Re: disk freeze?

On Mon, Apr 25, 2005 at 08:14:43AM -0700, Ted Faber wrote:
> This seems to be my week for "me,too" ing Randy, but I'm seeing similar
> hard lockups.  The machine locks to the point where it won't hear the
> soft power switch - I have to yank the plug out to reboot it. 
> 
> Needless to say, this is really annoying.
> 
> I did come in this morning, cvsup a new kernel and make & install it (no
> lockups yet, but it's only been running a few minutes).  When I rebooted
> after make kernel, I got a lockup on reboot (!) and a message indicating
> that one thread was not a sole owner of a lock.  (Sorry, I didn't copy
> the text verbatim). And again a hard lock up to the pull the plug out of
> the machine stage, though there was a panic and an attempt to rebbot
> the machine.
> 
> It seems to happen more frequently on my desktop which NFS mounts most
> of its filesystems (with -L).
> 
> All kinds of configs, etc, available on request.  I'm not set up for
> dumps or debugging, but I can get set up if it helps.

These lockups are continuing.  I am effectively a Windows user.

pun:/var/crash$ uname -a
FreeBSD pun.isi.edu 6.0-CURRENT FreeBSD 6.0-CURRENT #12: Wed Apr 27 18:28:51 PDT 2005     root_at_pun.isi.edu:/usr/src/sys/i386/compile/PUN i386

The kernel was compiled immediately after cvsup.

I've built a debug kernel and on the last lockup was able to get into
the debugger and try to force a panic, but it never dumped me a core to
/var/crash  - locked up and rebooted.  I did get a file in /var/crash
called bounds that's only 2 bytes long, an ASCII "2" and a newline.

I suspect that the file system is deadlocking somewhere, but I don't
know how to go about ruling that in or out.

I can get to the debugger, so any clues what to look for on the next
lockup, given that a crash dump is unlikely?

I'm happy to provide any info I can.

-- 
Ted Faber
http://www.isi.edu/~faber           PGP: http://www.isi.edu/~faber/pubkeys.asc
Unexpected attachment on this mail? See http://www.isi.edu/~faber/FAQ.html#SIG