n3wtn9/spark-oozie

spark-oozie

experimenting with launching spark jobs with oozie

look in job.properties for variable arguments oozie.wf.application.path=${nameNode}/user/${user.name}/${examplesRoot}/apps/spark

copy the following files to hdfs in oozie's application path

workflow.xml -> workflow.xml oozie-examples.jar -> lib/oozie-examples.jar data.txt -> input-data/text/data.txt

added

spark-assembly-.jar to hdfs:///user/oozie/share/lib/lib_/spark/

added to oozie-site.xml

oozie.service.SparkConfigurationService.spark.configurations spark.yarn.historyServer.address=http://SPH-HOST:18088, spark.eventLog.dir=hdfs://NN:8020/user/spark/applicationHistory, spark.eventLog.enabled=true

to use spark-yarn action in oozie with hdp

remove the invalid classpath in /etc/hadoop/conf/mapred-site.xml under mapreduce.application.classpath :/usr/hdp/${hdp.version}/hadoop/lib/hadoop-lzo-0.6.0.${hdp.version}.jar ref: https://cwiki.apache.org/confluence/display/Hive/Hive+on+Spark%3A+Getting+Started