System crashes at high CPU load Shuttle Barebone SN78SH7 with GeForce 8200 onboard graphic

[markusd112]

I am using a Shuttle Barebone SN78SH7 with an onboard graphic card nvidia GeForce 8200. The monitor is connected via HDMI / DVI.

When the system is on minimal workload everything is fine, but when I am starting several cpu intensive tasks, like virtualbox, intensive datatransfer via LAN, remote login via XDMCP, etc. at the same time, the system becomes unstable and crashes: on the screen is only a colored noise visible and the system hangs completely.
After reboot the text messages from the bios and from grub is displayed well, but when ubuntu 9.10 switches to graphics mode, there will appear blinking colored blocks (looks like a text mode?!) all over the monitor and the system hangs.

I have to completely remove the power cord for a minute and then everything works fine again.

ProblemType: Bug
Architecture: i386
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: markus 3222 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'NVidia'/'HDA NVidia at 0xfe020000 irq 21'
   Mixer name : 'Nvidia MCP78 HDMI'
   Components : 'HDA:10ec0888,12970000,00100001 HDA:10de0002,10de0101,00100000'
   Controls : 40
   Simple ctrls : 20
CheckboxSubmission: e27141b8feed9a0134eefdd87f008818
CheckboxSystem: 558fbfb2a1258711a37bb7e23c5d4e6e
Date: Sat Dec 12 07:33:25 2009
DistroRelease: Ubuntu 9.10
HibernationDevice: RESUME=
IwConfig:
 lo no wireless extensions.

 eth0 no wireless extensions.

 vboxnet0 no wireless extensions.
MachineType: Shuttle Inc SN78S
NonfreeKernelModules: nvidia
Package: linux-image-2.6.31-16-386 2.6.31-16.53
ProcCmdLine: root=UUID=34f6e725-3efe-4797-8503-b81b269df690 ro quiet splash
ProcEnviron:
 LANGUAGE=de_DE.UTF-8
 PATH=(custom, no user)
 LANG=de_DE.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-16.53-386
RelatedPackageVersions: linux-firmware 1.25
RfKill:

SourcePackage: linux
Uname: Linux 2.6.31-16-386 i686
WifiSyslog:
 Dec 12 07:23:28 localhost kernel: [ 1430.025041] device eth0 left promiscuous mode
 Dec 12 07:23:29 localhost kernel: [ 1431.381025] device eth0 entered promiscuous mode
 Dec 12 07:23:29 localhost kernel: [ 1431.405038] device eth0 left promiscuous mode
 Dec 12 07:23:29 localhost kernel: [ 1431.429547] device eth0 entered promiscuous mode
WpaSupplicantLog:

XsessionErrors:
 (gnome-settings-daemon:3237): GLib-CRITICAL **: g_propagate_error: assertion `src != NULL' failed
 (gnome-settings-daemon:3237): GLib-CRITICAL **: g_propagate_error: assertion `src != NULL' failed
 (polkit-gnome-authentication-agent-1:3277): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed
 (nautilus:3271): Eel-CRITICAL **: eel_preferences_get_boolean: assertion `preferences_is_initialized ()' failed
dmi.bios.date: 07/09/2009
dmi.bios.vendor: Phoenix Technologies, LTD
dmi.bios.version: 6.00 PG
dmi.board.name: FN78S
dmi.board.vendor: Shuttle Inc
dmi.board.version: V10
dmi.chassis.type: 3
dmi.chassis.vendor: Shuttle Inc
dmi.chassis.version: H7
dmi.modalias: dmi:bvnPhoenixTechnologies,LTD:bvr6.00PG:bd07/09/2009:svnShuttleInc:pnSN78S:pvrV10:rvnShuttleInc:rnFN78S:rvrV10:cvnShuttleInc:ct3:cvrH7:
dmi.product.name: SN78S
dmi.product.version: V10
dmi.sys.vendor: Shuttle Inc

[Steve (smalenfant)]

I'm experiencing the same exact "the system becomes unstable and crashes: on the screen is only a colored noise visible and the system hangs completely" when I run Handbrake on Fedora Core 12.

Could not fail with cpuburn or other cpu/memory testing utility.

[Jeremy Foshee (jeremyfoshee)]

Markus,
     Would you mind testing this against the latest Lucid Alpha? Also, if it is still an issue in Lucid, would you test on a vanilla kernel from upstream?

Thanks!

-JFo

[markusd112 (markusd112)]

Jeremy, can you give me a hint please, how to do that? Thanks.

[Russell Wing (wingfamily)]

I have the same problem- crashes with random coloured screen, need to switch off at power socket on wall to reset.

So far I have not found an obvious trigger, but could possibly be load based as stated above.

[markusd112 (markusd112)]

I have sent back my shuttle barebone to my dealer who has returned it to shuttle. The have tested it, but without finding any problem... After the barebone has shipped back to me, it works stable for several weeks, but then the same problem starts again. After that I have given it back to my dealer who returns my money.

Now I have bought another pc, that has no problem any longer.

[Russell Wing (wingfamily)]

Did you buy another Shuttle SN78SH7- therefore confirming that it was the hardware? Or did you buy a different type of machine- and therefore it could still be software based? I am suspicious of the integrated NVidia 8200 drivers, and wonder whether to test with another video card installed.

[markusd112 (markusd112)]

I have bought another computer, not SN78SH7.

I have tried a dedicated graphics card, but the system was still unstable and nearly unusable. I assume, that it was a hardware issue, because the system has shown some more courious behavior: after changing some bios settings, it hasn't reboot. I had to switch off the power to reboot.

When I switch the fan to manual speed and set it to maximum, the system became nearly stable. But after some months, the instability comes again, it seems that some dust has decreased to cooling power, so the system reaches again some thermal problems.

[Russell Wing (wingfamily)]

Markus- thanks for your feedback, certainly it does not sound as if the video card will solve the problem. Mine is less frequent than yours though- perhaps once a week, but the symptoms sound the same.
I have tried CPU burn and run Memtest overnight and this works fine which made me doubt hardware as the problem. That was why I was suspicious of the video card, especially given the display corruption. I also wonder if the drivers are stable given that Nvidia integrated drivers seem to be problematic from my experience compared to their cards.

[Russell Wing (wingfamily)]

There is thread on the Shuttle forums that suggest the power supply may be failing:

http://us.shuttle.com/scgforum/tm.aspx?m=4337&mpage=1&key=RAID,SN78SH7&#4337

I might try a swap out if it happens again....

[markusd112 (markusd112)]

It's worth a try. But the problem described there is not exactly the problem as here... But who knows... Please keep us informed about the result. Thx.

[Russell Wing (wingfamily)]

I've done some more testing and it seems to be the RAM related. I had 2 x 2GB sticks (DDR 2 800, PC 6400) installed in ganged mode. Removing the stick closest to the front of the machine makes the machine stable again- it's runs happily at full load for extended periods and continuously for 2 days now.

I am not sure if one of the memory sticks is at fault, they both seem to work and pass Memtest, but only when one is installed. So that leaves the motherboard or software I guess?

What's strange is that the machine had worked fine for 7-8 months and become less stable over the last 4 months.

I've found a reports of similar issues with SN78SH7 shuttles.
http://www.sudhian.com/index.php?/forums/viewthread/105923/
Interestingly I also have a QUAD core AMD Phenom X4 940 processor which is referenced in one the postings in the thread above.

So at the moment my best guess is a hardware issue... not sure whether it is worth the effort to buy more RAM to test with as I am more suspicious of the motherboard or memory socket and the machine is out of warranty.....

Is there any way that it could be related to the Kernel and the way the memory controller on the AMD processor is working???

[Brad Figg (brad-figg)]

This bug was filed against a series that is no longer supported and so is being marked as Won't Fix. If this issue still exists in a supported series, please file a new bug.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.