deleting instance doesn't update scheduler immediately

[Vish Ishaya]

Previously deleting an instance would update the scheduler resources fairly quickly. There is now a delay when deleting an instance until the scheduler makes the resources available again. This appears to be due to the fact that the delete code path used to call resource tracker to update the compute_node record but this no longer happens.

[Vish Ishaya (vishvananda)]

This appears to be a regression introduced in https://review.openstack.org/#/c/36363/

[Chris Behrens (cbehrens)]

Broke in git hash dce64683291ba2cdb5e6617e01ccc2909254acb4

Proposed fix:

http://paste.openstack.org/show/85184/

[oleksii shyman (oshyman)]

I will fix that issue, but the thing is - I can't reproduce that bug.
I use nova cli to boot some instances, I use nova cli to delete them, and it updates fairly quickly.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/105589

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/105589
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=3085aa183d9602c3725ca699b7f1a92a4a4a5226
Submitter: Jenkins
Branch: master

commit 3085aa183d9602c3725ca699b7f1a92a4a4a5226
Author: oleksii <email address hidden>
Date: Tue Jul 8 13:44:25 2014 -0700

    Update scheduler after instance delete

    Previously deleting an instance would update the scheduler resources
    fairly quickly. There is now a delay when deleting an instance until the
    scheduler makes the resources available again. This appears to be due to
    the fact that the delete code path used to call resource tracker to
    update the compute_node record but this no longer happens.

    Closes-Bug: 1336080
    Change-Id: I837b70ce06843ff83481d4985c9c08b74d20cab0