← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1368259] Re: unable to delete volume after attach with mount point fails on instance

 

** 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/1368259

Title:
  unable to delete volume after attach with mount point fails on
  instance

Status in OpenStack Compute (Nova):
  Invalid

Bug description:
  Description of problem:
  Openstack on rhel 7.0 VM box
  rhos 5.0  version .

  After creating instance , we try to attach volume to it .
  The volume created by tempest code :
  def test_attach_detach_volume_to_instance(self):
          # Volume is attached and detached successfully from an instance
          mountpoint = '/dev/vdc'
          _, body = self.client.attach_volume(self.volume['id'],
                                              self.server['id'],
                                              mountpoint)

  After volume has been created , we are attaching it to instance .
  The result of this action is that cinder shows that the volume attached to instance .
  But in nova - instace we do not see the volume in lsblk output  .
  We get error from logs :
  /var/log/nova/nova-compute.log:136:2014-09-08 02:59:39.573 31416 DEBUG nova.virt.disk.api [req-c0039d74-1c31-4e19-8719-bf3fb8eca505 8826f60585d244339c6a51dd7ce7b26f fbf97ed492214e9b8bb9f47206879a9c] Unable to mount image /var/lib/nova/instances/5531d5a9-cefd-4873-bbca-bdd4336ca5a8/disk with error Error mounting /var/lib/nova/instances/5531d5a9-cefd-4873-bbca-bdd4336ca5a8/disk with libguestfs (mount_options: /dev/sda on / (options: ''): mount: /dev/sda is write-protected, mounting read-only


  I attached another volume with auto mount  point .
  When we delete the instacne the cinder is binded to deleted instance and we can not delete the volume.

  
  Version-Release number of selected component (if applicable):

  
  How reproducible:

  #1 Create instance  
  #2 create volume with tempest code based on : 
  def test_attach_detach_volume_to_instance(self):
          # Volume is attached and detached successfully from an instance
          mountpoint = '/dev/vdc'
          _, body = self.client.attach_volume(self.volume['id'],
                                              self.server['id'],
                                              mountpoint)
  #2 the volume is attached , but we dont see it in instance (cirros image)
  #3 attach another volume with auto mount point to instance
  #4 delete the instance 
  #5 at this point cinder attachment is to "not existing" instance and we are not able to delete the volume.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1368259/+subscriptions