Use Hadoop 2.7 for newer Spark versions with yarn.

Author: lagerspetz

spark/init.sh

@@ -130,7 +130,7 @@ else
       elif [[ "$HADOOP_MAJOR_VERSION" == "2" ]]; then
         wget http://s3.amazonaws.com/spark-related-packages/spark-$SPARK_VERSION-bin-cdh4.tgz
       else
-        wget http://s3.amazonaws.com/spark-related-packages/spark-$SPARK_VERSION-bin-hadoop2.4.tgz
+        wget http://s3.amazonaws.com/spark-related-packages/spark-$SPARK_VERSION-bin-hadoop2.7.tgz
       fi
       if [ $? != 0 ]; then
         echo "ERROR: Unknown Spark version"