From owner-freebsd-net Wed Jan 3 6:39: 0 2001 From owner-freebsd-net@FreeBSD.ORG Wed Jan 3 06:38:56 2001 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from iguana.aciri.org (iguana.aciri.org [192.150.187.36]) by hub.freebsd.org (Postfix) with ESMTP id CA9B837B400 for ; Wed, 3 Jan 2001 06:38:55 -0800 (PST) Received: (from rizzo@localhost) by iguana.aciri.org (8.11.1/8.11.1) id f03Eckw80927; Wed, 3 Jan 2001 06:38:46 -0800 (PST) (envelope-from rizzo) From: Luigi Rizzo Message-Id: <200101031438.f03Eckw80927@iguana.aciri.org> Subject: Re: Dummynet problem In-Reply-To: <87vgrx3d0d.fsf@pf39.warszawa.sdi.tpnet.pl> from Slawek Zak at "Jan 2, 2001 9:38:58 pm" To: zaks@prioris.mini.pw.edu.pl (Slawek Zak) Date: Wed, 3 Jan 2001 06:38:46 -0800 (PST) Cc: freebsd-net@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL43 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: rizzo@iguana.aciri.org Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi, i have just run a test locally (on a 4.2R system, queues with weight 1 and 10, transfers on different tcp port) and the results are exactly what one would expect -- one flow gets 10 times the bw of the other one. So i believe you have done some mistake in your config or your measurement (e.g. some other bottleneck in the net limiting one flow to 60Kbit, leaving a full 60k to the other no matter how weight are assigned). Note that running this kind of experiments requires a bit of care -- with a 10:1 speed ratio, one of the transfer might complete much faster than the other leaving full bw to the the other flow for 90% of the time, which in the end causes both flow to show approx the same speed. cheers luigi > > it should not be equal provided the 'high weight' flow has sufficient > > traffic going. > > Both FTP transfers I've used for testing were around 60Kbps each. One done by > user dnld1, the other one by other user. > > > Can you do an 'ipfw zero' before the transfer, and provide the output of > > Sure. I've `ipfw zero'ed after both transfers were started. > > > ipfw show > > 00100 0 0 allow ip from any to any via lo0 > 00100 226 327495 queue 10 tcp from any to any uid dnld1 in > 00200 0 0 deny ip from any to 127.0.0.0/8 > 00200 677 338406 queue 11 ip from any to any > 65535 0 0 allow ip from any to any > > > ipfw queue show > > 00001: 128.000 Kbit/s 0 ms 50 sl. 0 queues (1 buckets) droptail > mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000 > q00010: weight 1 pipe 1 50 sl. 1 queues (1 buckets) droptail > mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000 > BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp > 0 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12088 488 693234 0 0 0 > q00011: weight 10 pipe 1 50 sl. 168 queues (64 buckets) droptail > mask: 0xff 0xffffffff/0xffff -> 0xffffffff/0xffff > BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp > 0 tcp AAA.BBB.CCC.DDD/22 XXX.YYY.ZZZ.QQQ/1022 1 44 0 0 0 > 4 tcp XXX.YYY.ZZZ.QQQ/12092 AAA.BBB.CCC.DDD/49186 318 12724 0 0 0 > 5 tcp XXX.YYY.ZZZ.QQQ/12091 AAA.BBB.CCC.DDD/49185 463 18524 0 0 0 > 5 tcp XXX.YYY.ZZZ.QQQ/12089 AAA.BBB.CCC.DDD/49184 5 204 0 0 0 > 15 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12088 1 40 0 0 0 > 23 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12084 30 2153 0 0 0 > 25 tcp AAA.BBB.CCC.DDD/49183 XXX.YYY.ZZZ.QQQ/12086 207 306124 0 0 0 > 30 tcp AAA.BBB.CCC.DDD/49182 XXX.YYY.ZZZ.QQQ/12085 4 1455 0 0 0 > 34 tcp XXX.YYY.ZZZ.QQQ/12084 AAA.BBB.CCC.DDD/21 34 1828 0 0 0 > 46 tcp XXX.YYY.ZZZ.QQQ/12088 AAA.BBB.CCC.DDD/21 24 1220 0 0 0 > 46 tcp XXX.YYY.ZZZ.QQQ/1022 AAA.BBB.CCC.DDD/22 2 84 0 0 0 > 48 tcp AAA.BBB.CCC.DDD/49186 XXX.YYY.ZZZ.QQQ/12092 317 469668 0 0 0 > 52 tcp XXX.YYY.ZZZ.QQQ/12086 AAA.BBB.CCC.DDD/49183 207 8284 0 0 0 > 53 tcp XXX.YYY.ZZZ.QQQ/12085 AAA.BBB.CCC.DDD/49182 5 204 0 0 0 > > > ipfw pipe show > > 00001: 128.000 Kbit/s 0 ms 50 sl. 0 queues (1 buckets) droptail > mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000 > q00010: weight 1 pipe 1 50 sl. 1 queues (1 buckets) droptail > mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000 > BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp > 0 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12088 488 693234 0 0 0 > q00011: weight 10 pipe 1 50 sl. 168 queues (64 buckets) droptail > mask: 0xff 0xffffffff/0xffff -> 0xffffffff/0xffff > BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp > 0 tcp AAA.BBB.CCC.DDD/22 XXX.YYY.ZZZ.QQQ/1022 1 44 0 0 0 > 4 tcp XXX.YYY.ZZZ.QQQ/12092 AAA.BBB.CCC.DDD/49186 318 12724 0 0 0 > 5 tcp XXX.YYY.ZZZ.QQQ/12091 AAA.BBB.CCC.DDD/49185 463 18524 0 0 0 > 5 tcp XXX.YYY.ZZZ.QQQ/12089 AAA.BBB.CCC.DDD/49184 5 204 0 0 0 > 15 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12088 1 40 0 0 0 > 23 tcp AAA.BBB.CCC.DDD/21 XXX.YYY.ZZZ.QQQ/12084 30 2153 0 0 0 > 25 tcp AAA.BBB.CCC.DDD/49183 XXX.YYY.ZZZ.QQQ/12086 207 306124 0 0 0 > 30 tcp AAA.BBB.CCC.DDD/49182 XXX.YYY.ZZZ.QQQ/12085 4 1455 0 0 0 > 34 tcp XXX.YYY.ZZZ.QQQ/12084 AAA.BBB.CCC.DDD/21 34 1828 0 0 0 > 46 tcp XXX.YYY.ZZZ.QQQ/12088 AAA.BBB.CCC.DDD/21 24 1220 0 0 0 > 46 tcp XXX.YYY.ZZZ.QQQ/1022 AAA.BBB.CCC.DDD/22 2 84 0 0 0 > 48 tcp AAA.BBB.CCC.DDD/49186 XXX.YYY.ZZZ.QQQ/12092 317 469668 0 0 0 > 52 tcp XXX.YYY.ZZZ.QQQ/12086 AAA.BBB.CCC.DDD/49183 207 8284 0 0 0 > 53 tcp XXX.YYY.ZZZ.QQQ/12085 AAA.BBB.CCC.DDD/49182 5 204 0 0 0 > > Irrelevant udp/icmp traffic was snipped, IP's were masked to protect the > innocent ;) `ipfw pipe show' and `ipfw queue show' look both very similar - hmm. > > Best regards, /S > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-net" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message