volume create race condition protection

[Edward Hope-Morley]

Creating a volume from another volume or snapshot is currently prone to being interrupted by a concurrent delete of the dependant volume or snapshot. Concern over this was also raised within the new taskflow implementation:

https://github.com/openstack/cinder/blob/master/cinder/volume/flows/create_volume/__init__.py#L463

[Edward Hope-Morley (hopem)]

Applying this patch and then doing a create from snapshot and delete snapshot concurrenlty will trigger the expected exception.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/56442

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/56442
Committed: http://github.com/openstack/cinder/commit/4f6e5fcc252799e2b9207b6ef2b58b52a7a93563
Submitter: Jenkins
Branch: master

commit 4f6e5fcc252799e2b9207b6ef2b58b52a7a93563
Author: Edward Hope-Morley <email address hidden>
Date: Thu Nov 14 19:00:00 2013 +0000

    Adds lock for create from vol/snap to avoid race conditions

    This patch protects create from volume/snapshot by using a
    lockfile to protect the operation from concurrent deletes of
    the volume/snapshot used in the create operation.

    Currently, if a volume/snapshot is deleted while a volume is
    being created from it that delete may complete during the
    create operation thus leaving the new volume in error or stuck
    state. This lock will ensure that:

    (a) if a create of VolA from snap/volB is in progress, any
        delete requests for snap/volB will wait until the create
        is complete.

    (b) if a delete of snap/volA is in progress, any create from
        snap/volA will wait until snap/volA delete is complete.

    Co-authored-by: Takashi Natsume <email address hidden>
    Closes-Bug: 1251334
    Change-Id: Ie4bc0af789ab232593f55aa2f6b34345eb9b9929