All,

I have a hadoop cluster in EC2 created by whirr, which facilitates access to 
the job tracker and name node via a ssh-implemented SOCKS proxy.  Whirr updates 
the local hadoop configuration in hadoop-site.xml with a setting for 
"hadoop.socks.proxy", along with the actual URIs for the job tracker and name 
node.  Is there a way to plug this into the oozie job properties or directly 
into the workflow xml?  Or, do oozie-run jobs require direct access to the job 
tracker and name node?

Thanks,
Evan Pollan

Reply via email to