Hi, thanks for the detailed feedback here. I do think it's worth addressing the question Stephen raised as far as what we use for the IV[1]; whether LSN or something else entirely, and if so what. The choice of LSN here is fairly fundamental to the existing implementation, so if we decide to do something different some of this might be moot.
Best, David [1] https://www.mail-archive.com/pgsql-hackers@lists.postgresql.org/msg154613.html
