Is this a BUG REPORT or FEATURE REQUEST?:
@kubernetes/sig-storage-bugs

What happened:
Volume manager calls isPodTerminated to check if we should unmount a volume. In it, it's checking if the Pod deletion timestamp is set and if all the containers are in terminated or waiting state.

I don't think we should actually check for waiting state because it indicates that the container is trying to be started. This could manifest as a race condition where we end up unmounting a volume while the container is being started up for a deleted pod.

This check was added as part of #41456. It took the notRunning util function from pod status manager and made a copy of it in the volume manager. Checking for waiting state may make sense for pod status manager, but not necessarily for how volume manager uses it.

What you expected to happen:
We should only unmount the volume if we are sure there's no chance of the container running.