Article 61038 of comp.sys.sun.admin:
Path: mri.com!peerfeed.ncal.verio.net!news.he.net!news.pagesat.net!news1.good.net!news.goodnet.com!not-for-mail
From: invalid@upt.org (Tom Jackiewicz)
Newsgroups: comp.sys.sun.admin
Subject: Re: Ethernet blocked by UltraSparc
Date: 20 Jan 1998 15:47:45 GMT
Organization: GoodNet
Lines: 30
Message-ID: <6a2gv1$avo$1@news.goodnet.com>
References: <69l5vg$on$1@gorm.zfn.uni-bremen.de> <34BE6272.3C0F@sarenet.es>
NNTP-Posting-Host: sideburns.upt.org
X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]
Xref: mri.com comp.sys.sun.admin:61038

: > We have a problem here with some PCs and a UltraSparc
: > connected to the same ethernet hub. When the Ultra transmits
: > large datafiles the PCs can't access the ethernet anymore.
: > 
: > Is this a known problem and is there a way to solve it?
: > For example slow down the transmission rate of the Ultra?
: 
: 	Humm. If it's a speed problem, try using an ethernet
: switch instead of a hub. As long as the Ultra and the PCs
: aren't talking to the same machine, the Ultra shouldn't disturb
: the PC communications. If both types of machines talk to the 
: same machine, you can try putting a fast ethernet card on
: the server, and a 10/100 Mbps switch. Of course, connect the
: Ultra to a 10 Mbps port, not a 100 Mbps one.

I've got similar issues when connecting my Sparc 20 to the network
(at 100mb full duplex) via a Cabletron switch. When the Network
Appliance file server gets really active (we have an F330/230 rated
for 1500 ops, but we sometimes get 2000-2200), all the interfaces
on the Sparc 20 just completely stall and I can't even ping the
machine.

Has anyone seen something like this before and have any ideas? 
Why would a high level of NFS traffic (appear to) cause things 
like that to happen?

--
TOM JACKIEWICZ	            'AIX does everything Unix does, only not as well'
Lead System Engineer,      I get mail everywhere. This one's from the office.
GoodNet                              http://www.good.net --- invalid@good.net