Flow of execution in Map Reduce

Question

I knew that the reducer cannot start until the Mapper is run fully. I want to know if the reduce in the Reducer can run while the Shuffle or Sort is still running /not fully completed.

It seems your question is completely answered here: http://stackoverflow.com/a/11673808/2700344 — leftjoin, Mar 31 '17 at 06:56
Possible duplicate of [When do reduce tasks start in Hadoop?](http://stackoverflow.com/questions/11672676/when-do-reduce-tasks-start-in-hadoop) — vefthym, Mar 31 '17 at 07:05

score 1 · Accepted Answer · answered Mar 31 '17 at 15:37

1

A reducer can start even if "mapper task have partially completed". The param controlling it is mapreduce.slowstart

Follow this. http://bytepadding.com/big-data/map-reduce/understanding-map-reduce-the-missing-guide/

answered Mar 31 '17 at 15:37

KrazyGautam

2,839
2
21
31

But in most of the blogs, including one mentioned by @leftjoin indicate that map tasks must be complete before the reduce task starts. – Keshav Pradeep Ramanath Apr 02 '17 at 15:56

Flow of execution in Map Reduce

1 Answers1