What happens if the output directory already exists?

In MapReduce parallel data flow, input is divided into:

Which phase ensures all values of a key go to the same reducer?

In MapReduce model, computation is moved to:

Speculative execution is used to:

If a Mapper fails, Hadoop:

Hadoop benchmarks like TeraSort are used to:

Which benchmark generates large datasets?

Correct sequence of benchmark execution:

In WordCount program, Mapper outputs:

Reducer in WordCount performs:

Hadoop Streaming allows:

In Hadoop Streaming, input/output is handled through:

Hadoop Pipes is mainly used for:

Which statement is TRUE about MapReduce?