openSUSE 10.3 hangs frequently - ACPI issue?
 

Hi there,

Ever since upgrading to openSUSE 10.3, I've been having problems with the system hanging within a couple of minutes of starting up.

The first few boots, it was lasting anywhere from 2 to 10 hours, but then on the fifth or so boot, it lasted a mere couple of minutes (sometimes not even getting past the KDE loading splash screen).

This only happens when the GUI is loaded; it works fine in console mode. The only way I've discovered to get a stable system is to boot with the following options:

apm=off acpi=off noresume

Not being the greatest hardware expert on the planet, I know the ACPI has something to do with power management. When it's disabled, my system is sluggish, particularly on graphic operations. But if I don't disable it, the system dies within a couple of minutes.

Hardware wise, the memory is definitely fine as is the graphics card; the CPU has an incredibly powerful cooler on it (installed a while ago after I kept getting messages in MCELOG about the temperature being too high) and the power supply is pretty high-powered too. The case has five external fans, and the power supply has two fans. That said, the machine gets pretty hot during use. It's an x86_64 bit processor, on an nVidia chipset.

Any thoughts as to why this is happening, and how I can manage to get my system running on full again?

Cheers!

Don't do acpi=off and apm=off.

Try doing just acpi=off , or just apm=off.

If this solves your problem, then stick with it! Not using acpi / apm is perfectly fine, some systems aren't compatible.

Consequences? and Konqueror?
 

Has your problem anything to do with Konqueror?
I am having system freezes when running Konqueror;
I suspect that it is having problems with non-standard web pages.

Also, can you tell me the consequences of each of the boot suggestions.
I might do something like this.

Colin

Hey guys,

Thanks for the swift replies!

duryodhan: Thanks, I've now discovered that it only needs "acpi=off" to boot. The problem is that when I boot with this option, my system is very sluggish (particularly when running graphics-intensive programs, despite having a very high-spec system); alternatively, if I leave it off, the system runs at high speed until it hangs after two or three minutes.

colinlc: I don't think this is a specific program that's causing it. When the system hangs on this problem, it *really* hangs - the mouse doesn't move, no command keys work - I can't even Ctrl+Alt+F1 to another terminal. Even the "Magic SysRq" key combinations seem to stop. The lights on my RAM show minor activity but the system is completely unresponsive (can't be accessed remotely by SSH once it's hanged, either)

From my limited understanding of hardware, I gather that ACPI is the successor to APM. Certainly when I boot with ACPI enabled, I get a lot of ACPI messages appearing on my boot screen (none of which look particularly bad). But given that my system is very modern, surely it should support ACPI?

In any case, although I can get my system to boot with the "acpi=off" option, it's running much slower than it should - is this normal, and is there any way to rectify it?

Thanks for your attention.
My system also comes to the same "very dead" stand still.
i.e. no mouse, Alt F4 does nothing, Ctrl+Alt+Del, does nothing.
In my case it is "press the reset button".

But, I had only noticed it with Konqueror, so I just switched
to Firefox and had the same hang. System or internet or device?

Regards,
Colin
PS: see
http://acpi.sourceforge.net/
So, the system ought to be better?

Have you tried pressing Ctrl + Alt + F1, to log into a text console? All in all I'd say it's very unlikely to be a program such as Konqueror or Firefox; in my experience a single user application doesn't bring your system to a complete halt. It's more likely just coincidence that you've noticed it whilst using Konqueror.

Have you tried booting the Failsafe option from your boot menu?

No, I haven't booted in Fail-safe mode. It might work, but then what do I do? My laptop won't start in kde GUI mode - starts in terminal mode and
I have to log in and 'startx' to get kde up. Still hangs after awhile.
Like you I have modern 64 bit machines.

Interestingly SuSE 10.3 has given me different problems on my two machines.
e.g. both messed up my Kile/LaTeX, but in totally different ways.

I also found more useful acpi discussion at:
http://forums.suselinuxsupport.de/lo...hp/t12780.html

Colin

Hi there,

I looked at the discussion thread you mentioned - the thrust of it seemed to be that this was a KDE issue, not a GNOME issue. I think that this is potentially wrong in general, but it certainly was in my case. I booted in to GNOME and after a while the same thing happened. Generally I find that it is only when I am doing something taxing on the system that it hangs; for example, in GNOME I ran the glxgears program and this caused it to hang; in KDE, running a 3D screensaver seems to have the same effect.

I appear to have 'fixed' my system somewhat in going through the BIOS and disabling just about everything that looked optionalm superfluous or unnecessary. Any options related to "ACPI" or "APIC" I generally disabled also. The system now boots fine without the ACPI=OFF parameter but still runs more sluggishly than it should (I suspect I have simply achieved in the BIOS what I used to achieve by passing that parameter).

I will go through the BIOS at some stage and try gradually re-enabling options until I find the one that was causing the crash. I would like to know, though, why disabling ACPI or APIC would make the system visibly slower? What is it that these achieve that makes it run more smoothly? And is this always a motherboard issue, or can it be caused by other devices? How come I've only experienced real problems in openSUSE 10.3, but never really before this, if it's a hardware compatability issue?

Hi,
I gave up. I put my BIOS back to the original settings (as I had had with earlier versions of SuSE Linux). Then I tar/zipped my 'mail' directory, 'exported' my contacts, and backed up my own data. Then I made a "new" install of 10.3. It asked me if I wanted to retain the current 'users' and I said yes; and the "new" installation trashed everything else. Things actually went better than earlier versions of SuSE. The only thing not working now is my kmail - apparently I should have saved something secret! Watch your mail settings because they are not saved and the 'settings' edit dialog seems to be different (maybe using different jargon?). Anyway, I can receive but not send.
I think that 'upgrade' is unreliable; so I'd just forget it and go for a new installation. It did not trash my MS Windows junk (not that I use it much.)
Since then I have not had a single crash/hang.
Good luck,
Colin

Hi MoghNX01,
Bad news - I am still getting hangs with this clean installation.
But, only when browsing the web!
How are you getting on?
Regards,
Colin

Possible solution for "hanging" issue?
 

Hi everybody,

My openSuSE 10.3 started to hang few days ago as well. Sometimes it happened on booting screen, sometimes few minutes after booting, sometimes it lasted up to few hours and it didn't matter what I was doing at the moment. Fail-safe mode worked absolutely fine so did WindowsXP which installed on another hard drive.

Tried the usual - disabling and uninstalling Compiz-Fusion, 3D, reinstalling NVIDIA drivers (either from repository and from NVIDIA web-site), regenerating xorg.conf etc. No fun at all, system was still hanging with no reaction to mouse or keyboard.

At the end I figured out that it was either processor or graphic card overheating. Sorry, can't be more specific as I replaced thermal compound on both of them at the same time (I know, just out of curiosity I should have tried that on one at a time, but was rather desperate to get my system back to normal :)

For the sake of truth I have to mention that I've also reset the memory modules, probably swapped them while doing that.
No hangs since then even with Compiz.

Hope this info will be of any help.

There's thermal compound on the graphics card you have? I can't see anywhere on mine... in either case I fitted a new fan with new compound not long back, and it stopped processor overheating messages I was receiving. The new problem still persists though... maybe two different causes creating the same problem?

I've got quite cheap and old PNY GeForce 5700LE with heatsink and fan. It used to have "pre-applied" compound on chip, that hard stuff they put there on a factory (not thermal epoxy). Removed it ages ago and applied some KPT-8 I have from my trip to Russia. Anyway, it was more likely to be CPU issue than GPU, although you should check your graphic card temperature if that's possible at all.

Alas, both are within normal tolerances; one of the first things I checked, and I'm kind of wishing it was a heating issue as at least then I'd know what the problem was. This only seems to happen when I have certain BIOS options enabled now; specifically, "LIMIT CPUID VAL" set to 3, and enabling Intel SpeedStep technology as well as HyperThreading. The problem is that without these enabled, the system doesn't seem quite as responsive, although it doesn't halt either...

I don't believe its heat
 

I think it is something in the software which mistakenly BELIEVES it is heat.

I'm getting the same problem on an Athlon x86_64 Dual Core Processor 3800+ running on an ABIT motherboard with an nVidia GeForce 7300 video card with 256MB memory and a WD SATA 300GB HD. This is not a cheap box so I think you can disregard any concerns about the quality of your hardware.

I'm also multibooting with XP, Ubuntu and openSUSE 10.2 on three different hard drives on this box, and have NO problems with those installations, so I think you all can safely eliminate hardware being the problem.

I noticed it for the first time about a week after upgrading when I attempted to burn a DVD in this machine. After I hard slammed the box with the power button, I got a weird message, something about the hard drive overheating, that I have not gotten since with any of the hangs/hard slams I have done since. I have also THOROUGHLY checked out the hard drive and it is NOT overheating.

Question to all posters on this thread: what have you added to your x86_64 openSUSE installation in order to read DVD movies?