VM Backup and Restore

This guide demonstrates using OADP to back up a virtual machine, make a destructive change, and then restore the VM to its previous state. This assumes OADP is installed and configured with the kubevirt plugin and a valid BackupStorageLocation.

Prerequisites

• OADP operator installed with BackupStorageLocation showing Available
• OpenShift Virtualization installed
• RWX-capable StorageClass available
• A VolumeSnapshotClass for your CSI driver (oc get volumesnapshotclass) — required for CSI snapshot-based VM backups

Create a Test Virtual Machine

1. Create a namespace for the test:

   oc new-project vm-backup-test
   

2. Go to Virtualization -> VirtualMachines -> ensure you are in the vm-backup-test project -> click "Create VirtualMachine"

3. Select "From template" and choose "Red Hat Enterprise Linux 9"
4. Name the VM backup-test-vm
5. Click "Customize VirtualMachine"

Set Login Credentials

1. Open the Scripts tab

2. Under Cloud-init, set a password for the guest user (RHEL cloud images have no usable default password):

   #cloud-config
   user: cloud-user
   password: Pass123!
   chpasswd:
     expire: false
   

Add a Data Disk

1. Click on the "Disks" tab
2. Click "Add disk"

3. Configure the data disk:

   • Name: data-disk
   • Source: Blank
   • Size: 5 GiB
   • Type: Disk
   • StorageClass: your default StorageClass
   • Access Mode: ReadWriteMany (RWX) 11. Click Add

Start the VM

1. Click "Create VirtualMachine"
2. Wait for the VM status to show "Running"

Write Test Data

1. Open the VM console from the WebUI:

   • Virtualization -> VirtualMachines -> click backup-test-vm -> Console tab 15. Log in as cloud-user / Pass123! 16. Format and mount the data disk, then write test data:

   sudo mkfs.xfs /dev/vdb
   sudo mkdir -p /mnt/data
   sudo mount /dev/vdb /mnt/data
   echo "OADP backup test - original data" | sudo tee /mnt/data/testfile.txt
   cat /mnt/data/testfile.txt
   

   You should see: OADP backup test - original data

2. Confirm the data is written:

   ls -la /mnt/data/
   

Take a Backup

1. Create a Backup CR to capture the running VM and its disks:

   apiVersion: velero.io/v1
   kind: Backup
   metadata:
     name: backup-test-vm-backup-1
     namespace: openshift-adp
   spec:
     includedNamespaces:
       - vm-backup-test
     snapshotMoveData: true
     ttl: 720h0m0s
   

Backup Storage Location

The storageLocation field is omitted because the DPA's backupLocations entry is set as default: true. Velero automatically uses the default BackupStorageLocation. Verify with: oc get backupstoragelocations -n openshift-adp

Note

This backs up all resources in the vm-backup-test namespace. Namespace-scoped backup ensures the VirtualMachine, its DataVolumes, PVCs, and associated secrets are all captured together.

Crash-Consistent Backups

OADP with the kubevirt and csi plugins can back up running VMs using CSI volume snapshots. The resulting backup is crash-consistent — equivalent to an unexpected power loss. This is sufficient for most workloads. If you need application-consistent backups (e.g., databases), either stop the VM first or use the QEMU guest agent for filesystem freeze/thaw.

1. Apply the backup:

   oc apply -f backup.yaml
   

2. Watch the backup progress:

   oc get backup backup-test-vm-backup-1 -n openshift-adp -w
   

   Wait for the PHASE to show Completed.

3. Verify the backup contents:

   oc get backup backup-test-vm-backup-1 -n openshift-adp -o jsonpath='{.status.phase}'
   

Make a Destructive Change

1. Open the VM console and modify the data:

   sudo mount /dev/vdb /mnt/data
   echo "THIS DATA HAS BEEN MODIFIED" | sudo tee /mnt/data/testfile.txt
   echo "extra-file-that-shouldnt-exist" | sudo tee /mnt/data/extra.txt
   cat /mnt/data/testfile.txt
   

   You should see: THIS DATA HAS BEEN MODIFIED

2. Stop the VM:

   virtctl stop backup-test-vm -n vm-backup-test
   

Delete the VM

1. Delete the VM and all PVCs in the namespace to simulate a disaster:

   oc delete vm backup-test-vm -n vm-backup-test
   oc delete pvc --all -n vm-backup-test
   

   Template/console VMs do not consistently label disks with app=backup-test-vm. Deleting all PVCs in the test namespace ensures root and data disks are removed so restore must recreate them.

2. Confirm the VM and PVCs are gone:

   oc get vm,pvc -n vm-backup-test
   

   Should return no VirtualMachine or PVC resources.

Restore from Backup

1. Create a Restore CR pointing to the backup:

   apiVersion: velero.io/v1
   kind: Restore
   metadata:
     name: backup-test-vm-restore-1
     namespace: openshift-adp
   spec:
     backupName: backup-test-vm-backup-1
     includedNamespaces:
       - vm-backup-test
     restorePVs: true
   

   oc apply -f restore.yaml
   

2. Watch the restore progress:

   oc get restore backup-test-vm-restore-1 -n openshift-adp -w
   

   Wait for the PHASE to show Completed.

Verify the Restore

1. Confirm the VM exists again:

   oc get vm backup-test-vm -n vm-backup-test
   

2. Start the restored VM if needed:

   If the VM was backed up with runStrategy: Always, it will auto-start after restore. Wait for the VMI to become Ready:

   oc get vmi -n vm-backup-test -w
   

   Only use virtctl start if the restored VM is in a Stopped state:

   virtctl start backup-test-vm -n vm-backup-test
   

   Wait for it to be running:

   oc get vmi backup-test-vm -n vm-backup-test -w
   

3. Open the VM console and verify the original data is restored:

   sudo mount /dev/vdb /mnt/data
   cat /mnt/data/testfile.txt
   ls /mnt/data/
   

   You should see: - testfile.txt contains: OADP backup test - original data (the original content) - extra.txt does not exist (the modification is gone)

   The VM has been fully restored to the state captured in the backup.

Summary

Step	What Happened
Create VM	RHEL 9 VM with a data disk, test data written
Backup	OADP captured the VM definition, disks, and PVC data
Destructive change	Modified data and added files to prove the change
Delete VM	Simulated a disaster by removing the VM entirely
Restore	OADP recreated the VM and restored disk contents
Verify	Original data is back, modifications are gone

Cleanup

1. Delete the test resources:

   oc delete vm backup-test-vm -n vm-backup-test --ignore-not-found
   oc delete pvc --all -n vm-backup-test --ignore-not-found
   oc delete backup backup-test-vm-backup-1 -n openshift-adp
   oc delete restore backup-test-vm-restore-1 -n openshift-adp
   oc delete project vm-backup-test