Spark 2.3

install-spark2-3/README.md

@@ -7,15 +7,15 @@ This script action is limited at the moment only to basic Spark services in comm
 
 1. Create HDInsight Spark cluster version 3.6 (Spark 2.1.1)
 
-2. Run script action: `install-spark2-3-v01.sh` on this cluster. 
+2. Run script action: `install-spark2-3-v04.sh` on this cluster. 
 
     Go to Azure portal > open cluster blade > open Script Actions tile > click Submit new and follow instructions. The script action is provided in this repository.
 
-    When you are on the "Submit script action" blade, you will see "Bash script URI" field. You need to make sure that the `install-spark2-3-v01.sh` is stored in an Azure Storage Blob, and make the link public.
+    When you are on the "Submit script action" blade, you will see "Bash script URI" field. You need to make sure that the `install-spark2-3-v04.sh` is stored in an Azure Storage Blob, and make the link public.
 
     OR
 
-    You can just add https://raw.githubusercontent.com/hdinsight/script-actions/master/install-spark2-3/install-spark2-3-v01.sh to "Bash script URI".
+    You can just add https://raw.githubusercontent.com/hdinsight/script-actions/master/install-spark2-3/install-spark2-3-v04.sh to "Bash script URI".
 
 3. Update class path in the cluster configuration. 
 

install-spark2-3/install-spark2-3-v01.sh → install-spark2-3/install-spark2-3-v04.sh

@@ -11,12 +11,12 @@ HADOOP_YARN_DIR="$CURRENT_DIR/hadoop-yarn"
 ## Download & Install Binary
 cd "/tmp"
 curl "http://apache.claz.org/spark/$newspark/$newspark-bin-hadoop2.7.tgz" | tar xzf -
-cd "$newspark"
+cd "$newspark-bin-hadoop2.7"
 rm -r "jars/hadoop"* "conf"
 ln -s "$SPARK_CONF_DIR" "conf"
 cd ..
 rm -r "$SPARK_DIR"
-mv "$newspark" "$SPARK_DIR"
+mv "$newspark-bin-hadoop2.7" "$SPARK_DIR"
 
 # Create symlinks
 sudo ln -sfn "$SPARK_DIR/yarn/$newspark-yarn-shuffle.jar" \