Automatic merge from submit-queue (batch tested with PRs 44362, 44421, 44468, 43878, 44480)
Delete EmptyDir volume directly instead of renaming the directory.
**What this PR does / why we need it**:
The volume operation executor can handle duplicate requests on the same volume now, so it is not necessary to rename the directory anymore. This change can cause pod deletion to take longer for large emptydir volumes because now the pod waits for the volume to be deleted until it continues pod cleanup. But this is actually required for local disk scheduling so that we don't schedule new pods that need emptydir volumes on the node if the previous emptydir has not be fully reclaimed yet.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#43534
**Special notes for your reviewer**:
**Release note**:
NONE
cc @kubernetes/sig-storage-pr-reviews
c6bfa43c58