On Sat, 2020-06-20 at 06:51 +0300, Toomas Kristin wrote:
> Thank you for help. hot_standby_feedback did what I expected (no lag on replication
> and no session termination) and case is closed. Only it is foggy for me how it can
> be when no logs about vacuuming from that time window when session is active on standby host.
Perhaps heap only tuple chain pruning.
Have you got "n_tup_hot_upd" > 0 somewhere in "pg_stat_user_tables"?
Then any reader can "micro-vacuum" blocks with such tuples.
Yours,
Laurenz Albe
--
Cybertec | https://www.cybertec-postgresql.com