kafka简述与集群配置

一、kafka简述

 1、简介

kafka是一个高吞吐的分布式消息队列系统。特点是生产者消费者模式，先进先出（FIFO）保证顺序，自己不丢数据，默认每隔7天清理数据。消息列队常见场景：系统之间解耦合、峰值压力缓冲、异步通信。

2、集群介绍

（1）Kafka架构是由producer（消息生产者）、consumer（消息消费者）、borker(kafka集群的server，负责处理消息读、写请求，存储消息，在kafka cluster这一层这里，其实里面是有很多个broker)、topic（消息队列/分类相当于队列，里面有生产者和消费者模型）、zookeeper(元数据信息存在zookeeper中，包括：存储消费偏移量，topic话题信息，partition信息) 这些部分组成。

（2）kafka里面的消息是有topic来组织的，简单的我们可以想象为一个队列，一个队列就是一个topic，然后它把每个topic又分为很多个partition，这个是为了做并行的，在每个partition内部消息强有序，相当于有序的队列，其中每个消息都有个序号offset，比如0到12，从前面读往后面写。一个partition对应一个broker，一个broker可以管多个partition，比如说，topic有6个partition，有两个broker，那每个broker就管3个partition。这个partition可以很简单想象为一个文件，当数据发过来的时候它就往这个partition上面append，追加就行，消息不经过内存缓冲，直接写入文件，kafka和很多消息系统不一样，很多消息系统是消费完了我就把它删掉，而kafka是根据时间策略删除，而不是消费完就删除，在kafka里面没有一个消费完这么个概念，只有过期这样一个概念。

（3）producer自己决定往哪个partition里面去写，这里有一些的策略，譬如如果hash，不用多个partition之间去join数据了。consumer自己维护消费到哪个offset，每个consumer都有对应的group，group内是queue消费模型（各个consumer消费不同的partition，因此一个消息在group内只消费一次），group间是publish-subscribe消费模型，各个group各自独立消费，互不影响，因此一个消息在被每个group消费一次。

3、leader负载均衡机制

当一个broker停止或者crashes时，所有本来将它作为leader的分区将会把leader转移到其他broker上去，极端情况下，会导致同一个leader管理多个分区，导致负载不均衡，同时当这个broker重启时，如果这个broker不再是任何分区的leader,kafka的client也不会从这个broker来读取消息，从而导致资源的浪费。

kafka中有一个被称为优先副本（preferred replicas）的概念。如果一个分区有3个副本，且这3个副本的优先级别分别为0,1,2，根据优先副本的概念，0会作为leader 。当0节点的broker挂掉时，会启动1这个节点broker当做leader。当0节点的broker再次启动后，会自动恢复为此partition的leader。不会导致负载不均衡和资源浪费，这就是leader的均衡机制。

在配置文件conf/ server.properties中配置开启（默认就是开启）：

auto.leader.rebalance.enable true

二、集群配置

1、zookeeper安装与配置

（1）下载并解压

    去下载https://mirrors.tuna.tsinghua.edu.cn/apache/zookeeper/

   在node01 /opt/bigdata/下 解压 tar -zxvf zookeeper-3.4.6.tar.gz

（2）编辑配置

         配置hosts vim /etc/hosts

1. 192.168.172.73 node03 2. 192.168.172.72 node02 3. 192.168.172.71 node01 

      配置zookeeper环境变量

1. export ZOOKEEPER_HOME=/opt/bigdata/zookeeper-3.4.6 #zookeeper安装路径 
    
2. export PATH=$ZOOKEEPER_HOME/bin:$PATH
    

      配置zoo.cfg

在/opt/bigdata下，复制cp zookeeper-3.4.5/conf/zoo_sample.cfg  zookeeper-3.4.5/conf/zoo.cfg 

编辑：vim zookeeper-3.4.5/conf/zoo.cfg 

1. # The number of milliseconds of each tick 2.  tickTime=2000 3. # The number of ticks that the initial  4. # synchronization phase can take 5.  initLimit=10 6. # The number of ticks that can pass between  7. # sending a request and getting an acknowledgement 8.  syncLimit=5 9. # the directory where the snapshot is stored. 10. # do not use /tmp for storage, /tmp here is just  11. # example sakes. 12.  dataDir=/opt/bigdata/data/zookeeper/zkdata #zookeeper数据存放路径 13.  dataLogDir=/opt/bigdata/data/zookeeper/zkdatalog #zookeeper日志存放路径 14. # the port at which the clients will connect 15.  clientPort=2181        ##zookeeper对外通信端口 17.  server.1=node01:2888:3888 18.  server.2=node02:2888:3888 19.  server.3=node03:2888:3888 20. # 21. # Be sure to read the maintenance section of the  22. # administrator guide before turning on autopurge. 23. # 24. # http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance 25. # 26. # The number of snapshots to retain in dataDir 27. #autopurge.snapRetainCount=3 28. # Purge task interval in hours 29. # Set to "0" to disable auto purge feature 30. #autopurge.purgeInterval=1 

分别在node01、node02、node03下/opt/bigdata/data/zookeeper/zkdata

vim myid 新建myid文件，内容分别为1、2、3保存

（3）节点分发

在node01 /opt/bigdata下 scp远程复制，分别分发到node02、node03对应目录下

1.  scp -r zookeeper-3.4.6 node02:`pwd` 2.  scp -r zookeeper-3.4.6 node03:`pwd`
    

（4）启动zookeeper集群

分别在node01、node02、node03下执行 zkServer.sh start命令启动zookeeper

稍等片刻，分别在node01、node02、node03下执行zkServer.sh status命令，查看状态

1.  [root@node01 ~]# zkServer.sh status 2.  JMX enabled by default 3.  Using config: /opt/bigdata/zookeeper-3.4.5/bin/../conf/zoo.cfg 4.  Mode: leader 5.  [root@node02 bigdata]# zkServer.sh status 6.  JMX enabled by default 7.  Using config: /opt/bigdata/zookeeper-3.4.5/bin/../conf/zoo.cfg 8.  Mode: follower 9.  [root@node03 ~]# zkServer.sh status 10.  JMX enabled by default 11.  Using config: /opt/bigdata/zookeeper-3.4.5/bin/../conf/zoo.cfg 12.  Mode: follower
    

3、kafka安装与配置

（1）下载并解压

wget http://mirror.bit.edu.cn/apache/kafka/1.1.0/kafka_2.11-1.1.0.tgz 去下载

在node01上  /opt/bigdata/下 解压

tar zxvf kafka_2.11-1.1.0.tgz  

（2）编辑配置

在/opt/bigdata/下 vim kafka_2.11-1.1.0/config/server.properties编辑配置

这里重点修改三个参数broker.id标识本机、log.dirs是kafka接收消息存放路径、

zookeeper.connect指定连接的zookeeper集群地址

其他参数保持默认即可，也可自己根据情况修改

1. ############################# Server Basics ############################# 2. # The id of the broker. This must be set to a unique integer for each broker. 3. broker.id=1 4. ############################# Socket Server Settings ############################# 5. listeners=PLAINTEXT://:9092 7. # The port the socket server listens on 8. #port=9092 10. # Hostname the broker will bind to. If not set, the server will bind to all interfaces 11. #host.name=localhost 13. # Hostname the broker will advertise to producers and consumers. If not set, it uses the 14. # value for "host.name" if configured.  Otherwise, it will use the value returned from 15. # java.net.InetAddress.getCanonicalHostName(). 16. #advertised.host.name=<hostname routable by clients> 18. # The port to publish to ZooKeeper for clients to use. If this is not set, 19. # it will publish the same port that the broker binds to. 20. #advertised.port=<port accessible by clients> 22. # The number of threads handling network requests 23. num.network.threads=3 25. # The number of threads doing disk I/O 26. num.io.threads=8 28. # The send buffer (SO_SNDBUF) used by the socket server 29. socket.send.buffer.bytes=102400 31. # The receive buffer (SO_RCVBUF) used by the socket server 32. socket.receive.buffer.bytes=102400 34. # The maximum size of a request that the socket server will accept (protection against OOM) 35. socket.request.max.bytes=104857600 37. ############################# Log Basics ############################# 38. # A comma seperated list of directories under which to store log files 39. log.dirs=/opt/bigdata/kafka_2.11-1.1.0/kafka-logs 41. # The default number of log partitions per topic. More partitions allow greater 42. # parallelism for consumption, but this will also result in more files across 43. # the brokers. 44. num.partitions=1 46. # The number of threads per data directory to be used for log recovery at startup and flushing at shutdown. 47. # This value is recommended to be increased for installations with data dirs located in RAID array. 48. num.recovery.threads.per.data.dir=1 50. ############################# Log Flush Policy ############################# 51. # The number of messages to accept before forcing a flush of data to disk 52. #log.flush.interval.messages=10000 54. # The maximum amount of time a message can sit in a log before we force a flush 55. #log.flush.interval.ms=1000 57. ############################# Log Retention Policy ############################# 58. # The following configurations control the disposal of log segments. The policy can 59. # be set to delete segments after a period of time, or after a given size has accumulated. 60. # A segment will be deleted whenever *either* of these criteria are met. Deletion always happens 61. # from the end of the log. 63. # The minimum age of a log file to be eligible for deletion 64. log.retention.hours=168 66. # A size-based retention policy for logs. Segments are pruned from the log as long as the remaining 67. # segments don't drop below log.retention.bytes. 68. #log.retention.bytes=1073741824 70. # The maximum size of a log segment file. When this size is reached a new log segment will be created. 71. log.segment.bytes=1073741824 73. # The interval at which log segments are checked to see if they can be deleted according  74. # to the retention policies 75. log.retention.check.interval.ms=300000 77. # If log.cleaner.enable=true is set the cleaner will be enabled and individual logs can then be marked for log compaction. 78. log.cleaner.enable=false 80. ############################# Zookeeper ############################# 81. zookeeper.connect=node01:2181,node02:2181,node03:2181 83. # Timeout in ms for connecting to zookeeper 84. zookeeper.connection.timeout.ms=6000 

（3）节点分发

在 /opt/bigdata下 远程复制到node01、node02对应路径下，当然复制时需要ssh免登录

1.  scp -r  kafka_2.11-1.1.0 node02:`pwd` 2.  scp -r  kafka_2.11-1.1.0 node03:`pwd`
    

分别修改server.properties对应的broker.id为2、3即可

（4）启动kafka集群

kafka集群启动前要启动zookeeper集群，若zookeeper集群没启动，首先启动

在/opt/bigdata下 ，三个节点分别执行如下命令，启动kafka集群

./kafka_2.11-1.1.0/bin/kafka-server-start.sh -daemon ./kafka_2.11-1.1.0/config/server.properties &

（5）基本操作

1)、创建topic

./kafka_2.11/bin/kafka-topics.sh --create --zookeeper node02:2181,node03:2181,node04:2181 --replication-factor 3 --partitions 6 --topic kfk_test

2)、列出创建的topic

./kafka_2.11/bin/kafka-topics.sh --list --zookeeper node02:2181,node03:2181,node04:2181

3)、生成数据

./kafka_2.11/bin/kafka-console-producer.sh -broker-list node02:9092,node03:9092,node04:9092 --topic kfk_test

4)、消费生产数据

1.  kafka 0.9版本之前用zookeeper 2.  ./kafka_2.11/bin/kafka-console-consumer.sh --zookeeper node02:2181,node03:2181,node04:2181 --from-beginning --topic kfk_test 4.  kafka 0.9版本之后不推荐zookeeper方式,仍然支持，但逐渐会被取消，推荐bootstrap-server方式 5.  ./kafka_2.11/bin/kafka-console-consumer.sh --bootstrap-server node02:9092,node03:9092,node04:9092 --from-beginning --topic kfk_test
    

5)、查看指定topic信息

./kafka_2.11/bin/kafka-topics.sh --describe --zookeeper node02:2181,node03:2181,node04:2181 --topic kfk_test

信息如下：

1.  Topic:kfk_test  PartitionCount:6 ReplicationFactor:3 Configs: 2.  Topic: kfk_test     Partition: 0 Leader: 1 Replicas: 1,3,2 Isr: 2,3,1 3.  Topic: kfk_test     Partition: 1 Leader: 2 Replicas: 2,1,3 Isr: 2,3,1 4.  Topic: kfk_test     Partition: 2 Leader: 3 Replicas: 3,2,1 Isr: 2,3,1 5.  Topic: kfk_test     Partition: 3 Leader: 1 Replicas: 1,2,3 Isr: 2,3,1 6.  Topic: kfk_test     Partition: 4 Leader: 2 Replicas: 2,3,1 Isr: 2,3,1 7.  Topic: kfk_test     Partition: 5 Leader: 3 Replicas: 3,1,2 Isr: 2,3,1 

可以看到6个分区，每个分区3个副本

partiton： partion id  分区id
leader：当前负责读写的lead broker id ，就是server.properties的broker.id
replicas：当前partition的所有replication broker  list 
isr：（in-sync replicas）relicas的子集，只包含出于活动状态的broker，离线或挂掉的broker不在此列表

6)、删除指定的topic

删除kafka中的topic

./kafka_2.11/bin/kafka-topics.sh --delete --zookeeper node02:2181,node03:2181,node04:2181 --topic kfk_test

删除zookeeper中的topic信息

1.  rmr /brokers/topics/kfk_test 3.  rmr /config/topics/kfk_test 5.  rmr /admin/delete_topics/kfk_test 7.  rmr /consumers/kfk_test-group 

 删除topic数据相关的目录

rm -rf /var/local/kafka/data/kfk_test*