Re: wheezy, message log being spammed by segfaults from apt_check.py
On Thursday 14 January 2016 11:33:10 Gene Heskett wrote:
> On Thursday 14 January 2016 08:43:53 Johann Klammer wrote:
> > On 01/14/2016 04:00 AM, Gene Heskett wrote:
> > > Greetings all;
> > >
> > > Intermittently but sometimes several times an hour from the times
> > > recorded. The error portion of the logged line:word wrapped
> > >
> > > segfault at 0 ip b7614966 sp bf9b64c8 error 6 in
> > > libc-2.13.so[b759b000+15e000
> > >
> > > No clue what an error 6 means. No errors shown in the man page,
> > > just mild horn blowing.
> > >
> > > And the segfault goes away if root runs it, so I have to assume
> > > root is not running it. Does anyone have a clue what I should be
> > > looking at and what to check for?
> > >
> > > Thanks.
> > >
> > > Cheers, Gene Heskett
> >
> > segfaults are always bad. It's dereferencing a NULL pointer here.
> > the exact numbers are in some intel processor manual,
> > but won't help you unless you want to debug the program.
> > It's probably an at or cron job running this script.
> > Perhaps something involved with auto-updating.
>
> On a hunch, I just ran synaptic via tdesudo, then checked the log
> after every operation, like reload, and the message appears at the
> very end of the reload as its redrawing the file listings
>
> But, when I ran it from synaptic-pkexec the log message appeared at
> the end of the reload as it was redrawing the file list. Same as
> above.
>
> Since I use synaptic to keep all my machines up to date, I'll go check
> the logs on the other three. ssh'd into each so I only have one
> terminal screen, but I ran synaptic-pkexec, updated all 3 machines,
> without generating a segfault, so its local to this machine. The first
> machine I checked, has also had TDE installed, the other 2 are still
> on KDE for a user facing gui. Other than that, all 4 machines were
> installed from the same hybride-iso image you can get from
> linuxcnc.org. Its based on debian wheezy, but with a pinned kernel on
> the other 3 machines, the kernels haveing been patched with the RTAI
> kit for realtime usage. All 4 machines are running a
> 3.4-9-rtai-686-pae kernel, which unfortunately for this machine, pae
> doesn't work after the rtai patch, but this is the only machine with
> sufficient (8Gb) memory to make use of the pae.
>
> There is a 3.16-something or other kernel that it wants me to update
> to, and which is installed, but its a 64 bit kernel and linuxcnc will
> not run on a 64 bit kernel for realtime stuff. So I always reboot to
> the one I know works as I quite often develop the gcode I write, on
> this machine, verifying it runs in the simulator & then shipping it to
> the machine that will make the part.
>
> But I think I'll reboot to that kernel image and find out. Later.
>
Rebooted to that kernel image(3.16.0-0.bpo.4-amd64 #1 SMP Debian
3.16.7-ckt20-1+deb8u2~bpo70+1 (2016-01-03) x86_64 GNU/Linux) & ANAICT,
everything is working and all 8Gb of my memory shows up on the htop
screen. Even linuxcnc's simulator looks to be running ok.
> But I've no clue where to look next. Synaptic itself does not appear
> to be effected by it. If it actually mis-behaved somehow, that would
> be a better clue.
Fine tuning the clues, the error seems to be logged (I have the tail on
the log and synaptic running on the same screen this time) and the
segfault appears nominally 1 second after the file listing has been
redrawn. How much of that is file system commit time to flush, IDK.
IOW, a different, much newer kernel has no effect on this.
A data point.
Thanks all.
Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
Reply to: