获取镜像

# 查看可用的稳定版本
sudo docker search flink
sudo docker pull fhueske/flink-sql-client-training-1.7.2:latest
sudo docker pull flink:1.10-scala_2.11
sudo docker pull wurstmeister/zookeeper
sudo docker pull wurstmeister/kafka:0.11.0.1
sudo docker pull elasticsearch:6.7.0
sudo docker image ls |grep flink

Flink环境准备

此次安装是基于Docker进行的，因此你只需要安装了Docker即可。不需要依赖Java、Scala环境、或是IDE。注意：Docker默认配置的资源可能不太够，会导致运行Flink Job时卡死。因此推荐配置Docker资源到3-4GB，3-4个CPU cores。
此次安装的环境使用Docker Compose来安装，包含了所需的各种服务的容器，包括：

• Flink SQL Client：用来提交query，以及可视化结果。
• Flink JobManager 和 TaskManager：用来运行Flink SQL任务。
• Apache Kafka：用来生成输入流和写入结果流。
• Apache Zookeeper：Kafka的依赖项。
• ElasticSearch：用来写入结果。

  服务编排

  mkdir -p /share/flink && cd /share/flink
  rm -rf /share/flink/docker-compose.yml
  vi /share/flink/docker-compose.yml

  编排文件如下：

  version: '2'
  services:
  sql-client:
    image: fhueske/flink-sql-client-training-1.7.2:latest
    depends_on:
      - kafka
      - jobmanager
      - elasticsearch
    environment:
      FLINK_JOBMANAGER_HOST: jobmanager
      ZOOKEEPER_CONNECT: zookeeper
      KAFKA_BOOTSTRAP: kafka
      ES_HOST: elasticsearch
  jobmanager:
    image: flink:1.10-scala_2.11
    hostname: "jobmanager"
    expose:
      - "6123"
    ports:
      - "8081:8081"
    command: jobmanager
    environment:
      - JOB_MANAGER_RPC_ADDRESS=jobmanager
  taskmanager:
    image: flink:1.10-scala_2.11
    expose:
      - "6121"
      - "6122"
    depends_on:
      - jobmanager
    command: taskmanager
    links:
      - jobmanager:jobmanager
    environment:
      - JOB_MANAGER_RPC_ADDRESS=jobmanager
  zookeeper:
    image: wurstmeister/zookeeper
    ports:
      - "2181:2181"
  kafka:
    image: wurstmeister/kafka:0.11.0.1
    ports:
      - "9092"
    depends_on:
      - zookeeper
    environment:
      HOSTNAME_COMMAND: "route -n | awk '/UG[ \t]/{print $$2}'"
      KAFKA_ZOOKEEPER_CONNECT: zookeeper:2181
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
  elasticsearch:
    image: elasticsearch:6.7.0
    environment:
      - cluster.name=docker-cluster
      - bootstrap.memory_lock=true
      - "ES_JAVA_OPTS=-Xms512m -Xmx512m"
      - discovery.type=single-node
    ports:
      - "9200:9200"
      - "9300:9300"
    ulimits:
      memlock:
        soft: -1
        hard: -1
      nofile:
        soft: 65536
        hard: 65536

  服务构建

  cd /share/flink
  sudo docker-compose -f docker-compose.yml build --no-cache # 不带缓存构建（只创建镜像，不会启动容器）
  sudo docker-compose -f docker-compose.yml up -d  # 构建后运行
  sudo docker-compose -f docker-compose.yml up --build # 跟踪方式构建，可用于调试
  sudo docker-compose -f docker-compose.yml stop # 停止
  sudo docker-compose -f docker-compose.yml down # 移除

  验证

  cd /share/flink
  # 查看进程
  sudo docker-compose -f docker-compose.yml ps
  # 进入Flink SQL Client
  sudo docker-compose exec sql-client ./sql-client.sh
  # jobmanager
  sudo docker-compose exec jobmanager /bin/bash
  # 批量WordCoun示例
  ./bin/flink run examples/batch/WordCount.jar
  # 查看日志
  sudo docker logs flink
  sudo docker logs -f -t --tail=50 flink
  # 查看网络
  sudo docker network ls
  sudo docker inspect flink

  参考

  语雀：Flink SQL训练环境搭建
  https://www.yuque.com/studys/gk4weq/rh1ldu