Skip to content

Latest commit

 

History

History
83 lines (77 loc) · 3.33 KB

clone-from-volumesnapshot-source.md

File metadata and controls

83 lines (77 loc) · 3.33 KB
Raw

Data Volume cloning from a VolumeSnapshot source

Introduction

Current CDI PVC cloning methods offer smart cloning which uses a 1:1 cloning method that creates a temporary snapshot of the source PVC for each intended clone.
Some storage systems are designed to scale better with a 1:N method where a single snapshot of the source PVC is used to create multiple clones.
One examples is ceph, where this would be about utilizing RBD layering.
A much simpler example would be storage that can take around 5-10 minutes to snapshot a 200GB volume.

Note: Data Volume cloning from snapshot source can work together with namespace transfer and size expansion

Prerequisites

  1. The source snapshot and target PVC share the same provisioner
  2. The user creating the DataVolume has permission to create the datavolumes/source resource in the source namespace
  3. Storage supports expansion (if the user attempts clone to larger target)

Flow Description

  • DataVolume is created with a Snapshot source

  • Check if restore from said snapshot to desired target is possible (as described in Prerequisites):

  • If possible:

    • Create the restore PVC
    • Set the claim reference of the PV to point to the new target PVC (namespace-transfer)

  • If not possible:

    • Attempt host-assisted cloning between 2 PVCs where CDI creates a temporary restore PVC (which will be cleaned up) from the snapshot to act as the source.

    Note: below k8s 1.29 (which has sourceVolumeMode on snapshots) it is advised to annotate the snapshots sources not created by CDI with cdi.kubevirt.io/storage.import.sourceVolumeMode
    This is because otherwise CDI cannot infer the volume mode to create a temporary restore.

Example

To kick off the process, we need a source volume snapshot.
We create a simple Data Volume that will act as a source for the snapshot.

apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: golden-snap-source
  namespace: golden-ns
spec:
  source:
      http:
         url: http://mirrors.nav.ro/fedora/linux/releases/33/Cloud/x86_64/images/Fedora-Cloud-Base-33-1.2.x86_64.qcow2
  storage:
    accessModes:
    - ReadWriteOnce
    resources:
      requests:
        storage: 8Gi

Next, we will create a VolumeSnapshot of the resulting PVC

apiVersion: snapshot.storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
  name: golden-volumesnapshot
  namespace: golden-ns
spec:
  volumeSnapshotClassName: csi-rbdplugin-snapclass
  source:
    persistentVolumeClaimName: golden-snap-source

Double check snapshot is ready to use (not a must, but nice to know)

$ kubectl get volumesnapshot golden-volumesnapshot 
NAME                    READYTOUSE   SOURCEPVC            SOURCESNAPSHOTCONTENT   RESTORESIZE   SNAPSHOTCLASS                            SNAPSHOTCONTENT                                    CREATIONTIME   AGE
golden-volumesnapshot   true ...

Now we will go ahead and create the target DataVolume that clones the above VolumeSnapshot

apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: cloned-datavolume
  namespace: default
spec:
  source:
    snapshot:
      namespace: golden-ns
      name: golden-volumesnapshot
  storage:
    storageClassName: rook-ceph-block
    resources:
      requests:
        storage: 9Gi