From owner-freebsd-hackers Sun Aug 4 23: 9:35 2002 Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A432737B401 for ; Sun, 4 Aug 2002 23:09:16 -0700 (PDT) Received: from services.webwarrior.net (overlord-host99.dsl.visi.com [209.98.86.99]) by mx1.FreeBSD.org (Postfix) with ESMTP id CF3E343E70 for ; Sun, 4 Aug 2002 23:09:15 -0700 (PDT) (envelope-from friar_josh@webwarrior.net) Received: from twincat.vladsempire.net (12-218-27-215.client.mchsi.com [12.218.27.215]) by services.webwarrior.net (Postfix) with ESMTP id 45DB1838209; Mon, 5 Aug 2002 01:09:14 -0500 (CDT) Date: Wed, 28 Nov 2001 13:39:04 +0000 From: Josh Paetzel To: John Vinters Cc: hackers@FreeBSD.ORG Subject: Re: FreeBSD performing worse than Linux? Message-ID: <20011128133904.C550@twincat.vladsempire.net> References: <3C052E84.3CF81084@jbg.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <3C052E84.3CF81084@jbg.co.uk>; from johnv@jbg.co.uk on Wed, Nov 28, 2001 at 06:35:48PM +0000 Lines: 55 Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Wed, Nov 28, 2001 at 06:35:48PM +0000, John Vinters wrote: > > I've (reasonably) recently installed 4.3-Release on a system running > Samba and a few light telnet apps, and noticed similar performance > problems. > > The SMB sessions would randomly change speed, and telnet sessions would > suffer from occasional "hesitation" (this is on a Dual PIII-700 MHz > machine with 1 Gb of RAM, which is currently very lightly loaded). > > I managed to track the problem down to the duplex settings on both the > Ethernet cards (AT-2500 TX, Realtek 8139 based, AFAIK) and the 10/100 > Switch. Forcing both the cards and the switch to particular settings > cured the problem, and lead to a massive performance increase. > > FTP seems to be particularly badly affected by the constant collisions > (causing backoff). The problem can be tricky to find as the switch > wasn't perceptably showing collisions on the collision LED, but viewing > the switch stats showed a different story! > > I've noticed similar problems with Linux and certain cards (it was a > while ago). > > > John Vinters > > Well, I am seeing dismal ftp performance on my 4.x boxes. I have a network of 4 machines, three of which are running -STABLE from Nov 22. The other machine is running NetBSD 1.5.2 Release. One of the FreeBSD machines has a base 10 cards in it and has reasonable performace with ftp transfer rates around 1.1Megs/sec. The NetBSD machine is a sparcstation 10 with an onboard intel base 10 adapter, and it too sees reasonable ftp performance. The other two -STABLE boxes have 100tx cards in them. One is a Linksys LNE100TX, and the other is an intel Pro 10/100B/100+. The hub for this network is an 8 port SOHOware autosensing affair. Both of the 100 cards auto-negotiate to 100tx half-duplex. I can get appoximately 1.5Megs/sec out of them using ftp. I have tried swapping cables, swapping ports, and replacing the hub with a crossover cable and manually configuring the cards for either full or half duplex operation. None of these steps makes any difference at all. I can reliably duplicate my transfer speeds on a 600 meg file with a std. deviation of less than a half a second no matter what network configuration I use. My next step will be to try some different NICs, but I don't have anything here that is 100tx based to swap with. I have gotten proper transfer rates out of these machines in the past, but I don't remember if the network cards have changed since then. I rarely move large files around at all, and so only looked into this as a curiosity when seeing this thread. I also intend to try some NFS mounts out to see if this is a protocol issue or not. Josh To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message