[openstack-dev] [OpenStack-Dev] accuracy regarding rechecks

John Griffith john.griffith at solidfire.com
Mon Jul 29 01:25:08 UTC 2013


Hey Everyone,

I thought I'd send a note out to ask folks to try and be a bit more
diligent when it comes to assigning bugs to recheck/reverify entries in the
gate.

I've noticed a few bugs that are frequently listed for recheck, even though
they have little or nothing to do with the failure.  Just a tip, if you see
your patch fail every gate test, it's very likely that it's not a
"transient error" or one of the bugs listed.  It's more likely related to
your patch or has something to do with Jenkins.

Anyway, I did some quick looking into bug 1131030 [1] at the request of
some folks who pointed out Cinder was causing the gates to fail.

Checking the logs for the last 8 rechecks (I stopped after 8), NONE of
these were actually volume create failures.  Almost all of them were
actually floating IP failures.  Keep in mind that the volume test is
dependent upon instance creation and floating IP assignment.  If either of
these fails, the volume test is listed as failed as well because the
instance and access via floating IP is required in order to test
volume-attach.  The other offender seems to be aggregates failure, however
I believe this has been addressed.

Anyway, I wanted to ask that folks take a little extra time with the
console logs for failed gate checks.  Gathering info on rechecks etc is
very valuable, but it's only as valuable as the info that we enter.

If you have questions regarding diagnosis of cinder/volume test failures in
the gate I'm happy to help out, just let me know.

Thanks,
John


[1] https://bugs.launchpad.net/cinder/+bug/1131030
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130728/3310b1ef/attachment.html>


More information about the OpenStack-dev mailing list