From owner-freebsd-isp Fri Feb 20 08:32:36 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA18070 for freebsd-isp-outgoing; Fri, 20 Feb 1998 08:32:36 -0800 (PST) (envelope-from owner-freebsd-isp@FreeBSD.ORG) Received: from s02.admin.cantv.net (s02.admin.cantv.net [161.196.66.41]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA18032 for ; Fri, 20 Feb 1998 08:32:28 -0800 (PST) (envelope-from lem@cantv.net) Received: from lemtop.hanoi.com.ve ([161.196.235.251]) by s02.admin.cantv.net (8.8.8/8.8.8) with SMTP id MAA13432; Fri, 20 Feb 1998 12:29:18 -0400 (GMT-0400) Message-Id: <3.0.5.32.19980220122303.0094f100@pop.cantv.net> X-Sender: lem@pop.cantv.net X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.5 (32) Date: Fri, 20 Feb 1998 12:23:03 -0400 To: "Michael L. VanLoon -- HeadCandy.com" From: =?iso-8859-1?Q?=22Luis_E=2E_Mu=F1oz=22?= Subject: Re: fault tolerant :)) setup Cc: bag@sinbin.demos.su (Alex G. Bulushev), agdolla@datanet.hu (Gabor Dolla), freebsd-isp@FreeBSD.ORG In-Reply-To: <199802200850.AAA20546@MindBender.serv.net> References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 12:50 AM 20/02/1998 -0800, Michael L. VanLoon -- HeadCandy.com wrote: [snip] >There have to be hardware products out there, somewhere, that work >this way, because this is they way NT clustering server works. You >must have two SCSI controllers connected to one or more (SCSI) drives. >I believe both servers are allowed to work and do dual service while >they are well. The good server takes over when the bad one fails. At least in the Digital (DEC, Compaq, whatever it is today :) cluster servers, the servers can talk to each other via the Net and via the SCSI bus. One of the hardest problems is finding out if the other server is really dead. If they only talk via the network, simply unplugging one of the NICs will cause problems as both will try to go live simultaneously. Digital also has another solution (hairier) called a 'True Cluster'. Machines in a True Cluster configuration share memory via a special bus, so you have actually three ways to talk to the other servers. Also, work can be resumed *exactly* where it was left and in a 100% transparent transition, at least this is what I've read. >Don't know if they had to do any magic to NTFS to make this work, or >if the magic happens inside the clustering services on both machines. I *think* the FS has to help. -lem To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message