Setting up and a quick execution of Spark
There are two different ways to set up Spark, build it from the source or download and extract it. Both ways are explained in the following sections.
Building from source
Download the source code from the link, http://spark.apache.org/downloads.html, which is also shown in the following screenshot:
You will require Maven 3.3.6 and Java 7+ to compile Spark 2.1.0. Also, you need to update the MAVEN_OPT
as the default setting will not be able to compile the code:
exportMAVEN_OPTS="-Xmx2g -XX:ReservedCodeCacheSize=512m"
Use the following command to trigger the build. It will compile Spark 2.1.0 with Hadoop Version 2.4.0:
./build/mvn -Pyarn -Phadoop-2.4 -Dhadoop.version=2.4.0 -DskipTests clean package
Downloading Spark
Download the latest version (2.1.0) using the same link (http://spark.apache.org/downloads.html) as given for Building from source section. Select the Spark release version if the user wants to install anything other than the latest version and...