Deploy one high-log-level mainnet geth node attached to a functioning CL

[tersec]

For nimbus-eth1 to investigate network peering. It should therefore be possible to find out its enode to connect to it specifically.

In principle, one Nethermind node would be nice too, but if Nimbus can solidly peer with Geth that's in practice enough to start with.

The main goal here is not number of nodes, it's having a detailed logging as possible over a long enough timespan (ideally 2 days or more) to be able to reason about details of peering behavior, including descoring, disconnecting, how long it takes to reconnect after being disconnected, et cetera.