we now disconnect peers with VERSION timestamps older than the genesis block.
So we cannot assume that we will have time to detect the successful connection before we get disconnected because of misbehavior when the version message is processed.

See change below in test_node.py

In our codebase we discourage sending any message before version or before verack. Keeping these pings makes the test fail intermittently (often) because disconnection can happen before the handshake timeout so the node never logs the expected messages below (expected_timeout_logs).

By keeping this one ping we can maintain the same log messages. Without it, we would lose the coverage for the message count before the handshake times out. We would need to replace:

socket no message in first 3 seconds, 1 0 peer=1

with:

socket no message in first 3 seconds, 0 0 peer=1

I didn't see any intermittent test failures by running it 100 times. So I suggest we keep it unless we start seeing test failures on CI

used by the version message below. Anything lower than the genesis timestamp would cause a disconnection

the previous two nodes got disconnected before this one was connected