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.
Lite image (less than 300MB) based on alpine to run zeppelin without interpreters.
Deps:
- alpine
- java openjdk8
- bash
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
Heavy image (2 GB) based on ubuntu to run interpreters:
- spark
- md
- sh
- python
- r
- jdbc
Deps:
- java openjdk-8-jre-headless
Use this image to build conda env pack for dev purposes.
See ./dev/docker-compose.yaml
Build:
docker build --build-arg=zeppelin_version=0.10.1 -t zep .
Sources: