Error in starting hadoop Job Tracker

Question

I tried to run a simple program in hadoop using Windows-Cygwin.

I am able to start the namenode .

The jobtracker start however fails with exception :

<         


        

Answer 1

I faced the same issue when working on the "Pseudo Distributed" examples as at this page: http://hadoop.apache.org/docs/r1.1.2/single_node_setup.html#PseudoDistributed

It turned out that hadoop simply wasn't picking up my conf files. The examples at the link above assume you are running in your install of hadoop (i.e. /Usr/jane/hadoop-1.1.2). I was trying to run the examples in another directory. I'm sure you could configure hadoop to recognize other 'conf' directories, but I took the easy route and just started running in my hadoop directory.

This thread helped me figure it out: https://issues.apache.org/jira/browse/HDFS-2515

Answer 2

The problem is the following lines should on into mapred-site.xml and NOT hdfs-site.xml,

<property> 
<name>mapred.job.tracker</name> 
<value>localhost:9101</value> 
</property>

By the way why are you trying to run Hadoop in Windows? For development? You don't have a linux machine or reluctant to install one?

One more thing, you usually put this property in core-site.xml not hdfs-site.xml,

<property>
<name>fs.default.name</name>
<value>hdfs://localhost:9100</value>
</property>