> I understand the --drop-slot part. But I don't understand what it means
> to ignore a missing replication slot when running --start.
I'm not sure I do either, honestly. I followed the Principle of Least Surprise in making it a no-op when those switches are used and the slot doesn't exist, over "no one will ever do that". Because someone will.
I'm happy to hear suggestions on what to do in that case beyond exit cleanly.
> The same option should be added to pg_receivewal as well.