xfs
[Top] [All Lists]

ADD 797165 - rationalise uuid kernel code for XFS

To: dxm@xxxxxxxxxxxxxxxxxxxx
Subject: ADD 797165 - rationalise uuid kernel code for XFS
From: pv@xxxxxxxxxxxxxxxxxxxxxx (dxm@xxxxxxxxxxxxxxxxx)
Date: Mon, 2 Oct 2000 15:15:05 -0700 (PDT)
Cc: linux-xfs@xxxxxxxxxxx
Reply-to: sgi.bugs.xfs@xxxxxxxxxxxxxxxxx
Sender: owner-linux-xfs@xxxxxxxxxxx
 Submitter : nathans                   Status : open                        
 Assigned Engineer : dxm               Priority : 4                         
*Modified Date : 10/02/00             *Modified User : dxm                  
*Modified User Domain : melbourne     *Description :
stick this in a bug so it isn't forgotten...

On Jul 21,  8:40am, Nathan Scott wrote:
> Subject: Re: LVM vs. kiobuf I/O
> > >         - Kernel uuid generation support
> > ...
> ...
> 
> I think there's a bunch of stuff which can be rationalised
> there too...

.....


==========================
ADDITIONAL INFORMATION (ADD)
From: daniel moore <dxm@xxxxxxxxxxxxxxxxxxxxxxxx>
Date: Oct 02 2000 03:15:05PM
[pvnews version: 1.71]
==========================

 => I was wrong last time. Glen added code to print uuid code from
 => xfs_db, not change it.
 => 
 => So we have two conflicing tasks here:
 => 
 => 1. turning an existing filesystem into two distinct filesystems.
 =>    because of how they are created - they would be different
 =>    devices, but have the same uuid.
 => 
 => 2. prevent mounts of the same device when it has multiple paths
 => 
 => I think we can make one work automatically and the other require
 => some operator intervention.

(irix6.5f-melb:irix:62215a)

Looks like he modified the write command so its possible to write the uuid
in the superblock (write wouldn't handle > 64 bit before). I made a similar
change to linux-xfs.

In the same mod he fixed up the code to change the UUID so it would work
properly on dirty filesystems (where log recovery could wipe out the
changed uuid).

His first change wouldn't work with the new log uuid code - it only
lets you write to individual superblocks. My "uuid" command sets all
superblocks and the log.

The second change would get dropped with the code to allow the kernel
to change the uuid on mount.

We don't lose any functionality here - we move task one to userspace
with a nice interface to pick a new uuid and we make task two work.


-----------------------------------------------------
 Daniel Moore                  dxm@xxxxxxx
 R&D Software Engineer         Phone: +61-3-98348209
 SGI Performance Tools Group   Fax:   +61-3-98132378
-----------------------------------------------------

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