I am in process of setting up a cluster. Plan is to make it production setup once the necessary load testing is done.
I will start with :
3 master nodes.
2 coordinating nodes.
5 data nodes.
The number of data nodes will be increased if needed.
From what I know I can use machines with less RAM and Disk space (read 4 GB RAM and 50GB Disk space) for Master nodes and Coordinators.
The data nodes need to have heavier spec'ed (8 GB RAM and 1 TB Disk).
Also is the number of coordinators right? Or should I add more?
I am not an expert in this. I will like the experts to chip in if there are potential issues.