File locations for job node set-ups
When you have a job node set-up, all the job node computers need to have access to the same data location folder. This is because the job nodes will write files directly to the folder rather than passing through the master node (which would be a bottleneck for big jobs). Furthermore, the user running the server must be the same for all the job nodes and it needs to act as the same user when accessing the folder no matter whether it is a job node or a master node.
The data location should be added after the job nodes have been configured and attached to the master node. In this way, all the job nodes will inherit the configurations made on the master node.
One relatively common problem faced in this regard is root squashing which often needs to be disabled, because it prevents the servers from writing and accessing the files as the same user - read more about this at http://nfs.sourceforge.net/#faq_b11.
You can read more about Job Distribution.