Create a Snapshot

A snapshot is the state of a Kubernetes volume at any given moment in time.

Snapshot Management via SUSE Storage UI

To create a snapshot of an existing cluster, follow these steps:

In the top navigation bar of the SUSE Storage UI, click Volume.
Click the name of the volume for which you want to create a snapshot. This leads to the volume detail page.
Click the Take Snapshot button.

Once the snapshot is created, you can view it in the list of snapshots for the volume before the Volume Head.

Snapshot Management with Custom Resources (CRs)

This section demonstrates how to create, list, restore and delete Longhorn snapshots directly via kubectl using the Custom Resources (CRs) of SUSE Storage.

SUSE Storage uses its own Snapshot CRD under the longhorn.io API group (for example, v1beta2), not the generic Kubernetes VolumeSnapshot from snapshot.storage.k8s.io.

Create a snapshot

Prepare the manifest - Create a file named longhorn-snapshot.yaml with the following content:

apiVersion: longhorn.io/v1beta2
kind: Snapshot
metadata:
  name: longhorn-test-snapshot
  namespace: longhorn-system
spec:
  volume: pvc-840804d8-6f11-49fd-afae-54bc5be639de   # replace with your actual Longhorn volume name
  createSnapshot: true

Apply the manifest:
```
kubectl apply -f longhorn-snapshot.yaml
```
Expected output:
```
snapshot.longhorn.io/longhorn-test-snapshot created
```
If the volume is detached, a brief warning about the engine not running appears. SUSE Storage automatically retries and the snapshot completes when the volume is attached.

List snapshots

To list all snapshots for a specific volume, use the following command:

kubectl get snapshots.longhorn.io -l longhornvolume=pvc-840804d8-6f11-49fd-afae-54bc5be639de -n longhorn-system

Delete a snapshot

To delete a specific snapshot, use the following command:

kubectl delete snapshot.longhorn.io longhorn-test-snapshot -n longhorn-system

Expected output:

snapshot.longhorn.io "longhorn-test-snapshot" deleted

SUSE Storage automatically handles the cleanup of the underlying data.