From owner-freebsd-arch Mon Aug 7 9:34:18 2000 Delivered-To: freebsd-arch@freebsd.org Received: from mercury.Sun.COM (mercury.Sun.COM [192.9.25.1]) by hub.freebsd.org (Postfix) with ESMTP id 692A137B9A6 for ; Mon, 7 Aug 2000 09:34:16 -0700 (PDT) (envelope-from Jean-Marc.Fenart@france.sun.com) Received: from sunfra.France.Sun.COM ([129.157.188.1]) by mercury.Sun.COM (8.9.3+Sun/8.9.3) with ESMTP id JAA07336 for ; Mon, 7 Aug 2000 09:34:13 -0700 (PDT) Received: from sunchorus.France.Sun.COM (sunchorus [129.157.173.1]) by sunfra.France.Sun.COM (8.8.8+Sun/8.8.8/ENSMAIL,v1.7) with ESMTP id SAA23740 for ; Mon, 7 Aug 2000 18:34:11 +0200 (MET DST) Received: from bagana (bagana [129.157.173.76]) by sunchorus.France.Sun.COM (8.8.8+Sun/8.8.8) with SMTP id SAA23960; Mon, 7 Aug 2000 18:34:10 +0200 (MET DST) Message-Id: <200008071634.SAA23960@sunchorus.France.Sun.COM> Date: Mon, 7 Aug 2000 18:34:10 +0200 (MET DST) From: Jean-Marc Fenart Reply-To: Jean-Marc Fenart Subject: HA-NFS To: freebsd-arch@FreeBSD.org Cc: Jean-Marc.Fenart@france.sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: pf/ud41y1hI+ONJx3MeNbg== X-Mailer: dtmail 1.3.0 @(#)CDE Version 1.3.2 SunOS 5.7 sun4u sparc Sender: owner-freebsd-arch@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Hello I'm looking for an implementation of the HA-NFS (Higly Available Network File Server) on top of a BSD Unix. This is an extension of the standard NFS server implementation which consists in providing a continuation of service for an NFS client, in case of a failure on the NFS server. This is done by using a secondary/backup NFS server which is able in case of a failure on the primary server to (1) takeover the IP address of the failed primary, (2) re-mount the filesystem accessed by the failed primary (to do this requires an underlying disk mirroring provided by instance by a volume manager), (3) update its cache of non-idempotent operations with the state of the primary at the time of the failure (the cache of non idempotent ops can be on a shared disks/logged file system or msg checkpointed between the primary and the secondary...). This is totaly transparent for an NFS client, the NFS protocol on the wire being unchanged and all modifications limited to the servers. The original paper for it is a Usenix-91 (HA-NFS Anuman Bhide and all). Does someone has a knowledge of such an implementation on top of BSD Unix ? Thank's for your help - JM Fenart - To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message