fs-backup cleanup hangs when failing

[Lyndon-Li]

Velero with default node-agent concurrency (1)
Create a PVB
When the PVB is in InProgress status, find some way to fail the PVB (don't restart node-agent)
Run another fs-backup or data mover
The other fs-backup or data mover fails to start and the log complains Data path instance is concurrent limited requeue later

The problem is with the failing PVB, the cleanup hangs with below logs:

time="2024-10-28T10:46:46Z" level=error msg="Async fs backup data path failed" controller=PodVolumeBackup error="Failed to ...
time="2024-10-28T10:46:46Z" level=info msg="Action finished" backup=velero/test14 controller=podvolumebackup logSource="pkg/uploader/provider/kopia.go:91" podvolumebackup=velero/test14-n2nlw
time="2024-10-28T10:46:46Z" level=info msg="Closing FileSystemBR" backup=velero/test14 controller=podvolumebackup logSource="pkg/datapath/file_system.go:145" podvolumebackup=velero/test14-n2nlw user=test14-n2nlw
<no more logs for this VGDP thread>