本文小编为大家详细介绍“docker如何搭建kafka集群”，内容详细，步骤清晰，细节处理妥当，希望这篇“docker如何搭建kafka集群”文章能帮助大家解决疑惑，下面跟着小编的思路慢慢深入，一起来学习新知识吧。
一、原生Docker命令
1. 删除所有dangling数据卷（即无用的Volume，僵尸文件）
docker volume rm $(docker volume ls -qf dangling=true)
2. 删除所有dangling镜像（即无tag的镜像）
docker rmi $(docker images | grep "^<none>" | awk "{print $3}"3. 删除所有关闭的容器
docker ps -a | grep Exit | cut -d ' ' -f 1 | xargs docker rm
二、镜像选择
环境为M1版本的mbp:
Zookeeper采用zookeeper
Kafka采用wurstmeister/kafka
Kafka-Manager采用scjtqs/kafka-manager
Mysql采用mysql/mysql-server
三、集群规划
1. 新建docker网络
docker network create docker-net --subnet 172.20.10.0/16
docker network ls
2. 集群规划
| hostname | Ip addr | port | listener |
|---|---|---|---|
| zook1 | 172.20.10.11 | 2184:2181 | |
| zook2 | 172.20.10.12 | 2185:2181 | |
| zook3 | 172.20.10.13 | 2186:2181 | |
| kafka1 | 172.20.10.14 | 内部9092:9092，外部9192:9192 | kafka1 |
| kafka2 | 172.20.10.15 | 内部9093:9093，外部9193:9193 | kafka2 |
| Kafka3 | 172.20.10.16 | 内部9094:9094，外部9194:9194 | Kafka3 |
| 本机（宿主机Mbp） | 172.20.10.2 | ||
| kafka manager | 172.20.10.10 | 9000:9000 |
四、Zookeeper集群安装
1. 新建文件zk-docker-compose.yml
version: '3.4' services: zook1: image: zookeeper:latest restart: always hostname: zook1 container_name: zook1 #容器名称，方便在rancher中显示有意义的名称 ports: - 2183:2181 #将本容器的zookeeper默认端口号映射出去 volumes: # 挂载数据卷 - "/Users/konsy/Development/volume/zkcluster/zook1/data:/data" - "/Users/konsy/Development/volume/zkcluster/zook1/datalog:/datalog" - "/Users/konsy/Development/volume/zkcluster/zook1/logs:/logs" environment: ZOO_MY_ID: 1 #即是zookeeper的节点值，也是kafka的brokerid值 ZOO_SERVERS: server.1=zook1:2888:3888;2181 server.2=zook2:2888:3888;2181 server.3=zook3:2888:3888;2181 networks: docker-net: ipv4_address: 172.20.10.11 zook2: image: zookeeper:latest restart: always hostname: zook2 container_name: zook2 #容器名称，方便在rancher中显示有意义的名称 ports: - 2184:2181 #将本容器的zookeeper默认端口号映射出去 volumes: - "/Users/konsy/Development/volume/zkcluster/zook2/data:/data" - "/Users/konsy/Development/volume/zkcluster/zook2/datalog:/datalog" - "/Users/konsy/Development/volume/zkcluster/zook2/logs:/logs" environment: ZOO_MY_ID: 2 #即是zookeeper的节点值，也是kafka的brokerid值 ZOO_SERVERS: server.1=zook1:2888:3888;2181 server.2=zook2:2888:3888;2181 server.3=zook3:2888:3888;2181 networks: docker-net: ipv4_address: 172.20.10.12 zook3: image: zookeeper:latest restart: always hostname: zook3 container_name: zook3 #容器名称，方便在rancher中显示有意义的名称 ports: - 2185:2181 #将本容器的zookeeper默认端口号映射出去 volumes: - "/Users/konsy/Development/volume/zkcluster/zook3/data:/data" - "/Users/konsy/Development/volume/zkcluster/zook3/datalog:/datalog" - "/Users/konsy/Development/volume/zkcluster/zook3/logs:/logs" environment: ZOO_MY_ID: 3 #即是zookeeper的节点值，也是kafka的brokerid值 ZOO_SERVERS: server.1=zook1:2888:3888;2181 server.2=zook2:2888:3888;2181 server.3=zook3:2888:3888;2181 networks: docker-net: ipv4_address: 172.20.10.13 networks: docker-net: external: name: docker-net
2. 执行脚本部署zookeeper至Docker:
docker compose -f ./zk-docker-compose.yml up -d
五、Kafka集群安装
1. 新建文件kafka-docker-compose.yml
version: '2' services: kafka1: image: docker.io/wurstmeister/kafka restart: always hostname: kafka1 container_name: kafka1 ports: - 9093:9093 - 9193:9193 environment: KAFKA_BROKER_ID: 1 KAFKA_LISTENERS: INSIDE://:9093,OUTSIDE://:9193 #KAFKA_ADVERTISED_LISTENERS=INSIDE://<container>:9092,OUTSIDE://<host>:9094 SKAFKA_ADVERTISED_LISTENERS: INSIDE://kafka1:9093,OUTSIDE://localhost:9193 KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: INSIDE:PLAINTEXT,OUTSIDE:PLAINTEXT KAFKA_INTER_BROKER_LISTENER_NAME: INSIDE KAFKA_ZOOKEEPER_CONNECT: zook1:2181,zook2:2181,zook3:2181 ALLOW_PLAINTEXT_LISTENER : 'yes' JMX_PORT: 9999 #开放JMX监控端口，来监测集群数据 volumes: - /Users/konsy/Development/volume/kafka/kafka1/wurstmeister/kafka:/wurstmeister/kafka - /Users/konsy/Development/volume/kafka/kafka1/kafka:/kafka external_links: - zook1 - zook2 - zook3 networks: docker-net: ipv4_address: 172.20.10.14 kafka2: image: docker.io/wurstmeister/kafka restart: always hostname: kafka2 container_name: kafka2 ports: - 9094:9094 - 9194:9194 environment: KAFKA_BROKER_ID: 2 KAFKA_LISTENERS: INSIDE://:9094,OUTSIDE://:9194 #KAFKA_ADVERTISED_LISTENERS=INSIDE://<container>:9092,OUTSIDE://<host>:9094 KAFKA_ADVERTISED_LISTENERS: INSIDE://kafka2:9094,OUTSIDE://localhost:9194 KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: INSIDE:PLAINTEXT,OUTSIDE:PLAINTEXT KAFKA_INTER_BROKER_LISTENER_NAME: INSIDE KAFKA_ZOOKEEPER_CONNECT: zook1:2181,zook2:2181,zook3:2181 ALLOW_PLAINTEXT_LISTENER : 'yes' JMX_PORT: 9999 #开放JMX监控端口，来监测集群数据 volumes: - /Users/konsy/Development/volume/kafka/kafka2/wurstmeister/kafka:/wurstmeister/kafka - /Users/konsy/Development/volume/kafka/kafka2/kafka:/kafka external_links: - zook1 - zook2 - zook3 networks: docker-net: ipv4_address: 172.20.10.15 kafka3: image: docker.io/wurstmeister/kafka restart: always hostname: kafka3 container_name: kafka3 ports: - 9095:9095 - 9195:9195 environment: KAFKA_BROKER_ID: 3 KAFKA_LISTENERS: INSIDE://:9095,OUTSIDE://:9195 #KAFKA_ADVERTISED_LISTENERS=INSIDE://<container>:9092,OUTSIDE://<host>:9094 KAFKA_ADVERTISED_LISTENERS: INSIDE://kafka3:9095,OUTSIDE://localhost:9195 KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: INSIDE:PLAINTEXT,OUTSIDE:PLAINTEXT KAFKA_INTER_BROKER_LISTENER_NAME: INSIDE KAFKA_ZOOKEEPER_CONNECT: zook1:2181,zook2:2181,zook3:2181 ALLOW_PLAINTEXT_LISTENER : 'yes' JMX_PORT: 9999 #开放JMX监控端口，来监测集群数据 volumes: - /Users/konsy/Development/volume/kafka/kafka3/wurstmeister/kafka:/wurstmeister/kafka - /Users/konsy/Development/volume/kafka/kafka3/kafka:/kafka external_links: - zook1 - zook2 - zook3 networks: docker-net: ipv4_address: 172.20.10.16 networks: docker-net: external: name: docker-net
2. 执行脚本部署kafka至Docker:
docker compose -f ./kafka-docker-compose.yml up -d
3. listeners 和 advertised.listeners
listeners: 学名叫监听器，其实就是告诉外部连接者要通过什么协议访问指定主机名和端口开放的 Kafka 服务。
advertised.listeners：和 listeners 相比多了个 advertised。Advertised 的含义表示宣称的、公布的，就是说这组监听器是 Broker 用于对外发布的。
比如说：
listeners: INSIDE://172.17.0.10:9092,OUTSIDE://172.17.0.10:9094 advertised_listeners: INSIDE://172.17.0.10:9092,OUTSIDE://<公网 ip>:端口 kafka_listener_security_protocol_map: "INSIDE:SASL_PLAINTEXT,OUTSIDE:SASL_PLAINTEXT" kafka_inter_broker_listener_name: "INSIDE"
advertised_listeners 监听器会注册在 zookeeper 中；
当我们对 172.17.0.10:9092 请求建立连接，kafka 服务器会通过 zookeeper 中注册的监听器，找到 INSIDE 监听器，然后通过 listeners 中找到对应的 通讯 ip 和 端口；
同理，当我们对 <公网 ip>:端口 请求建立连接，kafka 服务器会通过 zookeeper 中注册的监听器，找到 OUTSIDE 监听器，然后通过 listeners 中找到对应的 通讯 ip 和 端口 172.17.0.10:9094；
总结：advertised_listeners 是对外暴露的服务端口，真正建立连接用的是 listeners。
4.安装kafka-manager
4.1 新建文件kafka-manager-docker-compose.yml
version: '2' services: kafka-manager: image: scjtqs/kafka-manager:latest restart: always hostname: kafka-manager container_name: kafka-manager ports: - 9000:9000 external_links: # 连接本compose文件以外的container - zook1 - zook2 - zook3 - kafka1 - kafka2 - kafka3 environment: ZK_HOSTS: zook1:2181,zook2:2181,zook3:2181 KAFKA_BROKERS: kafka1:9093,kafka2:9094,kafka3:9095 APPLICATION_SECRET: letmein KM_ARGS: -Djava.net.preferIPv4Stack=true networks: docker-net: ipv4_address: 172.20.10.10 networks: docker-net: external: name: docker-net
4.2 执行脚本部署kafka-manager至Docker:
docker compose -f ./kafka-manager-docker-compose.yml up -d
4.3 配置Cluster
5. 测试kafka
5.1连接容器
进入kafka容器的命令行：
docker exec -ti kafka /bin/bash
进入kafka所在目录：
cd opt/kafka_2.13-2.8.1/
5.2 创建Replication为2，Partition为2的topic
bin/kafka-topics.sh --create --zookeeper zook1:2181 --replication-factor 2 --partitions 2 --topic partopic
5.3 查看topic的状态
在kafka容器中的 opt/kafka_2.12-1.1.0/ 目录下输入
bin/kafka-topics.sh --describe --zookeeper zook1:2181 --topic partopic