If the master instance of the hot standby system fails, a standby instance automatically assumes the role of master instance. The downtime for this procedure is very brief (a restart of the master instance generally takes much longer).
The cluster software constantly monitors the master instance of the hot standby system.
1. If the master instance fails, the cluster software selects one of the standby instances as the new master instance and sends that information to the standby instance.
2. The new standby instance adopts the virtual server name of the old master instance, with which the hot standby system is addressed externally.
3. The storage system gives the new master instance write authorization for the entire log area.
4. The new master instance imports the last un-imported redo log entries from the old master instance and redoes the respective transactions.
5. The database instance transfers the new master instance from the STANDBY to the ONLINE operational state.
Once you have handled the error in the master instance, you can add the old master instance to the hot standby system as a standby instance.