Can anyone explain these /var/log/messages
A few days ago the power went out, and I don't have a UPS so my system went down as well. It's Debian etch with 2.6.18 kernel (i386) on a Pentium 4. I should mention that after I restart I need to manually `modprobe -r ehci-hcd' which is a workaround I've found to stop the kernel from randomly crashing because it doesn't like my USB hard disk (/dev/sdc).
Anyway, now after a day or so of uptime I check /var/log/messages and get a lot of things like Code:
Feb 23 22:21:54 localhost kernel: c0149f2b Code:
Feb 23 23:51:10 localhost kernel: <1>BUG: unable to handle kernel NULL pointer dereference at virtual address 00000024 Code:
Feb 24 04:25:02 localhost kernel: <1>BUG: unable to handle kernel paging request at virtual address 1303599d Any ideas are appreciated in helping me understand this, |
well it's some sort of memory accessing error by the look of it, and some potentially useful debugging info. i'd guess there are acutally more messages above what you've typed though, which you've not shown us. might mean you have some dodgy ram, or there's a bug in the kernel you're running. if it's arisen from a crash, i'd be looking at the hardware myself.
|
Aside from the boot up messages, the rest are `-- MARK --'. I don't think it's hardware or a bad kernel, as I've been using both for quite a while. The uptime was at about ~70 days before the power went out with the same hardware and kernel, and never had any problems.
|
well sudden losses of power cause hardware problems. you aren't going get get a software bug arising froma power outage, that's for sure (until i'm corrected!). The errors say there are problems accessing pages of memory within the kernel, and memory is in RAM or swap...
|
Quote:
Power outage means hardware problems until proven otherwise. Have you done the basics - fsck (all partitions), memtest ???. For me, having to "work around" a (presumably) on-board USB problem would be sounding alarm bells re the motherboard. How you you test all the components on it ??? - you can't. For a required system I'd replace it, and hope nothing else was broken as well. I once lost everything due to lightning strike - mobo, 3 disks, DVD burner, video card. Only saved the floppy and a CD drive (and the memory surprisingly) - yes, I *had* a UPS. Another occasion, only mobo went. I try to have at least 2 machines with the same motherboard, so I can swap components to isolate failures. |
I had a similar problem...
My machine worked without issues on 'old' dsitributions like mandriva 2006, ubuntu etc. My problems began with new distro's like suse 10.2 , mandriva 2007, ubuntu 6.10 which had me worried about my machine being broken. My sessions would be very lucky to last as long as 1 hour without complete freezing and a flick of the power button. (the reason why I had cycled through so many distro's thinking that they were buggy - I was just trying to get a stable session!!!) The other common thread with the freezing was that the new distro's were kernel-2.6.18.x, and I saw the "printing eip:" as a commonality. Since my printer is USB, I disabled the Parallel port in the machine's BIOS. My machine now behaves flawlessly since disabling the parallel port. #uptime 10:00am up 45:27 Goodluck! |
I've got this worked out now. One of two 512 MB memory sticks had been damaged. I learned my lesson though, and when I was at the shop buying my replacement memory I picked up an APC UPS unit. A highly recommended investment!
|
All times are GMT -5. The time now is 02:04 PM. |