Based on my experience I would recommend following :-
a) Always have only one extract process running
b) Never put any conditions in extract
c) If you have any conditions put it in pump process
d) Split one table per replicat (if it has more than 500K DML(s) in a day)
e) Split one table into multiple replicate (if it has huge DML(s) in a day) (increase INITRANS for table and indexes in that case)
a) Always have only one extract process running
b) Never put any conditions in extract
c) If you have any conditions put it in pump process
d) Split one table per replicat (if it has more than 500K DML(s) in a day)
e) Split one table into multiple replicate (if it has huge DML(s) in a day) (increase INITRANS for table and indexes in that case)
No comments:
Post a Comment