Apache Kylin a Distributed Analytics Engine to provide SQL interface and OLAP on Hadoop. It was originally developed by eBay Inc. In our previously published guides we discussed how to install Apache Hue on Ubuntu Server. We have separate guides on how to install Apache Hadoop, how to install Ambari, how to install Apache Pig. In This Guide, We Will Show the Steps on How To Install Kylin on Ubuntu 16.04 Server. Server May Be Cloud Server, VPS or Dedicated Server. We can integrate Tableau 8, MS Excel and Power BI, Apache Flink, Apache Hue etc with the total installation. Kylin requires you having access to a Hadoop CLI, with full permissions to HDFS, Hive, hbase and map-reduce. To avoid permission issue, we suggest you using
root account. Obviously you can use Cloudera or Hortonworks distribution instead of our way.
For our this guide, we need Hadoop, Ambari, Hue to be installed. Obviously, doing the whole work manually is time consuming.
How To Install Kylin on Ubuntu 16.04 Server
We can start Hadoop with the below commands :
Ambari homepage will be at
http://your_sandbox_ip:8080. We need Hbase service to be running. Hbase Master port [Default 60000] and Zookeeper [Default 2181] need to be forwarded to Host OS. We can install Kylin in two ways – full manual installation and using scripts.
Steps of manual installation is written on official website of Apache Kylin :
That manual way of installation possibly less frequently used. There is Docker image of Kylin too.
Install Kylin on a beefy node that has Hadoop client and all the configuration files. The steps are really few. Download latest Kylin binaries :
KYLIN_HOME pointing to where you have extracted the above download. You can run
bin/check-env.shto print out the detailed information of your environment.
To start Kylin, run :
To stop Kylin, run :
Runtime logs will be at
logs/kylin.log. You will get the web GUI at
http://hostname:7070/kylin. Default username/password is
ADMIN/KYLIN. There will be nothing much. You can follow official guide to get started :
Commonest problem many peoples face is related to permission. I found their old GitHub documents quite useful :