Date: Tue, 6 Nov 2001 14:15:23 -0500
From: Andrew Klaassen <ak@dkp.com>
To: linux-xfs@oss.sgi.com, linux-raid@vger.kernel.org
Subject: Re: Oops - XFS mount after replacing wrong RAID5 drive
Message-ID: <20011106141523.B5048@dkp.com>
Mail-Followup-To: linux-xfs@oss.sgi.com, linux-raid@vger.kernel.org
References: <20011105103521.A3864@dkp.com> <1004974636.7318.5.camel@jen.americas.sgi.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1004974636.7318.5.camel@jen.americas.sgi.com>
User-Agent: Mutt/1.3.23i
Sender: owner-linux-xfs@oss.sgi.com
Precedence: bulk

On Mon, Nov 05, 2001 at 09:37:16AM -0600,
Steve Lord wrote:

> On Mon, 2001-11-05 at 09:35,
> Andrew Klaassen wrote:

> > I think I may have just replaced the wrong drive after a SW
> > RAID5 drive failure.
> > 
> > And then I mounted the XFS filesystem read-write.  (Doh!)
<snip>
> > How do I mount the filesystem without writing anything at all to
> > the array?

> mount -o ro,norecovery

Steve Lord is good to me
And so I thank Steve Lord
For giving me the things I need
Like XFS mounts that only read.
Steve Lord is good to me.
 - to the tune of Johnny Appleseed

;)

Well, things are going very, very well.  We've already rescued
the critical data, and it looks like we'll be able to rescue
everything.  Here's what I did, in case anyone is interested;
the goal of everything was to make sure absolutely no writing
would happen, so that we could back out any changes in case
anything went wrong.

With the server off:
 - I cloned the original hdp, the drive I had originally
replaced (which was showing signs of failure but not completely
failed yet), and put the clone into the server.
 - I disconnected hdn (the drive that had completely failed).

Using an install CD - just using the Alt-F2 shell from the text
installer, not using rescue mode, so that it wouldn't try to
initialize any RAID devices that I didn't want initialized:
 - I marked hdn as a failed-drive in the raidtab.
 - I changed the array mount options to ro,norecovery,noauto,...
in the fstab.
 - I added an append="single ..." line to lilo.conf
and ran lilo, just to be safe.

After rebooting:
 - I ran mkraid -f on the array to re-initialize the raid
superblocks.
 - I mounted the array using the ro,norecovery options.

That's about $1/2 million you just helped me rescue.  Thanks. :)

Andrew Klaassen