Map every restored disk before rebuilding a VM from its restore point

Source facts

An Azure VM restore point contains individual disk restore-point identifiers. Microsoft’s recovery procedure creates disks from those identifiers, then creates a new VM and attaches the restored disks. A target disk can be created in a different zone from its disk restore point. When the VM restore point was copied between regions, every constituent disk copy must finish before the VM can be restored. Microsoft Learn.

Applicability

Start with a selected restore point and an approved replacement-VM recovery plan. Identify the intended recovery time, every required disk, and the target region and zone. Review the source restore point’s coverage and consistency separately; reconstruction does not resolve data that was excluded from capture.

DSE recommendation

Prepare a disk-restore-point-to-target-disk map before creating the replacement VM. Have the application owner confirm the OS and data-disk roles and the intended attachment arrangement. For a cross-region recovery, require completion evidence for the entire disk set before beginning reconstruction. Keep the original VM and recovery artifacts outside any cleanup operation until the replacement is accepted.

Verification

Inspect the new disks and VM attachments against the approved map. Boot the replacement in the authorized recovery environment and test the application’s expected data and service behavior. Record the restore point, constituent disk identifiers, target resources, and result together. Investigate an omitted or incorrectly attached disk explicitly instead of accepting a successful VM creation response as proof that the complete workload was restored.

Official references

Microsoft Learn: Manage VM restore points. Source reviewed September 9, 2026.