[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-devel
Subject: Re: KDE Gear projects with failing CI (master) (12 September 2023)
From: Ben Cooksley <bcooksley () kde ! org>
Date: 2023-09-21 7:58:31
Message-ID: CA+XidOFOF4_VZLaLJ5HWhsY6P6+QhogaBkoRG54Yg_HbasRq1Q () mail ! gmail ! com
[Download RAW message or body]
On Thu, Sep 21, 2023 at 9:10 AM Alexander Semke <alexander.semke@web.de>
wrote:
> On Mittwoch, 20. September 2023 13:04:00 CEST Ben Cooksley wrote:
>
> > > = FAILING UNIT TESTS =
> > >
> > > konsole: (2nd week)
> > >
> > > * https://invent.kde.org/utilities/konsole/-/pipelines/484169
> > >
> > > * freebsd_qt515 tests are failing
> >
> > Sysadmin has been asked to take a look into this one, as by all accounts
> > there is no reason for this to suddenly start failing out of the blue
> with
> > no changes in Konsole itself.
> > I'm suspecting regressions within either our FreeBSD setup, or somewhere
> in
> > Frameworks (KParts specifically) at the moment.
> I'm fixing the tests for Cantor now which started to fail after I pushed a
> change to master that is not related to the failing tests now. The tests
> are
> not stable against different timings and the async. communication with the
> external process needs to be handled better. So, I assume something was
> changed for the freeBSD setup that affects the performance and the timings
> during the test executions. Maybe newer or older hardware in use?
>
Likely it relates to the load on the underlying CI worker node at the time
the job is being run. There is no guarantee you're the only job running at
the time as Linux or Windows could be running jobs (or they both could be)
and if they're both running that will definitely impair FreeBSD
performance. For most unit tests they're not load/performance sensitive so
this isn't an issue.
All of the mainline Linux/FreeBSD/Windows/Android jobs provided by Sysadmin
run on AMD Ryzen 7 7700 machines, with NVMe storage and plenty of RAM (the
host machines have either 64GB or 128GB depending on the physical node,
although the FreeBSD nodes only have 16GB allocated to them) so there is
certainly no hardware limitations or age issues here.
>
> Regards,
> Alexander
>
Cheers,
Ben
[Attachment #3 (text/html)]
<div dir="ltr"><div dir="ltr">On Thu, Sep 21, 2023 at 9:10 AM Alexander Semke \
<<a href="mailto:alexander.semke@web.de">alexander.semke@web.de</a>> \
wrote:<br></div><div class="gmail_quote"><blockquote class="gmail_quote" \
style="margin:0px 0px 0px 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex">On Mittwoch, 20. September 2023 13:04:00 CEST Ben \
Cooksley wrote:<br> <br>
> > = FAILING UNIT TESTS =<br>
> ><br>
> > konsole: (2nd week)<br>
> ><br>
> > * <a href="https://invent.kde.org/utilities/konsole/-/pipelines/484169" \
rel="noreferrer" target="_blank">https://invent.kde.org/utilities/konsole/-/pipelines/484169</a><br>
> ><br>
> > * freebsd_qt515 tests are failing<br>
><br>
> Sysadmin has been asked to take a look into this one, as by all accounts<br>
> there is no reason for this to suddenly start failing out of the blue with<br>
> no changes in Konsole itself.<br>
> I'm suspecting regressions within either our FreeBSD setup, or somewhere \
in<br> > Frameworks (KParts specifically) at the moment.<br>
I'm fixing the tests for Cantor now which started to fail after I pushed a<br>
change to master that is not related to the failing tests now. The tests are<br>
not stable against different timings and the async. communication with the<br>
external process needs to be handled better. So, I assume something was<br>
changed for the freeBSD setup that affects the performance and the timings<br>
during the test executions. Maybe newer or older hardware in \
use?<br></blockquote><div><br></div><div>Likely it relates to the load on the \
underlying CI worker node at the time the job is being run. There is no guarantee \
you're the only job running at the time as Linux or Windows could be running jobs \
(or they both could be) and if they're both running that will definitely impair \
FreeBSD performance. For most unit tests they're not load/performance sensitive \
so this isn't an issue.</div><div><br></div><div>All of the mainline \
Linux/FreeBSD/Windows/Android jobs provided by Sysadmin run on AMD Ryzen 7 7700 \
machines, with NVMe storage and plenty of RAM (the host machines have either 64GB or \
128GB depending on the physical node, although the FreeBSD nodes only have 16GB \
allocated to them) so there is certainly no hardware limitations or age issues \
here.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <br>
Regards,<br>
Alexander<br></blockquote><div><br></div><div>Cheers,</div><div>Ben \
</div></div></div>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic