What is the best system configuration for each node for a MapR Cluster with 3 nodes for R & D purpose?
As far as my knowledge and working experience on the MapR Admin , For R&D and POC purpose the following configurations are good enough to deal with data that can be limit up to 1 or 2 TB( Assume this much data that you are dealing with).
One machine - HP Server ( Blade servers) with 1 TB of storage and 32 GB of RAM
Rest of two machines - HP Workstation machines with 1 TB of storage each with 16 GB of RAM on each machine.
The Configurations always depends on the data that we dealt with,
1)The data that we dealt with
2)The amount of data that is getting stored to the File system
3)replication factor that we specified
4)Amount of data for newly coming streaming and for batch processing
Better understanding check this issue How to Choose Cluster Size?
Retrieving data ...