Date: Fri, 30 Jul 1999 04:13:39 -0400 (EDT) From: Alfred Perlstein <bright@rush.net> To: Greg Lehey <grog@lemis.com> Cc: Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>, Dag-Erling Smorgrav <des@flood.ping.uio.no>, current@FreeBSD.ORG Subject: Re: An oddity after a panic Message-ID: <Pine.BSF.3.96.990730040938.20420M-100000@cygnus.rush.net> In-Reply-To: <19990730170958.C56925@freebie.lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 30 Jul 1999, Greg Lehey wrote:
> On Friday, 30 July 1999 at 9:31:15 +0200, Dag-Erling Smorgrav wrote:
> > Jeroen Ruigrok/Asmodai <asmodai@wxs.nl> writes:
> >> [note I already did chflags]
> >> I'd love to hear some ideas about how to solve this.
> >
> > # chflags -R 0 /usr/obj/work
> > # rm -rf /usr/obj/work
>
> That doesn't seem to be the answer. And the directory where the
> problem occurred doesn't normally get any schg flags set.
>
> On Friday, 30 July 1999 at 8:41:27 +0200, Jeroen Ruigrok/Asmodai wrote:
> > Hi,
> >
> > I was cleansing /usr/obj after a panic to see if I could reproduce it and
> > I have this slight oddity:
> >
> > [root@daemon:/usr/obj] (32) # rm -rf work/
> > rm: work/FreeBSD/src/gnu/usr.bin/: Directory not empty
> > rm: work/FreeBSD/src/gnu: Directory not empty
> > rm: work/FreeBSD/src: Directory not empty
> > rm: work/FreeBSD: Directory not empty
> > rm: work/: Directory not empty
> >
> > [root@daemon:/usr/obj] (33) # cd work/FreeBSD/src/gnu/usr.bin/
> > [root@daemon:/usr/obj/work/FreeBSD/src/gnu/usr.bin] (34) # ll
> > total 2
> > 1 drwxr-xr-x 3 root wheel - 512 Jul 30 08:31 ./
> > 1 drwxr-xr-x 3 root wheel - 512 Jul 30 08:31 ../
> >
> > [root@daemon:/usr/obj/work/FreeBSD/src/gnu/usr.bin] (35) # cd ..
> > [root@daemon:/usr/obj/work/FreeBSD/src/gnu] (36) # ll
> > total 3
> > 1 drwxr-xr-x 3 root wheel - 512 Jul 30 08:31 ./
> > 1 drwxr-xr-x 3 root wheel - 512 Jul 30 08:33 ../
> > 1 drwxr-xr-x 3 root wheel - 512 Jul 30 08:31 usr.bin/
>
> That link count for usr.bin looks funny. It should be 2, and that's
> probably why you can't remove it.
>
> > I have no idea what's wrong here.
> > After the panic however I did get some softupdates inconsistencies, such
> > as: DUPs, Block counts, even an internal error: dups with -p.
>
> Yes, it looks like you have a broken file system. The easy part is to
> cut off the directory tree a bit further up (say, mv usr.bin to
> lost+found/andbroken). Actually getting rid of it and reclaiming any
> lost space will probably be more difficult and involve you in deeper
> knowledge of fsdb than you really wanted.
I've seen this happen with a DPT controller and regular scsi
when under a lot of I/O load for long periods of time, you'll
definetly want to use fsdb to ditch the dir, however if you
root around in there it'd be great to see the values in the
inodes pointed to by usr.bin and the "." and ".." entries contained
within.
It doesn't seem possible, but when I had this unreoveable directory
problem one of the inode's fields had overflowed to some
ridiculously high (or negative) number.
-Alfred Perlstein - [bright@rush.net|bright@wintelcom.net]
systems administrator and programmer
Wintelcom - http://www.wintelcom.net/
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?Pine.BSF.3.96.990730040938.20420M-100000>