Consider removing P2P duplicate connection disconnect

[heifner]

The net_plugin attempts to disconnect all but one of duplicate connections to a peer. The idea is that if peer A is configured to connect to peer B and peer B is configured to connect to peer A; each peer A and B will only have one connection to each other.

PeerA
p2p-peer-address = peerb:9876

PeerB
p2p-peer-address = peera:9876

PeerA <- connected -> PeerB
PeerB <- connected -> PeerA

The deduplication logic of net_plugin has never worked reliably. There are currently three different schemes that have been tried: proto_dup_goaway_resolution, proto_dup_node_id_goaway, and current scheme.

The current scheme still leaves duplicates in some race conditions. A simpler approach was recently attempted. See #1108. However, this approach had a race condition so that it would sometimes disconnect both connections; reverted in #1186. The issue is that the connection and handshake messages can be sent from each node at the same time. Both nodes can consider the other a duplicate at the same time and both disconnect.

In addition, EOSUSA has reported that the current de-duplication logic is causing continuous disconnect/reconnect from their proxy.

With the addition of the new block nack feature of net_plugin, the benefit from removing duplicate connections is reduced. At this point, I wonder if we shouldn't just remove this feature as it seems to be causing more harm than good.

[bhazzard]

We considered this and came to a solution instead.