Allow users to configure Rufio on a specific node

Rufio currently runs on a random worker node that `kube-scheduler` assigns to it.
Users may want to control which node/s Rufio gets scheduled based on which nodes would have network access to the servers' BMCs.
There should be a way for users to configure this through clusterconfig

Start by tracing how clusterconfig currently configures Rufio and how kube-scheduler places it on worker nodes. Define the supported node-selection configuration, then verify that the resulting Rufio deployment is scheduled only on the configured node or nodes.

Written by the indexing model from the issue text.