[Gluster-devel] Regression tests marked as SUCCESS when they shouldn't be

[Justin Clift]

On 16 Apr 2015, at 03:56, Jeff Darcy <jdarcy at redhat.com> wrote:
>> Noticing several of the recent regression tests are being marked as SUCCESS
>> in Jenkins (then Gerrit), when they're clearly failing.
>> 
>> eg:
>> 
>>  http://build.gluster.org/job/rackspace-regression-2GB-triggered/6968/console
>>  http://build.gluster.org/job/rackspace-regression-2GB-triggered/6969/console
>>  http://build.gluster.org/job/rackspace-regression-2GB-triggered/6970/console
>>  http://build.gluster.org/job/rackspace-regression-2GB-triggered/6966/console
>> 
>> Is this something to do with your patch to try and get failures finishing
>> faster?
> 
> Yes, I think it probably is.  At line 221 we echo the status for debugging,
> but that means the result of main() is the result of the echo (practically
> always zero) rather than the real result.  All we need to do is take out
> that echo line.

Well, if the echo is important for debugging, we can save the real result
status earlier and then "return $REAL_STATUS" type of thing.

+ Justin

--
GlusterFS - http://www.gluster.org

An open source, distributed file system scaling to several
petabytes, and handling thousands of clients.

My personal twitter: twitter.com/realjustinclift