I am just pasting the statement from the Slide Guide
When data is split, each chunk is striped across storage pools and replicated to other nodes. When data is replicated, MapR-FS will attempt to write the replicas to different storage pools. In other words, MapR-FS tries to keep your data distributed as much as possible. However, MapR-FS will try to keep your data within the same node."
This statement seems to be contradictory. When it is saying MapR-FS tries to keep you data distributed as much as possible means, I can understand that it will replicate the data across different nodes of different racks.
The immediate statement says Map-FS will try to keep your data within the same node."
I didn't understand this statement.