← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1684250] Re: Timeout waiting for vif plugging callback for instance

 

Reviewed:  https://review.openstack.org/585661
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=bef7e9a717f3bca45b5edd0ebc65ab1dde5942d9
Submitter: Zuul
Branch:    master

commit bef7e9a717f3bca45b5edd0ebc65ab1dde5942d9
Author: Lucian Petrut <lpetrut@xxxxxxxxxxxxxxxxxxxxxx>
Date:   Wed Jul 25 10:17:50 2018 +0300

    Hyper-V + OVS: plug vifs before starting VMs
    
    At the moment, the Hyper-V driver will bind OVS ports only after
    starting the VMs. This is because of an old limitation of the OVS
    Windows port, which has been addressed in OVS 2.5.
    
    This also means that the Nova driver fails when waiting for neutron
    vif plug events, since this step is performed later. So deployers
    are forced to set "vif_plugging_is_fatal", which means that the
    ports will be down, at least for a short time after the instance is
    reported as ACTIVE. This also breaks some Tempest tests.
    
    Since the Windows OVS version has been fixed, we can now bind the
    ports before actually starting the VMs. Note that this fix has been
    merged in the out-of-tree compute-hyperv some time ago.
    
    Co-authored-by: Alin Balutoiu <abalutoiu@xxxxxxxxxxxxxxxxxxxxxx>
    
    Change-Id: I0116a593c29808a53423cc9bbe098f2661fcb584
    Closes-Bug: #1684250


** 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/1684250

Title:
  Timeout waiting for vif plugging callback for instance

Status in compute-hyperv:
  Fix Released
Status in OpenStack Compute (nova):
  Fix Released

Bug description:
  At the moment, the Hyper-V driver will bind OVS ports only after
  starting the VMs. This is because of an old limitation of the OVS
  Windows port, which has been addressed in OVS 2.5.

  This also means that the Nova driver fails when waiting for neutron
  vif plug events[1], since this step is performed later. So deployers
  are forced to set "vif_plugging_is_fatal", which means that the ports
  will be down, at least for a short time after the instance is reported
  as ACTIVE. This also breaks some Tempest tests.

  [1] Trace: http://paste.openstack.org/show/DNKXXMcfqP9KcLY3Da3v/

To manage notifications about this bug go to:
https://bugs.launchpad.net/compute-hyperv/+bug/1684250/+subscriptions