yahoo-eng-team team mailing list archive

[Rodolfo Alonso <1931244@xxxxxxxxxxxxxxxxxx>]

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

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1931244

Title:
  ovn sriov broken from ussuri onwards

Status in Ubuntu Cloud Archive:
  Fix Released
Status in Ubuntu Cloud Archive ussuri series:
  Fix Released
Status in Ubuntu Cloud Archive victoria series:
  Invalid
Status in Ubuntu Cloud Archive wallaby series:
  Fix Released
Status in Ubuntu Cloud Archive xena series:
  Fix Released
Status in neutron:
  Fix Released
Status in neutron package in Ubuntu:
  Fix Released
Status in neutron source package in Focal:
  Fix Released
Status in neutron source package in Hirsute:
  Fix Released
Status in neutron source package in Impish:
  Fix Released

Bug description:
  I have an Openstack Ussuri 16.3.2 deployment using OVN. When I create
  a vm with one or more sriov ports it fails with:

  2021-06-08 11:38:31.939 526862 WARNING nova.virt.libvirt.driver [req-
  c4be797e-7d7e-4e73-8406-f74ae51db192 696c98b722a44d229e16b6d6474a27d4
  0b9102977dcc4d4ab662b48494bb3110 - 2e0bf6ec95c047d986a61f7570222149
  2e0bf6ec95c047d986a61f7570222149] [instance: 7ab9b374-51eb-
  4e94-8055-c69e8a7d76b3] Timeout waiting for [('network-vif-plugged',
  'c2b7c68d-c465-4ca2-869a-59bc73b2b595'), ('network-vif-plugged',
  'a50de16a-29ac-4dca-9cb6-0247a932fbf3')] for instance with vm_state
  building and task_state spawning.: eventlet.timeout.Timeout: 300
  seconds

  A bit of analysis shows that nova-compute did its thing and sits there
  waiting on network-vif-plugged. The sriov-agent then notices new VFs
  configured and sends a get_devices_details_list() rpc call to neutron
  and neutron never responds. Reverting to 16.3.1 fixes the issue.
  Taking a closer look at 16.3.2 by reverting patches lead to [1] as the
  culprit.

  [1]
  https://github.com/openstack/neutron/commit/7cf9597570f288d27768dc5ff7be04824d09f8bc

  === Ubuntu SRU details ===
  [Impact]
  [Test Case]
  See above.
  I think for testing we can run standard regression testing with OVN/neutron deployments plus tempest testing.

  For now we are planning to revert the commit as a stop-gap to prevent
  further upgrades from being regressed.

  [Regression Potential]
  There is regression potential in that the patch being reverted contributes partial fixes to the following related bugs. Considering most openstack users are on ussuri at this point and 16.3.2 has not been available for very long the revert that we are proposing would seem to have the least amount of regression potential.
  https://bugs.launchpad.net/neutron/+bug/1894117
  https://bugs.launchpad.net/neutron/+bug/1903008

To manage notifications about this bug go to:
https://bugs.launchpad.net/cloud-archive/+bug/1931244/+subscriptions