Setting up a local Tajo cluster

Apache Tajo™ provides two run modes: local mode and fully distributed mode. Here, we explain only the local mode where a Tajo instance runs on a local file system. A local mode Tajo instance can start up with very simple configurations.

First of all, you need to add the environment variables to conf/

# Hadoop home. Required
export HADOOP_HOME= ...

# The java implementation to use.  Required.
export JAVA_HOME= ...

To launch the tajo master, execute

$ $TAJO_HOME/bin/


If you want to how to setup a fully distributed mode of Tajo, please see Cluster Setup.


By default, Catalog server which manages table meta data uses Apache Derby as a persistent storage, and Derby stores data into /tmp/tajo-catalog-${username} directory. But, some operating systems may remove all contents in /tmp when booting up. In order to ensure persistent store of your catalog data, you need to set a proper location of derby directory. To learn Catalog configuration, please refer to Catalog Configuration.