Date: Tue, 28 Oct 2003 09:56:06 -0800 From: Wes Peters <wes@softweyr.com> To: Dan Strick <strick@covad.net>, pmes@btinet.net Cc: current@freebsd.org Subject: Re: UFS file system problem in either stable or current Message-ID: <200310280956.06279.wes@softweyr.com> In-Reply-To: <200310281642.h9SGgGBs000368@ice.nodomain> References: <200310281642.h9SGgGBs000368@ice.nodomain>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday 28 October 2003 08:42 am, Dan Strick wrote:
> On Wed, 22 Oct 2003 06:23:20 -0500, Peter Schultz wrote:
> > Dan Strick wrote:
> > > There seems to be an inconsistency between release 4.9-RC and 5.1
> > > ufs support. If I fsck the same ufs (type 1 of course) file system
> > > on both releases, each claims that the other has left incorrect
> > > summary data in the superblock. Presumably only one can be
> > > correct. I just don't know which to blame.
> > > ...
> >
> > There is no problem AFAIK, you just have to fsck with the matching
> > executable. A lot has changed with FreeBSD 5, spend some time with
> > the -current archive and you will learn more. I'm sure you noticed
> > how your findings are consistently inconsistent.
>
> Thanks for the pointer. I eventually found at least part of the
> discussion in the -current archive. I interpret the discussion as
> follows:
>
> The 5.x UFS1 file system turns out to be slightly incompatible with
> earlier UFS file systems. The problem is only that it keeps the
> summary data in a different location in the superblock, but that is
> sufficient to make the file systems incompatible. There seems to be
> no interest in making them compatible.
I need them to be compatible if possible. I'm working on a few related
problems, so I'm becoming more familiar with the code. I'm willing to
help if you have an approach in mind. Do you know why the change was
made?
> Suggestion: if FreeBSD 5 used a different clean flag and FreeBSD 4/5
> always cleared the other's clean flag whenever they rewrote a
> superblock, the file systems would automatically be refscked whenever
> you switched between operating systems but not after a normal reboot.
If somebody does so, please don't use the fs_state field; I have a local
patch that uses that for a different (incompatible) purpose that I'd like
to commit soon.
--
Where am I, and what am I doing in this handbasket?
Wes Peters wes@softweyr.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200310280956.06279.wes>