Umesh Kacha created SPARK-6995:
----------------------------------
Summary: Spark driver program throws exception cant get Master
Kerberose principal for use as renewer
Key: SPARK-6995
URL: https://issues.apache.org/jira/browse/SPARK-6995
Project: Spark
Issue Type: Question
Components: Java API
Affects Versions: 1.1.0
Environment: Linux,IntelliJ
Reporter: Umesh Kacha
am having thousands of files residing in HDFS which I would like to
read/process/store back to HDFS using Spark. I have the following Java Spark
driver program:
SparkConf sparkConf = new
SparkConf().setAppName("MyApp").setMaster("spark://myhost1.com");
JavaSparkContext sc = new JavaSparkContext(sparkConf);
JavaRDD<MySchemaPOJO> schemaRDD = sc.textFile("hdfs://path/to/file").map(
new Function<String,MySchemaPOJO>() {
public MySchemaPOJO call(String line) throws Exception {
String[] fields = line.split("\\t");
MySchemaPOJO schemaPOJO = new MySchemaPOJO();
//set required fields on POJO
return schemaPOJO;
}
});
System.out.println("POJO list are " + schemaRDD.collect());
schemaRDD.saveAsTextFile("hdfs://path/to/save/file");
sc.stop();
I have started Master/Worker on myhost1.com and I am running this driver
program on IntelliJ on another host myhost2.com. When I start driver program
everything works fine until schemaRDD.collect() and it throws exception as
Exception in thread 'main' java.io.IOException : Cant get Master
Kerberose principal to use as renewer.
I did kinit on both host where I am running driver and master/worker. My driver
program can connect to Master also only when actual actions gets called on RDD
it throws above exception. Please guide I am new to Spark.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
