From owner-freebsd-cluster@FreeBSD.ORG  Wed Jan 12 09:43:30 2005
Return-Path: <owner-freebsd-cluster@FreeBSD.ORG>
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 5799E16A4CE; Wed, 12 Jan 2005 09:43:30 +0000 (GMT)
Received: from mccinet.ru (relay.cell.ru [212.119.96.41])
	by mx1.FreeBSD.org (Postfix) with ESMTP
	id EFBB143D3F; Wed, 12 Jan 2005 09:43:28 +0000 (GMT)
	(envelope-from dolgop@mccinet.ru)
Received: from [212.1.235.150] (HELO server.dep624)
  by mccinet.ru (CommuniGate Pro SMTP 4.2.7)
  with ESMTP-TLS id 15308640; Wed, 12 Jan 2005 12:43:27 +0300
From: Evgeny Dolgopiat <dolgop@mccinet.ru>
To: Maxim Sobolev <sobomax@portaone.com>
Date: Wed, 12 Jan 2005 12:43:56 +0300
User-Agent: KMail/1.5.4
MIME-Version: 1.0
Content-Type: text/plain;
  charset="us-ascii"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200501121243.56305.dolgop@mccinet.ru>
cc: freebsd-net@freebsd.org
cc: "Rogier R.Mulhuijzen" <drwilco@drwilco.net>
cc: freebsd-cluster@freebsd.org
Subject: New failure detection algorithm for ng_one2many.
X-BeenThere: freebsd-cluster@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
Reply-To: evg_dolgop@mail.ru
List-Id: Clustering FreeBSD <freebsd-cluster.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-cluster>,
	<mailto:freebsd-cluster-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-cluster>
List-Post: <mailto:freebsd-cluster@freebsd.org>
List-Help: <mailto:freebsd-cluster-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-cluster>,
	<mailto:freebsd-cluster-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Wed, 12 Jan 2005 09:43:30 -0000

I wrote new failure detection algorithm based on heartbeat signal for
ng_one2many node. Features:

- automatic detection of failures;
- automatic detection of recoveries;
- detection of point of failure (see diagnostics in man page);
- configurable timing parameters of failure and recovery detection;
- you can create your own heartbeat packet or use default;
- you can set your rules for detecting that incoming packet is hearbeat  
packet;
- heartbeat algorithm can be used for different network layers (not only 
ethernet layer).

Patches for src and man page at 
http://www.watson.org/~ilmar/download/ng_one2many.tgz
These patches for CURRENT, but you can compile patched files in 5.3.