We are testing HDInsight Spark Cluster (YARN) and triying use SparkLauncher (setDeployMode: cluster) to send jobs to the cluster from another machine (backend service application without hadoop or related) out the cluster in the same network
Testing with a pc and another virtual Machine with Hortonworks standalone version works right
In azure hdisight we haven't seem this port open and seems only is used in standalone hadoop installation.But now, we are testing this approach in the HDinsight cluster unsuccessfully.
Someone has used this configuration?
We have tested livy with succes but we'd like analyze this other way.