1

Following the instructions for "Initializing Ganglia on a Job Flow" I get my cluster up but don't see any Ganglia process running (on 8157).

http://docs.amazonwebservices.com/ElasticMapReduce/latest/DeveloperGuide/init_Ganglia.html

elastic-mapreduce --create --alive --name "Tom's Daily Hive 8x Flow" --instance-type c1.medium --num-instances 8 --availability-zone us-east-1a --bootstrap-action "s3://elasticmapreduce/bootstrap-actions/install-ganglia" --stream

There is a 0K file in: /tmp/ganglia-installed

Any suggestions? Thanks!

Tom Emmons
  • 103
  • 1
  • 7
  • OK ... so it turns out the instructions at Amazon are a bit light on the details.
    Try here http://ecs-network.serv.pacific.edu/ecpe-293a/projects/commoncrawl-tutorial
    and here http://docs.amazonwebservices.com/ElasticMapReduce/latest/DeveloperGuide/UsingtheHadoopUserInterface.html#AccessingtheHadoopUserInterfacetoMonitorJobStatus2 and you will do much better.     – Tom Emmons Apr 12 '12 at 16:39

2 Answers2

0

OK ... so it turns out the instructions at Amazon are a bit light on the details.

Try here http://ecs-network.serv.pacific.edu/ecpe-293a/projects/commoncrawl-tutorial and here http://docs.amazonwebservices.com/ElasticMapReduce/latest/DeveloperGuide/UsingtheHadoopUserInterface.html#AccessingtheHadoopUserInterfacetoMonitorJobStatus2

and you will do much better.

I am light on port forwarding experience, and what they have in the EMR instructions led me down many wrong paths (guesses).

Tom Emmons
  • 103
  • 1
  • 7
0

Your create cluster command looks OK. Set up FoxyProxy on Firefox like they mention in the same document you are referring to in your question.

After the set up, go to localhost:8157/ganglia, that should take you to your monitoring graphs!

Mark Grover
  • 4,070
  • 22
  • 21