ResourceManager

Even though Hops YARN allows to distribute the ResourceManager in order to have scheduling run on one node (the Scheduler) and resource tracking running on several other nodes (the ResourceTrackers) the configuration of the resource manager is similar to the configuration of Apache Hadoop YARN. When running in distributed mode all the nodes participating in the resource management should be configured the same way as a ResourceManager would be configured. They will then automatically detect each other and elect a leader to be the Scheduler.

Adding/Removing a ResourceManager

As the ResourceManagers automatically detect each other through the database, adding a new ResourceManager consist simply in configuring and starting a new node as it would be done for the first ResourceManager. The same way removing a ResourceManager consists simply of stopping the ResourceManager. If this ResourceManager was the scheduler, this will induce some delay in the ongoing scheduling as a new scheduler needs to be elected, and needs to recover the cluster state, but all scheduling operations will resume as soon as these operations have finished.