From owner-freebsd-cluster@FreeBSD.ORG Sun Apr 20 17:22:39 2003 Return-Path: Delivered-To: freebsd-cluster@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 55C6E37B401 for ; Sun, 20 Apr 2003 17:22:39 -0700 (PDT) Received: from walnut.he.net (walnut.he.net [64.71.137.114]) by mx1.FreeBSD.org (Postfix) with ESMTP id F3EEF43FA3 for ; Sun, 20 Apr 2003 17:22:38 -0700 (PDT) (envelope-from kmacy@fsmware.com) Received: from localhost (kmacy@localhost) by walnut.he.net (8.8.6p2003-03-31/8.8.2) with ESMTP id RAA04985 for ; Sun, 20 Apr 2003 17:22:44 -0700 Date: Sun, 20 Apr 2003 17:22:44 -0700 (PDT) From: Kip Macy X-Sender: kmacy@walnut.he.net To: freebsd-cluster@freebsd.org Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: iSCSI and clustering with FreeBSD X-BeenThere: freebsd-cluster@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Clustering FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 21 Apr 2003 00:22:39 -0000 I'm currently setting up a small cluster of 8 dual processor machines. I will be running opengfs over iSCSI. For the time being, this translates to meaning that everything but the iSCSI target will be running Linux. I see a number of things that FreeBSD really needs to be on equal footing that I would like to see get going and would like to contribute towards, but realize that I don't have the time to do it myself and still pay my mortgage. I'm wondering if people are interested in getting together, or are simply too busy to do anything but bemoan their absence. 1) Shared storage: iscsi initiator A while back, I signed up to write an iscsi initiator. After having written an iscsi target, I now realize that I don't have the time to write an initiator by myself (at least not unpaid). Time required for development: 3 man months 2) Shared file system: OpenGFS or LUSTRE OpenGFS is a shared disk file system for Linux that was initially written for IRIX but was later ported to Linux. Hence, it has a reasonably clean separation from OS dependencies. Porting it too FreeBSD looks feasible. It would have to remain a module as it is GPL. LUSTRE looks like it scales better but requires more infrastructure. Time required for port: ~6 man months 3) Process management: bproc I'm sure everyone on this list is familiar with it. Time required for port: 3-9 man months - the kernel patch is fairly small, but I'm not certain how many Linux dependencies the system has. 5) Process migration: openMosix This is one colossal 45,000 line patch for Linux. It is obviously GPL and could not be put in a module, so it would have to be done from scratch. Time required for port: 2 man years??? On question that I ask myself, even as someone who strongly prefers FreeBSD to Linux, is "why does it matter if clustering is done on Linux or on FreeBSD?". Linux has a huge momentum advantage in this area. The only thing that FreeBSD obviously adds is that companies could add their own mods without being forced to open-source them. -Kip