I have been trying for a long time to make this work, but I cannot see why it fails. I have a .sh script starting a Hadoop mapreduce job:
yarn --config /etc/h
