neutron: VirtualInterface isn't cleaned up before rescheduling during allocation failure

[Matt Riedemann]

This is a follow-on to bug 1602357. That fixed the case that we delete the VirtualInterface objects in the nova db when deallocating networks for an instance or a single port.

But if we fail to allocate networking from the start, we also do a cleanup on the ports we've created and/or updated, but we aren't deleting the VIFs we've created, here:

https://github.com/openstack/nova/blob/92a388a1e34559b2ce69d31fdef996ff029495a6/nova/network/neutronv2/api.py#L847

That also needs to happen because we could do something like:

1. create/update port1, create vif1, ok
2. create/update port2, fails - we deallocate port1 and port2 but not vif1

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/342416

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/342416
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=7a8f1369c5dd9d6a6260ba0c168d06e9f0016d33
Submitter: Jenkins
Branch: master

commit 7a8f1369c5dd9d6a6260ba0c168d06e9f0016d33
Author: Matt Riedemann <email address hidden>
Date: Thu Jul 14 16:51:35 2016 -0400

    neutron: destroy VIFs if allocating ports fails

    Following on 92a388a1e34559b2ce69d31fdef996ff029495a6 we also
    need to rollback VirtualInterfaces that we create if allocation
    fails since a reschedule with a pre-existing port will fail
    allocation with a duplicate VIF constraint error.

    A new test is added for the rollback section being changed
    since there isn't an obvious existing test that this could be
    worked into.

    Change-Id: I04ca4fff319912ce2a8dc07308a3aec1c24dc58c
    Closes-Bug: #1603197

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/nova 14.0.0.0b3 development milestone.