*BSD News Article 5481


Return to BSD News archive

Path: sserve!manuel!munnari.oz.au!uunet!mcsun!sun4nl!tuegate.tue.nl!svin09!wzv!gvr.win.tue.nl!guido
From: guido@gvr.win.tue.nl (Guido van Rooij)
Newsgroups: comp.unix.bsd
Subject: [386bsd] fsck fails in rc, works otherwise
Message-ID: <3891@wzv.win.tue.nl>
Date: 23 Sep 92 16:21:02 GMT
Sender: news@wzv.win.tue.nl
Organization: Guido's home 486 box
Lines: 27

Untill today i had 386bsd 0.1 up and running without any problems. However
my noname io card went to heavan, and it was replaced by another one 
(different manufacturer). And from now on, init hangs during fsck -p from rc.
I have 2 harddisks, and my fstab looks as follows:
/dev/wd0a		/		ufs	rw 1 1
/dev/wd0h		/usr		ufs	rw 1 2
/dev/wd1h		/disk2		ufs	rw 1 2

wd0a is checked, but when fsck starts to check both other disks in parallel,
the system seems to hang. When pressing ctrl-c rc says: reboot interrupted, as
it should, so fsck does not hang, but that is all....The only escape is
ctrl-alt-del. When I interrupt rc during the check of wd0a, and do the fsck -p
from the shell given to me by init, all goes well.
I tried both the unpatched and patched version of init..same result.
I think the systems locks up during io with the harddisk, but I dont know wich
one (though the second one seems logical since wd0a is checked okay). I can't
explain why this only happens when fsck is run form rc.
I repeat that with the other IO card I never had this problem. I also checked
several other IO cards (3 manufacturers in total), giving the same strange results. 
I checked rc, wich was okay, I replaced init with a fresh version, giving the
same result, and since fsck does work otherwise, also fsck is okay. Also
the kernel didnt change due to some bits mysteriously changing (i have 2 copies
and they are stil the same).

Has anyone had the same thing and found a workaround?

-Guido