Scheduler should handle instance being deleted

[Phil Day]

Currently if an instance is deleted whilst the scheduler is looking for a suitable host it will log an error due to the InstanceNotFound exception being raised when it goes to update the scheduled_at value.

This isn't really an error as a user can delete an instance at any time, and some tests frameworks seem to do this very quickly.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/33122

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/33122
Committed: http://github.com/openstack/nova/commit/79039a7f6ffce2d5cc2e92eee03de94a3c2292f8
Submitter: Jenkins
Branch: master

commit 79039a7f6ffce2d5cc2e92eee03de94a3c2292f8
Author: Phil Day <email address hidden>
Date: Fri Jun 14 22:21:22 2013 +0100

    Handle instance being deleted while in filter scheduler

    Adds handling of InstanceNotFound when trying to update the
    scheduled_at time of an instance.

    Users can delete an instance at any time (and many test
    cases delete very quickly), so this should be a warning
    not an error

    Fixes bug 1191126

    Change-Id: I72b296668f60b1de71ff8c6d7bbb749023d89f12