I've developed a Spark application using the 1.2.0-SNAPSHOP branch that leverages Spark Streaming and Hive and can run it locally with no problem (I need some fixes in the 1.2.0 branch). I successfully launched my EC2 cluster by specifying a git commit hash from the 1.2.0-SNAPSHOT branch as the version (as documented in the scripts) and was able to run non-Hive Spark jobs against it. However, when I try and deploy my Hive application to it, I am getting an exception (java.lang.ArrayStoreException: org.apache.spark.sql.execution.SparkStrategies$CommandStrategy?) creating my HiveContext. I suspect that after the spark-ec2 launch script causes the source code clone of the git repo/commit point, the assembly isn't being built with Hive support (e.g. a -Phive profile on the mvn build command as described in building Spark with Maven in the documentation).
Is anyone aware of a way I can get Hive support added into my compiled assembly on the master/slaves using the spark_ec scripts (or where it would make sense to hack the existing scripts to get it working). Thanks, Curt -- View this message in context: http://apache-spark-user-list.1001560.n3.nabble.com/Problems-launching-1-2-0-SNAPSHOT-cluster-with-Hive-support-on-EC2-tp19221.html Sent from the Apache Spark User List mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@spark.apache.org For additional commands, e-mail: user-h...@spark.apache.org