← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1827628] Re: Cannot model affinity (and/or anti) with placement "limits" parameter

 

Reviewed:  https://review.opendev.org/658110
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=2f9e972ba3358fc5bc9bdc06faf47b21d509e20f
Submitter: Zuul
Branch:    master

commit 2f9e972ba3358fc5bc9bdc06faf47b21d509e20f
Author: Surya Seetharaman <suryaseetharaman.9@xxxxxxxxx>
Date:   Thu May 9 16:22:13 2019 +0200

    Disable limit if affinity(anti)/same(different)host is requested
    
    When max_placement_results is less than the total number of nodes in
    a deployment it may not be possible to use the affinity, anti-affinity,
    same host or different host filters as there is no guarantee
    for placement to return the expected hosts under such situations. This
    patch disables the max_placement_results parameter when nova queries
    placement for ``GET /allocation_candidates`` if the
    request_spec.scheduler_hints containts any of group, same_host or
    different_host keys.
    
    Change-Id: Ia2d5f80b6db59a8f6da03344aeaa6aa599407672
    Closes-Bug: #1827628


** Changed in: nova
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1827628

Title:
  Cannot model affinity (and/or anti) with placement "limits" parameter

Status in OpenStack Compute (nova):
  Fix Released
Status in OpenStack Compute (nova) queens series:
  Confirmed
Status in OpenStack Compute (nova) rocky series:
  Confirmed
Status in OpenStack Compute (nova) stein series:
  Confirmed

Bug description:
  It is currently not possible to use affinity/anti-affinity with the
  placement "limits" parameter since if you want your instance to land
  on a node where another instance lives you cannot rely on what
  placement GET/allocation_candidates would return and this could result
  in a no valid host. Current workaround is to unset the limits
  parameter.

  We already have the same-ish problem for disabled computes
  (https://bugs.launchpad.net/nova/+bug/1805984). We could just do a
  more generic solution that fits both these cases.

  The same problem is also applicable to
  https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#samehostfilter
  and
  https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#differenthostfilter

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1827628/+subscriptions


References