实验环境

主机名/IP	角色	系统
server01/192.168.1.201	controller	ubuntu22.04
server02/192.168.1.202	controller	ubuntu22.04
server03/192.168.1.203	broker	ubuntu22.04
server04/192.168.1.204	broker	ubuntu22.04

All Nodes

安装kafka

apt install openjdk-11-jre-headless -y
wget https://dlcdn.apache.org/kafka/3.8.0/kafka_2.13-3.8.0.tgz
tar -zxvf kafka_2.13-3.8.0.tgz -C /usr/local/etc

server01

修改配置文件

vim /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties

process.roles=controller
node.id=1
controller.quorum.voters=1@192.168.1.201:9093,2@192.168.1.202:9093
listeners=CONTROLLER://192.168.1.201:9093
controller.listener.names=CONTROLLER
listener.security.protocol.map=CONTROLLER:PLAINTEXT

server02

vim /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties

process.roles=controller
node.id=2
controller.quorum.voters=1@192.168.1.201:9093,2@192.168.1.202:9093
listeners=CONTROLLER://192.168.1.202:9093
controller.listener.names=CONTROLLER
listener.security.protocol.map=CONTROLLER:PLAINTEXT

server03

vim /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties

process.roles=broker
node.id=3
controller.quorum.voters=1@192.168.1.201:9093,2@192.168.1.202:9093
listeners=PLAINTEXT://192.168.1.203:9092
inter.broker.listener.name=PLAINTEXT
controller.listener.names=CONTROLLER
listener.security.protocol.map=CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT

server04

vim /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties

process.roles=broker
node.id=4
controller.quorum.voters=1@192.168.1.201:9093,2@192.168.1.202:9093
listeners=PLAINTEXT://192.168.1.204:9092
inter.broker.listener.name=PLAINTEXT
controller.listener.names=CONTROLLER
listener.security.protocol.map=CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT

server01

生成uuid

/usr/local/etc/kafka_2.13-3.8.0/bin/kafka-storage.sh random-uuid

# 这个生成的uuid后面要用在其他节点上
viJ79TulTXaN7yG53fdp7w

All Nodes

初始化集群数据目录

/usr/local/etc/kafka_2.13-3.8.0/bin/kafka-storage.sh format -t viJ79TulTXaN7yG53fdp7w -c /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties
/usr/local/etc/kafka_2.13-3.8.0/bin/kafka-server-start.sh  /usr/local/etc/kafka_2.13-3.8.0/config/kraft/server.properties

检查节点状态

/usr/local/etc/kafka_2.13-3.8.0/bin/kafka-metadata-quorum.sh --bootstrap-server 192.168.1.203:9092 describe --status
/usr/local/etc/kafka_2.13-3.8.0/bin/kafka-metadata-quorum.sh --bootstrap-server 192.168.1.204:9092 describe --status

ClusterId:              viJ79TulTXaN7yG53fdp7w
LeaderId:               1
LeaderEpoch:            2
HighWatermark:          4014
MaxFollowerLag:         0
MaxFollowerLagTimeMs:   0
CurrentVoters:          [1,2]
CurrentObservers:       [3,4]

输出信息表示：

1. ClusterId：集群的ID
2. LeaderId：当前的控制器领导者节点的 ID。在这个例子中，领导者节点 ID 是 1
3. LeaderEpoch：显示当前领导者的任期号。任期号每次选举领导者时都会增加
4. HighWatermark：表示所有副本已确认的最高位移
5. MaxFollowerLag：显示最大跟随者的滞后量。这里是 0，表示跟随者与领导者同步
6. MaxFollowerLagTimeMs：显示最大跟随者滞后的时间(ms)
7. CurrentVoters：列出当前参与投票的节点 ID
8. CurrentObservers：列出当前的观察者节点 ID。观察者不会参与投票，仅用于同步数据