Point out the wrong statement.

Increasing the number of reduces increases the framework overhead, but increases load balancing and lowers the cost of failures

It is legal to set the number of reduce-tasks to zero if no reduction is desired

The framework groups Reducer inputs by keys (since different mappers may have output the same key) in the sort stage