Kylin persists all data (meta data and cube) in HBase; You may want to export the data sometimes for whatever purposes
(backup, migration, troubleshotting etc); This page describes the steps to do this and also there is a Java app for you to do this easily.


  1. Cleanup unused cubes to save storage space (be cautious on production!): run the following command in hbase CLI:
hbase org.apache.hadoop.util.RunJar /${KYLIN_HOME}/lib/kylin-job-(version).jar org.apache.kylin.job.hadoop.cube.StorageCleanupJob --delete true
  1. List all HBase tables, iterate and then export each Kylin table to HDFS;

  2. Copy the export folder from HDFS to local file system, and then archive it;

  3. (optional) Download the archive from Hadoop CLI to local;

  4. Cleanup the export folder from CLI HDFS and local file system;

Kylin provide the “” (currently only exist in “minicluster” branch) for you to do the
step 2-5 in one run; Please ensure the correct path of “” has been set in the sys env; This Java uses the sandbox config by default;