[usability] Refactor kubeadm.cfg usage

[leifmadsen]

While working on another problem (#176), I realized I needed to add changes to kubeadm.cfg while instantiating a cluster. We already do this with the IPv6 deployment, and I can think of a few other things that would be nice to change here too.

I think there are 2 approaches here:

1. Add a template with all the kubeadm options, and build it out with some defaults that can override
2. Provide some standard files, but allow an override file to be provided

I'm not entirely sure how to do approach 2 right now (maybe create a new inventory directory where vms.generated.local will go, and provide some directories like group_vars/, etc? Not sure if we can do files/ and templates/ there, but I'd guess no). Approach 1 feels a bit more natural, although maybe slightly more cumbersome.

We'd want to make sure we're supported at least the IPv6 and control plane listen all functionality, and combine that in. Maybe it makes more sense if we provide examples of inventories for the vms.generated.local and some drop in variable files that would provide the configuration for IPv6, control plane all interfaces, etc.

We can get the current working kubeadm.cfg of a running cluster with the kubeadm config view command, per https://kubernetes.io/docs/reference/setup-tools/kubeadm/kubeadm-config/