This content has been marked as final. Show 3 replies
For the database connections to be failed over to the next available instance, you would have to configure TAF(Transparent application failover).
For a very basic example - here is how you would change your tns connect descriptor to
(description= (load_balance=on) (failover=on) (address= (protocol=tcp) (host=sales1-server)
(address= (protocol=tcp) (host=sales2-server) (port=1521)) (connect_data=
(failover_mode= (type=select) (method=basic))))
This connect descriptor needs to be changed in OiD where content services picks up the database connection url from.
Hope that helps
Thanks for the reply. We have been using a database connection string like this but still had problems. We have since discovered that the problem seems to be with the S2S servlet that allows trusted authentication. If a RAC node is shutdown then the S2S servlet will log exceptions and return errors when trying to create a session for a user until the servlet has used up all of it's invalid cached connections to the database. After this point it appears to start authenticating users correctly again. Our system cannot have this kind of period of authentication failures, so hopefully there is a resolution somewhere for this issue!