Sasi Peter <sape@iq.rulez.org> wrote:
>>> FWIW I have to agree with Rik on this matter.
>>> Changing the on-disk format every few weeks is very counterproductive for
>>> the people using resierfs. That is the only reason for me _not_ using
>>> reiserfs on 100 computers @ work.
> As far as I know Linus hasn't made any decision regarding this matter and
> the last thing I read was that he'd like a journaling fs in the 2.4
> kernel.
>That is right, and I suppose it is because he also considers ReiserFS to
>be production quality.
Here's a thought concerning the on-disk compatibility debate with
ReiserFS: I was recently in contact with Chris Mason about some
unrelated issue (i.e., making the ReiserFS timestamp data 64-bit) and
he noted that ReiserFS's object oriented "item handler" architecture
makes it fast and safe to update most old filesystem structures to new
formats on the fly (this includes just about everything from basic
stuff like stat data to the actual index and directory structures.)
Frankly, I'd like to see ext2/ext3 do *any* of these things now or in
the immediate future. There are a number of incompatible ext2 versions
in the works for adding B-trees, ACL storage, etc. to ext2, but
unlike with ReiserFS, all of these would break the on-disk format (not
to mention that ReiserFS *already* has very mature implementations
and/or superior alternatives to these features already.)
Maybe the best course of action would be to get the ReiserFS team to
commit to some specific release schedule for incompatible disk format
changes (i.e., those that can't be upgraded dynamically as Chris
claims they can.) As I understand it, the disk structure in 3.6.8 has
been made flexible to allow for a transparent upgrade/downgrade
between 3.6.x and 4.0. To me, this is a much more specific plan for
the future than ext2/ext3 (or XFS/JFS for that matter) have been able
to offer.
Given this, I have to agree that ReiserFS is ready for prime time
(*far* more than ext3 is at this point), possibly as the "preferred"
replacement for ext2 (as SuSE and Mandrake (?) have already stated.)
Having read this whole debate, we should probably try to put the pride
factor aside (i.e. the up-and-coming flame war between the developers
of ReiserFS and ext2/ext3) and just focus on preparing the best
candidate (on both technical advantage and development progress
grounds) for inclusion in 2.4 or 2.5.
Just my 2 cents worth...
- Matt Yourst
-------------------------------------------------------------
Matt T. Yourst Massachusetts Institute of Technology
yourst@mit.edu 617.225.7690
513 French House - 476 Memorial Drive - Cambridge, MA 02136
-------------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:22 EST