From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34830) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WGz3Q-0005Jg-9X for qemu-devel@nongnu.org; Fri, 21 Feb 2014 17:58:53 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WGz3L-0007Cj-5w for qemu-devel@nongnu.org; Fri, 21 Feb 2014 17:58:48 -0500 Received: from lnantes-156-75-100-125.w80-12.abo.wanadoo.fr ([80.12.84.125]:58672 helo=paradis.irqsave.net) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WGz3L-0007Cf-09 for qemu-devel@nongnu.org; Fri, 21 Feb 2014 17:58:43 -0500 Date: Fri, 21 Feb 2014 23:58:42 +0100 From: =?iso-8859-1?Q?Beno=EEt?= Canet Message-ID: <20140221225842.GB13587@irqsave.net> References: <1393017681-12794-1-git-send-email-benoit.canet@irqsave.net> <5307D7EB.2000508@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <5307D7EB.2000508@redhat.com> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH V19 00/12] Quorum block filter List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: =?iso-8859-1?Q?Beno=EEt?= Canet , kwolf@redhat.com, qemu-devel@nongnu.org, stefanha@redhat.com, mreitz@redhat.com The Friday 21 Feb 2014 =E0 15:49:15 (-0700), Eric Blake wrote : > On 02/21/2014 02:21 PM, Beno=EEt Canet wrote: > > Hi this is the latest version of the quorum block filter. > >=20 > > This version fix the qmp events and a small rebase conflict. > >=20 > > tested: > >=20 > > tested that json qmp string are correct and parse well with python > >=20 > > Ran qemu-iotest including 081 ok > >=20 > > quorum command line ok > > quorum command line snapshot transaction ok > > quorum QMP by reference ok > > quorum QMP by reference and snapshot transaction ok > > quorum QMP at once ok > > quorum QMP at once and snapshot transaction ok >=20 > Generic question: >=20 > When the quorum detects mismatch but there was still a majority, is > there a way to request that the quorum correct the odd man out? That > is, if I have a 2/3 quorum, and files A and B read '1' while C reads > '2', can the mere act of reading that sector be used to write '1' into = C? >=20 > Obviously, this should be an optional mode, not enabled by default; but > I could envision creating a self-healing redundant array via this metho= d. That's part of the optional features I want to add. I choose to start small to get it merged faster. I also plan to modify drive-mirror to do a bdrv_swap when the mirroring i= s finished in order to replace a faultly quorum file. Best regards Beno=EEt >=20 > --=20 > Eric Blake eblake redhat com +1-919-301-3266 > Libvirt virtualization library http://libvirt.org >=20