Subject: Re: FFS journal
To: Bill Studenmund <wrstuden@netbsd.org>
From: Manuel Bouyer <bouyer@antioche.eu.org>
List: tech-kern
Date: 07/05/2006 22:38:53
On Wed, Jul 05, 2006 at 01:31:19PM -0700, Bill Studenmund wrote:
> On Wed, Jul 05, 2006 at 10:20:21PM +0200, Manuel Bouyer wrote:
> > On Wed, Jul 05, 2006 at 12:49:54PM -0700, Jason Thorpe wrote:
> > > 
> > > On Jul 5, 2006, at 12:01 PM, Manuel Bouyer wrote:
> > > 
> > > >How can we mont it read-only right now ?
> > > 
> > > Right now we have ordered MD writes.  With a journal, you generally  
> > > don't do that
> > 
> > I think I explicitely said that we need to keep ordered MD writes for this
> 
> Uhm, the idea of a journal is that you no longer have to order MD writes. 
> As I understand it, that is the _point_. To retain ordering when we have a 
> journal is defeating the purpose.

The point could also be to make fsck faster, and that is usually the feature
the average user sees.

> 
> > > (that's what you have the journal for!)
> > 
> > It's also to make fsck faster. I'd prefer to have at last the option to
> > keep ordered writes and have fsck deal with the replay, so that we're
> > guaranteed to always be able to read the filesystem (and eventually repair
> > it) even if the journal is corrupted.
> 
> If you lose the journal, you need to do a full fsck before touching the 
> file system.

But with unrdered metadata writes you may end up with an unrecoverable
filesystem.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--