No, my question was about running different types of compute nodes to offer different types of virtualization for users (for example running a KVM node and an LXC node, and be able to request instances of one or the other from the API).
But rereading the bug description, I think that's unrelated... but rather a packaging improvement to make sure we don't have crazy depends on a generic "compute node".
No, my question was about running different types of compute nodes to offer different types of virtualization for users (for example running a KVM node and an LXC node, and be able to request instances of one or the other from the API).
But rereading the bug description, I think that's unrelated... but rather a packaging improvement to make sure we don't have crazy depends on a generic "compute node".