Devstack runs failing due to "Failure creating NET_ID for 36033ee931044cb5bc76570da0eff97b"

[Ben Nemec]

This problem has popped up on a number of changes recently. Devstack fails to complete successfully so the change fails without ever running any tests.

Example error.txt: http://logs.openstack.org/37/61037/9/check/check-tempest-dsvm-neutron/6d98c79/logs/error.txt.gz

[Matt Riedemann (mriedem)]

Also seeing it here:

http://logs.openstack.org/19/60719/4/check/gate-tempest-dsvm-neutron-large-ops/1f73559/logs/error.txt.gz

http://logs.openstack.org/19/60719/4/check/check-tempest-dsvm-neutron/4c21e32/logs/error.txt.gz

[Matt Riedemann (mriedem)]

Doesn't look like logs/error.txt.gz is indexed so I'm not sure what we can key off of for an elastic-recheck query.

[Matt Riedemann (mriedem)]

I opened bug 1265906 for indexing on logs/error.txt.gz since we'd need that to query against this failure it seems.

[Matt Riedemann (mriedem)]

I'm not seeing anything jump out at me from the recent neutron commit history, but there are a few in the recent devstack commit history that make me wonder about regressions:

https://review.openstack.org/61854
https://review.openstack.org/63493
https://review.openstack.org/62306
https://review.openstack.org/63641

Given the timing, the first one seems to be the most suspicious.

[Ben Nemec (bnemec)]

Looks like this is probably caused by https://bugs.launchpad.net/python-neutronclient/+bug/1265926 A fix (https://review.openstack.org/#/c/64910/) is in the gate waiting to be merged. Going to close this bug for now, will reopen if that doesn't fix the problem.

[Clint Byrum (clint-fewbar)]

I think rather than close the bug, it should be marked as a duplicate. Doing so.

[Steve Martinelli (stevemar)]

Starting to see this again here: https://review.openstack.org/#/c/69107/

log for reference: http://logs.openstack.org/07/69107/4/check/check-tempest-dsvm-neutron/34de6dc/logs/error.txt.gz