问题
I have a only map program which have more than 10 maps. My cluster have 5 nodes, my question is: can I scheduling running which map tasks to which node? For example, let No.1 map running in node1, No.3 map running in node3 and No.9 map running in node5. If I can do so, then what change in my code and configuration file is needed. By the way, I use HDFS. Thanks!
回答1:
Affinity can't be set in the Hadoop for the map task to node mapping. The scheduler automatically decides based on the data locality/slot status where to run a Map task.
来源：https://stackoverflow.com/questions/15647815/how-to-scheduling-hadoop-map-tasks-in-different-node-cluster