Hardware (?) trouble

eyescream · 2009-08-24 21:00:33

Hello!

I've recently run into some trouble over at Ubuntu and couldn't pin-point the problem and since I'd been meaning to try Arch I figured that would be the time. But apprently I am having the same problem again, which leads to me to believe it to be a hardware issue.

I'd experienced occasional freezes shortly after or during booting while using Ubuntu and eventually my system crashed and wouldn't boot properly after. When it booted again after a few tries, pretty much all of my apps only returned segmentation faults right away or after a short amount of time. After formatting it seemed fine, yet it happened again. However, I could not find a pattern. The freezes seemed to occur at different events during/after the boot.

I thought it might be related to Ubuntu, Gnome or some other software I was using (possibly some sort of conflict) so I went ahead and installed Arch Linux and KDE but after a short time I encountered problems. The first thing that happened was Konsole not working. I tried around and eventually changed into a virtual terminal, when I changed back everything froze. I was forced to hard reboot and then it wouldn't boot properly anymore.

For a more detailed description and the errors encountered see http://ubuntuforums.org/showpost.php?p= … ostcount=4

For the other stuff that previously happened on Ubuntu see http://ubuntuforums.org/showthread.php? … ost7787976

I've run fsck and memtest several times without any errors.

My hardware:

XFX 512 GTS 250
Intel Q9550 775 2833 BOX1333 12M
Corsair CMPSU-520HX 520W ATX2
Creative X-Fi Titanium PCIe (which I'm not using, though)
Asus P5Q-Pro P45
Kingston 4GB 1066-555 Dual Channel
Western Digital 320 GB SAT2 WD3200AAKS (two of those)
Lite-On DH-16D2S DVD drive

I am running 2.6.30-ARCH #1 SMP PREEMPT Mon Aug 17 16:06:45 CEST 2009 x86_64

I am pretty much at a loss here. I have no idea what it could be and I can't see any connection either.
I'd really appreciate any bit.

AdrenalineJunky · 2009-08-24 21:28:14

is your processor overclocked alot? or your memory? that could cause errors, have you tried running memcheck?

eyescream · 2009-08-24 21:44:40

I did not overclock anything and memtest found no errors. I ran it 3 times over the course of 2 days.

eyescream · 2009-08-24 22:06:29

I just rebooted after only running Firefox, Banshee and a terminal and it wouldn't shut down.
It returned this (just parts of it) and kept repeating it:

RIP: <lots of numbers and f's>
RSD: <lots of numbers and f's>
RAX: <lots of numbers and f's>
RDX: <lots of numbers and f's>

Call Trace:
[<fffff<numbers>>] ? _atomic_dev_<something>
dputx0c2/02c

basically the whole screen was filled with those two things and it kept repeating. The keyboard was still responsive.

Teoulas · 2009-08-24 22:24:25

Sounds like a hardware problem to me, although it's hard to tell what exactly the problem is. Have you tried running a live linux distro and see if you still have similar problems? Also, try removing RAM sticks and/or extra hard drives.

Last edited by Teoulas (2009-08-24 22:24:56)

eyescream · 2009-08-24 22:40:38

I had not thought of trying a Live CD. I'm going to try that tomorrow along with removing my second hard drive and trying both memory bars alone.
Thanks for the suggestions.
Since it's late I'll be reporting back tomorrow evening with the results.

eyescream · 2009-08-25 20:31:51

I was unable to test anything. It just wouldn't show any errors.
I spent the better part of the last 3 hours trying around, rebooting about 30 times, doing stuff I usually do when errors occur..nothing. I did however remove both of my ram sticks as well as my gpu and put them back in just in case something might have been wrong there.

I guess all I can do for now is wait for the errors to appear again and see whether I can diagnose it any further.

eyescream · 2009-09-05 09:56:48

After my last post I reformatted again and set up my system as I usually do.
In 7 days I encountered no errors or problems whatsoever but yesterday and today, a cold boot resulted in a freeze during boot again.
Both times it froze on "Starting Network", it kept saying "[Busy]". My keyboard and mouse lights turned off and a hard-reboot was the only option.
Now after that hard-reboot it booted normally.

I am clueless as to what could be broken and how to find that out. I can't reproduce the errors.
If there is any pattern to it, I should be getting more freezes in the next days and then major errors again, which should disappear after a complete reformat..until they return after about a week.

I have no idea what could cause errors in this manner. It is as if they are gone after a reformat, then they slowly build up until my whole system is unusable for a while and then either go away again or render my system completely useless until I reformat.
Also, when the errors appear, I am usually forced to hard-reboot due to errors while booting/shutting down that seem to change all the time.

R00KIE · 2009-09-05 13:44:07

I would say that you should check the voltages you are feeding to your RAM, maybe it needs a little more but if memtest didn't find anything then that might not be the problem.

Another thing that can cause problems is the PSU but yours seems to be good so maybe another dead end. Without any more ideas I would say check the cabling, remove and reinsert your power and data cables, try reinserting the memory modules again, check for dust clogging the heatsinks and check if all fans are working properly.

Other than that I would say some kind of hardware failure that is giving some warning before failing completely.

eyescream · 2009-09-05 17:30:10

Thank you for your reply!

I checked everything again like you suggested. Everything seems to be in order and properly plugged in, no clogging or anything of that kind either.
I've also checked all temperatures and they are normal.
I don't think it's my PSU either. I haven't encountered anything that would suggest it's PSU related and it's a rather good one.

I did however find my RAM voltage to be lower than the suggested one for my type. I read up on it and went through forums to find the best/most common settings for it and tried them.
I changed the voltage from 2.00 to the recommended (required?) 2.10 and found something interesting.
Whenever I tried to boot with a voltage of 2.10 it would freeze _every_ time. Sometimes on "Starting Network", sometimes on the login screen, sometimes right after logging in and sometimes a few seconds after logging in.
So I ran another two memtests with both voltages, 2.00 and 2.10. Neither found any errors, yet with the recommended 2.10 it will always freeze even though I am using the same settings as most other people do.
I should note that I previously changed my RAM clock from 800mhz to 1066mhz (which is the standard clock). I've tried around with both clocks and errors occurred with either of them.

The only other things I did in the BIOS were changing the boot priorities and enable AHCI.

Any ideas why the RAM shows no errors with either of the voltages during memtest but always freezes with the recommended one?

Gen2ly · 2009-09-05 22:45:01

I can't help you with the RAM diagnosis as I don't have any experience with that. But reading the thread I did notice the trend of "Starting Network" "Starting Network" and Firefox. Could this be a problem related to your network card? Faulty hardware of any type can effect all system hardware, primarily in this case the network card effecting the motherboard? Is your network card properly secured? I hate to propose this idea as it involves spending money but you might want to consider a trying new network card. This is the card I've used and had good experience with and only sets me back $15 after shipping.

Last edited by Gen2ly (2009-09-06 02:48:14)

R00KIE · 2009-09-06 01:28:37

How about the ram timings, are they correct for the frequency you are using? Can't think of anything else besides some hardware failure that makes it impossible or very hard for the system to work properly.

eyescream · 2009-09-06 09:25:09

I've checked the timings and everything more than once just to be sure, so that should be fine. I guess I will experiment around a bit more with that, though.

Hmm, network card. Hadn't even considered that. I'll be testing some stuff regarding that later, then. Thanks for the idea!

Also, today when I booted my pc it froze again during the first cold boot (it froze about 5 seconds after logging in). That's three days in a row, always during the first boot of the day.
Never happened with such frequency before. Though, after a hard-reboot everything is working fine, as usual (for now).

eyescream · 2009-09-06 14:28:09

I've tried changing the voltage to 2.10 again with different, slower timings. Same result. It would freeze while booting or after logging in.
Then I disabled the network daemon and changed the RAM voltage to 2.10 again and used the default/recommended timings.

First boot: It booted, I logged in and then the screen went black, my mouse and keyboard lights went out and my fans started spinning up - freeze.
Second boot: It got stuck on "Removing Leftover Files" while booting.
Third boot: Froze after logging in.

Judging by these results I guess it has nothing to do with my network card as it still freezes during boot regardless.

I don't think it is my mainboard as the errors would have to be more frequent and prominent, I guess.
I don't think it is my CPU either, same reasoning as above.
GPU seems unlikely because of the freezes and effects/games work normally.

My RAM checks out in memtest but won't work with the correct voltage.
My PSU is a good one that should be sufficient, but I don't know what PSU problems look like, only heard it can get ugly.

No idea. This is so frustrating. Even more so since this machine is only 2 months old.

thn81 · 2009-09-06 15:22:48

Did you check your harddisks with the diagnosis tool from Western Digital?
[url]http://support.wdc.com/product/download … 06&lang=de[\url]
Maybe one of your disks is about to fail and corrupts the data when being read or written. If this doesn't help you should try a completely different operating system, e.g. FreeBSD or Solaris or Windows (if you own a copy), and see if it also fails. If it does, then some piece of your hardware is definitely broken.

eyescream · 2009-09-08 19:33:00

I haven't checked with the diagnostics from WD because I don't have a Windows installation at the moment.
When I have time I will install Windows and run the diagnostic tools then and see how it goes.

I don't see how it could be related to linux. I've got linux running on 3 machines and only this one is having problems. I'd been using the exact same setup on this machine as on another one. Same image used to install, same apps installed for the most part, basically all of the non-system related settings were identical (compiz etc).

Also, I've tried increasing some of my voltages (CPU, SB, SN) but no matter what settings I tried it just wouldn't work and I was forced to revert to my old settings.
So I guess that's a dead end as well.

ugkbunb · 2009-09-08 20:02:49

When booting into windows I would always get BSOD if I had enabled my BIOS power management feature AHCI... It would never reporduce it self in linux but always when booting my windows partition... perhaps try disabling that.

eyescream · 2009-09-08 21:11:08

When I started experiencing problems I had AHCI disabled. At some point I enabled it but then also tried disabling it again but it didn't make any difference.
Shouldn't have anything to with that, I think.
Thanks for the idea, though!

eyescream · 2009-09-09 11:42:44

Today on my first boot I received some errors as opposed to the usual freezes. After a hard-reboot it booted normally and everything appears to be working so far.
So, here's what it said:

Googling for it didn't really return any useful results.

Does that mean anything to anyone?

Mindaugas · 2009-09-09 20:31:22

I have similar problems. I'll seek for results as well.

R00KIE · 2009-09-09 23:13:33

Just a thought, like someone else mentioned before already about the network, try booting with a minimal amount of deamons and modules and see if you still have problems, then add one deamon at a time and check for problems, specially deamons that deal with hardware, like network (wired or wireless), cpu frequency scaling and the likes, also I would try not to use proprietary drivers for a while.

Also if you need some extra firmware or module for some piece of hardware (it doesn't seem you do) try not to load the module(s) and/or use the firmware.

I know that it will suck a bit because the open source drivers most probably will not do everything you want but at least hopefully you take something that can be wrecking havoc out of the equation.

@Mindaugas,eyescream
Try to find common hardware components and firmware/bios versions, it may help to figure out the problem if you can pinpoint something in common.

eyescream · 2009-09-10 13:28:00

R00KIE wrote:

Just a thought, like someone else mentioned before already about the network, try booting with a minimal amount of deamons and modules and see if you still have problems, then add one deamon at a time and check for problems, specially deamons that deal with hardware, like network (wired or wireless), cpu frequency scaling and the likes, also I would try not to use proprietary drivers for a while.

The problem with that is that it is not really possible to do it that way. There seems to be no pattern to the errors and freezes, so there is no way to be sure when something works or doesn't work.
Yesterday I started my PC and what you can see in the screenshots above happened, after that I had no further problems whatsoever that day.
Today I started my PC and it booted perfectly without any problems.

Only once, before I formatted, I had major errors that wouldn't go away and would occur during about 70% of all boots and using the system was impossible because nothing was working anymore.
Now, once it boots it works perfectly the way it's supposed to (unless it freezes right after logging in).

Trying things one at a time like that would take months. I've already tried several software related things over the past couple of weeks (mainly after a fresh format to see whether I'd encounter any errors again) but nothing worked, the errors always returned.

Here's my list of daemons: DAEMONS=(syslog-ng network netfs crond alsa hal fam gdm)

I don't see how any of them could be responsible for my problems though. At least not alone.

Mindaugas, I've listed my hardware above, is there anything we have in common?
Or anything else that's not hardware-related?

eyescream · 2009-09-13 17:03:24

Since I've run out of ideas and haven't had the time yet to run the hard drive diagnostics, I decided to play around with some BIOS options again.

I put the RAM timings and voltage on auto and it would freeze again every time. Changed the voltage back to 2.00 and it worked just fine. 2.00 seems to be the only voltage with which it will boot.

Then I disabled AHCI and it wouldn't boot giving me some error about not finding my hard drive. Is that normal or can you only disable that after a format?
AHCI can't really be the issue though, seeing as I hadn't enabled it when the errors first started and enabled it only long after that.

On another note, usually when I reboot (after using my PC for a few hours) it will reboot without freezes.
Yet after I'd upgraded the kernel I rebooted and it froze.

So for the past 2 weeks freezes have only occurred when cold-booting and rebooting after a kernel upgrade.
I have no idea whether that means anything or not.

eyescream · 2009-09-26 13:12:02

Well. So I've managed to rid of the freezes by setting my RAM to 800 mhz again and using 1.9V. (I didn't get around to testing whether it would also work with 1066 mhz)

After about 10 boots without freezes I decided to format. (I was worried about my filesystems due to the amount of hard-reboots and all)
I formatted and everything went smoothly, for half a day.

I was watching a video in vlc (in fullscreen mode) and when I rewound a few seconds my PC froze.
It showed my desktop, without anything there, all I could see was my wallpaper, no panel, nothing.
I had to hard-reboot. During boot it then said that my home partition had failed the fsck and that I had to repair it.

Instead of doing that I decided to install windows to run the diagnostic tools for my hard drives.
I put the windows cd in my drive and it loaded all the drivers and stuff as usual but right before the actual setup started it showed a BSOD instead. I've rebooted several times but it's always the same.

I've tried an Ubuntu live cd but that works just fine.

Any ideas?

Mindaugas · 2009-09-26 13:36:56

Try to scan hard disk drive with MHDD.

Arch Linux

#1 2009-08-24 21:00:33

Hardware (?) trouble

#2 2009-08-24 21:28:14

Re: Hardware (?) trouble

#3 2009-08-24 21:44:40

Re: Hardware (?) trouble

#4 2009-08-24 22:06:29

Re: Hardware (?) trouble

#5 2009-08-24 22:24:25

Re: Hardware (?) trouble

#6 2009-08-24 22:40:38

Re: Hardware (?) trouble

#7 2009-08-25 20:31:51

Re: Hardware (?) trouble

#8 2009-09-05 09:56:48

Re: Hardware (?) trouble

#9 2009-09-05 13:44:07

Re: Hardware (?) trouble

#10 2009-09-05 17:30:10

Re: Hardware (?) trouble

#11 2009-09-05 22:45:01

Re: Hardware (?) trouble

#12 2009-09-06 01:28:37

Re: Hardware (?) trouble

#13 2009-09-06 09:25:09

Re: Hardware (?) trouble

#14 2009-09-06 14:28:09

Re: Hardware (?) trouble

#15 2009-09-06 15:22:48

Re: Hardware (?) trouble

#16 2009-09-08 19:33:00

Re: Hardware (?) trouble

#17 2009-09-08 20:02:49

Re: Hardware (?) trouble

#18 2009-09-08 21:11:08

Re: Hardware (?) trouble

#19 2009-09-09 11:42:44

Re: Hardware (?) trouble

#20 2009-09-09 20:31:22

Re: Hardware (?) trouble

#21 2009-09-09 23:13:33

Re: Hardware (?) trouble

#22 2009-09-10 13:28:00

Re: Hardware (?) trouble

#23 2009-09-13 17:03:24

Re: Hardware (?) trouble

#24 2009-09-26 13:12:02

Re: Hardware (?) trouble

#25 2009-09-26 13:36:56

Re: Hardware (?) trouble

Board footer