From owner-freebsd-bugs Fri Dec 22 03:08:47 1995 Return-Path: owner-bugs Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id DAA19777 for bugs-outgoing; Fri, 22 Dec 1995 03:08:47 -0800 (PST) Received: from zibbi.mikom.csir.co.za (zibbi.mikom.csir.co.za [146.64.24.58]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id DAA19770 for ; Fri, 22 Dec 1995 03:08:35 -0800 (PST) Received: (from rbezuide@localhost) by zibbi.mikom.csir.co.za (8.6.12/8.6.9) id NAA27283 for bugs@freebsd.org; Fri, 22 Dec 1995 13:07:52 +0200 From: R Bezuidenhout Message-Id: <199512221107.NAA27283@zibbi.mikom.csir.co.za> Subject: Possible TCP/UDP routing kernel bug? To: bugs@freebsd.org Date: Fri, 22 Dec 1995 13:07:52 +0200 (SAT) X-Mailer: ELM [version 2.4 PL24 ME8b] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-bugs@freebsd.org Precedence: bulk Seasons Greetings ... I have the following senario and I don't know what is going wrong ?? All machines are running FreeBSD 2.1 and are connected with a 10MB ethernet (test setup) PC1 ---------- Router1 ----------------- Router2 ----- PC2 I am using netcat wich basically can do a TCP or UDP stream from one machine to the other. On PC1 start netcat: ./nc -v -v -l -p 6789 On PC2 start netcat: yes AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA | ./nc -v -v 6789 Router2 starts to degrade, almost immediately in such a fashion that even pressing ENTER on a console (all machines are running in console mode) does not return a prompt. I can't even get a load average, but I know it is more or less 1.5 The two routers are running the followin: gated, xntpd. I have tried using different motherboards (all machines are 486DX2-66 with 8 meg ram and 16 meg swap). There doesn't seem to be any hardware defects. If I unplug the network cable, the Router2 starts doing what every I might have typed, until I replug the ether net. As far as I know FreeBSD should be able to and have previously handled the load?? This is true for both UDP and TCP .... With an ftp "loop" of ftp'ing the kernel from PC1 to PC2 the samething happens, only that the kernel doesn't degrade that much .. systat shows almost NO idle time and all other processes go up in %CPU (this means the kernel is using up most of the CPU because the processes have a weighted CPU % ?) On Router2 +/- 850 iterrupts between ed0 and ed1, remember this is with ftp, when using netcat I CAN'T DO ANYTHING !!! Any suggestions ?? Thanx -- ######################################################################## # # # Reinier Bezuidenhout Company: Mikomtek CSIR, ZA # # # # Network Engineer - NetSec development team # # # # Current Projects: NetSec - Secure Platform firewall system # # http://www.mikom.csir.co.za # # # # E-mail: rbezuide@mikom.csir.co.za # # # ########################################################################