_check_deploy_timeouts can attempt to spawn more processes then there are threads available

[Chris Krelle]

in ironic/conductor/manager.py the function _check_deploy_timeouts could attempt to spawn more worker threads than there are available in the thread pool, resulting in spurious errors being logged (but not causing any service failure). This should be avoided by adding a query filter for provision_updated_at, sort oldest first, and use a reasonably small LIMIT (eg, smaller than rpc_thread_count)

This bug is based on review: https://review.openstack.org/#/c/72395

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/78651

[Openstack Gerrit (openstack-gerrit)]

Reviewed: https://review.openstack.org/78651
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=d473bdabfdaf8fa2538de1056e804355a40979dc
Submitter: Jenkins
Branch: master

commit d473bdabfdaf8fa2538de1056e804355a40979dc
Author: Yuriy Zveryanskyy <email address hidden>
Date: Thu Mar 6 18:00:41 2014 +0200

    Add worker threads limit to _check_deploy_timeouts task

    New nodes filter parameters added to db.sqlalchemy.api. This
    parameters used in _check_deploy_timeouts periodic task for
    limit of worker threads which can be started simultaneously.
    New config option 'periodic_max_workers' added to conductor
    group.

    Closes-Bug: #1285793
    Change-Id: I646540e334dec05682640c05cedb315e0ee355bc