When performing a "local delete" of an instance, we
need to delete the allocations that the instance has
against any resource providers in Placement.
It should be noted that without this change, restarting
the nova-compute service will delete the allocations
for its compute node (assuming the compute node UUID
is the same as before the instance was deleted). That
is shown in the existing functional test modified here.
The more important reason for this change is that in
order to fix bug 1756179, we need to make sure the
resource provider allocations for a given compute node
are gone by the time the compute service is deleted.
This adds a new functional test and a release note for
the new behavior and need to configure nova-api for
talking to placement, which is idempotent if
not configured thanks to the @safe_connect decorator
used in SchedulerReportClient.
Closes-Bug: #1679750
Related-Bug: #1756179
Change-Id: If507e23f0b7e5fa417041c3870d77786498f741d