You are designing a data solution for Spark workload in HDInsight. Which of the options are correct in case of tuning the cluster for performance? (Multiple choice)

Answer options:

A.At Physical Layer, run a cluster with larger-sized VMs.
B.At Physical Layer, run a cluster with smaller-sized VMs.
C.Use smaller YARN containers.
D.Use Bigger YARN containers.

Correct Answers: A and C
For the HDInsight cluster, there are three layers that can be tuned to increase the containers and use all the available throughput. They are the Physical Layer, YARN layer, and Workload Layer. At Physical Layer, run clusters with larger-sized VMs; a larger cluster will have more space to run YARN containers. At the YARN layer, each YARN container size should be smaller; we have more space to create more YARN containers. 
PHYSICAL LAYER 
Image Source: Microsoft 

YARN LAYER
Image Source: Microsoft 

Options A and C correct: In the physical layer, VM size should be larger to accommodate more number of YARNS, while at YARN layer the yarn container size should be less to utilize the maximum resources at the physical Layer.
Options B and D are incorrect: This is exactly opposite to what is expected. So, it not only helps in tuning but also affects adversely.
Reference: 
To know more, please refer to the docs below: 

https://docs.microsoft.com/en-us/azure/storage/blobs/data-lake-storage-performance-tuning-guidance#general-considerations-for-an-hdinsight-cluster