0

I had this strange error where my mapreduce job always got stuck at map 0% reduce 0%. So I went on to do some research and found this link which did solve my problem when I ran the hadoop tasktracker & on the namenode, but according to my knowledge Namenode only supports a JobTracker and in some cases Secondary NameNode. TaskTrackers are for DataNodes right? Please correct me if I am wrong in my understanding. Why do I have to start tasktracker daemon on the NameNode everytime?

And Is there a permanent solution to make this automatic because everytime I have manually start off the TaskTracker on the NameNode.

Community
  • 1
  • 1
Devi Prasad
  • 78
  • 2
  • 13
  • You should start the task traker only once.by the script in hadoop/bin/start-all.sh.If task tracker is not starting you should check the logs. – sp_user123 Feb 16 '15 at 07:03

1 Answers1

1

It is also possible that you're Mapping Task is stuck in an infinite loop. The percentage is the percent of Map Tasks that are completed. If every time you submit a Map Job the mapping is stuck in an infinite loop, its never gonna be able to be completed.

mross1080
  • 144
  • 1
  • 7
  • The Mapping task runs fine when I enable the TaskTracker on the NameNode, and that too manually. What am I missing in the configuration? – Devi Prasad Feb 16 '15 at 06:11
  • If the map task is running fine with the configuration you set up then what's the problem now? – mross1080 Feb 17 '15 at 18:25
  • 1
    Everytime I have to manually start up the tasktracker on the namenode. Can this be done through some configuration settings? Am I missing some configuration settings? – Devi Prasad Feb 18 '15 at 06:10