Date: Fri, 18 Oct 1996 10:17:19 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: jas@flyingfox.COM (Jim Shankland) Cc: hardware@freebsd.org Subject: Re: collision count reported by netstat -I? Message-ID: <199610181717.KAA27086@GndRsh.aac.dev.com> In-Reply-To: <199610181607.JAA04414@saguaro.flyingfox.com> from Jim Shankland at "Oct 18, 96 09:07:38 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> While investigating some network performance problems, I've noticed > that the collision count reported by netstat -I for SMC9332DST cards > running in 10 Mbit mode tends to be much higher than I would expect. > I do not believe I'm dealing with defective cards, as this happens > on multiple cards, and on networks that seem to be performing OK. > > As a specific example, a workstation with one of the SMC cards, > located on a moderately busy wire, reports about a 3.5% > collision rate (Opackets/Collisions). The principal router on > that wire (also a FreeBSD box), sporting a 3C509 card, reports a > collision rate of just under 1 in a *millino*. As it happens, > virtually all traffic to and from the workstation, plus quite a > bit more, passes through the router. > > Does anyone have any thoughts as to what might be the cause of this > huge difference? There appears to be a bug in the if_de.c driver that is reporting too many collisions. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation, Inc. Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610181717.KAA27086>