Docker部署双Kafka Broker通信故障排查与求助

API 探路者

2026-8-15

Kafka Docker集群Broker间通信失败及配置错误解决

问题背景

尝试用Docker Compose搭建两个Kafka Broker实现互相通信，但先启动的Broker无法定位另一个Broker，同时使用kafkajs发送消息时提示KafkaJSNonRetriableError: Replication-factor is invalid。

初始Docker Compose配置

version: '3.8'
services:
  zookeeper:
    image: "bitnami/zookeeper:latest"
    ports:
      - "2181:2181"
    environment:
      - ALLOW_ANONYMOUS_LOGIN=yes
      - ZOOKEEPER_CLIENT_PORT=2181
      - ZOOKEEPER_TICK_TIME=2000
    tmpfs: "/datalog"
  kafka-1:
    image: "bitnami/kafka:latest"
    container_name: "kafka-1"
    ports:
      - "9094:9094"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=1
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_LISTENERS=PLAINTEXT://:9094
      - KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9094
      - KAFKA_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes
  kafka-2:
    image: "bitnami/kafka:latest"
    container_name: "kafka-2"
    ports:
      - "9095:9095"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=2
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_LISTENERS=PLAINTEXT://:9095
      - KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9095
      - KAFKA_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes

初始错误日志

kafka-2                      | [[2022-10-24 15:41:47,131] INFO [Controller id=2, targetBrokerId=1] Client requested connection close from node 1 (org.apache.kafka.clients.NetworkClient)
kafka-2                      | [2022-10-24 15:41:47,232] INFO [Controller id=2, targetBrokerId=1] Node 1 disconnected. (org.apache.kafka.clients.NetworkClient)
kafka-2                      | [2022-10-24 15:41:47,232] WARN [Controller id=2, targetBrokerId=1] Connection to node 1 (/127.0.0.1:9094) could not be established. Broker may not be available. (org.apache.kafka.clients.NetworkClient)
kafka-2                      | [2022-10-24 15:41:47,232] WARN [RequestSendThread controllerId=2] Controller 2's connection to broker 127.0.0.1:9094 (id: 1 rack: null) was unsuccessful (kafka.controller.RequestSendThread)
kafka-2                      | java.io.IOException: Connection to 127.0.0.1:9094 (id: 1 rack: null) failed.
kafka-2                      |  at org.apache.kafka.clients.NetworkClientUtils.awaitReady(NetworkClientUtils.java:70)
kafka-2                      |  at kafka.controller.RequestSendThread.brokerReady(ControllerChannelManager.scala:291)
kafka-2                      |  at kafka.controller.RequestSendThread.doWork(ControllerChannelManager.scala:245)
kafka-2                      |  at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:96)

发送消息时同步出现错误：
KafkaJSNonRetriableError: Replication-factor is invalid

修改配置后的新问题

添加内部通信监听器后，新配置如下：

version: '3.8'
services:
  dynamodb-local-express-bp:
    command: "-jar DynamoDBLocal.jar -sharedDb -dbPath ./data"
    image: "amazon/dynamodb-local:latest"
    container_name: dynamodb-local-express-bp
    ports:
      - "8000:8000"
    volumes:
      - "./docker/dynamodb:/home/dynamodblocal/data"
    working_dir: /home/dynamodblocal
  zookeeper:
    image: "bitnami/zookeeper:latest"
    ports:
      - "2181:2181"
    environment:
      - ALLOW_ANONYMOUS_LOGIN=yes
      - ZOOKEEPER_CLIENT_PORT=2181
      - ZOOKEEPER_TICK_TIME=2000
    tmpfs: "/datalog"
  kafka-1:
    image: "bitnami/kafka:latest"
    container_name: "kafka-1"
    expose:
      - 9092
    ports:
      - "9094:9094"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=1
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_CFG_LISTENERS=PLAINTEXT://:9094,INTERNAL://:9092
      - KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9094,INTERNAL://kafka-1:9092
      - KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=PLAINTEXT:PLAINTEXT,INTERNAL:PLAINTEXT
      - KAFKA_CFG_INTER_BROKER_LISTENER_NAME=INTERNAL
      - KAFKA_CFG_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_CFG_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes
  kafka-2:
    image: "bitnami/kafka:latest"
    container_name: "kafka-2"
    ports:
      - "9095:9095"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=2
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_CFG_LISTENERS=PLAINTEXT://:9095,INTERNAL://:9092
      - KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9095,INTERNAL://kafka-1:9092
      - KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=PLAINTEXT:PLAINTEXT,INTERNAL:PLAINTEXT
      - KAFKA_CFG_INTER_BROKER_LISTENER_NAME=INTERNAL
      - KAFKA_CFG_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_CFG_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes
      # - KAFKA_BROKER_ID=2
      # - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      # - KAFKA_LISTENERS=PLAINTEXT://:9095
      # - KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9095
      # - KAFKA_DEFAULT_REPLICATION_FACTOR=2
      # - KAFKA_NUM_PARTITIONS=2
      # - ALLOW_PLAINTEXT_LISTENER=yes

出现启动失败错误：

[2022-10-24 18:59:06,286] ERROR [KafkaServer id=2] Fatal error during KafkaServer startup. Prepare to shutdown (kafka.server.KafkaServer)
kafka-2                      | java.lang.IllegalArgumentException: requirement failed: Configured end points kafka-1:9092 in advertised listeners are already registered by broker 1
kafka-2                      |  at scala.Predef$.require(Predef.scala:281)
kafka-2                      |  at kafka.server.KafkaServer.$anonfun$createBrokerInfo$3(KafkaServer.scala:512)
kafka-2                      |  at kafka.server.KafkaServer.$anonfun$createBrokerInfo$3$adapted(KafkaServer.scala:510)
kafka-2                      |  at scala.collection.mutable.ResizableArray.foreach(ResizableArray.scala:62)
kafka-2                      |  at scala.collection.mutable.ResizableArray.foreach$(ResizableArray.scala:55)
kafka-2                      |  at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:49)
kafka-2                      |  at kafka.server.KafkaServer.createBrokerInfo(KafkaServer.scala:510)
kafka-2                      |  at kafka.server.KafkaServer.startup(KafkaServer.scala:331)
kafka-2                      |  at kafka.Kafka$.main(Kafka.scala:109)
kafka-2                      |  at kafka.Kafka.main(Kafka.scala)

问题解决

1. 初始通信失败原因

初始配置中，KAFKA_ADVERTISED_LISTENERS设置为127.0.0.1:9094/9095，导致Broker在Docker容器内部互相访问时，尝试连接自身容器的127.0.0.1而非另一个Broker的容器地址，无法建立跨容器连接。

2. 修改配置后的错误原因

kafka-2的INTERNAL监听器地址错误设置为kafka-1:9092，与kafka-1的内部注册地址重复，导致ZooKeeper中端点冲突，启动失败。

正确配置方案

调整两个Kafka节点的内部监听器地址，确保每个节点的内部地址唯一，同时保留外部访问的监听器：

version: '3.8'
services:
  dynamodb-local-express-bp:
    command: "-jar DynamoDBLocal.jar -sharedDb -dbPath ./data"
    image: "amazon/dynamodb-local:latest"
    container_name: dynamodb-local-express-bp
    ports:
      - "8000:8000"
    volumes:
      - "./docker/dynamodb:/home/dynamodblocal/data"
    working_dir: /home/dynamodblocal
  zookeeper:
    image: "bitnami/zookeeper:latest"
    ports:
      - "2181:2181"
    environment:
      - ALLOW_ANONYMOUS_LOGIN=yes
      - ZOOKEEPER_CLIENT_PORT=2181
      - ZOOKEEPER_TICK_TIME=2000
    tmpfs: "/datalog"
  kafka-1:
    image: "bitnami/kafka:latest"
    container_name: "kafka-1"
    expose:
      - 9092
    ports:
      - "9094:9094"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=1
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_CFG_LISTENERS=PLAINTEXT://:9094,INTERNAL://:9092
      - KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9094,INTERNAL://kafka-1:9092
      - KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=PLAINTEXT:PLAINTEXT,INTERNAL:PLAINTEXT
      - KAFKA_CFG_INTER_BROKER_LISTENER_NAME=INTERNAL
      - KAFKA_CFG_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_CFG_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes
  kafka-2:
    image: "bitnami/kafka:latest"
    container_name: "kafka-2"
    expose:
      - 9092
    ports:
      - "9095:9095"
    depends_on:
      - zookeeper
    restart: always
    environment:
      - KAFKA_BROKER_ID=2
      - KAFKA_ZOOKEEPER_CONNECT=zookeeper:2181
      - KAFKA_CFG_LISTENERS=PLAINTEXT://:9095,INTERNAL://:9092
      - KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9095,INTERNAL://kafka-2:9092
      - KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=PLAINTEXT:PLAINTEXT,INTERNAL:PLAINTEXT
      - KAFKA_CFG_INTER_BROKER_LISTENER_NAME=INTERNAL
      - KAFKA_CFG_DEFAULT_REPLICATION_FACTOR=2
      - KAFKA_CFG_NUM_PARTITIONS=2
      - ALLOW_PLAINTEXT_LISTENER=yes

关键修改点

将kafka-2的INTERNAL监听器地址改为kafka-2:9092，与kafka-1的kafka-1:9092区分，避免端点冲突。
给kafka-2添加expose: -9092，确保容器内部可访问该端口。
通过INTER_BROKER_LISTENER_NAME指定INTERNAL为Broker间通信的监听器，让Broker通过容器内部地址通信，解决跨容器连接问题。

复制因子错误解决

Broker集群正常启动后，KafkaJSNonRetriableError: Replication-factor is invalid错误会自动消失，此时集群有2个可用Broker，满足default.replication.factor=2的配置要求。

内容的提问来源于stack exchange，提问作者blu

Docker部署双Kafka Broker通信故障排查与求助

问题背景 #

初始Docker Compose配置 #

初始错误日志 #

修改配置后的新问题 #

问题解决 #

1. 初始通信失败原因 #

2. 修改配置后的错误原因 #

正确配置方案 #

关键修改点 #

复制因子错误解决 #