From: Juha Virtanen (jiivee@hut.fi)
Date: 04/26/93


From: jiivee@hut.fi (Juha Virtanen)
Subject: Re: Really weird panic
Date: 26 Apr 1993 22:10:28 GMT

In article <735802442.11310.0@moriarty.bh.andrew.cmu.edu> Frank T Lofaro <fl0p+@andrew.cmu.edu> writes:
><description about umounting filesystem and logging in after that deleted>

>Syncing hung (the sync process stuck forever in disk wait) and new getty's
>were nver spawned, but I could create files in the root fs (which was still
>mounted), and they were there even after a reboot!. Ps worked fine during all
>this, as did many command.

>Fairly interesting, but might indicated something a little screwy. :^)

My system acted once with 0.99.8 kernel quite in the same way. I
were idle for few hours when I came to it and tried to start
dialing with kermit (it was there running on one VC). I typed a
couple of letters and kernel said me "Trying to free block not in
datazone" and then kermit became zombie.

I was wondering what's going on and looked processes with procps
based top. I notices that one after another processes went into
"uninterruptiple sleep" -- as top's manpage says. First update,
then init, syslogd, some shells, cron based jobs (cron itself kept
running), X applications etc... And I couldn't sync!

After rebooting my box, both of my filesystems (root is minix and
otherone is extfs) were in good condition. There were nothing to
fix for (e)fsck though my system were up about an hour after
update stuck. Everything written to disk during this time never
reached the disk's surface...