I noticed in a presubmit job the diskmaker logs have a series of repeating errors:

I0314 01:50:57.882044   36668 cache.go:55] Added pv "local-pv-ea5b0e26" to cache
I0314 01:50:57.882171   36668 reconcile.go:97] "Looking for released PVs to cleanup" namespace="openshift-local-storage" name="tentothirty-overlapping-twentytofifty-1-2"
E0314 01:50:57.882288   36668 deleter.go:103] failed to get volume mode of path "/mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d": Directory check for "/mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d" failed: open /mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d: no such file or directory
I0314 01:50:57.882306   36668 reconcile.go:101] "Looking for symlinks to cleanup" namespace="openshift-local-storage" name="tentothirty-overlapping-twentytofifty-1-2"
2025-03-14T01:50:57.882Z	DEBUG	events	recorder/recorder.go:104	Error cleaning PV "local-pv-ea5b0e26": failed to get volume mode of path "/mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d": Directory check for "/mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d" failed: open /mnt/local-storage/tentothirty-overlapping-twentytofifty-1-2/nvme-Amazon_Elastic_Block_Store_vol0662dcb2581e2e50d: no such file or directory	{"type": "Warning", "object": {"kind":"PersistentVolume","name":"local-pv-ea5b0e26","uid":"7442678b-0860-434a-9260-1a8dbfd19d2c","apiVersion":"v1","resourceVersion":"49089"}, "reason": "VolumeFailedDelete"}

presubmit job:
https://prow.ci.openshift.org/view/gs/test-platform-results/pr-logs/pull/openshift_local-storage-operator/522/pull-ci-openshift-local-storage-operator-main-e2e-operator/1900324985017208832

diskmaker log:
https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/test-platform-results/pr-logs/pull/openshift_local-storage-operator/522/pull-ci-openshift-local-storage-operator-main-e2e-operator/1900324985017208832/artifacts/e2e-operator/gather-extra/artifacts/pods/openshift-local-storage_diskmaker-manager-jvpwg_diskmaker-manager.log

It fails to getVolMode here in deletePV:
https://github.com/openshift/local-storage-operator/blob/646a98497445b51ce0fc3c7455ae06cb4869ec33/vendor/sigs.k8s.io/sig-storage-local-static-provisioner/pkg/deleter/deleter.go#L150-L153

The PV is released but does not have deletionTimestamp:
https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/test-platform-results/pr-logs/pull/openshift_local-storage-operator/522/pull-ci-openshift-local-storage-operator-main-e2e-operator/1900324985017208832/artifacts/e2e-operator/gather-extra/artifacts/persistentvolumes.json

CleanupSymlinks() in diskmaker only processes PV's with deletionTimestamp. So how could it be that the symlink has been deleted while the PV has not?

I suspect this is just an issue in the order of cleanup functions in the e2e test job, cleanupSymlinkDir should run after cleanupLVSetResources:
https://github.com/openshift/local-storage-operator/blob/646a98497445b51ce0fc3c7455ae06cb4869ec33/test/e2e/localvolumeset_test.go#L99-L106

If cleanupSymlinkDir runs before the LVSet is deleted, we could get into this situation. Strictly speaking, we shouldn't even need cleanupSymlinkDir anymore after https://github.com/openshift/local-storage-operator/pull/504 -- we could try removing it from the test.

    4.19

    Unknown

    1. Run e2e-operator presubmit job
    2. Review diskmaker logs
    

    "no such file or directory" errors when trying to delete PV

    PV is deleted by diskmaker successfully

    When I noticed these errors, my branch was missing https://github.com/openshift/local-storage-operator/pull/522/commits/0b88cae58e65a6176b540713f7717e3971e8e810 -- just in case that's relevant to reproducing this.