For an Arc-enabled SQL MI, there are four storage classes that can be configured. Pay special attention to the backup location. Our PowerStore environment provided us the unique opportunity to present RWX (see table below) CLI access to our Arc-enabled SQL MIs. During an Always On Group failover, the new primary replica can then continue to write to the same backup location, solving for any data loss. In addition, PowerStore provides appliance level data protection with the Dynamic Resiliency Engine (DRE). DRE is a software-based approach to data redundancy that is more distributed, automated, and efficient than traditional RAID. The burden of selecting and balancing RAID levels is a thing of the past.
The –storage-class-x parameters in the Azure CLI command aligned with the SQL Server files, as shown in the following table.
Table 3. Storage class
--storage-class-data | SQL data .mdf files | RWO |
--storage-class-datalogs | SQL log .ldf files | RWO |
--storage-class-backups | SQL Server backup files. Full, Diff, T-log | RWX |
--storage-class-logs | SQL agent, errorlog, trace files, health logs, etc | RWO |
Kubernetes data access modes
Table 4. Kubernetes access modes
RWO | Read Write Once |
ROX | Read Only Many |
RWX | Read Write Many |
RWOP | Read Write Once Pod |
With PowerStore, we can have block and file storage configurations within the same appliance, which allows for NFS backup. With this option, we had the perfect location for the RWX –storage-class-backup requirement, all of this within a single PowerStore solution. PowerStore allows multiprotocol file access, which enables data mobility by allowing backup files from any version of SQL Server, to be accessed by NFS or SMB clients. This makes it easy to move data between traditional SQL Server instances and an Arc-enabled SQL MI.
As mentioned above, Arc-enabled SQL MI Business Critical deploys out of the box replica groups, spread across multiple Kubernetes nodes if available. This provides resiliency at many layers.
