CSIT Sporadic failures - tempest.api.network precommit_delete failure on security group delete

Description

this csit job:

https://logs.opendaylight.org/releng/jenkins092/netvirt-csit-1node-openstack-pike-upstream-stateful-nitrogen/3/log.html.gz

tempest output snippet:

{2}

tearDownClass (tempest.api.network.test_extensions.ExtensionsTestJSON) [0.000000s] ... FAILED

Captured traceback:
~~~~~~~~~~~~~~~~~~~
Traceback (most recent call last):

Environment

Operating System: All
Platform: All

Attachments

Activity

Show:

Kit Lou February 9, 2018 at 11:53 PM

Changing severity to major as we have confirmation that this is not a blocker.

JamO Luhrsen January 26, 2018 at 10:20 PM

Thanks .

any chance we can get this ported to pike? he's my try for that:
https://review.openstack.org/#/c/538352/

Jaime Caamaño Ruiz January 26, 2018 at 2:19 PM

For sake of completeness, this patch got rid of the problem also:

https://review.openstack.org/#/c/536935/

Zhijiang Hu January 26, 2018 at 1:41 AM

The reason why I think it was an blocker for OPNFV MS3.2 becasue I saw it happened many times in OPNFV functest (with Daisy installer which deploy OpenStack Pike + ODL carbon), for example, http://artifacts.opnfv.org/logs/functest/zte-pod3/2018-01-21_19-53-22/functest.log

Following the JIRA ticket I got the solution patch (https://review.openstack.org/#/c/533706/) and I have patched it dowstream into Daisy's repo first(https://gerrit.opnfv.org/gerrit/#/c/51031/). And functest for Daisy's sdn-nofeature scenario is stable now. So from Daisy Installer side, it is no longer a blocker for OPNFV MS3.2. Looking forward to getting the patch merged upstream.

JamO Luhrsen January 25, 2018 at 6:58 PM

, can we get someone to update this Jira with an explanation of why it is a blocker?

BTW, in case it's not clear in the comments above, this bug has it's fix coming on the openstack
side, not OpenDaylight side. I don't know how fast that will happen.