From owner-freebsd-questions@FreeBSD.ORG Mon Jul 25 23:32:25 2005 Return-Path: X-Original-To: freebsd-questions@freebsd.org Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D73B916A41F for ; Mon, 25 Jul 2005 23:32:25 +0000 (GMT) (envelope-from djh@nebcorp.com) Received: from ratchet.nebcorp.com (ratchet.nebcorp.com [205.217.153.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id B23F643D46 for ; Mon, 25 Jul 2005 23:32:25 +0000 (GMT) (envelope-from djh@nebcorp.com) Received: by ratchet.nebcorp.com (Postfix, from userid 1014) id 42CD6D9825; Mon, 25 Jul 2005 16:32:25 -0700 (PDT) Date: Mon, 25 Jul 2005 16:32:25 -0700 From: Danny Howard To: freebsd-questions@freebsd.org Message-ID: <20050725233225.GA24353@ratchet.nebcorp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.1i X-Loop: djhoward@uiuc.edu Subject: CARP Load-Balancing? (Or, HTTP peer load balancing?) X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 25 Jul 2005 23:32:26 -0000 Hello, I am looking at building a tiered web application (web layer, DB layer) and am wondering about configuring automatic failover for the web layer. If possible, I'd like to avoid having to implement a third firewall / balancing tier, but if I must, I shall do that ... I am looking at CARP. The documentation seems soarse, and it sounds like the "load balancing" option isn't much of an option. On the other hand, I could set multiple vhids, one per web server, and rig it up so that if one web server drops its interface, another will answer for it. Then "load balancing" is as simple as putting multiple web server IPs in for the A record ... (With CARP basically functioning as a "heartbeat" monitor ...) Does that sound sensible? I'm feeling a bit wary of this approach. Any suggestions for handling failover of HTTP among peers? Thanks, -danny