From owner-freebsd-geom@FreeBSD.ORG Mon Nov 17 08:31:00 2014 Return-Path: Delivered-To: freebsd-geom@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 758D9811 for ; Mon, 17 Nov 2014 08:31:00 +0000 (UTC) Received: from agora.rdrop.com (agora.rdrop.com [IPv6:2607:f678:1010::34]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 5290815B for ; Mon, 17 Nov 2014 08:31:00 +0000 (UTC) Received: from agora.rdrop.com (66@localhost [127.0.0.1]) by agora.rdrop.com (8.13.1/8.12.7) with ESMTP id sAH8UvKh068994 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 17 Nov 2014 00:30:58 -0800 (PST) (envelope-from perryh@pluto.rain.com) Received: (from uucp@localhost) by agora.rdrop.com (8.13.1/8.14.2/Submit) with UUCP id sAH8Uv0t068993; Mon, 17 Nov 2014 00:30:57 -0800 (PST) (envelope-from perryh@pluto.rain.com) Received: from fbsd81 by pluto.rain.com (4.1/SMI-4.1-pluto-M2060407) id AA22147; Mon, 17 Nov 14 00:28:05 PST Date: Mon, 17 Nov 2014 00:28:03 -0800 From: perryh@pluto.rain.com (Perry Hutchison) To: lists@natserv.net Subject: Re: Gmirror vs 10 Release-P12 Message-Id: <5469b193.lkIZWruTiJ873IFy%perryh@pluto.rain.com> References: <5465580A.6070609@natserv.net> <20141114161658.GA1719@mail.michaelwlucas.com> <5468F9D6.4020300@natserv.net> In-Reply-To: <5468F9D6.4020300@natserv.net> User-Agent: nail 11.25 7/29/05 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: mwlucas@michaelwlucas.com, freebsd-geom@freebsd.org X-BeenThere: freebsd-geom@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: GEOM-specific discussions and implementations List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 17 Nov 2014 08:31:00 -0000 Francisco Reyes wrote: > On 11/14/2014 11:16 AM, Michael W. Lucas wrote: > > I believe 9+ GPT partitioning doesn't work with gmirror on a whole > > disk. Pardon the narcissistic link, but: > > But it was a 10.0 Release > To recap.. > 10.0 Release.. gmirror works. That was either a bug in 10.0 Release, or an accident. > Updated to 10-P12... gmirror no longer worked. As expected. > For now went without gmirror altogether. The client needed the machine > ASAP. May try upgrading to 10.1 later afer it has been out longer. 10.1 should make no difference. The problem is that, per the GPT spec, the GPT partition data appears in both the first and last blocks of the physical drive. That conflicts with applying gmirror to the entire drive, since gmirror uses the last block of whatever provider it is hosted on. If you try to gmirror the whole drive, and GPT-partition the mirror, a strict BIOS will complain about the absence of the GPT data in the drive's last physical block and/or about the size in the GPT not matching the size reported by the drive. The proper way to use gmirror on a GPT disk is to mirror individual partitions, rather than the whole drive. Yes, this can be a nuisance -- if you have 5 partitions you need 5 mirrors rather than just one. Another approach is to make just one GPT partition, gmirror that, and then MBR-partition that mirror. However getting this scheme to work on the boot drive requires hacking the boot blocks.