From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:36048) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1avaq6-0004K2-Ey for qemu-devel@nongnu.org; Wed, 27 Apr 2016 21:33:59 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1avaq5-0000Hr-Bg for qemu-devel@nongnu.org; Wed, 27 Apr 2016 21:33:58 -0400 Date: Thu, 28 Apr 2016 09:33:55 +0800 From: Fam Zheng Message-ID: <20160428013355.GA29480@ad.usersys.redhat.com> References: <1460690887-32751-1-git-send-email-famz@redhat.com> <1460690887-32751-8-git-send-email-famz@redhat.com> <20160425004250.GA18010@ad.usersys.redhat.com> <20160427002025.GD834@ad.usersys.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Subject: Re: [Qemu-devel] [PATCH for-2.7 v2 07/17] rbd: Implement image locking List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: dillaman@redhat.com Cc: qemu-devel@nongnu.org, Kevin Wolf , qemu-block@nongnu.org, Jeff Cody , Markus Armbruster , Max Reitz , den@openvz.org, Paolo Bonzini , John Snow On Wed, 04/27 13:18, Jason Dillaman wrote: > On Tue, Apr 26, 2016 at 7:20 PM, Fam Zheng wrote: > > On Tue, 04/26 10:42, Jason Dillaman wrote: > >> On Sun, Apr 24, 2016 at 7:42 PM, Fam Zheng wrote: > >> > On Fri, 04/22 21:57, Jason Dillaman wrote: > >> >> Since this cannot automatically recover from a crashed QEMU client with an > >> >> RBD image, perhaps this RBD locking should not default to enabled. > >> >> Additionally, this will conflict with the "exclusive-lock" feature > >> >> available since the Ceph Hammer-release since both utilize the same locking > >> >> construct. > >> >> > >> >> As a quick background, the optional exclusive-lock feature can be > >> >> enabled/disabled on image and safely/automatically handles the case of > >> >> recovery from a crashed client. Under normal conditions, the RBD > >> >> exclusive-lock feature automatically acquires the lock upon the first > >> >> attempt to write to the image and transparently transitions ownership of > >> >> the lock between two or more clients -- used for QEMU live-migration. > >> > > >> > Is it enabled by default? > >> > > >> > >> Starting with the Jewel release of Ceph it is enabled by default. > > > > OK, then I'll leave rbd in this QEMU series for now. > > Without exposing some new API methods, this patch will, unfortunately, > directly conflict with the new Jewel rbd defaults and will actually > result in the image becoming read-only since librbd won't be able to > acquire the exclusive lock when QEMU owns the advisory lock. > > We can probably get the new API methods upstream within a week or two > [1]. If that's too long of a delay, I'd recommend dropping rbd > locking from the series for now. Yes you are right, I tried to mean "drop" with "leave" :) Thanks, Fam > > [1] http://tracker.ceph.com/issues/15632 > > -- > Jason