MRv1 uses the JobTracker to create and assign tasks to data nodes, which can become a resource bottleneck when the cluster scales out far enough (usually around 4,000 clusters).In general,Job tracker has to manage the resources and application
MRv2 (aka YARN, "Yet Another Resource Negotiator") has a Resource Manager for each cluster,which bothers about how many slots available? what if node fails? what is the capacity? and each data node runs a Node Manager. For each job, one slave node will act as the Application Master, monitoring resources/tasks, etc.
MRv2 (aka YARN, "Yet Another Resource Negotiator") has a Resource Manager for each cluster,which bothers about how many slots available? what if node fails? what is the capacity? and each data node runs a Node Manager. For each job, one slave node will act as the Application Master, monitoring resources/tasks, etc.
