ScaleIO Volumes Created before Newton, do not connect or disconnect after Newton upgrade

[Wade Holler]

Description
===========
ScaleIO volumes that were created pre stable/newton ( stable/mitaka ) in my case,
do not connect or disconnect to/from the hypervisor when nova-compute attempts the
connection after the upgrade to stable/newton.

This appears to be because the os_brick 1.2 to 1.6 change includes a new attribute
'scaleIO_volume_id'
This field is not created if the volumes were created at stable/mitaka or before.
In this case, this field is not found in nova.block_device_mappings 'connection_info'
column. This field is created for volumes created after stable/newton upgrade.

Steps to reproduce
==================
1. Install ScaleIO
2. Install openstack via openstack-ansible stable/mitaka
3. Create and connect ScaleIO volumes to instances
4. Upgrade to stable/newton
5. Shutdown a instance
6. Attempt start of a instance

Expected result
===============
The instance should start

Actual result
=============
The instance would not start.
http://pastebin.com/Pvd6HrcL

--after manually adding the 'scaleIO_volume_id' parameter to
nova.block_device_mapping.connection_info the connect / disconnect
works as expected.

Environment
===========
1. Exact version of OpenStack you are running. See the following
  stable/newton deployed via openstack-ansible 14.0.3

2. Which hypervisor did you use?
   libvirt+kvm

2. Which storage type did you use?
   scaleIO 2.0.1

3. Which networking type did you use?
   neutron with linux bridge

Logs & Configs
==============
http://pastebin.com/Pvd6HrcL

os_brick version with OSA stable/mitaka = 1.2
os_brick version with OSA stable/newton = 1.6

[Scott DAngelo (scott-dangelo)]

Added Brick and deleted Nova.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/640875
Committed: https://git.openstack.org/cgit/openstack/os-brick/commit/?id=6b60614afaf88819ce92a2d0ed269035a3c8b261
Submitter: Zuul
Branch: master

commit 6b60614afaf88819ce92a2d0ed269035a3c8b261
Author: Erik Olof Gunnar Andersson <email address hidden>
Date: Mon Mar 4 13:21:18 2019 -0800

    Fix VxFlexOs KeyError after upgrade

    A long time ago the driver was changed to store volume_id in the
    nova database and use that. Previously it would use the volume_name
    and make a call to get the volume_id.

    Both should be supported, but as the code path currently
    throws a KeyError it will never reach the backwardscompatible
    code.

    This makes sure that the driver works even after an upgrade
    from Mitaka.

    Change-Id: Ic389e1a8f0f43ef410eee89202fef19d554fce90
    Closes-Bug: #1648629

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/643092

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/643092
Committed: https://git.openstack.org/cgit/openstack/os-brick/commit/?id=05b6c9f9fad15354f598bbd332a5d56230ece1e0
Submitter: Zuul
Branch: master

commit 05b6c9f9fad15354f598bbd332a5d56230ece1e0
Author: Yury Kulazhenkov <email address hidden>
Date: Wed Mar 13 18:58:50 2019 +0300

    Fix ScaleIO KeyError after upgrade

    A long time ago the driver was changed to store volume_id
    in the nova database and use that. Previously it would use the
    volume_name and make a call to get the volume_id.

    Both should be supported, but as the code path currently throws
    a KeyError it will never reach the backwards compatible code.
    This makes sure that the driver works even after an upgrade from Mitaka.

    Depends-On: https://review.openstack.org/#/c/643130/
    Change-Id: Iab6a3ca6bf92dc5606c7723dc6c62bf07fa1d934
    Closes-Bug: #1648629

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/os-brick 2.8.1 release.