cinder reimage failure leaves the volume in downloading state

[Rajat Dhasmana]

When we try to reimage a volume and the image metadata doesn't pass the following checks, we end up with a volume in downloading state:

1. Image is not in 'active' state
2. Size of image is larger than size of volume
3. Volume size is smaller than the image 'minDisk' size
4. Image virtual size is greater than Volume size

In these cases, we don't roll back the volume state to the previous state.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/c/openstack/cinder/+/896149

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/cinder/+/896149
Committed: https://opendev.org/openstack/cinder/commit/a752012c92bb15cf264c617bb2071958a32a8ad1
Submitter: "Zuul (22348)"
Branch: master

commit a752012c92bb15cf264c617bb2071958a32a8ad1
Author: whoami-rajat <email address hidden>
Date: Thu Sep 21 17:41:16 2023 +0000

    Fix: Roll back volume status during reimage failure

    When we try to reimage a volume, we update the status of
    volume to 'downloading'.
    We later validate the image metadata (like image is 'active',
    image size is less than volume size, etc), and in case the
    validation fails, we currently don't revert the volume status
    back to original ('available', 'in-use' etc) and volume stays
    in 'downloading' state.

    This patch fixes this by catching the failure exception and
    doing a DB update to restore the volume status back to it's
    previous state.

    Closes-Bug: #2036994
    Change-Id: I05bf29e2a089b06398414b542b655a8083c9a21f

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/cinder 24.0.0.0rc1 release candidate.