backup-restore fails between ceph and other storage

[Lisa Li]

Test steps:
Backup storage is ceph.
1. Create a volume in ceph
2. backup-create above volume
3. Create a volume in lvm.
4. backup-restore backup in step 2 to volume created in step 3.

The test fails as it says the error message "can't find the image volume-8f282326-e2bc-4694-876c-d13cc26f71e1.backup.base".

Analysis:
When backup-create a ceph volume, it created a base image, and created a snap shot.
When backup-restore, it uses the snapshot to restore. But the logic is wrong in _restore_volume in file backup/drivers/ceph.py.

[Lisa Li (lisali)]

Changes:
https://review.openstack.org/198993

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/198993
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=e7c7355589c29c16af2bc73ce06164a72ff5af09
Submitter: Jenkins
Branch: master

commit e7c7355589c29c16af2bc73ce06164a72ff5af09
Author: lisali <email address hidden>
Date: Tue Jul 7 15:34:32 2015 +0800

    Fix restore point if backup base is diff-format in ceph

    In Ceph when backup base is diff-format and use this backup
    to restore volumes not rbd, need to find out the correct restore
    point.

    This is fixed in cinder/backup/drivers/ceph.py _diff_restore_allowed.

    Before the change: when backup is diff-format, and destination volume
    is not rbd, the function returns restore_point as None. As a result,
    backup restore uses wrong image/snapshot to do restore.

    With the change, we always get the right restore point, regardless
    of the kind of restore we'll do differential or not.

    Change-Id: I006c05d246f59fc6aff597543bedc68589d37576
    Closes-Bug: #1472088