kafka在zookeeper中存储结构

1.topic注册信息

/brokers/topics/[topic] :

存储某个topic的partitions所有分配信息

Schema:

 

{
    "version": "版本编号目前固定为数字1",
    "partitions": {
        "partitionId编号": [
            同步副本组brokerId列表
        ],
        "partitionId编号": [
            同步副本组brokerId列表
        ],
        .......
    }
}

 

Example:

{
"version": 1,
"partitions": {
"0": [1, 2],
"1": [2, 1],
"2": [1, 2],
}
}

说明：紫红色为patitions编号，蓝色为同步副本组brokerId列表

2.partition状态信息

/brokers/topics/[topic]/partitions/[0...N]  其中[0..N]表示partition索引号

/brokers/topics/[topic]/partitions/[partitionId]/state

Schema:

 

{
"controller_epoch": 表示kafka集群中的中央控制器选举次数,
"leader": 表示该partition选举leader的brokerId,
"version": 版本编号默认为1,
"leader_epoch": 该partition leader选举次数,
"isr": [同步副本组brokerId列表]
}

 

 

Example:

 

{
"controller_epoch": 1,
"leader": 2,
"version": 1,
"leader_epoch": 0,
"isr": [2, 1]
}

 

3. Broker注册信息

/brokers/ids/[0...N]

每个broker的配置文件中都需要指定一个数字类型的id(全局不可重复),此节点为临时znode(EPHEMERAL)

Schema:

 

{
"jmx_port": jmx端口号,
"timestamp": kafka broker初始启动时的时间戳,
"host": 主机名或ip地址,
"version": 版本编号默认为1,
"port": kafka broker的服务端端口号,由server.properties中参数port确定
}

 

 

Example:

 

{
"jmx_port": 6061,

"timestamp":"1403061899859"
"version": 1,
"host": "192.168.1.148",
"port": 9092
}

4. Controller epoch:

/controller_epoch -> int (epoch)

此值为一个数字,kafka集群中第一个broker第一次启动时为1，以后只要集群中center controller中央控制器所在broker变更或挂掉，就会重新选举新的center controller，每次center controller变更controller_epoch值就会 + 1;

5. Controller注册信息:

/controller -> int (broker id of the controller)  存储center controller中央控制器所在kafka broker的信息

Schema:

 

{
"version": 版本编号默认为1,
"brokerid": kafka集群中broker唯一编号,
"timestamp": kafka broker中央控制器变更时的时间戳
}

 

 

Example:

 

{
"version": 1,
"brokerid": 3,
"timestamp": "1403061802981"
}

 

Consumer and Consumer group概念: 

a.每个consumer客户端被创建时,会向zookeeper注册自己的信息;
b.此作用主要是为了"负载均衡".
c.同一个Consumer Group中的Consumers，Kafka将相应Topic中的每个消息只发送给其中一个Consumer。
d.Consumer Group中的每个Consumer读取Topic的一个或多个Partitions，并且是唯一的Consumer；
e.一个Consumer group的多个consumer的所有线程依次有序地消费一个topic的所有partitions,如果Consumer group中所有consumer总线程大于partitions数量，则会出现空闲情况;

举例说明：

kafka集群中创建一个topic为report-log   4 partitions 索引编号为0,1,2,3

假如有目前有三个消费者node：注意-->一个consumer中一个消费线程可以消费一个或多个partition.

如果每个consumer创建一个consumer thread线程,各个node消费情况如下，node1消费索引编号为0,1分区，node2费索引编号为2,node3费索引编号为3

如果每个consumer创建2个consumer thread线程，各个node消费情况如下(是从consumer node先后启动状态来确定的)，node1消费索引编号为0,1分区；node2费索引编号为2,3；node3为空闲状态

总结：
从以上可知，Consumer Group中各个consumer是根据先后启动的顺序有序消费一个topic的所有partitions的。

如果Consumer Group中所有consumer的总线程数大于partitions数量，则可能consumer thread或consumer会出现空闲状态。

 

Consumer均衡算法
当一个group中,有consumer加入或者离开时,会触发partitions均衡.均衡的最终目的,是提升topic的并发消费能力.
1) 假如topic1,具有如下partitions: P0,P1,P2,P3
2) 加入group中,有如下consumer: C0,C1
3) 首先根据partition索引号对partitions排序: P0,P1,P2,P3
4) 根据(consumer.id + '-'+ thread序号)排序: C0,C1
5) 计算倍数: M = [P0,P1,P2,P3].size / [C0,C1].size,本例值M=2(向上取整)
6) 然后依次分配partitions: C0 = [P0,P1],C1=[P2,P3],即Ci = [P(i * M),P((i + 1) * M -1)]

6. Consumer注册信息:

每个consumer都有一个唯一的ID(consumerId可以通过配置文件指定,也可以由系统生成),此id用来标记消费者信息.

/consumers/[groupId]/ids/[consumerIdString]

是一个临时的znode,此节点的值为请看consumerIdString产生规则,即表示此consumer目前所消费的topic + partitions列表.

consumerId产生规则：

StringconsumerUuid = null;
    if(config.consumerId!=null && config.consumerId)
      consumerUuid = consumerId;
    else {
      String uuid = UUID.randomUUID()
      consumerUuid = "%s-%d-%s".format(
        InetAddress.getLocalHost.getHostName, System.currentTimeMillis,
        uuid.getMostSignificantBits().toHexString.substring(0,8));

}
     String consumerIdString = config.groupId + "_" + consumerUuid;

Schema:

 

{
"version": 版本编号默认为1,
"subscription": { //订阅topic列表
"topic名称": consumer中topic消费者线程数
},
"pattern": "static",
"timestamp": "consumer启动时的时间戳"
}

 

 

Example:

{
"version": 1,
"subscription": {
"open_platform_opt_push_plus1": 5
},
"pattern": "static",
"timestamp": "1411294187842"
}

7. Consumer owner:

/consumers/[groupId]/owners/[topic]/[partitionId] -> consumerIdString + threadId索引编号

当consumer启动时,所触发的操作:

a) 首先进行"Consumer Id注册";

b) 然后在"Consumer id 注册"节点下注册一个watch用来监听当前group中其他consumer的"退出"和"加入";只要此znode path下节点列表变更,都会触发此group下consumer的负载均衡.(比如一个consumer失效,那么其他consumer接管partitions).

c) 在"Broker id 注册"节点下,注册一个watch用来监听broker的存活情况;如果broker列表变更,将会触发所有的groups下的consumer重新balance.

8. Consumer offset:

/consumers/[groupId]/offsets/[topic]/[partitionId] -> long (offset)

用来跟踪每个consumer目前所消费的partition中最大的offset

此znode为持久节点,可以看出offset跟group_id有关,以表明当消费者组(consumer group)中一个消费者失效,

重新触发balance,其他consumer可以继续消费.

9. Re-assign partitions

/admin/reassign_partitions

{

   "fields":[

      {

         "name":"version",

         "type":"int",

         "doc":"version id"

      },

      {

         "name":"partitions",

         "type":{

            "type":"array",

            "items":{

               "fields":[

                  {

                     "name":"topic",

                     "type":"string",

                     "doc":"topic of the partition to be reassigned"

                  },

                  {

                     "name":"partition",

                     "type":"int",

                     "doc":"the partition to be reassigned"

                  },

                  {

                     "name":"replicas",

                     "type":"array",

                     "items":"int",

                     "doc":"a list of replica ids"

                  }

               ],

            }

            "doc":"an array of partitions to be reassigned to new replicas"

         }

      }

   ]

}

 

Example:

{

  "version": 1,

  "partitions":

     [

        {

            "topic": "Foo",

            "partition": 1,

            "replicas": [0, 1, 3]

        }

     ]            

}

10. Preferred replication election

/admin/preferred_replica_election

{

   "fields":[

      {

         "name":"version",

         "type":"int",

         "doc":"version id"

      },

      {

         "name":"partitions",

         "type":{

            "type":"array",

            "items":{

               "fields":[

                  {

                     "name":"topic",

                     "type":"string",

                     "doc":"topic of the partition for which preferred replica election should be triggered"

                  },

                  {

                     "name":"partition",

                     "type":"int",

                     "doc":"the partition for which preferred replica election should be triggered"

                  }

               ],

            }

            "doc":"an array of partitions for which preferred replica election should be triggered"

         }

      }

   ]

}

 

例子:

 

{

  "version": 1,

  "partitions":

     [

        {

            "topic": "Foo",

            "partition": 1         

        },

        {

            "topic": "Bar",

            "partition": 0         

        }

     ]            

}

11. 删除topics
/admin/delete_topics

Schema:

{ "fields":

    [ {"name": "version", "type": "int", "doc": "version id"},

      {"name": "topics",

       "type": { "type": "array", "items": "string", "doc": "an array of topics to be deleted"}

      } ]

}

 

例子:

{

  "version": 1,

  "topics": ["foo", "bar"]

}

Topic配置

/config/topics/[topic_name]

例子

{

  "version": 1,

  "config": {

    "config.a": "x",

    "config.b": "y",

    ...

  }

}

文章转载来源:http://blog.csdn.net/lizhitao/article/details/23744675