On Sun, Sep 23, 2012 at 8:41 AM, Itamar Heim <iheim@redhat.com> wrote:
On 09/23/2012 05:33 PM, Josh Logan wrote:

On Sun, Sep 23, 2012 at 6:10 AM, Itamar Heim <iheim@redhat.com
<mailto:iheim@redhat.com>> wrote:

    On 09/22/2012 08:58 AM, Josh Logan wrote:


        I'm currently setting up an ovirt cluster and so far it looks
        good.  I
        like the integration with Foreman http://theforeman.org/ .

        I would like to use Ceph / rbd for my storage.  I saw some
        mention of
        patches coming in May, but I did not find any new posts.

        What is the status of this work?  Is there some patches I can
        try out?
        I have a working Ceph cluster and a working ovirt cluster, I
        just need a
        way to bring them together.

        Thanks, JOSH



    I don't remember any active work on this right now (for sure nothing
    like the gluster integration being done).
    but iiuc, ceph provides posixfs support - did you try creating a
    posixfs based storage domain?
    (you would need a "full" host (not ovirt-node) to install ceph
    client components on).

    Thanks,
        Itamar



I am doing my work on Fedora 17 hosts, not ovirt-node, since I know this
will need more OS support.

There are a few different Ceph filesystems.  But the posix based one is
the least ready for production.  The rbd filesystem is integrated into
qemu and libvirt is the most suited for VM images.

Are the Gluster patches available?  I would like to see what that
feature looks like and if I can modify them for Ceph.
If there is a better filesystem to investigate please let me know.

Thanks, JOSH


gluster as a native storage domain (rather than posixfs)  is still in reviews (and has patches only for vdsm side).
http://gerrit.ovirt.org/#/c/6856/

you can also use NFS in the meantime if relevant for ceph.


Thanks for the pointer.  I'll follow that and see what I learn.

The vdsm side may be similar since both are network disk device.  There are only 2 steps needed to start up a VM with rbd.

qemu-img create -f rbd rbd:data/host1 10G

Then to start the image for qemu add -drive file=rbd:data/host1,if=none,id=drive-virtio-disk0,format=raw

or within libvirt:
    <disk type='network' device='disk'>
      <driver name='qemu' type='raw'/>
      <source protocol='rbd' name='data/host1'/>
      <target dev='vda' bus='virtio'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/>
    </disk>

So the steps are simple, and maybe Gluster is more complex then I should use as an example.

Thanks, JOSH