VolumeSnapshot resources still remain after pass the E2E test #264
Description
What happened:
Run storage E2E test and passed the test, but some VolumeSnapshot resources remain.
What you expected to happen:
VolumeSnapshot resources are deleted after E2E test.
How to reproduce it (as minimally and precisely as possible):
Use https://github.com/zlabjp/kube-conformance .
Below is my testdriver.yaml
ShortName: ontap-block
StorageClass:
FromName: false
FromFile: </path/to/storage-class>
SnapshotClass:
FromName: true
DriverInfo:
Name: <Name My Storage Driver>
Capabilities:
persistence: true
fsGroup: true
exec: true
block: true
pvcDataSource: true
controllerExpansion: true
nodeExpansion: true
volumeLimits: false
singleNodeVolume: false
snapshotDataSource: true
topoligy: false
SupportedSizeRange:
Min: 1Gi
Max: 10Gi
RequiredAccessModes:
- ReadWriteOnce
Anything else we need to know?:
The main cause may be here.
snapshotWorker calls checkAndUpdateSnapshotClass to check VolumeSnapshotClass.
This logic prevent deleting VolumeSnapshot when VolumeSnapshotClass have already deleted.
On the E2E side, here is the logic to cleanup resources.
cleanupFunc := func() {
framework.Logf("deleting snapshot %q/%q", snapshot.GetNamespace(), snapshot.GetName())
err = dynamicClient.Resource(snapshotGVR).Namespace(updatedClaim.Namespace).Delete(snapshot.GetName(), nil)
if err != nil && !apierrs.IsNotFound(err) {
framework.Failf("Error deleting snapshot %q. Error: %v", snapshot.GetName(), err)
}
framework.Logf("deleting initClaim %q/%q", updatedClaim.Namespace, updatedClaim.Name)
err = client.CoreV1().PersistentVolumeClaims(updatedClaim.Namespace).Delete(updatedClaim.Name, nil)
if err != nil && !apierrs.IsNotFound(err) {
framework.Failf("Error deleting initClaim %q. Error: %v", updatedClaim.Name, err)
}
framework.Logf("deleting SnapshotClass %s", snapshotClass.GetName())
framework.ExpectNoError(dynamicClient.Resource(snapshotClassGVR).Delete(snapshotClass.GetName(), nil))
}
E2E doesn't wait for finishing to delete the VolumeSnapshot, and delete SnapshotClass.
So that, SnapshotResource can't be deleted.
I think snapshotWorker need not to check VolumeSnapshotClass when delete VolumeSnapshot( and VolumeSnapshotContent).
Environment:
- Kubernetes version (use
kubectl version): v1.17.2 - Cloud provider or hardware configuration: openstack
- OS (e.g:
cat /etc/os-release): container-linux-2303.3.0 - Kernel (e.g.
uname -a): 4.19.86-coreos - Install tools: original