How to setup custom Spark parameter in HDInsights

问题:

I am creating HDInsights cluster on Azure according to this desciption

Now I would like to set up spark custom parameter, for example spark.yarn.appMasterEnv.PYSPARK3_PYTHON or spark_daemon_memory in time of cluster provisioning.

Is it possible to setup using Data Factory/Automation Account? I can not find any example doing this.

Thanks

回答1:

You can use SparkConfig in Data Factory to pass these configurations to Spark.

For example:

"typeProperties": {
...
  "sparkConfig": {
     "spark.submit.pyFiles": "/dist/package_name-1.0.0-py3.5.egg",
     "spark.yarn.appMasterEnv.PYSPARK_PYTHON": "/usr/bin/anaconda/envs/py35/bin/python3"
  }
}

This way you can specify all Spark configs that are listed in docs here.