From owner-freebsd-net Wed Dec 4 21:24:22 2002 Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 42F5C37B401 for ; Wed, 4 Dec 2002 21:24:21 -0800 (PST) Received: from tp.databus.com (p70-227.acedsl.com [66.114.70.227]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8CB6C43EC2 for ; Wed, 4 Dec 2002 21:24:20 -0800 (PST) (envelope-from barney@tp.databus.com) Received: from tp.databus.com (localhost.databus.com [127.0.0.1]) by tp.databus.com (8.12.6/8.12.6) with ESMTP id gB55OEMG011821; Thu, 5 Dec 2002 00:24:14 -0500 (EST) (envelope-from barney@tp.databus.com) Received: (from barney@localhost) by tp.databus.com (8.12.6/8.12.6/Submit) id gB55OEYY011820; Thu, 5 Dec 2002 00:24:14 -0500 (EST) (envelope-from barney) Date: Thu, 5 Dec 2002 00:24:14 -0500 From: Barney Wolff To: Don Bowman Cc: "'Chuck Swiger'" , "'freebsd-net@freebsd.org'" Subject: Re: SO_DONTROUTE, arp's, ipfw fwd, etc Message-ID: <20021205052414.GA11711@tp.databus.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4i X-Scanned-By: MIMEDefang 2.26 (www . roaringpenguin . com / mimedefang) Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Wed, Dec 04, 2002 at 05:33:50PM -0500, Don Bowman wrote: > [client] > | > -------------------------- > | Load Balancer | > -------------------------- > | | > | | > [Redirector1] [Redirector2] > \ / > \ / > --------------------- > | | > [BSD1] [BSD2] > | | > ----------------------------- > | | | | | | | | | | > Telnet servers(A) Telnet (B) > > Thanks for the input, keep it coming! Talk me down from this ledge :) Well, if you really don't want to jump, just break the switch between the redirectors and the BSDs into two crossover cables, so redir1 talks only to BSD1 and redir2 talks only to BSD2. You have equal immunity to single failures, no obscure failure modes, and no need for custom kernel stuff. I'd bet anything that actually observed availability would be better, not worse, because of the decreased chance for interesting bugs taking the whole complex down. -- Barney Wolff http://www.databus.com/bwresume.pdf I'm available by contract or FT, in the NYC metro area or via the 'Net. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message