[Motmot] Segmentation faults

[John Schneider]

Hey Andrew,

I'll downgrade some computers tonight with that earlier kernel and try
some long term acquisition. I'm also going about it another way: I'm
currently running one computer with the current kernel but old fview
packages (generated by apt-on-cd) that I had laying about, too early
to tell if it fixes anything, but could narrow things down as well.
Will update tomorrow,

Jon

On Mon, Mar 30, 2009 at 5:12 PM, Andrew Straw <astraw at caltech.edu> wrote:
> John Schneider wrote:
>> Hey Andrew,
>>
>> The segmentation fault happens at variable times throughout
>> acquisition, sometimes several hours into it. I've rebooted,
>> un/replugged, and set the framebuffer to 5; it segfaulted sometime
>> last night.
>> As for why the camera would be closed, I'm at a loss.
>> Screensavers/power management have all been disabled, no other
>> programs are open when fview is going.
>> I seem to be able to reproduce this fairly regularly (unfortunately),
>> any other troubleshooting steps?
>
>
> Hi Jon,
>
> OK, I'm able to reproduce the bug on our end using Alice's FView setup,
> and I'm in the process of debugging it now. Given that it's a fairly
> infrequent event that triggers it, this process may be sped up if you
> can help. One of my hunches is that there's some recent update to the
> Ubuntu Hardy kernel that is causing these problems, because I have done
>  little that could account for this. So, if you could reboot your
> computers and in the GRUB boot screen select the 2nd most recent kernel
> to boot into, that would be great. Specifically, choose "2.6.24-22"
> instead of "-23". (No need to choose the "recovery mode".) Then run the
> cameras and see if the problem arises again.
>
> For what it's worth, we are running with the motmot camiface drivers
> (but not FView) daily for extended periods on a multi-camera tracking
> setup with 5 or 9 cameras and not seeing this bug. So, I guess it's
> something being triggered by FView itself, even if it's not FView
> directly with the bug.
>
> I'll keep replying to this list as the debugging process unfolds.
>
> -Andrew
>
> --
> Andrew D. Straw, Ph.D.
> California Institute of Technology
> http://www.its.caltech.edu/~astraw/
> _______________________________________________
> Motmot mailing list
> Motmot at code.astraw.com
> http://code.astraw.com/cgi-bin/mailman/listinfo/motmot
>