[Gluster-devel] Regression test failures

Vijay Bellur vbellur at redhat.com
Fri Feb 14 12:30:02 UTC 2014


Hi All,

We have noticed an increase in number of failures for pre-commit 
regression test runs recently. It is not productive for any of us to 
have obvious regression test failures in Jenkins. Here are some thoughts 
on how we can help ourselves better:

1. Run regression tests on your local setup before rfc.sh. This will 
help in catching obvious problems.

2. If the problem happens on build.gluster.org only and not in your 
local setup, please try to send a patch with DEBUG=1 or set -x in the 
failing testcase. This will provide more verbose output and help isolate 
a failure.

3. Folks who have access to build.gluster.org can help by 
troubleshooting/isolating a problem that happens on build.gluster.org alone.

There is a plan to increase the number of Jenkins instances to run more 
regression tests in parallel and help us scale better. Please reach out 
to me offlist if you or your organization is interested in supporting a 
jenkins CI instance for glusterfs.

Thanks,
Vijay




More information about the Gluster-devel mailing list