clone_image does not update provider_location

[Navneet]

The method clone_image does not return provider_location for the new volume made by cloning given image. Due to this cinder restart, in case of volumes present by cloning image, will error out as mount for volumes share fails. Also volume deletion leaves such volumes dangling in nfs share although it reports success in cinder client and cleans from database. The method clone_image needs to return a tuple (provider_location, cloned) to the manager.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/37097

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/37097
Committed: http://github.com/openstack/cinder/commit/632cdeca7923b7a06c3d7537b89a21d1790f230d
Submitter: Jenkins
Branch: master

commit 632cdeca7923b7a06c3d7537b89a21d1790f230d
Author: Navneet Singh <email address hidden>
Date: Sat Jul 13 12:25:30 2013 +0530

    Clone_image should return dict of vol properties, clone status.

    The method will work fine in case of drivers not dependent
    on volume properties like provider_location. It will fail to
    restart in case of nfs drivers and also leave volume created as
    result of clone_image functionality created in the
    nfs share in dangling state after deletion as provider_location
    is None. This fix requires dict of volume properties, cloned status
    to be returned which facilitates passing back provider_location
    in case of nfs drivers and hence resolves the issue.

    bug 1200708

    Change-Id: I590571e52d1c64b6dba7d7e76cd71badd74e51d1