1) Change the mode to max performance.
2) change log modes to ASYNC
3) Set transport=off
4) set apply=off
Perform 3 log switches on each of the two live instances.
Then record the state of dataguard for the configuration, each cluster database and each instance.
Here you will see that:
2) All databases are in ASYNC mode
3) All databases are in MAXPERFORMANCE

There are no alerts in the alert log.
We then shutdown the standby database and start getting alerts in the primary database alert log which, for live systems, are sent to production support.

Oracles answer to “why the alerts” is:

The reason is that when standby db is down, the client connection from EM (via data guard broker) which tried to connect to ( failed since the service name is not there when the db is down.

If we disable dg broker then the alerts stop.



Whenever we need to bring down a production standby we have to disable DG broker to ensure no alerts are generated for prod support by the live database.