From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wido den Hollander Subject: Re: RBD Backup Date: Thu, 22 Nov 2012 21:13:33 +0800 Message-ID: <50AE24FD.8090103@widodh.nl> References: <50ACD936.8060907@profihost.ag> <50ACDB85.4020006@widodh.nl> <50AE0535.8080904@profihost.ag> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from smtp01.mail.pcextreme.nl ([109.72.87.137]:52308 "EHLO smtp01.mail.pcextreme.nl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754373Ab2KVSiC (ORCPT ); Thu, 22 Nov 2012 13:38:02 -0500 In-Reply-To: <50AE0535.8080904@profihost.ag> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Stefan Priebe - Profihost AG Cc: "ceph-devel@vger.kernel.org" On 11/22/2012 06:57 PM, Stefan Priebe - Profihost AG wrote: > Hi, > > Am 21.11.2012 14:47, schrieb Wido den Hollander: >> The snapshot isn't consistent since it has no way of telling the VM to >> flush it's buffers. >> >> To make it consistent you have to run "sync" (In the VM) just prior to >> creating the snapshot. > > Mhm but between executing sync and executing snap is again time to store > data. > True. That is always a problem with snapshots. I always regard data written to disk in the last 30 seconds as being in the "danger zone". When you use libvirt and QCOW2 as a backing store for your virtual machine you can also snapshot with libvirt. It will not only snapshot the disk, but it will also store the memory contents from the virtual machine so you have a consistent state of the virtual machine. This has a drawback however, since when you give the VM 16GB of memory, you have to store 16GB of data. Right now this doesn't work yet with RBD, but there is a feature request in the tracker. I can't seem to find it right now. What you could do is: $ ssh root@virtual-machine "sync" $ rbd snap create vm-disk@snap1 $ rbd export --snap snap1 vm-disk /mnt/backup/vm-disk_snap1.img This way you have a pretty consistent snapshot. Wido >> rbd export --snap BACKUP image1 /mnt/backup/image1.img >> losetup /mnt/backup/image1.img >> >> kpartx -a /dev/loop0 >> >> Now you will have the partitions from the RBD image available in >> /dev/mapper/loop0pX > Works fine! > > Greets, > Stefan