replication_status not properly set on volume creation

[Gorka Eguileor]

The scheduler is currently always setting `replication_status` to "disabled" regardles of the real replication status.

Some drivers are updating it on volume creation, but even in that case the user will have received the API REST response with the `replication_status` set to "disabled" and will have to do a `show` of the volume once the creation has been completed to confirm if the volume is actually replicated or not.

And there are some drivers that aren't updating the status, so we should be setting it at the scheduler.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/400803

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/400803
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=7647f994f10ca2678244ef8034baaf99b3ce9482
Submitter: Jenkins
Branch: master

commit 7647f994f10ca2678244ef8034baaf99b3ce9482
Author: Gorka Eguileor <email address hidden>
Date: Thu Nov 17 18:24:30 2016 +0100

    Set replication_status on volume creation

    The scheduler is currently always setting `replication_status` to
    "disabled" regardles of the real replication status.

    Some drivers are updating it on volume creation, but even in that case
    the user will have received the API REST response with the
    `replication_status` set to "disabled" and will have to do a `show` of
    the volume once the creation has been completed to confirm if the volume
    is actually replicated or not.

    This patch changes this and allows the scheduler to correctly set the
    `replication_status` on volume creation based on the standard property
    `replication_enable`.

    This doesn't solve 100% of the cases, as a Cloud Administrator may not
    be using the `replication_enable` extra specs in volume types to set
    replication because it's using a backed that is replicating at
    backend/pool level.

    In that case there will be a discrepancy, just like we have now, until
    the driver completes the creation and returns the model update for the
    volume.

    Closes-Bug: #1643883
    Change-Id: Id4df2b7ad55e9b5def91329f5437da9caa185c30

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/cinder 10.0.0.0b3 development milestone.