問題文
A team is reviewing an Apache Spark application that reads sales files, aggregates them, and writes a summary. They want to state precisely which process decides how the work is split into tasks and which processes actually compute those tasks. Which statement correctly assigns these two responsibilities?
選択肢
- The cluster manager builds the plan and hands ready-made tasks to the executors, so the driver only submits the application and then waits for the final result to come back to it.
- The driver both plans and computes, while executors only hold cached data in memory, so raising the executor count would just enlarge the cache.
- Each executor plans its own portion independently after reading the source files.
- The driver builds the plan and schedules tasks, and the executors run those tasks on their own threads.