i have aws rds postgres read replica consuming rds instance. i'd consistent view on tables in particular schema such can query them etl purposes.
what i'm trying similar pg_dump, i'd rather consistent view of schema , query tables, rather writing them out separate file.
right we're using snapshots of production database, restoring them , streaming there, saving snapshot takes long time due load , read performance snapshotted database quite poor.
this can accomplished using same thing pg_dump does: use "repeatable read" transaction:
begin; set transaction isolation level repeatable read;
execute queries
commit;
