On Wed, 11 Jun 1997 12:55:33 -0400 
 Ken Hornstein <kenh@cmf.nrl.navy.mil> wrote:

 > Depends what the definition of "better" is.  Linux advocates claim
 > that non-synchronous metadata writes is a feature, and is a big
 > performance gain.  I'm not a filesystem guru, nor do I claim to be
 > one on TV, so I can't evaluate this claim on any technical merits.
 > What do other people think?

Async metadata writes do give "better performance", but IMO are the
wrong way to get better performance.  If you crash before your metadata
has been written, there is no consistent information from which you can
repair the file system.  Basically, it's bloody dangerous, especially if
you consider your data important.

Linux ext2fs also has slightly "better performance" because they don't
have fragments, thus simplifying the bmap code ... However, I consider
this to be a lose:

	(a) If you want better performance from bigger block sizes,
	    you could end up wasting lots of space.

	(b) If you want to use all of that disk you paid for, you have
	    to use small blocks, which hurt performance.

It's a matter of trade-offs.  Having weighed the issues for myself,
I consider FFS to be a better file system than ext2fs.  ext2fs is not
new technology... it's a ufs-generation file system, done later, with
fewer features.

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939