Volume Snapshots

Create Persistent Volume Snapshots

This procedure shows how to create Persistent Volume Snapshots. Read more about this feature in the official Kubernetes documentation.

Before you begin

The following items and configurations are required for this procedure:

  • Kubernetes version 1.17.x or higher

  • Konvoy version 1.5.x or higher

  • CSI driver that supports volume snapshots

Limitations

  • Only some CSI drivers support snapshots if underlying platform supports it

Setup Volume Snapshot Class

As StorageClass provides a way for administrators to describe the “classes” of storage they offer when provisioning a volume, VolumeSnapshotClass provides a way to describe the “classes” of storage when provisioning a volume snapshot. You can read more about the different options in this document.

The VolumeSnapshotClass depends on the infrastructure provider you are using. Refer to the sections below for the appropriate one.

AWS

Refer to the AWS EBS CSI documentation.

cat << EOF | kubectl apply -f -
apiVersion: snapshot.storage.k8s.io/v1beta1
kind: VolumeSnapshotClass
metadata:
  name: csi-vsc
driver: ebs.csi.aws.com
deletionPolicy: Delete
EOF

Azure

Refer to the Azure CSI documentation.

cat << EOF | kubectl apply -f -
apiVersion: snapshot.storage.k8s.io/v1beta1
kind: VolumeSnapshotClass
metadata:
  name: csi-vsc
driver: disk.csi.azure.com
deletionPolicy: Delete
parameters:
  incremental: "true"  # available values: "true", "false" ("true" by default)
EOF

GCP

Refer to the GCP CSI documentationgcp-snapshots.

cat << EOF | kubectl apply -f -
apiVersion: snapshot.storage.k8s.io/v1beta1
kind: VolumeSnapshotClass
metadata:
  name: csi-vsc
driver: pd.csi.storage.gke.io
deletionPolicy: Delete
EOF

Test Volume Snapshots

  1. Create a new PersistentVolumeClaim:

    cat << EOF | kubectl apply -f -
    apiVersion: v1
    kind: PersistentVolumeClaim
    metadata:
      name: snapshot-claim
    spec:
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 4Gi
    EOF
    
  2. Create a sample app using the persistentVolumeClaim:

    cat << EOF | kubectl apply -f -
    apiVersion: v1
    kind: Pod
    metadata:
      name: app
    spec:
      containers:
      - name: app
        image: centos
        command: ["/bin/sh"]
        args: ["-c", "while true; do echo \$(date -u) >> /data/out.txt; sleep 5; done"]
        volumeMounts:
        - name: persistent-storage
          mountPath: /data
      volumes:
      - name: persistent-storage
        persistentVolumeClaim:
          claimName: snapshot-claim
    EOF
    
  3. Validate the pod successfully wrote data to the volume. Note the timestamp of the first entry:

    kubectl exec -it app -- cat /data/out.txt
    

    The output should look similar to:

    Tue Aug 18 15:00:58 UTC 2020
    Tue Aug 18 15:01:03 UTC 2020
    Tue Aug 18 15:01:08 UTC 2020
    
  4. Create a VolumeSnapshot referencing the PersistentVolumeClaim name:

    cat << EOF | kubectl apply -f -
    apiVersion: snapshot.storage.k8s.io/v1beta1
    kind: VolumeSnapshot
    metadata:
      name: volume-snapshot
    spec:
      volumeSnapshotClassName: csi-vsc
      source:
        persistentVolumeClaimName: snapshot-claim
    EOF
    
  5. Wait for the READYTOUSE: true attribute of the VolumeSnapshot:

    kubectl get volumesnapshot volume-snapshot -w
    
  6. Delete the existing app and PersistentVolumeClaim:

    kubectl delete pod app && kubectl delete pvc snapshot-claim
    
  7. Restore a volume from the snapshot with a PersistentVolumeClaim referencing the VolumeSnapshot in its dataSource:

    cat << EOF | kubectl apply -f -
    apiVersion: v1
    kind: PersistentVolumeClaim
    metadata:
      name: snapshot-restored-claim
    spec:
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 4Gi
      dataSource:
        name: volume-snapshot
        kind: VolumeSnapshot
        apiGroup: snapshot.storage.k8s.io
    EOF
    
  8. Create another sample app using the restored persistentVolumeClaim:

    cat << EOF | kubectl apply -f -
    apiVersion: v1
    kind: Pod
    metadata:
      name: app
    spec:
      containers:
      - name: app
        image: centos
        command: ["/bin/sh"]
        args: ["-c", "while true; do echo \$(date -u) >> /data/out.txt; sleep 5; done"]
        volumeMounts:
        - name: persistent-storage
          mountPath: /data
      volumes:
      - name: persistent-storage
        persistentVolumeClaim:
          claimName: snapshot-restored-claim
    EOF
    
  9. Validate the new pod has the restored data by comparing the timestamp of the first entry to that of in step 3:

    kubectl exec -it app -- cat /data/out.txt
    

    The output should look similar to:

    Tue Aug 18 15:00:58 UTC 2020
    Tue Aug 18 15:01:03 UTC 2020
    Tue Aug 18 15:01:08 UTC 2020
    ...
    Tue Aug 18 15:05:40 UTC 2020
    Tue Aug 18 15:05:45 UTC 2020
    
  10. Delete the app, PersistentVolumeClaim and VolumeSnapshot :

    kubectl delete pod app && kubectl delete pvc snapshot-restored-claim && kubectl delete volumesnapshot volume-snapshot
    
  11. Delete the VolumeSnapshotClass:

    kubectl delete volumesnapshotclass csi-vsc