On Mon, 19 Nov 2018 17:19:22 +0100, Surya Seetharaman wrote:

On Mon, Nov 19, 2018 at 2:39 PM Matt Riedemann <mriede...@gmail.com <mailto:mriede...@gmail.com>> wrote:

    On 11/19/2018 3:17 AM, melanie witt wrote:
     > - Not directly related to the session, but CERN (hallway track) and
     > NeCTAR (dev ML) have both given feedback and asked that the
     > policy-driven idea for handling quota for down cells be avoided.
     > the "propose counting quota in placement" spec to see if there's
    any way
     > forward here

    Should this be abandoned then?


    Since there is no microversion impact to that change, it could be added
    separately as a bug fix for the down cell case if other operators want
    that functionality. But maybe we don't know what other operators want
    since no one else is at multi-cell cells v2 yet.

I thought the policy check was needed until the "propose counting quota in placement" has been implemented as a workaround and that is what the "handling down cell" spec also proposed, unless the former spec would be implemented within this cycle in which case we do not need the
policy check.

Right, I don't think that anyone _wants_ the policy check approach. That was just the workaround, last resort idea we had for dealing with down cells in the absence of being able to count quota usage from placement.

The operators we've discussed with (CERN, NeCTAR, Oath) would like quota counting not to depend on cell databases, if possible. But they are understanding and will accept the policy-driven workaround if we can't move forward with counting quota usage from placement.

If we can get agreement on the count quota usage from placement spec (I have updated it with new proposed details), then we should abandon the policy-driven behavior patch. I am eager to find out what everyone thinks of the latest proposal.


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe

Reply via email to