Ironic virt driver makes too many API calls for resource tracking

[Jim Rollenhagen]

Current implementation of functions related to resource tracking query Ironic for all nodes, then makes an API request to Ironic for each node. This doesn't scale at all and should be reduced to one call in most cases.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/108545
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=d799a29a5a4f013aa457045dfeb6e91f9af611ff
Submitter: Jenkins
Branch: master

commit d799a29a5a4f013aa457045dfeb6e91f9af611ff
Author: Chris Behrens <email address hidden>
Date: Fri Jul 11 13:49:01 2014 -0700

    Query full node details and cache

    Re-use this information to reduce a massive amount of queries to ironic
    api to get info about individual nodes on the periodic task that syncs
    resource usage.

    Partial-Bug: 1357027

    Change-Id: Ie6b6898297a6f2346745c04a37d34dbb647ff6e7

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/113036
Committed: https://git.openstack.org/cgit/openstack/ironic/commit/?id=c05159c9c7201d926d423ce6ca5e0e679e8c60ad
Submitter: Jenkins
Branch: master

commit c05159c9c7201d926d423ce6ca5e0e679e8c60ad
Author: Jim Rollenhagen <email address hidden>
Date: Fri Aug 8 22:08:50 2014 +0000

    Use cache in node_is_available()

    This should make start-up times much better.

    Closes-Bug: 1357027

    Change-Id: I98ce118667dd9b8bf93ad781bfa8dc5077f34b45