Integrate with hadoop
Hadoop is an optional component of zeppelin unless you need the following features
- Use hdfs to store notes.
- Use hdfs to store interpreter configuration
- Use hdfs to store recovery data
- Launch interpreter in yarn mode
Zeppelin 0.9 doesn't ship with hadoop dependencies, you need to include hadoop jars by yourself via the following steps
- Hadoop client (both 2.x and 3.x are supported) is installed.
$HADOOP_HOME/binis put in
PATH. Because internally zeppelin will run command
hadoop classpathto get all the hadoop jars and put them in the classpath of Zeppelin.