Comment 32 for bug 1353962

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master)

Reviewed: https://review.openstack.org/168158
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=af4ce3ee63e4e4af9113405396c3201eb3efe3c6
Submitter: Jenkins
Branch: master

commit af4ce3ee63e4e4af9113405396c3201eb3efe3c6
Author: Matt Riedemann <email address hidden>
Date: Thu Mar 26 13:46:40 2015 -0700

    Add debug logging to quota_reserve flow

    There are some places in the quota_reserve flow that could use some
    debug logging for helping to figure out the problem when we over quota:

    1. Log when we detect a desync in _is_quota_refresh_needed and force a
       refresh.
    2. There is a comment in quota_reserve after we call
       _refresh_quota_usages saying we assume the sync was OK and
       essentially everything is hunky dory. Log if we actually refreshed
       without a change to usage.
    3. Log the details from project/user quotas/usages when we go over
       quota in _calculate_overquota.
    4. Log if we update project_usages using values from user_usages in
       quota_reserve.
    5. Finally, before we raise the OverQuota exception in quota_reserve,
       the user_usages dict values are UserQuota objects which are sqla
       dict-like objects and don't stringify nicely for the logs, so
       convert those to actual dict values for logging.

    Related-Bug: #1353962

    Change-Id: Ic6e3d5fdf51c0a13c97509d0888a3bba00fb3193