kafka 入门笔记（#1）

单机测试

下载版本，解压

tar -xzf kafka_2.11-0.10.1.1.tgz

cd kafka_2.11-0.10.1.1

启动服务

Kafka用到了Zookeeper ，所以首先要启动zookeeper，先启动一个单实例的zk服务。

bin/zookeeper-server-start.sh config/zookeeper.properties &

启动Kafka 服务

bin/kafka-server-start.sh config/server.properties

创建 topic

创建一个名叫“test”的topic ，只有一个分区，一个副本

bin/kafka-topics.sh --create --zookeeper 127.0.0.1:2181 --replication-factor 1 --partitions 1 --topic test

可以通过list 命令查看已创建的topic：

bin/kafka-topics.sh --list --zookeeper 127.0.0.1:2181

test

发送消息

Kafka 使用一个简单的命令行producer，从文件中或者从标准输入中读取消息并发送到服务端。默认的每条命令将发送一条消息

运行producer并在控制台中输一些消息，这些消息将被发送到服务端：

bin/kafka-console-producer.sh --broker-list 127.0.0.1:9092 --topic test

This is a message /:Enter

This is another message /:Enter

Enter 键为push 一条记录，Ctrl + c 退出发送

启动consumer

Kafka也有一个命令行consumer可以读取消息并输出到标准输出

bin/kafka-console-consumer.sh --zookeeper 127.0.0.1:2181 --topic test --from-beginning

This is a message

This is another message

如果开两个命令终端，通过producer 输入message，consumer 立刻就可以读取出消息

搭建多broker 集群

上面启动了单broker，现在启动3个broker 组成的集群(机器A,B,C)

机器A:192.168.56.129 ;机器B:192.168.56.131 ;机器C:192.168.56.132

配置

ABC 均照前面安装好kafka

zookeeper单例：放到机器A(192.168.56.129:2181)

机器A 配置：

broker.id=0

advertised.host.name=192.168.56.129

zookeeper.connect=localhost:2181

修改机器B 中kafka 的配置文件

broker.id=1

advertised.host.name=192.168.56.131

zookeeper.connect=192.168.56.129:2181

修改机器C 中kafka 的配置文件

broker.id=1

advertised.host.name=192.168.56.132

zookeeper.connect=192.168.56.129:2181

为了避免出现 Should not set log end offset on partition 异常，需要将各个broker 的server.properties 中设置

advertised.host.name=192.168.xxx

启动

机器A上，启动zookeeper

bin/zookeeper-server-start.sh config/zookeeper.properties &

依次启动A，B，C上的 kafka broker

bin/kafka-server-start.sh config/server.properties &

创建一个拥有 3个副本的topic

bin/kafka-topics.sh --create --zookeeper 192.168.56.129:2181 --replication-factor 3 --partitions 1 --topic my-replicated-topic

检测节点信息

bin/kafka-topics.sh --describe --zookeeper 192.168.56.129:2181 --topic my-replicated-topic

Topic:my-replicated-topic	PartitionCount:1	ReplicationFactor:3	Configs:

Topic: my-replicated-topic	Partition: 0	Leader: 0	Replicas: 0,1,2	Isr: 0,1,2

Leader：负责处理消息的读写，leader 是从所有节点中随机选择的

Replicas：列出所有的副本节点，不管节点是否在服务中

Isr：是正在服务的节点

注意：broker 运行了3台，但是Isr检测到一直只有一台，而且无法写入消息，结果发现是防火墙的问题

如 : Replicas: 0,1,2	Isr: 0

如producer 写消息时：Error while fetching metadata with correlation id 11 : {my-replicated-topic2=LEADER_NOT_AVAILABLE}

这时需要把各台机器的 9092 端口打开

vi /etc/sysconfig/iptables

service iptables restart

消息读写

同单例；随意挑选机器，做为producer 向my-replicated-topic 写入消息，挑另外一台机器做consumer ，消费这些消息

bin/kafka-console-producer.sh --broker-list 127.0.0.1:9092 --topic my-replicated-topic

bin/kafka-console-consumer.sh --zookeeper 192.168.56.129:2181 --from-beginning --topic my-replicated-topic

容错能力

上面检测节点信息时，broker.id=0 是作为Leader，现在kill Leader node

在几秒后，就恢复正常，见其他broker 的日志：

	[2017-04-06 16:52:07,601] WARN [ReplicaFetcherThread-0-0], Error in fetch kafka.server.ReplicaFetcherThread$FetchRequest@11acd8a (kafka.server.ReplicaFetcherThread)

java.io.IOException: Connection to 192.168.56.129:9092 (id: 0 rack: null) failed

	at kafka.utils.NetworkClientBlockingOps$.awaitReady$1(NetworkClientBlockingOps.scala:83)

	at kafka.utils.NetworkClientBlockingOps$.blockingReady$extension(NetworkClientBlockingOps.scala:93)

	at kafka.server.ReplicaFetcherThread.sendRequest(ReplicaFetcherThread.scala:248)

	at kafka.server.ReplicaFetcherThread.fetch(ReplicaFetcherThread.scala:238)

	at kafka.server.ReplicaFetcherThread.fetch(ReplicaFetcherThread.scala:42)

	at kafka.server.AbstractFetcherThread.processFetchRequest(AbstractFetcherThread.scala:118)

	at kafka.server.AbstractFetcherThread.doWork(AbstractFetcherThread.scala:103)

	at kafka.utils.ShutdownableThread.run(ShutdownableThread.scala:63)

[2017-04-06 16:52:08,091] INFO Creating /controller (is it secure? false) (kafka.utils.ZKCheckedEphemeral)

[2017-04-06 16:52:08,093] INFO Result of znode creation is: OK (kafka.utils.ZKCheckedEphemeral)

[2017-04-06 16:52:08,094] INFO 1 successfully elected as leader (kafka.server.ZookeeperLeaderElector)

[2017-04-06 16:52:08,312] INFO [ReplicaFetcherManager on broker 1] Removed fetcher for partitions my-replicated-topic2-0 (kafka.server.ReplicaFetcherManager)

[2017-04-06 16:52:08,314] INFO [ReplicaFetcherThread-0-0], Shutting down (kafka.server.ReplicaFetcherThread)

[2017-04-06 16:52:08,315] INFO [ReplicaFetcherThread-0-0], Stopped  (kafka.server.ReplicaFetcherThread)

[2017-04-06 16:52:08,316] INFO [ReplicaFetcherThread-0-0], Shutdown completed (kafka.server.ReplicaFetcherThread)

[2017-04-06 16:52:08,363] INFO New leader is 1 (kafka.server.ZookeeperLeaderElector$LeaderChangeListener)

[2017-04-06 16:52:14,877] INFO Partition [my-replicated-topic,0] on broker 1: Shrinking ISR for partition [my-replicated-topic,0] from 0,1,2 to 1,2 (kafka.cluster.Partition)

再次检测节点信息：

[root@localhost kafka_2.11-0.10.1.1]# bin/kafka-topics.sh --describe --zookeeper 192.168.56.129:2181 --topic my-replicated-topic2

Topic:my-replicated-topic2	PartitionCount:1	ReplicationFactor:3	Configs:

Topic: my-replicated-topic2	Partition: 0	Leader: 1	Replicas: 0,1,2	Isr: 1,2

剩下两台机器，一个生产消息，一个消费消息，都没有问题

按照 http://www.aboutyun.com/thread-12882-1-1.html 思路进行操作。集群从单机多端口，改用多机器演示