CodePudding user response:
Refers to the graphs of writing your own function and then packaged into a jar file, each jobtracker scheduling is write your own jar file, please, is you of the jar to the HDFS, then each tasktracker executes download
Said the jobtracker node is insufficient memory to too much homework is to deal with too many different kinds of tasks, I don't understand is the function of each jar package, once submitted several different data analysis processing
memory has a variety of situations, see the log, sometimes is jt high concurrency, machine performance, have a plenty of the task of memory high performance not tt
Said the tasktracker node, the mobile computing convenient than mobile data
understand