test_floating_mangle_rules fails with mismatch error

[Armando Migliaccio]

http://logs.openstack.org/09/336609/1/gate/gate-neutron-python27-ubuntu-xenial/31d11e0/testr_results.html.gz

ft23.24: neutron.tests.unit.agent.l3.test_dvr_local_router.TestDvrRouterOperations.test_floating_mangle_rules_StringException: Empty attachments:
  stdout

pythonlogging:'': {{{WARNING [stevedore.named] Could not load neutron.agent.linux.interface.NullDriver}}}
stderr: {{{
neutron/agent/l3/agent.py:375: DeprecationWarning: L3_AGENT_MODE: moved to neutron_lib.constants
  connection, l3_constants.L3_AGENT_MODE)
}}}

Traceback (most recent call last):
  File "neutron/tests/unit/agent/l3/test_dvr_local_router.py", line 251, in test_floating_mangle_rules
    self.assertEqual(expected, actual)
  File "/home/jenkins/workspace/gate-neutron-python27-ubuntu-xenial/.tox/py27/local/lib/python2.7/site-packages/testtools/testcase.py", line 411, in assertEqual
    self.assertThat(observed, matcher, message)
  File "/home/jenkins/workspace/gate-neutron-python27-ubuntu-xenial/.tox/py27/local/lib/python2.7/site-packages/testtools/testcase.py", line 498, in assertThat
    raise mismatch_error
testtools.matchers._impl.MismatchError: !=:
reference = [('floatingip', '-d 15.1.2.3 -i fake_router -j MARK --set-xmark fake_mark'),
 ('FORWARD', '-s 192.168.0.1 -j $float-snat')]
actual = [('floatingip', '-d 15.1.2.3/32 -i fake_router -j MARK --set-xmark fake_mark'),
 ('FORWARD', '-s 192.168.0.1/32 -j $float-snat')]

[Brian Haley (brian-haley)]

This doesn't seem critical, it's regarding a failure for a patch that hasn't merged yet - https://review.openstack.org/#/c/336609/ - I'll work on fixing the patch to pass the tests it added.

I'm guessing it was found in the gate since it had previously passed jenkins before Kevin's iptables verification patch merged?

[Carl Baldwin (carl-baldwin)]

Logstash has a few hits but they're all pretty recent. Something merged recently?

http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22in%20test_floating_mangle_rules%5C%22%20AND%20voting%3A1

[Brian Haley (brian-haley)]

That test was being added in https://review.openstack.org/#/c/336609/ which never merged, perhaps someone based another change on-top of it?

Either way, I've pushed a new change that should fix the issue.

[Armando Migliaccio (armando-migliaccio)]

Ok I see what's happening, this is an example of two competing changes ending up in the gate interleaving and thus causing a failure in the merge phase..

[Carl Baldwin (carl-baldwin)]

I was worried that there were two other changes that saw the same failures that were found in logstash. But, I know why that happened now. They were approved to merge after this one was approved and so they were merged with 336609 by zuul assuming that it would be merged first.

So, Brian is correct, this unit test failure does not appear in master. We were just looking at the making of a big gate reset due to the bad unit test addition in https://review.openstack.org/#/c/336609/ . Good work Brian!