From owner-freebsd-fs@FreeBSD.ORG  Tue Jul 25 15:08:33 2006
Return-Path: <owner-freebsd-fs@FreeBSD.ORG>
X-Original-To: fs@freebsd.org
Delivered-To: freebsd-fs@FreeBSD.ORG
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 94CC916A4DE
	for <fs@freebsd.org>; Tue, 25 Jul 2006 15:08:33 +0000 (UTC)
	(envelope-from rick@snowhite.cis.uoguelph.ca)
Received: from ccshst09.cs.uoguelph.ca (ccshst09.cs.uoguelph.ca
	[131.104.94.206])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 2EDE243D46
	for <fs@freebsd.org>; Tue, 25 Jul 2006 15:08:30 +0000 (GMT)
	(envelope-from rick@snowhite.cis.uoguelph.ca)
Received: from snowhite.cis.uoguelph.ca (snowhite.cis.uoguelph.ca
	[131.104.48.1])
	by ccshst09.cs.uoguelph.ca (8.13.1/8.13.1) with ESMTP id k6PF8TZK030163;
	Tue, 25 Jul 2006 11:08:29 -0400
Received: (from rick@localhost)
	by snowhite.cis.uoguelph.ca (8.9.3/8.9.3) id LAA73358;
	Tue, 25 Jul 2006 11:09:41 -0400 (EDT)
Date: Tue, 25 Jul 2006 11:09:41 -0400 (EDT)
From: rick@snowhite.cis.uoguelph.ca
Message-Id: <200607251509.LAA73358@snowhite.cis.uoguelph.ca>
To: fs@freebsd.org
X-Scanned-By: MIMEDefang 2.52 on 131.104.94.206
Cc: peadar.edwards@gmail.com
Subject: Re: freebsd4.11 patch for nfs over tcp
X-BeenThere: freebsd-fs@freebsd.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Filesystems <freebsd-fs.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-fs>,
	<mailto:freebsd-fs-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-fs>
List-Post: <mailto:freebsd-fs@freebsd.org>
List-Help: <mailto:freebsd-fs-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-fs>,
	<mailto:freebsd-fs-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Tue, 25 Jul 2006 15:08:33 -0000

> I always point people at "Rethinking the TCP Nagle Algorithm" (Mogul &
> Minshall) when this comes up. It's a nice readable paper on the issues
> and some propsed solutions.

And, yes, it does cause performance problems for NFS. If an RPC is in the
send queue, it delays sending the request. For the case I was testing
that normally takes 5 seconds, it took about 3 minutes. You could
watch the server's disk led "blink <wait a bit>, blink...".

As an historical note, it was Tom Talpey who suggested turning off the
Nagle algorithm to me in 1989, when I was observing slow NFS over TCP on
a MicroVAXII (now, that means REALLY SLOOWWWW).

rick