From owner-freebsd-questions  Tue Aug  8  7:40:35 2000
Delivered-To: freebsd-questions@freebsd.org
Received: from smtp13.bellglobal.com (smtp13.bellglobal.com [204.101.251.52])
	by hub.freebsd.org (Postfix) with ESMTP id AB71F37B687
	for <freebsd-questions@FreeBSD.ORG>; Tue,  8 Aug 2000 07:40:28 -0700 (PDT)
	(envelope-from admin@chemcomp.com)
Received: from hermes.chemcomp.com (ppp11297.qc.bellglobal.com [206.172.147.10])
	by smtp13.bellglobal.com (8.8.5/8.8.5) with ESMTP id KAA07847;
	Tue, 8 Aug 2000 10:44:48 -0400 (EDT)
Received: from chemcomp.com (sky.chemcomp.com [192.1.1.62])
	by hermes.chemcomp.com (Postfix) with ESMTP
	id 4FDA816824; Tue,  8 Aug 2000 10:37:45 -0400 (EDT)
Message-ID: <39901B38.52F27A68@chemcomp.com>
Date: Tue, 08 Aug 2000 10:37:44 -0400
From: System Administrator <admin@chemcomp.com>
Organization: Chemical Computing Group, Inc.
X-Mailer: Mozilla 4.72 [en] (X11; I; FreeBSD 4.0-RELEASE i386)
X-Accept-Language: fr-CA, fr, en
MIME-Version: 1.0
To: dg@root.com
Cc: freebsd-questions@FreeBSD.ORG
Subject: Re: Ethernet "long event" error frames with FBSD machines
References: <200008072123.OAA04143@implode.root.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-questions@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Man, that was it! The darn switch is not full-duplex!!!

You guys are the best,

thanks!

David Greenman wrote:
> 
>    Sounds like the duplex is set wrong. Try setting both sides for 10/full.
> 
> -DG
> 
> David Greenman
> Co-founder, The FreeBSD Project - http://www.freebsd.org
> Manufacturer of high-performance Internet servers - http://www.terasolutions.com
> Pave the road of life with opportunities.
> 
> >I have a few FreeBSD machines installed on our local network here, and
> >they seem to perform very badly when doing large TCP transfers (e.g.
> >FTP). All machines are connected on a 10baseT/UTP network using a switch
> >(3Com SuperStackII Desktop Switch). But tests using nttcp give me
> >readings that are quite bad. The maximum bandwidth I can reach using TCP
> >is 1Mb/s. UDP tests are performing a lot better, giving 9 to 10Mb/s
> >transfers.
> >
> >After investigation on the switch's console, I see that the 3 FBSD
> >machines that I have on the network generate "long events" error that
> >are described in 3com's manual as:
> >
> >Late Events: This counter is incremented by one
> >each time a collision occurs after the valid packet
> >minimum time. A late event is an out- of- window
> >collision that may occur if your Ethernet LAN
> >exceeds the maximum size as defined in the IEEE
> >standard. A late event is also counted as a collision.
> >
> >I seem to be missing something here. No other machine on our network
> >behave so badly. I would really like to have advice on this, because it
> >may be the turning point of the mere presence of FreeBSD on our network.
> >I cannot reproduce the behavior on any other machine that FreeBSD
> >machines.
> >
> >What follows is the examples of the tests I ran, snapshot of the
> >Switch's console and my current kernel config. I run
> >
> >FreeBSD sky.chemcomp.com 4.0-RELEASE FreeBSD 4.0-RELEASE #8: Mon Aug  7
> >12:25:44 EDT 2000     root@sky.chemcomp.com:/usr/src/sys/compile/SKY
> >i386
> >
> >on all machines.
> >
> >Thanks,
> >
> >A.

<snipped>


-- 
Antoine Beaupre
System Administrator
Chemical Computing Group, Inc.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message