I had decided to try switching from Debian Etch on i386 to Xubuntu Gutsy on amd64 and am having stability issues.

Every so often, while I am at the desktop (hasn't happened while at an ssh terminal), the computer has completely locked up. Switching to my debian server, I can't ping or ssh into the box either, it appears to be a total system lockup. The computer does not respond to ACPI events (like the power button) or any keyboard commands (ctrl+alt+del, ctrl+alt+backspace, ctrl+alt+F1), so it doesn't appear to be the Xserver locking up.

I had been running Debian Etch on i386 on kernel 2.6.21 (SMP K7 version) with the only "tainted kernel" drivers being the nvidia from debian's nvidia-kernel-source and the ones from vmplayer 2. I was running for around a year with absolutely no lock-ups.

I just switched to Xubuntu 7.10 RC (Gutsy) last weekend. I installed the amd64 version. Since then I have had 2-3 of these lock-ups.

I have looked in the syslog, messages, dmesg (which got overwritten unfortunately), kern.log, daemon.log, faillog, debug and user.log in the /var/log directory, and everything is silent. That is to say, the last logs from before hitting the reset button show no errors, no kernel panics, nothing.

I am at a loss of trying to troubleshoot this issue without giving up on xubuntu amd64, which I'd rather not do.

Relevant hardware:
Epox 9npa ultra (nforce4)
Evga e-GeForce 7300 GS
200 GB sata 2 HD
1.75 MB PC2700 mushkin RAM - 1GB+.5GB+.25GB
AMD 3800+ Athlon 64 X2 Dual-Core

The things I can think may be the culprit software wise:

Kernel: 2.6.22-14-generic #1 SMP x86_64
Nvidia kernel module or the nvidia-glx-new
vmplayer 2.0.1 vs. 2.0.0
Just maybe the amd64 not being stable?

The other possibility is that somehow the amd64 kernel is stressing my hardware differently than the i386 one and that is causing an issue.

Anyone have and ideas and how to troubleshoot this? It is not very reproducible, there isn't one thing that I can put a finger on and say that it caused it. I was using the mouse wheel one time and another time I think I was dragging a window, but nothing special

Thanks,
Andrew

I have just remembered that I have had this problem before. ~2 years ago I tried debian amd64 and had the same problem. So there is definitely something wrong with the amd64 working on my hardware. It may be that my RAM is older (I know it isn't bad -- I have tested it several times and it was stable with i386), or something else.

I have 2 choices,

1. Go back to i386
2. Try to upgrade my RAM

Any opinions on if it is the RAM or not? If it were you, would you bother trying upgrading the RAM?

You say you tested the RAM - did you run memtest on it or you just derive it's good from the fact it was running 386 kernel correctly?

I ran memtest on it for over 12 hours straight - no errors.

I'll have the same thing happen on my Desktop (Dual Core AMD64). I suspected ram as well but the test ran successfully. My problem is very intermittent. It may run for a couple weeks or 45 minutes. I've never run a 32 bit distro on it to compare that way.

Correction:

just cracked my case again, and I was wrong on the ram.

I have:
PC3200/400 1 GB on the DIMM 1 slot
PC2700/333 512 MB on the DIMM 3 slot
PC2700/333 256 MB on the DIMM 4 slot

I am trying with only the PC3200 at the moment to see if that helps. Will report back later.

jschiwal - Sorry to hear you have had similar issues without a solution yet. So far I haven't lasted more than a day or two, so I hope I can get more accurate results. The couple of weeks could be very hard to troubleshoot. Perhaps there is something wrong in the amd64 kernel?

Update: it isn't the RAM

I just used the PC3200/400 1GM DIMM from my wife's computer which is stable, and immediately after my last post it locked up again.

I guess it is now down to:

1. linux amd64 is unstable
2. my motherboard/BIOS
3. my CPU

I'm starting to think that the linux amd64 is just unstable, at least with my hardware. I'll try a BIOS update, but I think I'll end up giving up on the amd64 kernel/distro.

There isn't any real reason to run 64 bit os, unless you have 4 gigs or more of ram

I am back on i386. The FTP site for Epox is horrendous, so I was unable to get the latest BIOS (reason enough for me to not buy from them again possibly). While I was trying to setup the i386 architecture install it froze again. I will keep people posted on my findings as I go to help anyone that may have the same issue.

Well, it had been almost a week without a lock up and I hoped I was in the clear, but this week I had one lock up and today have had two. I don't know what to do now as it was not the 64bit platform and not the RAM, so basically I guess I am stuck until hopefully somehow a kernel upgrade or a new NVidia driver or something fixes it.

Don't suppose someone out there has anything else to try?

FYI -- I did update my BIOS, apparently you don't want to use the Epox USA page, but the global one to get the latest BIOS. So that was not the issue either.

Looks like is the nvidia driver:

https://bugs.launchpad.net/ubuntu/+s...22/+bug/145112

No hangs with the vesa driver