Test would occasionally fail when the rate of migration is faster than the time it took for the map reduce task to finish. This will cause the mapReduce task to keep on triggering the staleConfig exception until it exceeded the retry limit.

Note: Throttling won't fix the racy nature of this test.