check_deploy_timeout fails glance auth during clean_up

[aeva black]

Forked from bug https://bugs.launchpad.net/ironic/+bug/1300589 which is really about a different problem.

(ironic.conductor.manager): 2014-04-01 03:31:17,856 ERROR Timeout reached when waiting callback for node f36bc755-df36-40fc-8d57-9504e194eadd
(glanceclient.common.http): 2014-04-01 03:31:17,879 ERROR Request returned failure status.
(ironic.conductor.utils): 2014-04-01 03:31:17,879 ERROR Cleanup failed for node f36bc755-df36-40fc-8d57-9504e194eadd after deploy timeout: HTTPUnauthorized (HTTP 401)

Digging through the trace for this, it appears to be caused by ironic.conductor.manager:_check_deploy_timeout() triggering a clean_up(), but then that failing when trying to authenticate with Glance.

[Yuriy Zveryanskyy (yzveryanskyy)]

Context in periodic tasks does not have auth_token, this causes auth error with Glance. Better way to fix I think - remove get Glance info from clean_up() (use stored info).

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/85703

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/85703
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=92eda07045ea4202053d6fd705a3bbce08e8769d
Submitter: Jenkins
Branch: master

commit 92eda07045ea4202053d6fd705a3bbce08e8769d
Author: Yuriy Zveryanskyy <email address hidden>
Date: Mon Apr 7 15:43:48 2014 +0300

    Remove usage of Glance from PXE clean_up()

    Save uuids of instance kernel and ramdisk from Glance into node driver_info
    ('pxe_kernel' and 'pxe_ramdisk'). Stored values used in PXE clean_up()
    method.

    Closes-Bug: #1300602
    Change-Id: I59b3dbd4f66c59060e5b229eae00b80fdc4702e5