Which of the following cluster configurations will fail to ensure the completion of a Spark application in the event of a worker node failure?

Note: Each configuration has roughly the same compute power, using 100GB of RAM and 200 cores.