Resource management using YARN is a key component of Hadoop 2 architecture and it has multiple possible scheduler configurations. [Marks: 5] (a) Give at least 2 reasons why you may consider a Fair scheduling policy as opposed to a Capacity based policy ? (b) Suppose you are using a Fair scheduler with only 2 queues with each having weight=10. Now suppose one of the queues (Queuel) has a sudden surge of important jobs and it is at 90% utilization of it's fair share for more than a threshold time (say, 1 min). The second queue (Queue2) has no surge. In what way can this be handled by the scheduler ? State your assumptions and highlight any optimizations the scheduler will perform. (c) Suppose you are using the Capacity scheduler and a queue is configured with min=10% and max=40%. Given the queue can have multiple users, what is the mechanism to make sure that the max allocation for a single user is within 20% ?

anagement Resource t using YARN is a key component of Hadoop 2 architecture and it has multiple possible scheduler configurations. [Marks: 51 (a) Give at least 2 reasons why you may consider a Fair scheduling policy as opposed to a Capacity based policy? (b) Suppose you are using a Fair scheduler with only 2 queues with each having weight-10. Now suppose one of the queues (Queuel) has a sudden surge of important jobs and it is at 90% utilization of it's fair share for more than a threshold time (say, 1 min). The second queue (Queue2) has no surge. In what way can this be handled by the scheduler? State your assumptions and highlight any optimizations the scheduler will perform (C) Suppose you are using the Capacity scheduler and a queue is configured with min=10% and max=40%.. Given the queue can have multiple users, what is the mechanism to make sure that the max allocation for a single user is within 20%? anagement Resource t using YARN is a key component of Hadoop 2 architecture and it has multiple possible scheduler configurations. [Marks: 51 (a) Give at least 2 reasons why you may consider a Fair scheduling policy as opposed to a Capacity based policy? (b) Suppose you are using a Fair scheduler with only 2 queues with each having weight-10. Now suppose one of the queues (Queuel) has a sudden surge of important jobs and it is at 90% utilization of it's fair share for more than a threshold time (say, 1 min). The second queue (Queue2) has no surge. In what way can this be handled by the scheduler? State your assumptions and highlight any optimizations the scheduler will perform (C) Suppose you are using the Capacity scheduler and a queue is configured with min=10% and max=40%.. Given the queue can have multiple users, what is the mechanism to make sure that the max allocation for a single user is within 20%