使用下面提到的hadoop服务构建一个轻量级的hadoop集群。 这些版本与Hortonworks HDP 2.3.4发行版的版本相匹配。
- Alpine Linux: latest
- Debian: 8.3
- Java: Open JDK 7
- hadoop : 2.7.1
- pig: 0.15.0
- Zookeeper: 3.4.6
- Hbase: 1.1.2
- Kafka: 0.9.0.1
- Kafka manager: latest
- Storm: 0.10.0
- Phoenix: 4.7.0
- Flume: 1.6.0
Note: Hadoop和storm建立在Debian Linux和Alpine Linux上的版本上
Dockerhub: https://hub.docker.com/r/pkeropen3/
- Install Docker
- Install Docker compose
The Docker project is tested on Ubuntu host. There is no guarantee that this will work on Windows. Docker项目在Window和macOS主机上进行了测试. 其中Window的使用docker-machine.
- 从docker hub Pull镜像
- OR
- 单独构建每个镜像
这将单独构建每个镜像。 然后花费一些时间,并且只需要完成一次。
- 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
查看使用正确的标签版本创建的所有镜像docker logs -f _CONTAINER_NAME_
跟踪特定容器的日志
如果需要修改某个镜像,可以单独修改的构建脚本。构建脚本位于每个镜像目录中。
./start.sh
(background process)
OR
docker-compose up
(foreground process)
docker-compose stop
根目录有一个.env文件, 是Window的路径设置.COMPOSE_CONVERT_WINDOWS_PATHS=1, 在Window平台运作docker,不然会报如下错误:
请通过以下README文档 , 获取有关每个镜像的详细信息
- Alpine base README
- Debian base README
- Hadoop README
- Pig README
- Zookeeper README
- Hbase README
- Kafka README
- Storm README
- Flume README
- https://hub.docker.com/r/wurstmeister/kafka/
- https://hub.docker.com/r/sheepkiller/kafka-manager/
- https://hub.docker.com/r/wurstmeister/storm/
- https://hub.docker.com/r/sequenceiq/hadoop-ubuntu/
- http://sookocheff.com/post/docker/containerizing-zookeeper-a-guided-tour/
- http://www.slideshare.net/fabiofumarola1/8a-setup-hbase
- https://docs.docker.com/engine/userguide/intro/
- https://docs.docker.com/compose/
- https://phoenix.apache.org/