Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 29 Nov 2002 19:09:15 +0200
From:      Vallo Kallaste <kalts@estpak.ee>
To:        current@freebsd.org
Subject:   Re: Harry Potter and the Disappearing Disklabel
Message-ID:  <20021129170915.GB2400@tiiu.internal>
In-Reply-To: <20021129164548.GA2400@tiiu.internal>
References:  <20021129093417.V7358-100000@volatile.chemikals.org> <20021129164548.GA2400@tiiu.internal>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Nov 29, 2002 at 06:45:48PM +0200, Vallo Kallaste
<kalts@estpak.ee> wrote:

> > I've seen one post similar to this, but not much else. I think maybe the
> > UFS2 problem had to do with Kirk's recent changes, but the disklabel
> > issue... I'm wary to reboot my machine! What in the hell could be causing
> > this? I'm tempted to point the finger at GEOM, but hate to say anything
> > like that.
> 
> Same here today. I had system from Nov 21, both world and kernel.
> Did buildworld, installworld and then rebooted with old 21Nov
> kernel. At boot fsck whined about /usr (ad0s1d) partition and died
> with incorrect superblock message leaving the system in single user.
> The /usr partition has UFS2 filesystem. Why the partition had to be
> fsck'ed? The system went down cleanly after build-installworld.
> I tried to fsck_ffs -b 32 /usr but it didn't like it either and died
> with signal 8. Floating point exception. I know the next alternate
> superblock _is_ there at 32, because I converted /usr to UFS2 only
> few days ago and remember the newfs command exactly.
> After the failed attempt of fsck_ffs -b 32 suddenly some fragment of
> recent -current talk popped in my mind and I remember there was talk
> about mount command doing some trickery. So I went with
> mount -t ufs -f /dev/ad0s1d /usr and voila the data was there.
> I'm almost sure that I can reproduce it, because I have the / and
> /usr dumps from the time I did UFS2 converting and the live-current
> cd burnt for this purpose (JPSNAP). It's possible to go back in time
> and fully restore the system as it were before.

One thought about the initial fsck issue. The system uptime was 8
days and almost all the time it did compilation/clearing up of my
workstation bundle port (~100 individual port). I did it because of
stability issues before, to control the kernel with only DISABLE_PSE
enabled. Because space in /usr is limited on this system, the
/usr/ports is mounted over ro NFS, but WRKDIRPREFIX, DISTDIR and
PACKAGES are set to local filesystem, so /usr periodically filled up
to ~95% and drained quickly (several concurrent rm -rf's) to 30%.
This is quite a stress to softupdates and filesystem in general, so
if there's a bug this explains the need for fsck after boot. Just a
thought.
-- 

Vallo Kallaste
kalts@estpak.ee

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021129170915.GB2400>