On Fri, May 16, 2014 at 08:18:36AM -0600, Eric Blake wrote: > On 05/16/2014 07:54 AM, Daniel P. Berrange wrote: > > >> <disk type='quorum' device='disk'> > >> <driver name='qemu' type='quorum'/> > >> <threshold value=2/> > >> <backingStore type='file'> > >> <format type='qcow2'/> > >> <source file='/var/lib/libvirt/images/file1.qcow2'/> > >> </backingStore> > >> <backingStore type='file'> > >> <format type='qcow2'/> > >> <source file='/var/lib/libvirt/images/file2.qcow2'/> > >> </backingStore> > >> <backingStore type='file'> > >> <format type='qcow2'/> > >> <source file='/var/lib/libvirt/images/file3.qcow2'/> > >> </backingStore> > >> <target dev='vda' bus='virtio'/> > >> </disk> > > > > It feels rather odd to have <backingStore> elements but no top level > > disk images. Really these are all top level images > > Unfortunately, we are allowed to have a quorum with mixed-mode sources - > I could have a quorum where file 1 is a local file, file 2 is a block > device, and file 3 is a gluster protocol. But since we encode the type > of file at the <disk type='...'> level, there is NO way to list three > different <source> elements for those three quorum members. I think > Benoit's proposal makes sense - a quorum is a node in the backing chain > with NO <source> element, but instead has MULTIPLE <backingStore> elements. Ok, I reluctantly agree. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list