[lug] automatic reboot with RH 7.1

[D. Stimits]

Bowen Zhou wrote:
> 
> Hi folks,
> 
> My problem is that my machine (Dell Precision 530, with RH 7.1) got
> suddenly automatic rebooting roughly about once a week, and kick me out
> gdm once every day! It drives me crazy, especially sometimes I am running
> simulations! Following is some exception from the /var/log/messages. Could
> you guys give me some clues about that?

Sudden reboot can be a bad power supply, or a very short/fast drop in
the power line (if no UPS), or heat problems. As far as software goes,
the kernel will spontaneously reboot any time there is a triple
exception, which can happen when ram is bad (try memtext86,
http://www.memtest86.com/) or the cpu or chipset overheats.

> 
> Many thanks...
> Bowen
> 
> Jul  2 12:39:07 svoice modprobe: modprobe: Can't locate module
> sound-service-0-3

In /etc/modules.conf, add:
alias sound-service-0-3 off
(or figure out if some kernel module provides this service, and alias it
to the module name)

> Jul  2 13:24:53 svoice gnome-name-server[1200]: input condition is: 0x11,
> exiting

Gnome is crashing. Not uncommon, but you won't necessarily know if it is
due to gnome or to hardware corrupting memory, so on.

> Jul  2 13:24:53 svoice gdm(pam_unix)[1054]: session closed for user zhoub
> Jul  2 13:24:53 svoice gdm[1054]: gdm_auth_user_remove: /home/zhoub is not
> owned by uid 0.
> Jul  2 13:24:53 svoice gdm[1054]: gdm_auth_user_remove: Ignoring
> suspiciously looking cookie file /home/zhoub/.Xauthority
> Jul  2 13:24:55 svoice modprobe: modprobe: Can't locate module
> char-major-81

Character device major 81 actually refers to a range of devices. Several
of them look like radio or TV capture cards. Did you configure your
system to work with a TV or radio card? You can enter in
/etc/modules.conf the line:
alias char-major-81 off
...and if you don't need the service, OR if you compiled it in (rather
than using a module), the message will go away without harming anything.
If you do need a module, it isn't finding the right one anyway, you
would instead need to find the name of the module that fulfills this
requirement, and alias it to the name of the module instead of "off".

> Jul  2 13:24:56 svoice gdm[24625]: gdm_slave_windows_kill_ioerror_handler:
> Fatal X error - Restarting :0
> Jul  2 13:24:58 svoice modprobe: modprobe: Can't locate module
> char-major-81
> Jul  2 13:24:59 svoice gdm[24629]: gdm_slave_windows_kill_ioerror_handler:
> Fatal X error - Restarting :0
> Jul  2 13:25:00 svoice modprobe: modprobe: Can't locate module
> char-major-81
> Jul  2 13:25:02 svoice gdm[24633]: gdm_slave_windows_kill_ioerror_handler:
> Fatal X error - Restarting :0
> Jul  2 13:25:03 svoice modprobe: modprobe: Can't locate module
> char-major-81
> Jul  2 13:25:05 svoice gdm[24637]: gdm_slave_windows_kill_ioerror_handler:
> Fatal X error - Restarting :0
> Jul  2 13:25:05 svoice gdm[1046]: Failed to start X server several times
> in a short time period; disabling display :0
> Jul  2 13:25:06 svoice modprobe: modprobe: Can't locate module
> char-major-81
> Jul  2 13:31:00 svoice syslogd 1.4-0: restart.
> Jul  2 13:31:00 svoice syslog: syslogd startup succeeded
> Jul  2 13:31:00 svoice kernel: klogd 1.4-0, log source = /proc/kmsg
> started.
> Jul  2 13:31:00 svoice kernel: Inspecting /boot/System.map-2.4.2-2smp


Eeeek! This is a terribly old and buggy kernel. You need a new one. Even
the precompiled/canned versions have updates if you use Redhat (see
ftp://updates.redhat.com). You could be hitting kernel bugs.

D. Stimits, stimits at idcomm.com