Faulty multipath devices left on compute nodes after deleting instances booted from Cinder volumes

Issue

After running a stress test which creates a Cinder volume from an image, boots that image as an instance and then deletes the instance and volume, the multipath command produces output like the following:

# multipath -ll
30000000000000000 dm-11 3PARdata,VV
size=38G features='1 queue_if_no_path' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=0 status=enabled
  |- 8:0:2:1 sdj 8:144 failed faulty running
  |- 7:0:2:1 sdh 8:112 failed faulty running
  `- 8:0:2:4 sdp 8:240 failed faulty running