This project contains multiple Docker images to run Apache Zeppelin on kuberneters.

We strongly recommand to use datatok helm charts.

Zeppelin interpreters will run on separate Docker containers.

Content

server

Lite image (less than 300MB) based on alpine to run zeppelin without interpreters.

Deps:

• alpine
• java openjdk8
• bash

interpreter-spark

Image based on ubuntu to run Spark zeppelin interpreter. The SPARK_HOME folder must be provided at runtime.

Interpreters:

• spark
• md
• sh

Deps:

• java openjdk-8-jre-headless

interpreter-r

Heavy image (2 GB) based on ubuntu to run interpreters:

• spark
• md
• sh
• python
• r
• jdbc

Deps:

• java openjdk-8-jre-headless

r-packer

Use this image to build conda env pack for dev purposes.

Usage

See ./dev/docker-compose.yaml

Build:

docker build --build-arg=zeppelin_version=0.10.1 -t zep .

Sources:

• https://github.com/apache/zeppelin/blob/v0.10.1/scripts/docker/zeppelin-server/Dockerfile