Primary database is down
Log into the Data Guard Broker on standby node and check the configuration (dgmgrl sys/<password>)
Issue the failover command.
Modify the listener.ora file on the new primary node to allow it to listen for connections from the old primary.
To prevent Grid Infrastructure from starting the database in the wrong mode the configuration needs to be modified by issuing the following clusterware commands:
On the old primary:
srvctl modify database -d dbtst -s mount
One the new primary:
srvctl modify database -d dbtstg -s open
Now that the system has failed over when the old primary node is back up and running it needs to be reinstated.
From the new primary node log into the Data Guard Broker and issue the reinstate command.
Now move to the old primary node and follow the instructions from the Data Guard Broker on the new primary.
With the database instance back in the mounted mode move back to the new primary and reissue the reinstate command.
Checking the configuration should show that the nodes have swapped roles at this time and are operating as expected.
Now that the nodes are up and running to return to normal service a switchover must be performed.
From the new standby (old primary) log into the Data Guard Broker and issue the switchover command.
Checking the configuration should show that the systems are back to normal
Modify the listener.ora file on the standby node.
Bounce listener on node and ensure it’s back to normal.
Finally update the Grid Infrastructure configuration startup mode.
On the primary:
srvctl modify database -d dbtst -s open
On the secondary:
srvctl modify database -d dbtstg -s mount