PureISCSIDriver should disconnect a volume from all hosts before trying to delete it

[Patrick East]

In some error conditions we will attempt to delete a volume which has been connected to a host in the Purity OE. When this delete then is requested we get an error from the Purity REST API and the volume is not deleted. The problem is that the volume is still there, but as far as Cinder is concerned it does not exist anymore.

To avoid leaving behind orphaned volumes when we have errors the PureISCSIDriver should remove the host connections to a volume in Purity before it deletes the volume.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/135816

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/135816
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=84378681d89c70185065df6958540f872c63bd3f
Submitter: Jenkins
Branch: master

commit 84378681d89c70185065df6958540f872c63bd3f
Author: Patrick East <email address hidden>
Date: Fri Sep 12 17:09:42 2014 -0700

    PureISCSIDriver needs to disconnect hosts before deleting volumes.

    Some error conditions can cause a situation where the volume is
    “connected” to a Purity host, but the volume failed to attach on the
    initiator side. If we then try to delete this volume it will cause
    errors and leave orphaned volumes behind on the Flash Array.

    The solution is to check for any connections and then remove them prior
    to making a call to delete the volume.

    Closes-Bug: 1388260
    Change-Id: I8f5b3b28900d79228fd91f2ad53535224f263657