From peidran@u.washington.edu Mon Mar 18 11:07:23 2002 Received: from mailscan6.cac.washington.edu (mailscan6.cac.washington.edu [140.142.33.14]) by lists.u.washington.edu (8.12.1+UW01.12/8.12.1+UW02.01) with SMTP id g2IJ7LDN028076 for ; Mon, 18 Mar 2002 11:07:21 -0800 Received: FROM mxu1.u.washington.edu BY mailscan6.cac.washington.edu ; Mon Mar 18 11:07:14 2002 -0800 Received: from mxout3.cac.washington.edu (mxout3.cac.washington.edu [140.142.32.19]) by mxu1.u.washington.edu (8.12.1+UW01.12/8.12.1+UW02.01) with ESMTP id g2IJ7Dc6020767 for ; Mon, 18 Mar 2002 11:07:13 -0800 Received: from mailscan-out2.cac.washington.edu (mailscan-out2.cac.washington.edu [140.142.33.17]) by mxout3.cac.washington.edu (8.12.1+UW01.12/8.12.1+UW02.01) with SMTP id g2IJ7D8J005306 for ; Mon, 18 Mar 2002 11:07:13 -0800 Received: FROM mailhost1.u.washington.edu BY mailscan-out2.cac.washington.edu ; Mon Mar 18 11:07:13 2002 -0800 Received: from serv9.lly.org (serv9.ischool.washington.edu [128.208.100.209]) by mailhost1.u.washington.edu (8.12.1+UW01.12/8.12.1+UW02.01) with ESMTP id g2IJ7DgK022266 for ; Mon, 18 Mar 2002 11:07:13 -0800 Received: from peidran by serv9.lly.org with local (Exim 3.33 #1 (Debian)) id 16n2TQ-0002S8-00 for ; Mon, 18 Mar 2002 11:07:12 -0800 Date: Mon, 18 Mar 2002 11:07:12 -0800 From: Peter Abrahamsen To: linux@u.washington.edu Subject: Re: Network slowdown with Linux Message-ID: <20020318190712.GC8538@serv9.lly.org> Reply-To: peidran@u.washington.edu References: <20020318025814.GA8333@rene.lly.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-Operating-System: Linux serv9 2.4.8-ac12 Hey all, This is a general reply - hopefully I'll address everyone who responded. On Mon, Mar 18, 2002 at 12:47:05AM -0800, 'irrelevant' M. Rogers wrote: > > RX packets:837901 errors:7582 dropped:185 overruns:0 frame:11346 > > TX packets:1157911 errors:2 dropped:0 overruns:0 carrier:4 > > collisions:0 > > RX bytes:664521335 (633.7 MiB) TX bytes:1308545792 (1.2 GiB) These statistics are for an approximately 12-hour period. While uploading to the server, I saw the errors-count go up by about 5/sec, and the frame count by something fairly comparable. I will try plugging another device into this wall port, and moving my host to a different port. The computer in question is on a static IP; I can try pump'ing an address, but I really don't expect that to make a difference. > Also, is your NIC set to auto-sense? Nearly all the switches on campus > are set to auto-negotiate. If you force it be 100 full duplex, the switch > port may (when the auto-negotiation fails) drop the connection down to > half duplex (can't recall if it's 10 half or 100 half). This will cause a > lot of errors because your machine will blithely babble on when collisions > happen. I'm not sure how to set that. Neither ifconfig nor dmesg mention anything; couldn't find anything in /proc. I'm using the Nov 6 tulip driver with a Lite-On PNIC-II. > BTW, does your routing table also have loopback and host entries > (something like this): No, if memory serves, the linux kernel started assuming loopback routing. I can ping localhost just fine. There's also a chance that this is some hardware IO/interrupt bullshit. I know my video card often doesn't play nice in linux, sometimes I get spurious interrupt messages, and sometimes it forgets about my keyboard for a few seconds. Thanks all for their help; I'll go do some more research. Peter -- Peter Abrahamsen University of Washington, Seattle, WA, USA Key 0x039922C0 : 259E 19C4 6FB4 1CA2 AC9D 75CE 8B5F 993D 0399 22C0 We are lost where the mind can't find us - Utterly lost. - Ikkyu .