Date: Thu, 13 Dec 2001 10:54:51 +0100 From: Thomas Zenker <thz@Lennartz-electronic.de> To: stable@freebsd.org Cc: jlemon@flugsvamp.com, dillon@apollo.backplane.com Subject: TCP stack still hosed? Message-ID: <20011213105451.A738@mezcal.tue.le>
next in thread | raw e-mail | index | archive | help
Hi, after making new kernels I have noticed a dramatical performance drop in the tcp stack. The last normally working kernel I have is from 20th november. The sources are up to date, our archives are updated nightly. I am using a scripted sysinstall for doing installations/updates on our embedded equipment. Testing the installation on a normal PC I can see the numbers on the speedometer :-) "host-242" is installed, the release is fetched from "mezcal". Both are connected to a switch 10/100TX, mezcal has a fxp running with 100baseTX, host-242 has a NetGear USB/ethernet adapter (10TX). Here are the numbers: 1. server new, client new: 10-15 Kb/s mezcal: FreeBSD 4.4-STABLE #23: Thu Dec 13 08:48:40 CET 2001 host-242: FreeBSD 4.4-STABLE #0: Wed Dec 12 13:37:50 GMT 2001 2. server old, client new: 140-160 Kb/s mezcal: FreeBSD 4.4-STABLE #21: Tue Nov 20 14:17:19 CET 2001 host-242: FreeBSD 4.4-STABLE #0: Wed Dec 12 13:37:50 GMT 2001 3. server old, client old: 225-245 Kb/s mezcal: FreeBSD 4.4-STABLE #21: Tue Nov 20 14:17:19 CET 2001 host-242: FreeBSD 4.3-STABLE #0: Tue Jul 17 10:58:55 GMT 2001 I have tcpdumps for all three cases. Strange are sometimes lots of duplicate acks. Who is interrested in the tcpdumps? Also I have tried to switch off newreno and delayed_ack, but with no visible effect. Thomas -- Thomas Zenker c/o Lennartz electronic GmbH Bismarckstrasse 136, D-72072 Tuebingen, Germany Phone: +49-(0)7071-93550 Email: thz@lennartz-electronic.de To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011213105451.A738>