py34 tests hang forever

Bug #1424440 reported by Brian Curtin on 2015-02-22
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack SDK
Fix Released
Critical
Terry Howe

Bug Description

py34 jobs in Jenkins are failing, and you can reproduce this if you "tox -r". This will cause the test suite to run forever and you'll end up with python3.4 processes you'll need to kill manually.

Brian Curtin (brian.curtin) wrote :

This works: git checkout b067250080e6cc434563a2b507abe6f99fc3e73d && tox -r
The next change hangs on py34: git checkout 40fc003f0fbff5c0216ca2eaea7c79a77db5cbf5 && tox -r

I can't explain how subsequent check and gate jobs were able to succeed after that went through, including that change by itself, but starting at 2015-02-21 21:15 the hang started showing up and has caused the last five reviews to run jobs to end up failing py34 after hanging for 40 minutes.

Brian Curtin (brian.curtin) wrote :

Hmm, maybe it's not just that commit by itself. When on master, if you just change requirements.txt so six is 1.7.0 instead of 1.9.0, the hang still occurs.

Brian Curtin (brian.curtin) wrote :

Now b06725 also fails.

Terry Howe (thowe-g) wrote :

I assumed this was a problem someone else caused.

Brian Curtin (brian.curtin) wrote :

Actually I think it's httpretty, now that I see they released a new version yesterday.

Brian Curtin (brian.curtin) wrote :

(hit send earlier, sorry for extra noise)
One of the few times I was able to get any traceback out of the tests, something was going on inside a list related test which was using httpretty. I'll try tomorrow to get a more specific diagnosis if possible, but we may have to put a cap on the version in test-requirements.txt.

Brian Curtin (brian.curtin) wrote :

Ok, I think it's related to https://github.com/gabrielfalcao/HTTPretty/issues/221 which references https://github.com/gabrielfalcao/HTTPretty/pull/193 (by Dave Stanek, so this may have come up in Keystone).

Change abandoned by Brian Curtin (<email address hidden>) on branch: master
Review: https://review.openstack.org/158164
Reason: never mind, this has to go through global-requirements first

Brian Curtin (brian.curtin) wrote :

Submitted a change to openstack/requirements to have the 0.8.7 release skipped https://review.openstack.org/#/c/158165/

Brian Curtin (brian.curtin) wrote :

global-requirements changing again for 0.8.8 https://review.openstack.org/#/c/163664/ - fixed temporarily in fixgates.txt until we remove httpretty

Changed in python-openstacksdk:
status: New → Fix Committed
Terry Howe (thowe-g) on 2015-06-02
Changed in python-openstacksdk:
assignee: nobody → Terry Howe (thowe-g)
status: Fix Committed → Fix Released
milestone: none → halfliberty
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers