test seems to fail occasionally nova.tests.conductor.test_conductor.ConductorLocalAPITestCase.test_vol_usage_update_again

[John Garbutt]

Full logs:
http://logs.openstack.org/32760/5/check/gate-nova-python27/24601/console.html

[David Jia (jiataotj)]

this bug affected my code review: https://review.openstack.org/#/c/30694/ more than three times, it used to be replaced by another bug, but it occurs again now

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/33289

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/33289
Committed: http://github.com/openstack/nova/commit/d73113a52c9317bef143ffc545c42054b67a49dc
Submitter: Jenkins
Branch: master

commit d73113a52c9317bef143ffc545c42054b67a49dc
Author: Dan Smith <email address hidden>
Date: Mon Jun 17 08:55:00 2013 -0700

    Fix overzealous conductor test for vol_usage_update

    The rest of the conductor tests are pretty unit-esque, in that
    they don't test the underlying functionality. Rather, they confirm
    that the various conductor APIs (public, local, rpc, and manger)
    all agree and that the plumbing between them works as expected.

    The test_vol_usage_update() test does too much, and as a result,
    gets all twisted in how it mucks with mox to stub out utcnow()
    without breaking the base test harness. Unfortunately, it seems
    to fail quite often, depending on testr ordering, and has been
    blocking the gate quite a bit lately.

    This patch refactors that test to be like all the rest of the
    conductor tests. It also removes the useless mirror test of
    test_vol_usage_update_again().

    Fixes bug 1190971

    Change-Id: I815b25f6585786a8dbd5eefcfb39e2b4790641cb