Verify multidisk backup coordination before striping an AVS guest filesystem

Source facts

Microsoft’s Azure VMware Solution guidance warns that, when a volume is striped across disks, the backup or disaster-recovery product must support backing up multiple virtual disks simultaneously. The filesystem needs the participating disks frozen during the snapshot or backup.

For a single VM, the source contrasts separate VMDK filesystems with storage-management tools that stripe across VMDKs. Its examples include LVM, MSSQL Filegroups, and Oracle ASM. Microsoft Learn.

Applicability

Map the guest filesystem or application data layout to its VMDKs and datastores. Review the exact backup product’s support for that arrangement before adopting striping. Do not treat performance examples from the source as a guarantee for the proposed workload.

DSE recommendation

DSE recommends making the protection review a prerequisite to the layout decision. Have the virtualization, storage, application, and backup owners identify the complete participating disk set and how a consistent point is captured. If that coordination cannot be established, reconsider the proposed layout instead of accepting an untested recovery dependency for a performance target.

Verification

Use an isolated representative guest to exercise the approved coordinated backup and restore procedure. Confirm every member disk belongs to the intended recovery point, then have the application owner validate the restored dataset. Preserve the disk-to-datastore map and coordination evidence; separate per-disk job successes are not the agreed test of the striped filesystem.

Official references

Microsoft Learn: Azure VMware Solution datastore performance considerations for Azure NetApp Files. Source retrieved September 9, 2026.