Open Stack

Hi Folks,

I'd like the following change to be considered for an FFE please:  https://review.openstack.org/#/c/34291/

This change adds a capability that was agreed as needed at the Havana design summit, is functionally complete (including the corresponding python-novaclient support (https://review.openstack.org/#/c/38847/), and has been through many review cycles since it was first posted in June.   The change includes both V2 and V3 api changes.

The only remaining review comment at the point when it got hit by FF was whether it was appropriate for this change to introduce a secheduler/api.py layer rather than following the existing (and exceptional) process of calling the scheduler/rpcapi.py methods directly.   Since this is the first query to call (rather than cast) into the scheduler it seems to me a sensible abstraction to add, and makes the scheduler consistent with all other services in Nova.

The change is low risk in that it only adds a new query path to the scheduler, and does not alter any existing code paths.

Thanks for the consideration,
Phil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130912/e8b74ef9/attachment.html>