Re: Kernel weirdness


Subject: Re: Kernel weirdness
From: Jim Cole (greyleaf@yggdrasill.net)
Date: Fri Sep 17 1999 - 03:12:49 MDT


Have you tried bringing it down by switching run levels? You should be able
to use "init 1" as root to drop down to single user mode. Also "init 0" or
"init 6" for halt or reboot.

The "D" you get in the process list means the process is in an
uninterruptible sleep. The process probably dropped into that state during
I/O and got hung there when the kernel panic hit. As far as I know, you
can't kill a process in this state. Even kill -9 will fail.

If the init's fail, the only other thing I can think of would be to
schedule a shutdown with a few minutes of delay and then take down the
network interface (e.g. ifconfig eth0 down -- with eth0 replaced as
appropriate). Its a long shot, but if the process is hung on the network
side, maybe that will free it up.

Jim

Tom Tarka wrote:

> Ok, here's some weirdness for you. I'll be verbose so I only have to
> say it once. Any help would be appreciated (the machine is still
> running
> and still needs help).
>
> HW: 400 MHz G4, linux installed on external 9G wide SCSI drive (no mac
> partitions)
>
> What's up:
>
> The load has gone to 5.99 and is holding there. Attempts to soft reboot
> have failed, and several procs in the proclist have
> the status "D".
>
> What happened:
>
> Started xxms, tried to play an mp3. Got a "cannot open device" error.
> Ran MAKEDEV audio in /dev
> Pressed play again. mp3 apparently plays but no sound out of my
> headphones.
>
> Meanwhile, I'm copying heaps of files from another machine on the net
> using scp
> and I get this message:
>
> Message from syslogd@ at Thu Sep 16 13:08:36 1999 ...
> kernel: Kernel panic: kernel access of bad area pc c0045000 lr c0044f08
> address 24 tsk scp/13444
>
> So I kill the client scp (on the other machine) and then attempt to kill
> the local scp and fail. Can't
> unmount the disk to run fsck on it because the scp process is still
> hanging around (with stat "D") and
> umount hangs.
>
> Attempt to shutdown:
>
> [root@ tommy]# shutdown -r -F 0
>
> Gdk-ERROR **: an x io error occurred
> aborting...
> Aborted
>
> Attempt to sync;sync;reboot --- hangs.
>
> Any ideas? I've still got an open xterm to the machine and it's still
> up, but I can't seem to reboot it (w/o a hard reboot, of course)
> and I was wondering what the cause of my problems might be (just copying
> too many files and the kernel buffer can't keep up? the
> disk can't keep up?)
>
> Thanks,
> t



This archive was generated by hypermail 2a24 : Fri Oct 01 1999 - 16:13:44 MDT