[yocto] Morty 2.2.1 build failure

[Jussi Kukkonen]

On 23 March 2017 at 22:11, Greg Wilson-Lindberg <GWilson at sakuraus.com>
wrote:

> Hi Armin & Paul,
>
> First there is nothing in the X log files.
> I'm building on a desktop machine with 12GB of memory and an 8 core
> processor running Kubuntu 14.04.
> I added the PARALLEL_MAKE parameter set to 4, no effect, still built with
> 8 threads, crashed.
> Set BB_NUMBER_THREADS to 4, built with 4 threads, still crashed.
>

Getting thrown out of your desktop environment probably indicates the core
problem is outside of Yocto -- this is still a fine place to discuss the
issue, just pointing it out. You could keep an eye on 'top' sorted by %MEM
while doing a build: my theory here is that something (like the monstrous
amount of IO) that the build does could trigger a desktop component to do
something stupid that cascades into your desktop shell exiting/crashing.
This could show up as something using more and more memory at some point
during the build.

Hope this helps, sorry for lack of definitive answers,
  Jussi



>
> Regards,
> Greg
>
> > -----Original Message-----
> > From: akuster [mailto:akuster at mvista.com]
> > Sent: Thursday, March 23, 2017 12:33 PM
> > To: Greg Wilson-Lindberg <GWilson at sakuraus.com>
> > Cc: Paul Eggleton <paul.eggleton at linux.intel.com>;
> yocto at yoctoproject.org
> > Subject: Re: [yocto] Morty 2.2.1 build failure
> >
> >
> >
> > On 03/23/2017 12:25 PM, Paul Eggleton wrote:
> > > That's really bizarre. There shouldn't be anything in a bitbake build
> > > that could cause anything like this (other than possibly how intensive
> > > it is, which might trigger out-of-memory or an underlying
> > hardware/software failure).
> >
> > I usually get my builds to reboot my system when I build on a loptop. I
> have
> > not seen this on a tower style work station. You may want to play with
> the
> > BB_NUMBER_THREADS=# and PARALLEL_MAKE = "-j #" settings in your
> > local.conf.
> >
> > - armin
> > >
> > > Assuming the X session is ending is there anything in your
> > > ~/.xsession-errors or /var/log/Xorg.*.log?
> > >
> > > Cheers,
> > > Paul
> > >
> > > On Friday, 24 March 2017 5:58:24 AM NZDT Greg Wilson-Lindberg wrote:
> > >> Hi Paul,
> > >>
> > >> I looked in the logs from the failure yesterday and couldn't find
> > >> anything, so I restarted the build and it ran for about another 1000
> > >> steps before crashing again. I still can't find anything in any of
> > >> the logs, syslog, kern.log, auth.log, etc.
> > >>
> > >> Cheers,
> > >> Greg
> > >>
> > >>> -----Original Message-----
> > >>> From: Paul Eggleton [mailto:paul.eggleton at linux.intel.com]
> > >>> Sent: Thursday, March 23, 2017 2:49 AM
> > >>> To: Greg Wilson-Lindberg <GWilson at sakuraus.com>
> > >>> Cc: yocto at yoctoproject.org; akuster808 <akuster808 at gmail.com>
> > >>> Subject: Re: [yocto] Morty 2.2.1 build failure
> > >>>
> > >>> On Thursday, 23 March 2017 9:23:24 AM NZDT Greg Wilson-Lindberg
> > wrote:
> > >>>> I tried it again this morning and it did get past the file not
> > >>>> found error.
> > >>>> But when it gets above step 4000 somewhere it gets an error that
> > >>>> kills not just the yocto build but the full login session. I'm left
> > >>>> staring at the Linux Login screen.
> > >>> Your machine isn't running out of RAM by any chance, triggering the
> > >>> OOM killer? Anything indicative in your system logs?
> > >>>
> > >>> Cheers,
> > >>> Paul
> > >>>
> > >>> --
> > >>>
> > >>> Paul Eggleton
> > >>> Intel Open Source Technology Centre
> > >
>
> --
> _______________________________________________
> yocto mailing list
> yocto at yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/yocto/attachments/20170324/84d8820b/attachment.html>