From owner-freebsd-isp  Sun Jan 20 21:26:42 2002
Delivered-To: freebsd-isp@freebsd.org
Received: from misery.sdf.com (misery.sdf.com [207.200.153.226])
	by hub.freebsd.org (Postfix) with ESMTP id A450D37B417
	for <freebsd-isp@freebsd.org>; Sun, 20 Jan 2002 21:26:37 -0800 (PST)
Received: from tom (helo=localhost)
	by misery.sdf.com with local-esmtp (Exim 2.12 #1)
	id 16SWMf-00059b-00; Sun, 20 Jan 2002 20:47:25 -0800
Date: Sun, 20 Jan 2002 20:47:23 -0800 (PST)
From: Tom Samplonius <tom@sdf.com>
To: Emre Bastuz <info@emre.de>
Cc: freebsd-isp <freebsd-isp@FreeBSD.ORG>
Subject: Re: Solaris vs. FreeBSD in High Traffic Environments
In-Reply-To: <3C227F39.5090702@emre.de>
Message-ID: <Pine.BSF.4.05.10201202040080.9032-100000@misery.sdf.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=X-UNKNOWN
Content-Transfer-Encoding: QUOTED-PRINTABLE
Sender: owner-freebsd-isp@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-isp.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-isp>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-isp>
X-Loop: FreeBSD.org


On Fri, 21 Dec 2001, Emre Bastuz wrote:

> Hi Tom,
>=20
> >   I think it is important to define what "high traffic" is.  10Mbps
> > sustained?  30Mbps?  Percentage of dynamic content?
>=20
> high traffic means about 300 GB/month worth of static content in this cas=
e.

  300GB/mo isn't too bad.  I hosted a site temporarily that did about
800GB in two weeks.  The client didn't want to pay anymore than $75/mo, so
most of the content was removed.

  I would say any server doing more than a 1000GB a month is "high
traffic".

> >>Themachine got offline twice within 48 hours (the Solaris box had never
> >>crashed), that=B4s why some tweeking on the kernel and other parameters
> >>was done. Since then the server is performing great.
> >   How did it crash?
=2E..
> I=B4m still curious what caused the outage though ...

  It was probably an mbuf shortage.  Probably.  There is also a chance it
is something unrelated to tuning, and something is wrong with the server.
You should check "netstat -m".  The peak value should be less than the max
value.

  You shouldn't need to go to 65,000 mbuf clusters for a 300GB/mo site.
You might be wasting a lot of kernel memory.

Tom


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