From owner-freebsd-current  Wed Nov 10 10: 9: 3 1999
Delivered-To: freebsd-current@freebsd.org
Received: from nuke.danger.ms (a204b210n122client59.hawaii.rr.com [204.210.122.59])
	by hub.freebsd.org (Postfix) with ESMTP id 3B74D1525E
	for <freebsd-current@freebsd.org>; Wed, 10 Nov 1999 10:08:58 -0800 (PST)
	(envelope-from billm@danger.ms)
Received: by nuke.danger.ms (Postfix, from userid 1010)
	id CAE7828B1C; Wed, 10 Nov 1999 08:09:26 -1000 (HST)
Received: from localhost (localhost [127.0.0.1])
	by nuke.danger.ms (Postfix) with ESMTP
	id C5BEE24D18; Wed, 10 Nov 1999 08:09:26 -1000 (HST)
Date: Wed, 10 Nov 1999 08:09:26 -1000 (HST)
From: Bill Marquette <billm@danger.ms>
To: Doug Ambrisko <ambrisko@whistle.com>
Cc: freebsd-current@freebsd.org
Subject: Re: No buffer space available errors
In-Reply-To: <199911101725.JAA68371@whistle.com>
Message-ID: <Pine.BSF.4.10.9911100801110.3899-100000@nuke.danger.ms>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

On Wed, 10 Nov 1999, Doug Ambrisko wrote:
> On one out of 8 machines, I ran into this problem.  My network is running
> at 100BaseTX.  I noticed that ifconfig showed OACTIVE flag set and I
> was running in autosense mode.  So I setup the media to 100BaseTX and now
> it works okay.
> 
> My guess is the autosense gets confused sometimes.

Doug, thanks for the tip, I'll give that a try as I haven't been forcing
either card to 10Mbit or TP connections.  I have added to the mystery
however and this is aggravatingly puzzling.

In my previous message I had mentioned that walking away fromt he console
for any length of time would cause the machine to generate the "No buffer
space" errors.  Since then I've turned off my screensaver (Matrix from
KDE) and the system has stabilized completely; well almost completely.  I
can still duplicate the buffer situation and I've still seen it occur when
doing fairly high bandwidth (for cable, not for ethernet) file transfers.
If I can manage to bring the xfer rate up to around 100K/sec total between
inbound and outbound I experience the "Out of buffer" messages within a
couple minutes.  netstat -m doesn't show anything unusual.  I am going to
try removing both NMBCLUSTERS and NBUF from my config and just use a
maxusers of 128 and see how that goes.

--Bill




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message