NetApp E-Series: Deleting a Cinder volume with a missing backend volume fails

[Michael Price]

If the backend volume is missing/deleted/removed, deleting the Cinder volume will fail, and cause it to move to error_deleting status. This is a regression.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/235974

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/235974
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=5cbf00b624e6c2bb8597d82640a1cb61b9405ba6
Submitter: Jenkins
Branch: master

commit 5cbf00b624e6c2bb8597d82640a1cb61b9405ba6
Author: Michael Price <email address hidden>
Date: Wed Sep 30 13:56:56 2015 -0500

    NetApp: E-Series fix deletion of missing volume

    When a Cinder volume is deleted but the backend volume is missing,
    an exception was being thrown, rather than swallowed and being logged.
    A KeyError was being expected, but a VolumeNotFound exception was
    actually being raised.

    This patch catches the VolumeNotFound exception, and logs a debug
    message if the volume could not be deleted from the backend (because
    it no longer exists). This patch restores compatibility with older
    versions where an exception would not be thrown.

    Closes-Bug: 1502166
    Change-Id: I47a8d9b408fc4ac7fae30f043b0dfca0bb763d8d

[Thierry Carrez (ttx)]

This issue was fixed in the openstack/cinder 8.0.0.0b1 development milestone.