Automation can no longer silently undo the resolver redundancy

A second name resolver was added recently so that the platform no longer loses every internal lookup when one machine stops. The machines themselves were updated to use both.

The automation that manages those machines, however, still had a single resolver written into its defaults. The next routine run would have quietly rewritten every machine back to one resolver and undone the redundancy, with nothing to indicate it had happened. The defaults now match reality, and a related setting that could have renamed the jump host's network card has been corrected too.