[Feature Request] Add support for VolumeGroupSnapshot

[itzhak-alayev-nonamesec]

/kind bug

What happened?
I created a VolumeGroupSnapshot/v1beta1 and got the following errors in csi-snapshotter container logs:

I0219 15:48:40.967917       1 groupsnapshot_helper.go:390] createGroupSnapshotWrapper: Creating group snapshot for group snapshot content groupsnapcontent-4c4b8f15-d38c-4151-ac8a-0a150f6203b2 through the plugin ...
E0219 15:48:41.021287       1 groupsnapshot_helper.go:736] checkandUpdateGroupSnapshotContentStatus [groupsnapcontent-bbdb460f-f3fb-47b7-9328-a7320384ff9e]: error occurred failed to take group snapshot of the volum
es [vol-0ffa2fe0e4d7b7111 vol-03c9cbe4b08181df3 vol-017a7135664a2e4e2]: "rpc error: code = Unimplemented desc = unknown service csi.v1.GroupController"
E0219 15:48:41.021311       1 groupsnapshot_helper.go:157] could not sync group snapshot content "groupsnapcontent-bbdb460f-f3fb-47b7-9328-a7320384ff9e": failed to take group snapshot of the volumes [vol-0ffa2fe0e4
d7b7111 vol-03c9cbe4b08181df3 vol-017a7135664a2e4e2]: "rpc error: code = Unimplemented desc = unknown service csi.v1.GroupController"

What you expected to happen?
Snapshots to be created and VolumeGroupSnapshotContent to be in Ready state.

How to reproduce it (as minimally and precisely as possible)?

1. Install ebs-csi-driver via an EKS Addon
2. Install the following Helm chart which has the latest v1beta1 VolumeGroupSnapshot CRDs - https://artifacthub.io/packages/helm/piraeus-charts/snapshot-controller
3. Create a VolumeGroupSnapshot.
4. Checkout the csi-snapshotter container logs in ebs-csi-controller Pod.

Anything else we need to know?:

Environment

• Kubernetes version (use kubectl version):
  Client Version: v1.31.0
  Kustomize Version: v5.4.2
  Server Version: v1.32.1-eks-8cce635

• Driver version: v1.39.0

[torredil]

Hi there, the EBS CSI Driver does not currently implement the VolumeGroupSnapshot RPC or support that capability. We will track this as a feature request.

/retitle [Feature Request] Add support for VolumeGroupSnapshot
/kind feature

In the mean time, could you share more details about your use case? Thank you!

[itzhak-alayev-nonamesec]

the usecase is:
a statefulset that creates dynamic PVCs, instead of having a VolumeSnapshot for each replica of the statefulset i thought why not use the VolumeGroupSnapshot to snapshot all replicas using a label.

[ConnorJC3]

/lifecycle frozen

Pre-emptively freezing so the bot doesn't close this. We should do this in the long term, but it is currently not feasible to implement in a manner compliant with the CSI Spec due to the way the AWS CreateSnapshots API works - the API will need to change before this can be implemented.