LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Hardware (http://www.linuxquestions.org/questions/linux-hardware-18/)
-   -   Temperature is OK but System Shuts Down Randomly (http://www.linuxquestions.org/questions/linux-hardware-18/temperature-is-ok-but-system-shuts-down-randomly-650919/)

xp_newbie 06-22-2008 03:11 PM

Temperature is OK but System Shuts Down Randomly
 
I just built a new Quad Core based system (see signature). :)

I immediately installed Ubuntu 8.0.4 on it. Installation seems to have gone OK without any noticeable problems.

However, in almost every run of the system ends up with surprise/sudden reboot. This is of course unacceptable.

How do I start troubleshooting where the problem is?

BTW, it seems that using the mouse (PS/2, not USB) triggers those shutdowns more than anything else.

Also BTW, I tried a temporary install of Vista Ultimate and the system seems much more stable (definitely more usable). However, even with Vista I had one case (only one case so far) in which using the mouse froze the entire system (yes, the system froze but didn't reboot).

Last BTW (for now :)) - The mouse works perfectly on the older system (see signature below). It is connected to both systems via a KVM switch.

I would appreciate any tip or help that would help me start troubleshooting my new system.

Thanks!

lazlow 06-22-2008 03:34 PM

I really hate KVM switches. I have seen more of them eat hardware than anything else.

Try plugging the mouse straight into the computer and see if that cures the problem.

silkstone 06-22-2008 04:17 PM

Hello from a new member!

I used a PS/2 mouse and keyboard on Feisty with no problem at all, but after a fresh install of Hardy the keyboard locked up at random intervals for no apparent reason. I've also seen reports of people having similar problems with PS/2 mice.

I switched to a USB mouse and keyboard and there were no further problems - now using wireless USB mouse/keyboard and that works fine too. It makes me wonder if there is a bug with Hardy and PS/2 devices.

rjlee 06-22-2008 04:59 PM

There are no easy solutions, but there are ways to get more information:

If the machine reboots without cause, this could mean any number of things from a faulty reboot button to a software driver error. Start to track down the issue by taking a look in the /var/log/messages file, and seeing if there is anything unusual just before the unexpected reboot.

To debug freezing and locking up, if you can get it to happen semi-regularly: If you have another machine around, then you might try installing sshd and enabling SSH in the Services menu. Then, when the machine freezes, try and ssh into the machine with your usual username and password (ssh in Linux, PuTTY in Windows). If that works, you can check the messages file and run top to see if anything is amiss. If you can't log in by SSH, then you also know that the system (and not just the user interface) has gone down; also check if the Num Lock key toggles the Num Lock light on and off, that will tell you if the kernel is processing interrupts or not.

Btw, I also have a machine that I've been trying to install Hardy on, and it won't boot with my PS/2 mouse attached.

xp_newbie 06-22-2008 07:24 PM

Quote:

Originally Posted by lazlow (Post 3191982)
I really hate KVM switches. I have seen more of them eat hardware than anything else.

Try plugging the mouse straight into the computer and see if that cures the problem.

Well, I tried just that as first step. It didn't help. :(

It would have been too easy if it helped (aside for the fact that I would then have had to find a solution for the basic need to use two PCs with one set of KVM).

Quote:

Originally Posted by silkstone (Post 3191982)
I used a PS/2 mouse and keyboard on Feisty with no problem at all, but after a fresh install of Hardy the keyboard locked up at random intervals for no apparent reason. I've also seen reports of people having similar problems with PS/2 mice.

Wow! as a new member you seem to be right on the money here, because this is basically the same exact thing that I experience (the other side of the KVM switch is connected to a Ubuntu 6.06 system and it works flawlessly there).

Quote:

Originally Posted by silkstone (Post 3191982)
I switched to a USB mouse and keyboard and there were no further problems - now using wireless USB mouse/keyboard and that works fine too. It makes me wonder if there is a bug with Hardy and PS/2 devices.

Well, if using USB mouse is indeed the solution to my problem, how do I solve the problem of needing a KVM switch?

Thanks!

xp_newbie 06-22-2008 07:34 PM

Quote:

Originally Posted by rjlee (Post 3192049)
Btw, I also have a machine that I've been trying to install Hardy on, and it won't boot with my PS/2 mouse attached.

Well, I was able to complete installation with the PS/2 mouse attached (via a KVM switch). I don't know what that means.

Quote:

Originally Posted by rjlee (Post 3192049)
If the machine reboots without cause, this could mean any number of things from a faulty reboot button to a software driver error. Start to track down the issue by taking a look in the /var/log/messages file, and seeing if there is anything unusual just before the unexpected reboot.

No random reboots anymore after increasing DDR2 voltage from 1.8V to 2.0V (no overclocking, just checking the manufacturers specs for the memory)


Quote:

Originally Posted by rjlee (Post 3192049)
To debug freezing and locking up, if you can get it to happen semi-regularly: If you have another machine around, then you might try installing sshd and enabling SSH in the Services menu. Then, when the machine freezes, try and ssh into the machine with your usual username and password (ssh in Linux, PuTTY in Windows). If that works, you can check the messages file and run top to see if anything is amiss. If you can't log in by SSH, then you also know that the system (and not just the user interface) has gone down; also check if the Num Lock key toggles the Num Lock light on and off, that will tell you if the kernel is processing interrupts or not.

I couldn't SSH to it... which could be explained by the fact the Num Lock key doesn't toggle the Num Lock light on and off... which means that the kernel is not even processing interrupts.

This problem is nastier than I had hoped. I wish it would simply a matter of replacing the PS/2 mouse with a USB one, but read this blog and you see that Ubuntu 8.04 has serious problems:

http://techxplorer.com/2008/04/23/fi...in-ubuntu-804/

Almost like Voodoo...

It seems like Ubuntu 8.04 requires very delicate selection of certain components + a very precise angle the moon... I don't experience those lockups in Vista (same exact hardware, dual-boot).

BTW, my graphics card is ASUS EN8600GT SILENT/HTDP/256M (based on GeForce 8600 GT 256MB 128-bit GDDR3 PCI Express x16 HDCP Ready SLI Supported):

http://www.newegg.com/Product/Produc...82E16814121076

Any other ideas?

Thanks!

strick1226 06-23-2008 11:43 AM

A few ideas:

1)It's possible you're having issues with your RAM config.

Have you tried operating on only a single pair of DIMM's (down to "only" 4GB)?

I've seen stability issues on way too many boards when attempting to run dual banks of double-sided DIMM's--worth trying.

I applaud you on your increasing the default RAM voltage after checking the defaults on the RAM itself and in BIOS--most people would not have thought to do so. There's way too many boards out there that run the RAM at below-spec voltage, causing all sorts of issues.

2) BIOS 14 is a bit old for this board. I'd recommend updating to 16, which includes AHCI and CPU microcode fixes, among other things.

http://www.uabit.com/index.php?optio..._name=IP35+Pro

3) Power supply issues?

Generally speaking, when you have spontaneous reboots and/or hangs on modern systems these days it's very often due to a flaky power supply rather than an IRQ conflict. I'd try swapping out another ATX power supply--even if only temporarily--to verify the issue still happens with a completely different power supply.

4) Flaky motherboard

A bad motherboard--possibly with defective PS/2 ports--could cause the problem as well. If it only happens with PS/2 port activity, it's entirely possible it's simply a matter of a defective system board, particularly in the area of the PS/2 ports--or any cold solder joints nearby.

Overall, the IP35 Pro seems to have a fairly good reputation, aside from the very first batch. Hopefully it's not something too complicated...

As for Ubuntu needing a "precise angle to the moon" in regard to hardware, I don't think it's nearly that treacherous. I've been running it for years on a number of machines in various configurations and have been very happy with the results. My latest box, a Xeon 3350 on an Asus Rampage Formula board, hasn't experienced a hiccup to date--and it's pretty cutting-edge stuff.

Hang in there. Let us know how things go!

xp_newbie 06-23-2008 06:15 PM

strick1226, thank you for your thoughtful reply. I will attempt to answer every question you posed here, then perhaps add some new input/insights:

Quote:

Originally Posted by strick1226 (Post 3192708)
Have you tried operating on only a single pair of DIMM's (down to "only" 4GB)?

Yes. Same exact behavior.

Quote:

Originally Posted by strick1226 (Post 3192708)
BIOS 14 is a bit old for this board. I'd recommend updating to 16, which includes AHCI and CPU microcode fixes, among other things.

Actually I originally had the problems with BIOS 16. I downgraded to BIOS 14 because BIOS 16 has a well documented problem when it comes to reporting temperature for Q6600 CPUs (which is what I have) - It reads at least 15C higher than reality. None of the BIOS changes in v16 make any noticeable improvement for the Q6600. The main changes in the latest are aimed at users of 45nm CPU's, and in optimizing for those CPUs, ABIT goofed up for the older 65nm type such as the Q6600.

Quote:

Originally Posted by strick1226 (Post 3192708)
3) Power supply issues?

I took care of selecting the best power supply there is for this motherboard. I also invested quite some research culling information regarding each and every component in my system to make sure that it has high rating in terms of Linux compatibility. Please don't tell me that when it comes to Ubuntu 8.04 even that doesn't help. :)

Quote:

Originally Posted by strick1226 (Post 3192708)
4) Flaky motherboard

How can it be flaky if works perfectly with Vista?

My bottom line is that I built this system from scratch, having only one copy of each component. If any component in this system is defective, how do I know which one it is? So that I can return it to Newegg for replacement?

Quote:

Originally Posted by strick1226 (Post 3192708)
Overall, the IP35 Pro seems to have a fairly good reputation, aside from the very first batch.

That is why I chose it - it was the best modern fit among the boards appearing both in this forum's HCL and Newegg. :)

Quote:

Originally Posted by strick1226 (Post 3192708)
As for Ubuntu needing a "precise angle to the moon" in regard to hardware, I don't think it's nearly that treacherous.

Hmmm... please read what this guy wrote in response to a similar call for help in ubuntuforums.org:

http://ubuntuforums.org/showpost.php...49&postcount=3

Quote:

Originally Posted by strick1226 (Post 3192708)
I've been running it for years on a number of machines in various configurations and have been very happy with the results.

Me too! But with Ubuntu 6.06, not 7.04 or 8.04... In fact, I built this system because my Ubuntu 6.06 on a P4/HT is becoming too old for the stuff I am doing (running multiple VMs concurrently, using VMWare).

Now, here is some good news. I just enabled the nVidia "restricted driver" and got amazing screen effect that I have never seen before. It also reduced to max resolution from 1600x1200 to 1280x1024 but I believe that I can deal with that later. But this is not what I wanted to point out. What I wanted to point out is that... since I enabled this, I haven't had (yet) a freeze. It's too early to declare victory, but the fact that I was able to write this entire reply on my new machine without any freeze is already good. :)

I will update this thread as I discover new things.

Thanks!

strick1226 06-24-2008 09:43 AM

xp_newbie,

Glad to hear you tried running the single bank of RAM, and already have experimented with the different BIOS levels. That stinks that ABIT hasn't provided another update with the proper Q6600 support--that's only one of the most popular CPU's of the past year, and the main reason so many people put together a new box... ah, well..

I neglected to mention in my first post that it's very, very unlikely your CPU is overheating--that's a killer heatsink, so you should be good to go :)

Even the most incredibly expensive and highly-rated power supply, just like any other part, can be defective. I've purchased 98% of my stuff from Newegg since ~2000, so I've nothing against them. However, read through the reviews of even the highest-rated stuff and you'll find some users mentioned their initial shipment of something was flaky, or DOA. It happens, unfortunately. I never consider anything new-in-the-box to be completely free of defects until I've had it running for a while, without issues.

From the sound of the problem, as you described it, it really hit me as a possiblity that it could be the power supply, so it's always worth trying another one--even an older one, so long as it can connect to the board properly--to rule it out. Usually spontaneous reboots are due to occasional drops in voltage on one of the rails, resulting in the same kind of reaction as pressing the reset button.

Things certainly sound promising now that you've installed the binary nVidia drivers. I've not a clue how this would have fixed your PS/2 port and reboot issue--indeed, if anything, people tend to experience the extra occasional issue _with_ the binary nVidia or ATI drivers rather than running X.org with an unaccelerated/non-restricted driver.

As for how Vista ran better than Ubuntu 8.04--well, I'm not a programmer, so that's anyone's guess. Worth noting, however, that the M$ Windoze kernel and the linux kernel handle hardware very differently--there's always a possibility one might be more forgiving of a hardware issue than the other. In my experience, however, I've found linux generally tends to be the more reliable OS on lots of hardware--the way kernel modules are included for various network cards, chipsets, etc. generally helps keep things updated much more easily than manually updating individuals drivers yourself in Windoze.

Before anyone flames me as an anti-M$ guy, I use both linux and Windoze, and appreciate each of them for what they can do in the right environment. :D

Glad to hear things sound promising--please keep us updated and let us know how things pan out! :)

xp_newbie 06-24-2008 12:55 PM

Quote:

Originally Posted by strick1226 (Post 3193582)
Even the most incredibly expensive and highly-rated power supply, just like any other part, can be defective.

I agree, but a faulty power supply should be equally faulty for Vista and Ubuntu 8.04. So I don't think that my (expensive) power supply is the problem.

But I hope that the hardware troubleshooting part of this thread is no longer relevant becuase... since I enabled the NVIDIA restricted driver (not even the latest one from nvidia.com, just the one that comes with Ubuntu 8.04), I have yet to experience even a single lockup. :)

Oh - and Compiz is beautiful. :)

BTW, the temperatures as they stand now are:
Quote:

PWM (aka Sys): 58C
Core 0: 38C
Core 1: 37C
Core 2: 38C
Core 3: 35C
GPU: 57C
So, you are right - the CPU temperature is the last one to be suspected. The only one raising concerns is the fanless GPU (ASUS EN8600GT).


I will keep you updated if I discover something new.

jiml8 06-24-2008 02:22 PM

Based upon all the evidence I have seen in this thread, I would vote against a power supply or mobo fault and for a problem with Ubuntu. A wild pointer in a kernel driver or a buffer overflow will take the kernel down every time. That Vista loads up without issues supports the idea that the problem is NOT hardware.

I must say that I tried to update my laptop's Kubuntu installation to 8.04. The updater crashed and presently my Linux partitions on the laptop are unavailable. That machine dual boots with XP so I have just been using XP for now; I'll get around to fixing linux on it later.

There are an assortment of different drivers available for mice and keyboards in every linux distro, including ubuntu. My only suggestion at this point is to force a different driver to load and see what happens.

You might also try a different distro. I have to say that I like both Fedora and Mandriva better than Ubuntu. You will find Mandriva to be nearly as easy to install and configure, and less of a PITA than Ubuntu when you decide you need to get "down and dirty" with the system.

I've been using Mandriva since Mandrake 7.2 (with no reinstalls...upgrades all along the way) and I have to say that it has become a very polished and reliable distro.

edit:

Since you have found a possible tie-in with the NVidia driver, you might try downloading the latest one from nvidia.com and installing it using the nvidia installer, bypassing the apt mechanism of ubuntu. This could get you going.

Alternatively, you might have some xorg issue, with its drivers.

Also, my experience with compiz/xorg/nvidia has been that you need to avoid xgl like the plague. If you have it enabled that could be the problem right there.

farslayer 06-24-2008 02:33 PM

Just to answer one of your earlier questions.. How do I connect a PS2/KVM to a PC via USB..

Just use a PS2 to USB converter.. NONE of my new servers have PS2 ports anymore, but the KVM in the server room is PS2 and does not support USB. I use these adapters and everything work just fine..


Just fyi,

Carry on !

strick1226 06-24-2008 02:35 PM

Note that the system froze in Vista as well. Only once, but it's my understanding Vista wasn't run for very long, so it's not an entirely fair comparison.

That having been said, it does seem to be a weird quirk of sorts -- perhaps it would be best to try some other distro's, as jiml8 suggests, and see if things are more reliable.

I think there have been two or three kernel updates to 8.04 during this month alone (currently running 2.6.24.16-generic), so it also might be worth it to try processing the system updates (which would include the kernel).

If you find another distro works better for you, then by all means use what you like, and what works best for you.

That's the nice thing about linux--lots of choices! In the M$ world, if you don't like Vista or XP, you're kinda stuck :)

xp_newbie 06-25-2008 12:15 AM

Quote:

Originally Posted by strick1226 (Post 3193840)
Note that the system froze in Vista as well. Only once, but it's my understanding Vista wasn't run for very long, so it's not an entirely fair comparison.

Sorry if I wasn't clear on this: The freezing problem on Vista occured only once - it disppeared completely once I increased some voltages in the BIOS. Then, I ran Vista and Ubuntu equally long, and while on Vista I experienced rock solid stability, I had the reported freezes on Ubuntu 8.04.

Then I replaced the display driver with the restricted one and now Ubuntu 8.04 is rock solid as well. :)

Bottom line: I now have a rock solid system now. Running rock solid both Vista and Ubuntu 8.04.

Oh, one last thing: The 3D cube in Compiz is very cute - and rock solid! :D


All times are GMT -5. The time now is 12:05 PM.