Greetings,
* Michael Holzman (michaelholzman@gmail.com) wrote:
> I have two applications A and B. A runs SELECT statements only and only on
> tableA. B actively updates tableB, A never looks into tableB. B has nothing
> to do with tableA.
In an ideal world, such distinct applications would probably do better
to live in independent PG clusters, which would address this issue with
VACUUM'ing and also allow you to do file-level backup/restore of each
independently, scale physical replicas independently, as well as manage
roles to be specific for each without seeing others, etc.
Thanks,
Stephen