LBAAS Intermittent gate failure TestHealthMonitorBasic.test_health_monitor_basic

[Michael Johnson]

We are seeing intermittent gate failures at an increasing rate for the neutron-lbaas scenario test.

Example gate run:
http://logs.openstack.org/50/259550/7/check/gate-neutron-lbaasv2-dsvm-scenario/d48b77d/

neutron_lbaas.tests.tempest.v2.scenario.test_healthmonitor_basic.TestHealthMonitorBasic.test_health_monitor_basic [456.618148s] ... FAILED

File "neutron_lbaas/tests/tempest/v2/scenario/test_healthmonitor_basic.py", line 45, in test_health_monitor_basic
File "neutron_lbaas/tests/tempest/v2/scenario/base.py", line 519, in _traffic_validation_after_stopping_server
File "neutron_lbaas/tests/tempest/v2/scenario/base.py", line 509, in _send_requests
socket.error: [Errno 104] Connection reset by peer

[Nate Johnston (nate-johnston)]

Confirmed that I see a spike in failures for this test starting between 13:40 and 13:45 EST, but after 14:50 the rate seems to go back down to normal levels. Please recheck and see if the issue is still going on; if so I will reclassify to 'Confirmed'.

http://logstash.openstack.org/#/dashboard/file/logstash.json?query=message:\%22AssertionError:%200%20%3D%3D%200%20:%20No%20IPv4%20addresses%20found%20in:%20[]\%22%20AND%20build_name:\%22gate-tempest-dsvm-neutron-linuxbridge\%22

[Nate Johnston (nate-johnston)]

Better URL:

http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22FAILURE%5C%22%20and%20build_name%3A%5C%22gate-neutron-lbaasv2-dsvm-scenario%5C%22

[Adam Harwell (adam-harwell)]

Definitely saw this today. http://logs.openstack.org/32/277532/1/check/gate-neutron-lbaasv2-dsvm-scenario/50dee30/console.html#_2016-02-08_20_04_52_372

[Adam Harwell (adam-harwell)]

This might be the more useful snippet: http://logs.openstack.org/32/277532/1/check/gate-neutron-lbaasv2-dsvm-scenario/50dee30/console.html#_2016-02-08_20_32_00_492

[Adam Harwell (adam-harwell)]

And in a different place: http://logs.openstack.org/02/277602/1/check/gate-neutron-lbaasv2-dsvm-scenario/6dc697b/console.html#_2016-02-09_06_35_34_128

Bad Status Line may be due to a listening bug on the *amp agent* side??