yahoo-eng-team team mailing list archive

[David Kranz <dkranz@xxxxxxxxxx>]

I am not seeing how this is a bug in tempest. Tempest is deleting the vm
only after nova reports that the volume is 'in-use' which seems fine.
It would be nice if there was a backtrace, log, or something associated
with this ticket. This might be a cinder issue but more likely nova and
the test is making a nova call.

** Changed in: tempest
       Status: New => Invalid

** Also affects: nova
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1335889

Title:
  Race condition in quickly attaching / deleting volumes

Status in Cinder:
  New
Status in OpenStack Compute (Nova):
  Confirmed
Status in Tempest:
  Invalid

Bug description:
  It seems that there is a race condition in the stress
  volume_attach_delete test. It creates VMs and volumes, attaches
  volumes and deletes everything.

  The test is waiting for volumes to be in 'in-use' state before
  deleting VMs. It seems that Nova/Cinder don't have time to register
  volumes as attached in their databases before VMs get deleted. Volumes
  are then left attached to deleted VMs and unable to be deleted.

To manage notifications about this bug go to:
https://bugs.launchpad.net/cinder/+bug/1335889/+subscriptions