xfs
[Top] [All Lists]

Re: XFS and Emacs don't play nice

To: Scott Jaderholm <scott@xxxxxxxxxxxxx>
Subject: Re: XFS and Emacs don't play nice
From: Steve Lord <lord@xxxxxxx>
Date: Tue, 10 Jul 2001 04:17:17 -0500
Cc: linux-xfs@xxxxxxxxxxx, walters@xxxxxxxxxxxxxxxxxx
In-reply-to: Message from Scott Jaderholm <scott@jaderholm.com> of "09 Jul 2001 23:34:18 MDT." <m3u20lpcth.fsf@localhost.localdomain>
Sender: owner-linux-xfs@xxxxxxxxxxx
First of all, I would encourage both of you to update to a more recent
kernel. There are some new rpms on the ftp site, I would recommend these:

ftp://oss.sgi.com/projects/xfs/download/testing/Release-1.0.1-PR3/RPMS/RHlinux-2.4.3/

They are based on the updated kernel rpm from redhat 7.1, and contain a lot
of xfs fixes. Secondly the cvs tree development tree is always a good thing
to try when you have problems.

As for what is happening to your file. This is not corruption per se, the
updated inode size made it out to disk, but the data did not. If you do
an xfs_bmap you will probably find there are are no extents in the file.
(xfs_bmap .emacs). It is also not in in kernel buffer corruption problem,
it is because the buffers were not written to disk before the crash, but
the inode size was written in a transaction.

Please try a later kernel as I have not seen reports like this in a
while.

Steve

> Sorry about the previous mostly blank message, I was trying to figure
> out how to add a CC header and pressed C-c (send).
> 
> Hi,
> 
> I am using XFS from the SGI RHL 7.1 installer.  I'm using the smp
> kernel that comes with that system,
> 
> Linux version 2.4.2-SGI_XFS_1.0smp (root@exclaim) (gcc version
> egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #1 SMP Fri Apr 27
> 17:52:42 CDT 2001
> 
> I am using FSF Emacs 21.0.102.
> 
> My system just froze and when I rebooted I found this
> 
> $ ls -l .emacs
> -rw-r--r--    1 scott    scott       29543 Jul  9 23:01 .emacs
> $ wc .emacs
>       0       0   29543 .emacs
> $ od .emacs                                                               ~/.
> backup
> 0000000 000000 000000 000000 000000 000000 000000 000000 000000
> *
> 0071540 000000 000000 000000 000000
> 0071547
> 
> I was editing .emacs in Emacs when my computer froze.  It looks like
> it's just 29kb of nul characters.  Colin Walters (who I've CCed) has
> experienced similiar things with files being edited in emacs, and he
> is using version 21.0.103.
> 
> Colin Walter is running Linux meta 2.4.3-xfs #1 SMP Fri Jun 1 15:56:17
> EDT 2001 i686 unknown
> 
> Colin thinks it has something to do with buffer cache corruption in
> the kernel.
> 
> If you could please CC us on any discussion related to this we would
> appreciate it, we're not subscribed to the list.
> 
> Sincerely,
> jsj
> 
> -- 
> the sky is tired of being blue



<Prev in Thread] Current Thread [Next in Thread>