model_update in failover_replication is not being applied to the db model

[Patrick East]

In failover_replication we get some updates from the volume driver method and then the idea is to apply them to the volume being failed over. In this process we end up sending a dictionary that looks something like:

{'model': {'provider_location': 'foo'}} which should really be just {'provider_location':'foo'} when passed into db.volume_update(..)

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/219875

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/219875
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=ca7050ec206284bf74cd06ab7b24178967c42bcc
Submitter: Jenkins
Branch: master

commit ca7050ec206284bf74cd06ab7b24178967c42bcc
Author: Patrick East <email address hidden>
Date: Wed Sep 2 13:58:02 2015 -0700

    Use consolidated update for failover_replication

    Right now it’s using a special built volume_updates dictionary that
    contains some fields *and* a model update. We then glom all of them
    together into an update dictionary for the volume model.

    This change simplifies them into a single model update more in-line with
    how the other volume API’s work for drivers.

    It also fixes a bug where the model_update property of the volume_update
    was not being applied to the db model.

    Change-Id: I8ad85bf0e62099a8ea4dd471dfcb64fa057f1021
    Closes-Bug: #1491587