snapshot is stuck in 'deleting' after reboot cinder-volume in cascade deleting process

[wanghao]

When cascade deleting volume and snapshot, reboot cinder-volume, after it snapshot is stuck in 'deleting' status and volume is gone. It's more worse that you can't delete this
snapshot even reset its status, since no volume is existing.

This should be fixed by adding process to handle the 'deleting' snapshot in init_host.

Reproduce step:

1. Create volume and snapshot.
2. Cascade deleting this volume.
3. When deleting is doing, reboot cinder-volume.
4. After this, snapshot is stuck in 'deleting' status.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/322418

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/322418
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=bc9d737c87cb21e308a7a3be66ddd792418503ff
Submitter: Jenkins
Branch: master

commit bc9d737c87cb21e308a7a3be66ddd792418503ff
Author: wanghao <email address hidden>
Date: Sat May 28 17:19:26 2016 +0800

    Fix snapshot stuck in 'deleting' after reboot c-v

    When cascade deleting volume and snapshot, reboot
    cinder-volume, snapshot will be stuck in 'deleting'
    after it, and it's more worse that user can't delete
    this snapshot even reset its status, since no deleting
    volume is exitsing any more after init host.

    Fix this bug by using cascade deleting volume
    when init host, as any volume that is in deleting
    status has already passed API checks and should
    be deleted with all its snapshots.

    Change-Id: Ic47290b8b7004b9e37d096f52af025931ab13c6c
    Closes-Bug: #1586604

[Doug Hellmann (doug-hellmann)]

This issue was fixed in the openstack/cinder 9.0.0.0b2 development milestone.