False OST failure

[Barak Korren]

On 26 May 2017 at 10:49, Milan Zamazal <mzamazal at redhat.com> wrote:
> Hi, check-merged has failed for my merged Vdsm patch yesterday:
> http://jenkins.ovirt.org/job/vdsm_master_check-merged-el7-x86_64/1811/
>
> The error was:
>
>   14:36:12 + lago copy-to-vm vdsm_functional_tests_host-el7 yd2rya /etc/yum/yum.conf

The following is not an error message is a warning that always shows
up because the user in mock is not a member of the 'lago' group (its
harmless though, b/c it is 'root' so doesn't need to be a member).

>   14:36:13 current session does not belong to lago group.

Note the huge time gap between these two lines:

>   20:31:02 Build timed out (after 360 minutes). Marking the build as failed.
>
> So it's a false alert, could the failure be prevented?

It not a false alert, something is happening and is causing the system
the be stuck for over 6 hours. Need to look for what is happening in
the other logs.

Note that with Lago stuff can be happening on the VMs without any
messages going the the STDOUT, so the Jenkins job output screen is not
the best place to look for reasons for failures on Lago.

BTW, this has nothing to do with OST other then also using Lago.

Adding Yaniv who is more familiar with the vdsm check-merged job.


-- 
Barak Korren
RHV DevOps team , RHCE, RHCi
Red Hat EMEA
redhat.com | TRIED. TESTED. TRUSTED. | redhat.com/trusted