Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 29 Mar 2002 16:56:10 +0100
From:      "Pelhrimovszky, Zsolt" <Pelhrimovszky.Zsolt@evosoft.hu>
To:        stable@FreeBSD.ORG, net@FreeBSD.ORG
Subject:   Mystic network behaviour
Message-ID:  <3CA48E9A.603B7040@evosoft.hu>

next in thread | raw e-mail | index | archive | help
Hi, 

in the past months we had a lot of network problems with our new FreeBSD
gateway. 
These were network slowdowns. Playing with sysctls (rfc1323, newreno,
syncookies, delayed_ack ..) didn't help. 
We upraded the system from 4.4 stable, up to 4.5 stable 2 weekly,
downgraded it to 4.5 Release p2, but the slowdown still remained. 

The gateway has two interfaces: tx0 and tx1. 
tx0 was connected to 10Mb hub (towards to our ISPs router with 512Kbit
leased line), tx1 to an 100MB switch (internal lan).
The throughput was usually so much, that it filled the 512k bandwith. 
There were enough mbufs and other nerwork resources, the system does
only gatewaying.  

The slowdown was caused by input errors on the tx0 interface (1-2 new
errors per second, output from netstat -ni). Reinitialising it with
ifconfig down and up stopped these errors for some time (from 2 minutes
up to 2-3 days). Setting manually the ifconfig parameters (media, and
mediaopts) also didn't help.

The final solution was, that we changed the 10Mb hub to an 100MB switch,
and now it works fine. 

Has anybody some ideas what was worng?
Is a defect somewere in the FreeBSD ip-stack, or network driver
implementation?  

Regards, 
	Zs. Pelhrimovszky

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?3CA48E9A.603B7040>