[BUG]: When VolumeGroup Snapshot ReadyToUse, the VolumeSnapshot don't have Status

[phuongatemc]

When VolumeGroup Snapshot Ready, the VolumeSnapshot don't have Status

Steps to reproduce the behavior:

• Using dellemc/csi-volumegroup-snapshotter:v0.2.0, I created VolumeGroup Snapshot of 3 PVCs with same label "volume-group=volumeGroup1" using controller code.
• After the DellCsiVolumeGroupSnapshot.Status ReadyToUse is true, the code tried to fetch the VolumeSnapshot and found that it misses the Status field.
  VG Snapshot's Status: {
  "snapshotGroupID": "217b52c106138b0f-c56b7d7100000003",
  "snapshotGroupName": "volumegroup1-102021-184622",
  "snapshots": "volumegroup1-102021-184622-0-pvol0,volumegroup1-102021-184622-1-pvol1,volumegroup1-102021-184622-2-pvol2",
  "creationTime": "2021-10-20T18:45:25Z",
  "readyToUse": true,
  "status": "Complete"
  }
  ...
  Missing Status of volumesnapshot 'volumegroup1-102021-184622-2-pvol2' in ns 'helmtest-vxflexos'.

[nb950]

please retest with VG image 0.3.0 , seems related to issue reported in #67

[phuongatemc]

I still saw this issue in 0.3.0 VolumeGroup Snapshotter.

[jooseppi-luna]

Hi @phuongatemc -- can you give us access to the cluster with the issue? Also, could you give us:

1. kubectl describe volumesnapshot xxx -n namespace output where status is missing
2. kubectl logs of the external-snapshotter pod?
3. The output of kubectl logs controller vg-snapshotter -n vxflexos
4. The version of external-snapshotter you are running
5. Please run the verify.sh script located in the scripts folder and send us the output.

[phuongatemc]

I think this may be a timing issue because if I created the VolumeGroup Snapshot manually using kubectl commands, I've never seen this issue. I only saw this issue when I used golang code to fetch the VolumeSnapshots after they are created in the VolumeGroup Snapshot (Status.ReadyToUse == true). And even with golang code, it is intermittent.

[nb950]

talked to customer , suggest wait for volumesnapshot status to be updated , it may take a little time because external-snapshotter is responsible for updating status . if this is indeed the fix please confirm

[phuongatemc]

I have implemented in my controller code to validate VolumeSnapshot and VolumeSnapshotContent and retry if they are not ready as discussed with Jai Kumar. So far I couldn't see the problem again.

[nb950]

created story action item to fix this in vgs controller.

[mjsdell]

Marking this as closed per the above comment.

[bharathsreekanth]

This issue was fixed as part of dell/csi-volumegroup-snapshotter#14.