YARN schedulers are the efficient algorithms written to manage cluster resources. YARN's ResourceManager service has a pluggable and pure scheduler component, that is, it does not monitor or track the applications running in the cluster. It is responsible only for allocation of resources to running applications.
You might be wondering, what is resource allocation and why it is important? Well, let's consider a simple scenario. Suppose an organization has a 100 nodes Hadoop-YARN cluster and there are N teams (for example, A, B, ….. N) using the same cluster. Each team has around 10-15 members and each team member can submit YARN applications on the cluster. In order to provide a shared multitenant and efficient cluster utilization, cluster resource allocation plays an important role for a cluster administrator. The cluster resources are divided among the different teams or team members based on a pluggable policy. While defining sharing parameters, the...
