From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:36694)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <benoit.canet@irqsave.net>) id 1THHd7-0005Hl-Ph
	for qemu-devel@nongnu.org; Thu, 27 Sep 2012 13:12:06 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <benoit.canet@irqsave.net>) id 1THHd6-0000Zg-Of
	for qemu-devel@nongnu.org; Thu, 27 Sep 2012 13:12:05 -0400
Received: from paradis.irqsave.net ([109.190.18.76]:41400)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <benoit.canet@irqsave.net>) id 1THHd6-0000ZR-Ga
	for qemu-devel@nongnu.org; Thu, 27 Sep 2012 13:12:04 -0400
Date: Thu, 27 Sep 2012 19:11:58 +0200
From: =?iso-8859-1?Q?Beno=EEt?= Canet <benoit.canet@irqsave.net>
Message-ID: <20120927171157.GA4407@irqsave.net>
References: <1348756198-10845-1-git-send-email-benoit@irqsave.net>
	<50647793.7000106@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <50647793.7000106@redhat.com>
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] [partial RFC 0/2] QCow2 deduplication write
 mechanism
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Avi Kivity <avi@redhat.com>
Cc: kwolf@redhat.com, pbonzini@redhat.com, qemu-devel@nongnu.org, stefanha@linux.vnet.ibm.com

> If I understood correctly, this does cluster-level dedup within a qcow2
> image.

yes

> What is the motivation here?  Reduce space usage if a guest copies file=
s
> internally?
First use case is people using LXC, OpenVZ or chroot inside vms they rent
to their "cloud" provider.

Second use case is applications generating duplicates blocks.
A well known CAD software does this and on a filer doing deduplication
on 4KB blocks the dedup ratio is around five.

>=20
> Why use cluster granularity?  If the guest uses smaller granularity, it
> will misalign the data wrt cluster boundaries, and deduplication will f=
ail.

Using cluster granularity would allow to use the qcow2 refcount and l1/l2
tables to track duplicated blocks.
The patch embryo does read the missing data if given data is misaligned.
To avoid hurting performance using a smaller cluster size with the adequa=
te
changes in qcow2 may help.

Regards

Beno=EEt