On Mon, Jun 15, 2020 at 12:40:03PM +0900, Fujii Masao wrote:
> BTW, I just wonder why each row in pg_replication_slots needs to have
> min_safe_lsn column? Basically min_safe_lsn should be the same between
> every replication slots.
Indeed, that's confusing in its current shape. I would buy putting
this value into pg_replication_slots if there were some consistency of
the data to worry about because of locking issues, but here this data
is controlled within info_lck, which is out of the the repslot LW
lock. So I think that it is incorrect to put this data in this view
and that we should remove it, and that instead we had better push for
a system view that maps with the contents of XLogCtl.
--
Michael