← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1236411] Re: failed volume detach at libvirt driver can set a wrong volume state

 

** Changed in: nova
       Status: Incomplete => Invalid

-- 
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/1236411

Title:
  failed volume detach at libvirt driver can set a wrong volume state

Status in OpenStack Compute (Nova):
  Invalid

Bug description:
  In libvirt driver, detach_volume, a DiskNotFound is raised if the
  volume is no longer attached to the guest. However the exception is
  then caught and self.volume_api.roll_detaching(context, volume_id) is
  called which will make the volume 'in-use'. The it is reraised and the
  volume detach is aborted.

  This doesn't seem correct, especially if two detach messages for the
  same volume manages to get to the compute manager, the volume would
  now be fully removed from host but remain 'in-use' in cinder db.

  It would seem more appropriate to just ignore (after logging)
  DiskNotFound and continuing to detach from the volume driver and
  billing.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1236411/+subscriptions