Hide Forgot
A common deployment of OpenStack is to have a controller node with all the services running and a compute node with nova-compute running Currently we just install the openstack-nova package on the compute nodes. That's fine, but since it's such a common deployment scenario it would be nice to be able to just install the nova-compute service. It doesn't save much in disk space, but it would help make things more clear. We could potentially do the same for all services, probably starting with nova-network (since that can be installed on all compute nodes in HA mode) but I don't think there's any great need for it yet. So, we should have: - the main body of python code in an openstack-nova-common package - the nova-compute service in openstack-nova-compute, which requires common - everything else in openstack-nova, which requires openstack-nova-compute The common package isn't completely necessary to do this, but it would seem odd to have most of the code in openstack-nova-compute.
This split was done for the Essex packages in Fedora 17 and EPEL