segmentation faults at boot time

starless · 08-18-2004, 10:04 PM

Hi all,
Almost new to the Linux world. Installed Slack some days ago: a success!
It worked great until tonight, when I couldn't ssh into it: I could only get a prompt by pressing <Ctrl+c>. But it was useless as every command I would type in returned a segmentation fault...
I rebooted the thing to realize that the boot process generates a lot of segmentation faults before hanging.
Partial output of the boot process:

Code:

...some segfault that I can't read...
...
INIT: version 2.84 booting
...
/dev/hda4 on / type reiserfs (rw)
/etc/rc.d/rc/S: line 162:	52 Segmentation fault	/bin/rm -f /etc/mtab*
...
mount: /dev/hda1 already mounted on /boot
mount: devpts already mounted on /dev/pts
nothing was mounted
/etc/rc.d/rc/S: line 243:	69 Segmentation fault	/bin/rm -f *
/etc/rc.d/rc/S: line 245:	70 Segmentation fault	/bin/rm -f /var/run/*pid /etc/nologin ...
...
Going multiuser...
request_module[ppp0]: fork failed, errno 1
request_module[ppp0]: fork failed, errno 1
*** HUNG HERE ***

Any idea of what might have caused this?
Thank you!

djbanaan · 08-19-2004, 04:14 AM

That looks pretty bad. I think your having hardware troubles, most likely faulty DIMM's. The segmentation fault (a.k.a. sigsegv, a.k.a. signal 11) occurs when a proces tries to access a part of the memory that is not available. This is usually directly related to faulty DIMM's.

Although it can also be a bug in the software you're using, this is less than likely in your case since it happens on every command.

Try swapping out your memory modules, if that's an option for you.

Memory failure is the most likely cause for this, but there are other possibilities. Check this document for more info. It contains a lot of information about this error.

starless · 08-19-2004, 08:10 PM

I think you're right. I realized that the amount of RAM reported is not what it used to be... I suddenly have 64MB more than before... Never noticed I was missing any!!
Well, I tried switching them (2x128) and using only one of them at a time in different slots but it never worked. I could succesfully boot off the Slack 10 disk 2 though. Issuing commands works fine. But after mounting /mnt/slack, /mnt/slack/boot and /mnt/slack/proc (am I missing something???) and chrooting to /mnt/slack, all the commands I run return a segmentation fault again. So I thought the fs got corrupted... but "reiserfs --check /dev/hda4" told me "No corruptions found"...
Note that this is a fairly fresh install and that all the important stuff resides on other partitions. I would just like to know what is going on here! Is there anything else I should check for before wiping it?
Again thanks!

Shade · 08-19-2004, 09:18 PM

It's strange that it works from the CDrom, but not your normal file system...
I still think it's a hardware problem, though.

Try using a live distro or boot disk with the memtest86 utility on it. Run it overnight and see what it finds.

--Shade

starless · 08-21-2004, 01:35 PM

Well, I put back both RAM modules and ran memtest86 for 30 hours
Cached: 256M, RsvdMem: 68K, MemMap: e820-Std, Cahce: on, ECC: off, Test: Std, Pass: 37 , Errors: 0.
I still get segmentation faults when running commands from / (/dev/hda4).
I got another king of error: /etc/rc.d/rc.M: line 1: /bin/cut: Text file busy
The boot process hangs after /etc/rc.d/rc.hotplug start :-(

Think I will let this one go and reinstall using only the oldest ram dimm that has worked forever...

Thanks for your help!