yes it does.
This is one of the annoying thing bout the master publisher failover to a secondary. If it happens then you basically are installing a new cluster node from scratch, which is why it just becomes a subscriber and then you tell it to be a master publisher.
The db has been wiped , which is why you don't put your insight db on a master publisher
Had discussions about why the master publisher / secondary publisher failover stuff was set up the way it is. Engineering said you couldn;t guarantee that a "fail-back" to the original publisher would work every time, the reason why it failed over to the secondary might still be there.