Open Stack

Hi,

On 21 Feb 2014, at 20:34, Aaron Rosen <aaronorosen at gmail.com<mailto:aaronorosen at gmail.com>> wrote:

Hi,

Yesterday, I pushed a patch to review and was surprised that several of the third party CI systems reported back that the patch-set worked where it definitely shouldn't have. Anyways, I tested out my theory a little more and it turns out a few of the 3rd party CI systems for neutron are just returning  SUCCESS even if the patch set didn't run successfully (https://review.openstack.org/#/c/75304/).

Here's a short summery of what I found.

Hyper-V CI -- This seems like an easy fix as it's posting "build succeeded" but also puts to the side "test run failed". Would probably be a good idea to remove the "build succeeded" message to avoid any confusion.

The Hyper-V CI is non-voting (as required for new third-party CIs) and this is the reason why any post from it will show “build succeeded”. As published in other threads, AFAIK the only way to get rid of this issue is to have the CI as voting.

Brocade CI - From the log files it posts it shows that it tries to apply my patch but fails:


2014-02-20 20:23:48 + cd /opt/stack/neutron
2014-02-20 20:23:48 + git fetch https://review.openstack.org/openstack/neutron.git refs/changes/04/75304/1
2014-02-20 20:24:00 From https://review.openstack.org/openstack/neutron
2014-02-20<https://review.openstack.org/openstack/neutron2014-02-20> 20:24:00  * branch            refs/changes/04/75304/1 -> FETCH_HEAD
2014-02-20 20:24:00 + git checkout FETCH_HEAD
2014-02-20 20:24:00 error: Your local changes to the following files would be overwritten by checkout:
2014-02-20 20:24:00     etc/neutron/plugins/ml2/ml2_conf_brocade.ini
2014-02-20 20:24:00     neutron/plugins/ml2/drivers/brocade/mechanism_brocade.py
2014-02-20 20:24:00 Please, commit your changes or stash them before you can switch branches.
2014-02-20 20:24:00 Aborting
2014-02-20 20:24:00 + cd /opt/stack/neutron

but still continues running (without my patchset) and reports success. -- This actually looks like a devstack bug  (i'll check it out).

PLUMgrid CI - Seems to always vote +1 without a failure (https://review.openstack.org/#/dashboard/10117) though the logs are private so we can't really tell whats going on.

I was thinking it might be worth while or helpful to have a job that tests that CI is actually fails when we expect it to.

Best,

Aaron

_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Thanks,

Octavian.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140221/971f6ff4/attachment.html>