yahoo-eng-team team mailing list archive

[Thierry Carrez <thierry.carrez+lp@xxxxxxxxx>]

** Changed in: neutron
       Status: Fix Committed => 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/1418097

Title:
  Kernel panic or ovs-vswitch crash when running DVR functional test

Status in OpenStack Neutron (virtual network service):
  Fix Released

Bug description:
  Still early in triage process. Running:
  tox -e dsvm-functional neutron.tests.functional.agent.test_l3_agent.TestDvrRouter.test_dvr_router_lifecycle_without_ha_with_snat_with_fips

  Results in a kernel panic or ovs-vswitchd crash, with this patch:
  https://review.openstack.org/#/c/151882/

  Rebasing one patch earlier results in smooth test runs.

  If I don't get a kernel panic, then ovs-vswitchd crashes, at which
  point most ovs-vsctl commands fail. Restarting the openvswitch service
  doesn't seem to solve it, only restarting the machine.

  Note: It's not necessarily that specific test that is causing issues,
  but it is one that reproduces the issue reliably. Running the non-DVR
  test case (L3AgentTestCase) works fine. So does running the OVS lib
  functional tests.

  My devstack VM is running on:
  * Fedora 20
  * openvswitch-2.3.1-1.fc20.x86_64
  * kernel-3.17.2-200.fc20.x86_64

  I checked with team mates, the same test on the same code passes with:
  * CentOS 7
  * openvswitch-2.1.2-2.el7.centos.1.x86_64
  * kernel-3.10.0-123.6.3.el7.x86_64

  This may happen as a result of a Neutron code change in the patch
  above, possibly coupled with a regression in a recent OVS version. I
  think it's clear to assume that this is not only a test-only issue,
  but could happen in production as well.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1418097/+subscriptions