I would like to add new machines to my existing cluster but they won't be similar to the current nodes. I have to scenarios I'm thinking of:
1. What are the implications (besides initial load balancing) of adding a new node to the cluster, if this node runs on a machine similar to all other nodes but has 10x storage ? 2.What are the implications of adding a new machine that is 10x "better" machine - 10x storage, faster disks, better CPU, more memory, etc. ? anything to consider besides factorizing the number of map / reduce tasks and the memory allocation for them on the "better" machine ? Thanks, Amit.