Question about image migration between Datastores and Ceph

Hi all

We want to try the new OpenNebula feature to migrate some images from one datastore to another.

In our case we are using Ceph as storage backend and that makes us wonder if it would be possible to include several Ceph datastores in your OpenNebula configuration. Not just a new Ceph pool within the same Ceph server, I mean a new Ceph storage configured in a different endpoint with different MON_HOST and ceph.conf file.

Right now we are playing around with several OpenNebula cloud clusters and Ceph endpoints and it would be great if it would be possible to migrate some resources like images between them. I think that OpenNebula uses rbd command to manage the images but (correct me if I’m wrong) it always uses /etc/ceph/ceph.conf file by default. So to support this the command should be changed by:

$ rbd ls -c /etc/ceph/ceph_ds2.conf -p one --id libvirt

and probably you should include, in your second Ceph DS configuration, a new variable for example:


that should be used by rbd commands if its defined…

The first question, is anyone else interested in this feature?

and the second question for OpenNebula developers :slight_smile: , are you willing to include this feature? or do you have some plans already to include something like this?

Cheers and thanks!


It’s already possible to configure two (or more) seperate Ceph clusters, I’ve got this running on a test cluster. There are two variables you need to set in your Ceph DS, besides the regular Ceph DS settings:


I have no experience with storage migration between Ceph clusters. Not sure if it’s possible either. But if it is this would be a nice feature to have (cross cluster migrations for example)

Hi @stefan

Ah that’s nice! I didn’t know that, we will try to add the conf variable and check this in our testing cluster.


I created a feature request:

Thanks!!! We’ll try to merge this for 5.4RC