From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263483AbUDBBPP (ORCPT ); Thu, 1 Apr 2004 20:15:15 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263479AbUDBBPP (ORCPT ); Thu, 1 Apr 2004 20:15:15 -0500 Received: from dial249.pm3abing3.abingdonpm.naxs.com ([216.98.75.249]:43656 "EHLO animx.eu.org") by vger.kernel.org with ESMTP id S263505AbUDBBOr (ORCPT ); Thu, 1 Apr 2004 20:14:47 -0500 Date: Thu, 1 Apr 2004 20:25:00 -0500 From: Wakko Warner To: linux-kernel@vger.kernel.org Subject: Slow network. Message-ID: <20040401202500.A16167@animx.eu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.3i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm looking for some pointers as to why this is happening. I have a machine with 2.4.25/2.6.4 (boots over network) that serves 2 hard disks to my network. It has a microstar ms-6163 systemboard (I think that's the right number), a 160gb maxtor as hda. hda contains an ext3 filesystem. Nic is a 3c905b with MBA. 256mb memory. 700mhz (100fsb) pIII. Reading a 2.9gb file off hda is took 1:51 minutes (~26mb/sec). When I do this over nfs, it takes over 8 minutes (~5.7mb/sec). I used to get around 11mb/sec transfer from this machine. The changes I made was a different board/cpu/memory/kernel version and used an 80wire ide cable for hda. The nic is the same, hda is the same. The old machine was a celeron 333a, 40wire ide cable, 192mb ram. I don't know what the board model was. With the old board, I saw about 32mb/sec on the same file (local) and 11mb/sec (nfs). The kernel was 2.4.20. Does anyone have any ideas? I looked at ifconfig, vmstat, /proc/interrupts but nothing looks out of the ordinary. I almost want to just downgrade the machine, but i like the microstar board better (faster cpu). -- Lab tests show that use of micro$oft causes cancer in lab animals