From owner-freebsd-current  Mon Nov 11 09:28:00 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id JAA27527
          for current-outgoing; Mon, 11 Nov 1996 09:28:00 -0800 (PST)
Received: from brasil.moneng.mei.com (brasil.moneng.mei.com [151.186.109.160])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA27514
          for <freebsd-current@freebsd.org>; Mon, 11 Nov 1996 09:27:57 -0800 (PST)
Received: (from jgreco@localhost) by brasil.moneng.mei.com (8.7.Beta.1/8.7.Beta.1) id LAA19552; Mon, 11 Nov 1996 11:24:39 -0600
From: Joe Greco <jgreco@brasil.moneng.mei.com>
Message-Id: <199611111724.LAA19552@brasil.moneng.mei.com>
Subject: Re: ufs is too slow?
To: danj@3skel.com (Dan Janowski)
Date: Mon, 11 Nov 1996 11:24:38 -0600 (CST)
Cc: bsdcur@shadows.aeon.net, jgreco@brasil.moneng.mei.com,
        freebsd-current@freebsd.org
In-Reply-To: <199611111704.MAA12463@fnur.3skel.com> from "Dan Janowski" at Nov 11, 96 12:04:35 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> I have seen that OpenBSD is doing something with lfs, but I
> am not sure what. It would be worth while to get lfs running
> for sure; if you ever wondered, a lot of that disk bandwith
> goes to filesystem overhead.

In the specific case of "lots of small articles being accessed all over
the place", yes...  in general I would have to disagree :-)

With the "-noatime" hack, you gain some of that bandwidth back by not
having to support all the crummy metadata updates that you probably
didn't care too much about anyways. 

> I once exchanged some e-mail with someone at BSDI and with
> Margo Seltzer, who was a principle for lfs. The apparent
> primary reason why lfs does not run here is that lfs does some
> wierd stuff with the ATT buffer code that is missing in
> 4.4-lite. I was not able to get a synopsis of what or how to
> get around it, but it didn't sound like lfs was broken, it's
> just missing some wheels.
> 
> Maybe we can all talk about it a little and figure out
> how hard it would be to get going. If we were running
> a kick-ass big/fast file system, FreeBSD would capture
> some more of the Int(er|tra)+Net market. In addition
> to which, the infamous 'make world' time would surely
> benefit.

While I agree with this analysis, I unfortunately lack the skills and
internals knowledge to even contemplate doing something like this...
otherwise I would be off writing "newsfs" in a corner somewhere.

... JG