for a (preferably automatic) failover to the secondary region with Azure Files for K8s volumes?