执行重选leader脚本

 选举前：{ \"controller_epoch\

选举后：{ \"controller_epoch\

   1、根据相应的策略选出leader   2、将选举的结果写入zk中，对应节点为brokers/topics/{topic}/partitions/{partition}/state   3、更新controllerContext中的partitionLeadershipInfo信息   4、给每个还存活的副本发送leaderAndIsr请求   5、给集群中每个broker发送updateMetadata请求

controller

watch

leaderElectionClient

LeaderElection脚本

topic_1-0

state

LeaderElectionCommand

LeaderElection脚本：{  \"partitions\": [    {      \"topic\": \"topic_1\

topic_1

kafka cluster

broker1

admin

校验数据

 写入zk节点数据：{  \"partitions\": [    {      \"topic\": \"topic_1\

leaderAndIsrRequest

0

调用electLeaders

partitions

KafkaController

写入

调用分区状态机处理分区状态

topics

{ \"version\

zk

brokers

broker3

broker2

preferred_replica_election

write