Date: Fri, 27 Jan 2006 21:45:39 -0600 From: Craig Boston <craig@xfoil.gank.org> To: "Matthew D. Fuller" <fullermd@over-yonder.net> Cc: current@FreeBSD.org, Kris Kennaway <kris@obsecurity.org> Subject: Re: /c: unmount pending error: blocks -68512 files 0 Message-ID: <20060128034539.GA44075@nowhere> In-Reply-To: <20060128013955.GH1388@over-yonder.net> References: <20060128000909.GA69458@xor.obsecurity.org> <43DABFCA.5030509@mac.com> <20060128005304.GA70164@xor.obsecurity.org> <20060128013955.GH1388@over-yonder.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Jan 27, 2006 at 07:39:55PM -0600, Matthew D. Fuller wrote:
> I have most of my fs's mounted read-only (including things like / and
> /usr/ports and such) which I'm often mount -u'ing to r/w and back. I
> see a similar error often ("kernel: /: update error: blocks 20 files
> 1" is the most recent I see in messages)if I don't take special care
> to sync sync sync and/or wait a while after using it before mount
> -u'ing it. It's been happening for a while, back to even before mount
> -u started blowing up with EFAULT.
I've been seeing this as well -- on a system with flash memory for its
configuration (/config). /etc is a malloc-backed md. Periodically a
script runs that compares /etc to /config; if any changes are detected
/config is remounted read-write and the changes sycned over, then it is
remounted read-only (all done with mount -u). This way power can be cut
at any time with a very low risk of the filesystems being dirty.
When I moved it from 5.x to 6.x I started seeing the "update error"
messages when it was remounted read-only, and if I ran fsck it would
find problems. I ended up having to modify the script to unmount
/config completely and then mount it read-only...
I don't think it's quite the same as Kris's problem though as I've never
seen it with a negative number.
Craig
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060128034539.GA44075>