In Dataguard configuration with have only one active node (physical standby) which will apply logs, second node is down. In event of disaster you commit switchover or failover to the physical standby and once it become primary database then you start the second node. In your case if dr-noed1 fails then you have to start dr-node2 and start managed recovery process there.
In a RAC environment, the Data Guard Broker will start the MRP on one instance (you can nominate which one is preferred) and if that fails, the Broker will re-start MRP on the surviving instance.
Oracle Certified Master DBA