From owner-freebsd-questions@FreeBSD.ORG Mon Jul 14 16:08:44 2008 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 9178A1065678 for ; Mon, 14 Jul 2008 16:08:44 +0000 (UTC) (envelope-from mvolaski@aecom.yu.edu) Received: from mx1.aecom.yu.edu (mx1.aecom.yu.edu [129.98.1.51]) by mx1.freebsd.org (Postfix) with ESMTP id 7B7E88FC13 for ; Mon, 14 Jul 2008 16:08:44 +0000 (UTC) (envelope-from mvolaski@aecom.yu.edu) Received: from draco.aecom.yu.edu (draco.aecom.yu.edu [129.98.1.160]) by mx1.aecom.yu.edu (Postfix) with ESMTP id D59159F0098; Mon, 14 Jul 2008 12:08:43 -0400 (EDT) X-AuditID: 816201a0-a629cbb0000015ac-8f-487b7a0b35cd Received: from smtp2.aecom.yu.edu (smtp2.aecom.yu.edu [129.98.1.62]) by draco.aecom.yu.edu (Symantec Mail Security) with ESMTP id 79C4F718004; Mon, 14 Jul 2008 12:08:43 -0400 (EDT) Received: from [129.98.90.227] (usseinstein.aecom.yu.edu [129.98.90.227]) by smtp2.aecom.yu.edu (Postfix) with ESMTP id 33E98B714; Mon, 14 Jul 2008 12:08:43 -0400 (EDT) Mime-Version: 1.0 Message-Id: In-Reply-To: <226ae0c60807140804g467facdn37756a244e0b7de@mail.gmail.com> References: <226ae0c60807140804g467facdn37756a244e0b7de@mail.gmail.com> Date: Mon, 14 Jul 2008 12:08:15 -0400 To: "David Robillard" From: Maurice Volaski Content-Type: text/plain; charset="us-ascii" ; format="flowed" X-Brightmail-Tracker: AAAAAA== Cc: freebsd-questions@freebsd.org Subject: Re: Would ZFS and gmirror work well together in a two-node failover cluster? 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, 14 Jul 2008 16:08:44 -0000 >Which type of connection do you intend to use for the shared storage >JBOD? SAN or direct attached SCSI? Don't forget to change the SCSI On each server, there would be a SATA JBOD attached internally with a 3Ware or Areca card. >Which means your two node cluster won't be active/active. You'll have >an active node and a failover node. That may be alright or it may not. Right, there would be a primary server,which contains the mounted drives and the failover, where only the ggated and gmirror see the drives (until a primary failure occurs and presumably something like heartbeat would then mount these drives and take over as the new primary). >Now if we come back to the problem at hand, mainly using zfs under >gmirror. I've never heard of anyone using this. It does sound a bit >strange to me since both zfs and gmirror will do mirroring. I would gmirror would be doing mirroring, but ZFS would be doing RAIDZ2. So altogether it's RAIDZ2 + RAID 1. Under Linux, such an arrangement is fairly easy to do with an ordinary filesystem and the local RAID would be handled at another layer, either in hardware or in software with md. DRBD would do the mirroring. -- Maurice Volaski, mvolaski@aecom.yu.edu Computing Support, Rose F. Kennedy Center Albert Einstein College of Medicine of Yeshiva University