Build a lightweight hadoop cluster with below mentioned hadoop services. These versions match what Hortonworks HDP 2.3.4 distribution has.
forked from: https://github.com/anair-it/hadoop-docker-lite
Dockerhub: https://hub.docker.com/r/anoopnair/
- Alpine Linux: latest
- Debian: 8.3
- Java: Open JDK 7
- hadoop : 2.7.1
- Zookeeper: 3.4.6
- Kafka: 0.9.0.1
- Kafka manager: latest
Note: Hadoop is built on Debian Linux and all others on Alpine Linux
- Install Docker
- Install Docker compose
The Docker project is tested on Ubuntu host. There is no guarantee that this will work on Windows.
- Pull from docker hub OR
- Build each image individually
This will individually build the above mentioned images. This will take sometime and is required to be done only once.
- Get zookeeper IP address using
docker inspect --format='{{.NetworkSettings.IPAddress}}' zookeeper
- Get hbase IP address using
docker inspect --format='{{.NetworkSettings.IPAddress}}' hbase
- Add zookeeper and hbase entries in the host /etc/hosts file
docker images
to see all images created with proper tag versiondocker logs -f _CONTAINER_NAME_
to follow logs for a specific container
If there is inconsistency in the image creation process, run the affected build script only. The build script is in each component directory.
./start.sh
(background process)
OR
docker-compose up
(foreground process)
docker-compose stop
Please go through below READMEs to get detailed information about each image
- https://hub.docker.com/r/wurstmeister/kafka/
- https://hub.docker.com/r/sheepkiller/kafka-manager/
- https://hub.docker.com/r/sequenceiq/hadoop-ubuntu/
- http://sookocheff.com/post/docker/containerizing-zookeeper-a-guided-tour/
- https://docs.docker.com/engine/userguide/intro/
- https://docs.docker.com/compose/