Properly handle volume cleanup with RBD driver

[Mike Perez]

If a volume gets deleted in the Ceph backend, and Cinder still knows about it in a 'deleting' state, cinder-volume on startup will try to clean it up. rbd.Image() will however raise a rbd.ImageNotFound exception in these cases. These should be disregarded so cinder can continue the delete if it's no longer in the backend.

Stack trace:
http://paste.openstack.org/show/53858/

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/58107

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/58107
Committed: http://github.com/openstack/cinder/commit/3de7da12d1098ef777305099e5f4a039e536bf99
Submitter: Jenkins
Branch: master

commit 3de7da12d1098ef777305099e5f4a039e536bf99
Author: Mike Perez <email address hidden>
Date: Sat Nov 23 10:50:15 2013 -0800

    Continue to delete volumes that DNE in rbd backend

    If a volume has already been deleted in the RBD backend, catch the
    exception and let Cinder continue removing it in the database.

    Closes-Bug: #1254318
    Change-Id: I99b591ce78d82c6eaabff0459d9c523e85c8e65f

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/havana
Review: https://review.openstack.org/58178

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/58178
Committed: http://github.com/openstack/cinder/commit/22f257bd6c0eff06e51662233b484dcfab91a93d
Submitter: Jenkins
Branch: stable/havana

commit 22f257bd6c0eff06e51662233b484dcfab91a93d
Author: Mike Perez <email address hidden>
Date: Sat Nov 23 10:50:15 2013 -0800

    Continue to delete volumes that DNE in rbd backend

    If a volume has already been deleted in the RBD backend, catch the
    exception and let Cinder continue removing it in the database.

    Closes-Bug: #1254318
    Change-Id: I99b591ce78d82c6eaabff0459d9c523e85c8e65f
    (cherry picked from commit 3de7da12d1098ef777305099e5f4a039e536bf99)