PackOne: Pack clouds, engines and data services into one light stack.

pip, packone
pip install pk1-jet==0.1.0b1



PackOne is used to simplify big data cluster deployment on the clouds include OpenStack, H3CloudOS, EVCloud and CSTCloud etc. It can bootstrap a cluster from scratch in a few clicks, and materialize the cluster into cloud images to boost following cluster creating procedures. Besides, PackOne can scale-out/in clusters by only one click.

This software is inspired by the "serverless" trend in cloud computing and big data processing, with the ambitions to bridge the IaaS to Apache Ambari seamlessly and coordinate Ambari Services into an elastic high-level workspace. Currently, It support to deploy and scale out Hadoop, Spark, Hive, Neo4j, MongoDB, Kylin, Redis and many other clustering software in several clicks or api callings.


Choose a linux (Centos 7.5 is verified) host which network can reach the target openstack group resource, and create a postgresql db with its information (db_user, db_passwd, db_host, db_port, db_name) collected. Then run:

pip install pk1

pip install -U pip setuptools

pk1 setup --database db_user:db_passwd:db_host:b_port:db_name

Start Service

pk1 start [--listening]

Step 1: Add Cloud (OpenStack, for example)

open, and fill the form like:

Step 2: Bootstrap an Ambari cluster

open, and fill the form like:

Step 3: Materialize/Scale clusters

open, select the target clusters, and click the following materialize.../scale... link:

Step 4: Boost cluster creation

Similar to Step 2, open, but choose a scale whose name without 'boostrap'.


pk1 stop


pk1 uninstall


National key Research Program of China: Scientific Big Data Management System (No.2016YFB1000600)