reburning if pi breaks

[jpfleischer]

two scenarios, a and b.
a: how can we reburn a manager which requires a new key on the manager being communicated to the nodes (documentation issue if this is already provided, but likely not.)

b: if worker breaks, then how do we reburn one or multiple workers without reburning the manager? using the manager's existing key to make sure these workers can be integrated into the cluster.