An Azure native disaster recovery service. Previously known as Microsoft Azure Hyper-V Recovery Manager.
Hello Alin Dascalu
Thank you for posting your query on Microsoft Q&A platform.
To confirm whether any recoverable copy of your Linux VM exists before May 4, 2026 (ideally between April 25 – May 4, 2026), you need to check four independent places in Azure, because each one stores point-in-time copies differently and they do not share data with each other:
- Azure Backup (Recovery Services vault → recovery points)
- VM Restore Points (Microsoft.Compute restore point collections)
- Managed Disk Snapshots
- Azure Site Recovery / disk copies / images
Below is exactly where to check each one in the portal, how to interpret the result, and the safest way to restore if a copy is found.
- Check if the VM was protected by Azure Backup (Recovery Services vault):
Azure Backup stores recovery points in a Recovery Services vault. The vault and the VM must be in the same region, so start by checking vaults in your VM's region.
Steps in the Azure portal:
- Go to the VM blade → Operations → Backup.
- If it shows "Enable backup", the VM was not protected by Azure Backup at the time.
- If it shows a vault name and a list of recovery points, Azure Backup is/was enabled.
- Alternatively, open Backup center → Backup instances → filter Datasource type = Azure Virtual Machines and search for the VM name. This shows protection across all vaults in the subscription.
- If a backup item exists, open it and review Restore points — filter by date range April 25 – May 4, 2026. Each recovery point shows the time, consistency type (Application-/File-system-/Crash-consistent), and tier (Snapshot vs. Vault).
Important nuance about Linux VMs: By default Azure Backup takes a file-system consistent backup on Linux (unless pre/post-scripts were configured for application consistency). Both are fully restorable.
Even if the VM has since been deleted, the recovery points can still exist in the vault as long as the backup item wasn't deleted with "Delete backup data". Check Backup center → Backup instances → "Soft-deleted" view as well deleted backup data is retained for an additional 14 days by default (longer with Enhanced soft delete).
- Check for VM Restore Points (separate from Azure Backup):
VM Restore Points are an independent Azure Compute feature. They live in Restore point collections under your subscription, not in a Recovery Services vault. A VM can have VM restore points even if Azure Backup was never enabled.
Steps in the Azure portal:
- In the global search bar, type "Restore point collections" and select it.
- Filter by the subscription and the resource group of your VM.
- Open any collection whose Source VM points to your VM — inside, each restore point shows the creation time and consistency mode (Application-consistent or Crash-consistent).
- Look for any restore point created on or before May 4, 2026.
Each VM restore point contains a disk restore point per attached disk (1 for OS + 1 per data disk). The first restore point is a full copy; subsequent ones are incremental but you can restore from any single restore point directly.
- Check for Managed Disk Snapshots:
Managed snapshots are read-only, point-in-time copies of a managed disk that exist as their own resource fully independent of the source VM. They're a very common manual or scripted backup mechanism.
Steps in the Azure portal:
- In the global search bar, type "Snapshots" and open the Snapshots service. This lists every snapshot in the subscription.
- Add the "Time created" column (or sort by it) and filter to dates on or before May 4, 2026.
- For each candidate, open the snapshot → on the Overview blade, check the Source disk property. If it matches your VM's OS or data disk resource ID, you have a usable point-in-time copy of that disk.
- Also browse to the Disk resource of the VM (or the original disk, if still present) → left menu → Snapshots tab to see linked snapshots quickly.
Tip: Snapshots survive VM deletion as long as the snapshot itself wasn't deleted. So even if the original VM is gone, the snapshot may still be there in the resource group (or in a different RG used for backups).
- Check other places where a "copy" may exist:
If Azure Backup, VM restore points, and snapshots all come back empty, also check:
- All resources blade → filter Type = Disk → look for disks with names like
<vmname>-osdisk-copy,*-backup, etc., that may have been created viaaz disk create --sourceor PowerShell as a manual disk copy. - All resources → filter Type = Image / Compute Gallery Image Version — a generalized or specialized image of the VM may have been captured.
- Azure Site Recovery (ASR) — open the Recovery Services vault → Replicated items. If ASR was enabled, replicated copies (with recovery points up to 72 hours, depending on the policy) may exist in the target region.
- Backup vault (different from Recovery Services vault) — if Azure Disk Backup was used, the recovery points live in a Backup vault, not a Recovery Services vault.
If none of the above show any item dated on or before May 4, 2026, then unfortunately no recoverable Azure-side copy exists for that timeframe. Azure does not keep any hidden/implicit backups of VMs recoverability requires one of the above mechanisms to have been configured beforehand.
- Safest restore method: Restore to a new VM / new disks (no overwrite)
If you find a recovery point in any of the above sources, do not use "Replace existing", because that overwrites the current VM's disks. Use one of these non-destructive options instead.
A. If the recovery point is in a Recovery Services vault (Azure Backup):
In the portal: Recovery Services vault → Backup items → Azure Virtual Machine → → Restore VM, pick the recovery point in the April 25 – May 4 range, then choose one of these:
- Create new virtual machine — fastest. Provide a new VM name, target resource group, VNet/subnet. The new VM is created in the same region as the source. The original VM is untouched.
- Restore disks — restores the disks to a resource group of your choice and generates an ARM template you can edit (size, NICs, extensions) before deploying a brand-new VM. This is the most flexible option for Linux because you can attach the OS disk to a rescue VM first to verify the filesystem before booting.
Avoid the Replace existing option — it is unsupported for many Linux scenarios and would overwrite the current VM's disks.
B. If the copy is a VM Restore Point:
Use the disk restore points to create new managed disks, then attach them to a new VM:
- Portal: open the Restore point collection → select the restore point → Create VM (or Create disk for each disk restore point) → deploy a new VM from the restored disks.
C. If the copy is a managed disk snapshot:
- Portal: open the Snapshot → Create disk → then Create VM from that disk (or attach as a data disk to a rescue VM for inspection first). The source disk/VM is not affected.
Reference documentation:
- Backup and restore options for VMs in Azure — overview of all the mechanisms above
- About the Azure VM restore process (Azure Backup) — recovery point tiers (Snapshot vs. Vault), OLR/ALR
- Restore VMs by using the Azure portal (Azure Backup) — Create new VM / Restore disks / Replace existing
- Use virtual machine restore points — VM restore point collections concept
- Manage Virtual Machine restore points — list/restore disk restore points
- Create a snapshot of a virtual hard disk — managed disk snapshots in the portal
- Overview of Recovery Services vaults — soft delete (14-day grace) for deleted backup data
Thanks,
Suchitra.