I have a computer that running Ubuntu 10.04.2 LTS.

Consistently after a couple hours the X desktop and GDM dies. It works fine on my other computers but this computer is an intel 2.4 GHz, 1.5 GB.

This can't be duplicated by anything specific, but I know it's not out of RAM at this point. The console keyboard and mouse are inoperational, and the monitor display seems to be going through test patterns, which turns out to be that the monitor is trying to sync to video output which is switching on about once a second. I can ssh into the box and see that many processes that were running are no longer running, for example gnome-session... and that once per second there is a new X process invoked with a different auth-for-gem parameter, per below:

/usr/bin/X :0 -nr -verbose -auth /var/run/gdm/auth-for-gdm-Kynptf/database -nolisten tcp
(one second later...)
/usr/bin/X :0 -br -verbose -auth /var/run/gdm/auth-for-gdm-GfBhRk/database -nolisten tcp
(one second later...)
/usr/bin/X :0 -br -verbose -auth /var/run/gdm/auth-for-gdm-PribK0/database -nolisten tcp

How can I troubleshoot this?

Here is some info from various logfiles:

# dmesg | grep agp
[ 4.345441] Linux agpgart interface v0.103
[ 4.377791] agpgart-intel 0000:00:00.0: Intel 830M Chipset
[ 4.378057] agpgart-intel 0000:00:00.0: detected 892K stolen memory
[ 4.381363] agpgart-intel 0000:00:00.0: AGP aperture is 128M @ 0xf0000000

# dmesg | grep -i vga
[ 0.000000] Console: colour VGA+ 80x25
[ 0.119552] vgaarb: device added: PCI:0000:00:02.0,decodes=io+mem,owns=io+mem,locks=none
[ 0.119619] vgaarb: loaded
[ 4.517038] vga16fb: initializing
[ 4.517045] vga16fb: mapped to 0xc00a0000
[ 4.517268] fb0: VGA16 VGA frame buffer device

# tail /var/log/messages
May 28 17:22:22 ubuntu pulseaudio[9271]: pid.c: Stale PID file, overwriting.
May 28 17:22:22 ubuntu pulseaudio[9271]: main.c: Unable to contact D-Bus: org.freedesktop.DBus.Error.NoServer: Failed to connect to socket /tmp/dbus-i7xEal70Ls: Connection refused

# tail /var/log/Xorg.0.log
(II) GLX: Initialized DRI2 GL provider for screen 0
(II) intel(0): Setting screen physical size to 270 x 203

Fatal server error:
Failed to submit batchbuffer: Input/output error

# tail /usr/adm/kern.log
May 28 17:45:54 ubuntu kernel: [577246.001043] [drm:i915_gem_idle] *ERROR* hardware wedged
May 28 17:45:55 ubuntu kernel: [577247.175163] [drm:i915_gem_entervt_ioctl] *ERROR* Reenabling wedged hardware, good luck
May 28 17:45:55 ubuntu kernel: [577247.488009] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung

Ehrm...harddisk dying on you?

and

to name two. The socket, by the way, is (to me) a dead (sorry for the word) give-away. Linux/Unix sockets are disk based.

Either that or the disk is flooding, if not sunk already...

Luck, dude

Thor
Edit : now I'm pretty sure it's the disk flooding up, just remembered the very same thing I had way-way back in the past...and that was the same behaviour, an on-and-off pulsing of the X...check the disk space...

There's no hard drive. This is a live USB version of ubuntu (created with the usb installer at pendrivelinux.com). The same problem happens on different USB thumbdrives I've tried, so it's nothing wrong with the thumbdrive. Also tried the same usb thumbdrive on other computers, and it only happens on this computer. Also note that Linux keeps running, it's only GDM that suddenly keeps dying. Also, the problem NEVER happens on this computer when running Ubuntu with a different Desktop Environment, such as Lubuntu. So appparently it's something that GDM (gnome) doesn't like about this computer after a few hours, and I need to find out what that is.

Wow, okay, those are important details. Gnome3? What are the video-differences between this computer and the others? USB drive...those are new(er) machines, so we can rule out the memory, perhaps the video card? Just thinking out loud. Because Gnome3 may not like the video hardware on that box...

Lemme watch this thread and try to think with you, if that's okay...

Thor

Thanks for your reply.
It's gnome 2.30.2
What other helpful details can I provide? Regarding the video card, you can see the dmesg output in the original post.

Hmm, I noticed the dbus died...this has to have a reason. Okay, let's go back to the hardware and catalog the differences. As you stated, the same stick "behaves" on other machines...and all/most of the sticks misbehave on that one machine, so...that's where the answer lies...

I'll try to sniff on the net...but you've got a real mistery here...

Oh well, this is belgium, land of Poirot...

Thor

Hi,

Just a quick idea (it's four in the morning here, so I cannot guarantee a clear head yet... ) are you allowed to test this with a different distro? Maybe (just maybe) there's some element in the current distro that does "not agree" with the hardware?

Thor

I've started running Mint, so I'll see how this works... would really prefer to run ubuntu 10.04 though.