Skip to navigation

Restore a snapshot

Restoring a snapshot takes a VM’s disks back to the moment the snapshot was taken. It lets you undo a failed upgrade on the same VM, build a separate copy of the VM, or roll back a single disk. Two of the three modes change the existing VM, so read the table below before you start.

Which restore mode to choose

You want to…ModeChanges the current VM?
Undo a failed upgrade or deploy on this VMReplace Current VMYes. All captured disks go back to the snapshot; later changes are lost.
Clone the server, or look at old data, without touching the originalCreate New VMNo
Roll back only one disk, for example a corrupted data volumeRestore Selected DiskYes, that disk. It goes back to the snapshot; other disks are untouched.

Rolling back can’t be undone. Before Replace Current VM or Restore Selected Disk, take a new snapshot of the current state if there’s any chance you’ll need it.

How it works

  1. You pick a snapshot on the VM’s Snapshots tab and click Restore.
  2. You choose a mode. The panel shows a Restore Preview and the Restore Impact.
  3. You click Confirm Restore. For an in-place restore, the VM is stopped, its disks are rolled back, and it starts again automatically. The VM shows Restoring meanwhile.
  4. A message says “Recovery restore completed successfully”. Restored data volumes are attached but not mounted.

Before you begin

  • A snapshot that is ready (Available on Tools → Snapshots).
  • Write access to the workspace.
  • For an in-place restore, plan for a short outage: the VM is stopped and started, and open connections drop.
  • For Create New VM, credit for the new VM and its volumes, which are billed as new resources. The new VM is created in the original VM’s location, with its own new IP address.
  • No other restore, backup or resize running on the VM.

Restore in the portal

1

Open the snapshot

Open the VM, click Snapshots, and scroll to Recovery Points. Click Details to check the captured disks and Mount Instructions.

2

Start the restore

Click Restore. A side panel shows the recovery point, its captured disks and three modes.

3

Choose and fill in a mode

ModeWhat to fill in
Replace Current VMNothing else
Create New VMNew VM Name (a name is suggested), New VM Plan (marked Eligible, Disk too small or Unavailable), and Restored Data Volume Names
Restore Selected DiskThe Captured Disk to roll back. The preview says, for example, “This will roll back vmroot-test-clone (vda) in place.”
4

Read the impact

ModeRestore Impact says
Replace Current VM”The current VM will be stopped, its root disk will be replaced, and captured data disks will be attached again.”
Create New VM”A new VM and new live block volumes will be created from this recovery point and billed as active resources.”
Restore Selected Disk”The selected disk will be rolled back in place. The VM may be stopped briefly and started again automatically.”
5

Confirm

Click Confirm Restore and wait for “Recovery restore completed successfully”.

Restore with the CLI or API

# Build a new VM from the snapshot, leaving the original alone
ibee vms snapshots restore SNAPSHOT_SET_ID --vm-id VM_ID \
--target-mode new_vm --target-vm-name web-01-clone --target-plan-id PLAN_ID --yes
# Follow the restore
ibee vms snapshots restore-status RESTORE_ID

--target-mode is replace (the default), new_vm or volume_only (with --selected-volume-id). The API call is POST /compute/cloud-vm-snapshots/{snapshot_set_id}/actions/restore; it returns a restore_id to poll at GET /compute/cloud-vm-snapshots/restores/{restore_id}. In new_vm mode you can also send vpc_id, subnet_id, network_connectivity and ssh_keys for the new VM. GPU VMs use ibee gpus snapshots and /compute/gpu-vm-snapshots/....

Verify

ModeCheck
Replace Current VMThe VM is Running with the same IP address and hostname. Files are as they were in the snapshot.
Create New VMA new VM appears in the list with its own IP address. Update DNS when you’re ready to switch.
Restore Selected DiskThe VM is Running; only the chosen disk is rolled back.

Mount restored data volumes

Restored data volumes are attached but not mounted. Open the snapshot’s Details, click Copy commands under Mount Instructions, and run them inside the VM:

sudo mkdir -p /mnt/data
sudo mount /dev/vdc /mnt/data
echo 'UUID=<filesystem-uuid> /mnt/data ext4 defaults,nofail 0 2' | sudo tee -a /etc/fstab

Don’t format (mkfs) a restored volume: it already holds your data.

Limits

Value
Snapshots you can restoreReady (Available) ones
In-place restoreStops the VM once, then starts it again
New VM locationThe original VM’s location
New VM root diskAt least the captured root disk’s size; plans offer 50 GB to 400 GB
Restores at a time1 per VM

Troubleshooting

You seeWhat it meansWhat to do
Only ready snapshot sets can be restored.The snapshot is still being created, or failedWait until it’s Available, or pick another.
A recovery restore is already in progress. Please wait.Another restore is running on this VMWait for it to finish.
A plan is marked Disk too smallIts root disk is smaller than the captured onePick a bigger plan.
Recovery restore failedThe restore stopped with an errorCheck Activity logs, then retry or contact support.
A data volume is in lsblk but emptyIt isn’t mounted yetRun the Mount Instructions.
You restored the wrong diskThe rollback is doneRestore again from a snapshot that has the data you want. Next time, snapshot first.