0

So I'm running a pretty basic (just a search for a simple expression) program via Hadoop streaming on my 3-node cluster. When I run the job, JobTracker informs me that only 6 maps and 6 reduces are running, with 2000 pending map and reduce jobs. Why is this happening? I have set the maximum number of mappers and reducers (60, 20 per node) in my conf/mapred-site.xml file. What do you suggest to solve this issue?

Alpha
  • 11
  • 3
  • 1
    How many maps/reduces have you chosen for the job? How many available map/reduce slots are shown on JobTracker UI? – Roman Bodnarchuk Jul 19 '13 at 18:50
  • I have chosen 20 maps and 20 reduces for the job, both in the command line that I am executing and in the config file. Available (pending) map slots are in the thousands, and available (pending) reduce slots are 20. But only 6 map/reduce are running at once. Please advise. – Alpha Jul 19 '13 at 21:36

0 Answers0