[Avocado-devel] acceptance-system-fedora and acceptance-system-centos failing sporadically with timeout?

[Beraldo Leal]

Hi all,

I will try to give my two cents:

On Fri, Feb 05, 2021 at 03:31:40PM -0500, John Snow wrote:
> On 2/5/21 11:43 AM, Philippe Mathieu-Daudé wrote:
> > Cc'ing Avocado team & John (Python inferior exit delay?).
> > 
> > On 1/28/21 11:10 AM, Thomas Huth wrote:
> > > On 28/01/2021 10.45, Claudio Fontana wrote:
> > > > 
> > > > is it just me, or is the CI sometimes failing with timeout?
> > > > 
> > > > Fedora:
> > > > https://gitlab.com/hw-claudio/qemu/-/jobs/986936506
> > > 
> > > I've sent a patch for that issue just yesterday:
> > > 
> > >   https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06852.html
> > > 
> > > > CentOS:
> > > > https://gitlab.com/hw-claudio/qemu/-/jobs/980769080
> > > 
> > > Never seen that one before - if you hit it again, could you please save
> > > the artifacts and have a look at the log file in there to see what's
> > > exactly the problem?
> > 
> > https://gitlab.com/philmd/qemu/-/jobs/1008007125
> > 
> >   (28/36)
> > tests/acceptance/replay_kernel.py:ReplayKernelNormal.test_ppc_mac99:
> > ERROR: Test reported status but did not finish (90.09 s)
> > 
> > Attached debug.log.
> > 
> 
> ¯\_(ツ)_/¯
> 
> I don't know what "reported status but did not finish" means.
> 
> The debug log looks like it passes, too, so... I don't know that this has
> much do with code I maintain yet. I'm sure the Avocado team will find me if
> I am wrong :)

Afaict, this happen when the process exceeds the deadline to finish.
Sometimes the test is finished but the "post test" stage is stucked
because of some reason.

Maybe setting 'runner.timeout.process_alive' to a higher number could help
here:

----
$ avocado config reference | grep process_alive -A 6

runner.timeout.process_alive
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The amount of time to wait after a test has reported status but the
test process has not finished

* Default: 60
* Type: <class 'int'>
----

But I might be wrong. I know that Cleber was working with this, so
probably he could help here too.

--
Beraldo