Subject: Re: LFS (was Thank you NetBSD)
To: None <,>
From: Jochen Kunz <>
List: netbsd-users
Date: 02/23/2005 23:56:04
On Wed, 23 Feb 2005 14:26:03 -0500
Thor Lancelot Simon <> wrote:

> > I have to admit that I have nearly no knowledge on file system
> > internas.
> [...]
> > NetBSD. AdvFS has _really_ advanced concepts and features compared
> > with any other FS knowen to me. At least somthing like XFS would be
> > nice.
> I'm sorry, I see more than a slight contradiction here.
At first sight this is a contradiction. You can understand it like:
I have no knowledge of the internas, I can only judge from a user point
of veiw. From this perspective we have a never ending "LFS is unusable"
story while other *ix (like) OS have solved this problem years ago. And
some of them solved it quite well. (OK, we have softupdates for FFS.)

Still, would LFS be able to do what AdvFS can? (snapshots, extending and
shrinking of filesystems, migration of file systems from one disk to an
other, multiple file systems per partition, one file system spread over
multiple partitions / disks, ... All this can, _must_ be done while the
file system is _mounted_read/write_.)

> The basic problem with the BSD LFS is that the quality of the
> implementation is so low that it's almost not usable.  That says
> precisely nothing about the utility of LFS as a general structure for
> a filesystem;
Aha! Thanks for this clarification. So the implementation is the
problem, not the basic design principle. This gives a complete new
aspect to this discussion! (For me at least.)

Well in this case NetBSD should follow its "It doesn't work unless it's
right" devise. I.e. drain the old LFS code and reimplement it from
scratch. Much like the transition to UVM. Sure, this is no week end

\begin{off topic}
> indeed, the Sprite papers are as persuasive now as they
> were a decade ago, and those folks had _real_ results from code that
> actually _worked_ (I know, I used it).
AHA! Someone who actually run Sprite, The Real Thing. :-)
I am looking for a project to exhibit at the anual Vintage Computing
Vestival Europa. I have been fascinated by Sprite for quite some time.
So the idea arose to build a Sprite cluster with some of the several
sun4c and PMAX machines I own. The old Sprite code is available for
download but the problem is how to get it on the machines. It would nice
if you could provide some hints on where and how to start.
\end{off topic}

> If you want to see an example of an LFS-like filesystem that's been
> wildly successful in the real world, WAFL is a pretty good one.
OK, I'll ask google for pointers to WAFL.