Configuring Yarn/MapReduce2 memory configuration

[#1]

I have a hadoop cluster on 2 nodes. Each node has 2 cpus and 24 gb memory.

So I am thinking of having 3 mappers and 3 reducers on each node. That's good?

So in Ambari -> Yarn -> Configs, can I have :

Memory allocated for all YARN containers on a node : 18 gb

Minimum container size (memory) : 1 gb

Maximum container size (memory) : 4 gb

If I configure as per above, then Ambari suggests other configurations based on these values - is that safe to follow?

Likewise in Ambari -> MapReduce2 -> Configs :

Not sure what to configure Map Memory, Reduce Memory, Appmaster Memory.

Appreciate the insights.

[#2 · Accepted solution]

Below is a good read on Determining HDP Memory configs:

http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.4/bk_installing_manually_book/content/determin...