Currently the spawned job is executed as a local job runner, the execution happens only in the node in which you fired the job. Specify the job tracker address by setting the property mapred.job.tracker
in your mapred-site.xml
inorder to make the execution distributed.
Execution in distributed mode might solve your outOfMemory issue
If you look at the environment variable HADOOP_CONF_DIR
, its values is empty set its value using the following command export HADOOP_CONF_DIR=/etc/hadoop/conf
. Make sure the value of the property mapred.job.tracker
which should point to your jobTracker in /etc/hadoop/conf/mapred-site.xml configuration