This note is unclear to me:
Unless no other witness produces blocks and witness participation is high enough, subsequent runs which connect to an existing witness node over the p2p network, or which get blockchain state from an existing data directory, need not have the --enable-stale-production flag.
Does it mean that if I have a private testnet on a single computer, I need to use this flag at all times when I run the witness node or just for the very first run of the witness node when the first block gets produced?
Also you effectively say that if the witness participation is
high enough I need this flag to be on - which makes little sense to me. Shouldn't it be the other way round?