Kafka源码分析 Topic与Partition使用

主题与分区

Topic的管理命令

1. 创建Topic

创建主题的命令，创建一个名为test的拥有10个partition，副本因子是3的Topic。

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --create --topic test --partitions 10 --replication-factor 3

2. 查看Topic

查看当前所有可用主题：

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

查看主题信息：

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --describe --topic test

查看多个主题详细信息用“，”隔开：

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --describe --topic test,test2

查找所有包含失效副本的分区：

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --describe --topic test --under-replicated-partitions

查找主题中没有leader副本的分区：

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --describe --topic test --unavailable-partitions

3. 修改Topic

增加分区命令：

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --alter --topic topic-config --partitions 3

不支持减少分区

4. 删除Topic

删除主题的命令

bin/kafka-topics.sh --zookeeper localhost:2181/kafka --delete test --if-exists

手动删除主题顺序执行，

1. 删除zookeeper节点 /config/topics/test
2. 删除zookeeper节点/brokers/topics/test 及其子节点
3. 删除集群中所有与主题test有关的文件 rm -rf /tmp/kafka-logs/test

kafka-topics.sh命令式创建Topic代码阅读分析

创建Topic流程：

使用kafka-topic.shell脚本，其内容是调用的TopicCommand类

#!/bin/bash
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
# 
#    http://www.apache.org/licenses/LICENSE-2.0
# 
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

exec $(dirname $0)/kafka-run-class.sh kafka.admin.TopicCommand "$@"

类的mian方法

def main(args: Array[String]): Unit = {

  val opts = new TopicCommandOptions(args)

  if(args.length == 0)
    CommandLineUtils.printUsageAndDie(opts.parser, "Create, delete, describe, or change a topic.")

  // should have exactly one action
  val actions = Seq(opts.createOpt, opts.listOpt, opts.alterOpt, opts.describeOpt, opts.deleteOpt).count(opts.options.has _)
  if(actions != 1)
    CommandLineUtils.printUsageAndDie(opts.parser, "Command must include exactly one action: --list, --describe, --create, --alter or --delete")

  opts.checkArgs()

  val time = Time.SYSTEM
  val zkClient = KafkaZkClient(opts.options.valueOf(opts.zkConnectOpt), JaasUtils.isZkSecurityEnabled, 30000, 30000,
    Int.MaxValue, time)

  var exitCode = 0
  try {
    if(opts.options.has(opts.createOpt))
      createTopic(zkClient, opts)
    else if(opts.options.has(opts.alterOpt))
      alterTopic(zkClient, opts)
    else if(opts.options.has(opts.listOpt))
      listTopics(zkClient, opts)
    else if(opts.options.has(opts.describeOpt))
      describeTopic(zkClient, opts)
    else if(opts.options.has(opts.deleteOpt))
      deleteTopic(zkClient, opts)
  } catch {
    case e: Throwable =>
      println("Error while executing topic command : " + e.getMessage)
      error(Utils.stackTrace(e))
      exitCode = 1
  } finally {
    zkClient.close()
    Exit.exit(exitCode)
  }

}

实例化一个Topic命令行操作对象，检测参数合法性，构造一个kafkazk的client，分辨操作的类型，这按createTopic(zkClient, opts)。

def createTopic(zkClient: KafkaZkClient, opts: TopicCommandOptions) {
  val topic = opts.options.valueOf(opts.topicOpt)
  val configs = parseTopicConfigsToBeAdded(opts)
  val ifNotExists = opts.options.has(opts.ifNotExistsOpt)
  if (Topic.hasCollisionChars(topic))
    println("WARNING: Due to limitations in metric names, topics with a period ('.') or underscore ('_') could collide. To avoid issues it is best to use either, but not both.")
  val adminZkClient = new AdminZkClient(zkClient)
  try {
    if (opts.options.has(opts.replicaAssignmentOpt)) {
      val assignment = parseReplicaAssignment(opts.options.valueOf(opts.replicaAssignmentOpt))
      adminZkClient.createOrUpdateTopicPartitionAssignmentPathInZK(topic, assignment, configs, update = false)
    } else {
      CommandLineUtils.checkRequiredArgs(opts.parser, opts.options, opts.partitionsOpt, opts.replicationFactorOpt)
      val partitions = opts.options.valueOf(opts.partitionsOpt).intValue
      val replicas = opts.options.valueOf(opts.replicationFactorOpt).intValue
      val rackAwareMode = if (opts.options.has(opts.disableRackAware)) RackAwareMode.Disabled
                          else RackAwareMode.Enforced
      adminZkClient.createTopic(topic, partitions, replicas, configs, rackAwareMode)
    }
    println("Created topic \"%s\".".format(topic))
  } catch  {
    case e: TopicExistsException => if (!ifNotExists) throw e
  }
}

获取topic，解析topic的配置，判断特殊参数if-not-exists，检测topicname，对含有_和.的进行waring告警。实例化一个AdminZkClient对象；检测是否通过参数replica-assignment进行了partition到broker分配的分配列表。如果制订了分区分配就按此方案进行创建。否则走以下分支，检查命令中是否partitions和replication-factor参数，获取参数的值。检查参数disable-rack-aware,考虑创建topic是否考虑机架因素。

此处分析没有设置分配方案的分支

def createTopic(topic: String,
                partitions: Int,
                replicationFactor: Int,
                topicConfig: Properties = new Properties,
                rackAwareMode: RackAwareMode = RackAwareMode.Enforced) {
  val brokerMetadatas = getBrokerMetadatas(rackAwareMode)
  // 生成分区分配方案
  val replicaAssignment = AdminUtils.assignReplicasToBrokers(brokerMetadatas, partitions, replicationFactor)
  createOrUpdateTopicPartitionAssignmentPathInZK(topic, replicaAssignment, topicConfig)
}

获取broker元数据信息

通过zk获取集群上所有的broker的ids列表，zk的path是/brokers/ids，结果会根据ids进行sorted，小的在前。获取在ZK上存储的每一个broker的数据

通过AdminUtils进行assignReplicasToBrokers分配方案的生成。

/**
 * Creates or Updates the partition assignment for a given topic
 * @param topic
 * @param partitionReplicaAssignment
 * @param config
 * @param update
 */
def createOrUpdateTopicPartitionAssignmentPathInZK(topic: String,
                                                   partitionReplicaAssignment: Map[Int, Seq[Int]],
                                                   config: Properties = new Properties,
                                                   update: Boolean = false) {
  validateCreateOrUpdateTopic(topic, partitionReplicaAssignment, config, update)

  if (!update) {
    // write out the config if there is any, this isn't transactional with the partition assignments
    zkClient.setOrCreateEntityConfigs(ConfigType.Topic, topic, config)
  }

  // create the partition assignment
  writeTopicPartitionAssignment(topic, partitionReplicaAssignment, update)
}

在zk上进行topic和分配方案的创建。

作为 Controller的broker对ZK上的pathroker/topic进行了watch。当zk写入成功时候。会触发监听handler。

case object TopicChange extends ControllerEvent {
  override def state: ControllerState = ControllerState.TopicChange

  override def process(): Unit = {
    if (!isActive) return
    val topics = zkClient.getAllTopicsInCluster.toSet
    val newTopics = topics -- controllerContext.allTopics
    val deletedTopics = controllerContext.allTopics -- topics
    controllerContext.allTopics = topics

    registerPartitionModificationsHandlers(newTopics.toSeq)
    val addedPartitionReplicaAssignment = zkClient.getReplicaAssignmentForTopics(newTopics)
    controllerContext.partitionReplicaAssignment = controllerContext.partitionReplicaAssignment.filter(p =>
      !deletedTopics.contains(p._1.topic))
    controllerContext.partitionReplicaAssignment ++= addedPartitionReplicaAssignment
    info(s"New topics: [$newTopics], deleted topics: [$deletedTopics], new partition replica assignment " +
      s"[$addedPartitionReplicaAssignment]")
    if (addedPartitionReplicaAssignment.nonEmpty)
      onNewPartitionCreation(addedPartitionReplicaAssignment.keySet)
  }
}

/**
 * This callback is invoked by the topic change callback with the list of failed brokers as input.
 * It does the following -
 * 1. Move the newly created partitions to the NewPartition state
 * 2. Move the newly created partitions from NewPartition->OnlinePartition state
 */
private def onNewPartitionCreation(newPartitions: Set[TopicPartition]) {
  info(s"New partition creation callback for ${newPartitions.mkString(",")}")
  partitionStateMachine.handleStateChanges(newPartitions.toSeq, NewPartition)
  replicaStateMachine.handleStateChanges(controllerContext.replicasForPartition(newPartitions).toSeq, NewReplica)
  partitionStateMachine.handleStateChanges(newPartitions.toSeq, OnlinePartition, Option(OfflinePartitionLeaderElectionStrategy))
  replicaStateMachine.handleStateChanges(controllerContext.replicasForPartition(newPartitions).toSeq, OnlineReplica)
}

partitionStateMachine的handleStateChanges

def handleStateChanges(partitions: Seq[TopicPartition], targetState: PartitionState,
                       partitionLeaderElectionStrategyOpt: Option[PartitionLeaderElectionStrategy] = None): Unit = {
  if (partitions.nonEmpty) {
    try {
      controllerBrokerRequestBatch.newBatch()
      doHandleStateChanges(partitions, targetState, partitionLeaderElectionStrategyOpt)
      controllerBrokerRequestBatch.sendRequestsToBrokers(controllerContext.epoch)
    } catch {
      case e: Throwable => error(s"Error while moving some partitions to $targetState state", e)
    }
  }
}

replicaStateMachine的handleStateChanges

  def handleStateChanges(replicas: Seq[PartitionAndReplica], targetState: ReplicaState,
                         callbacks: Callbacks = new Callbacks()): Unit = {
    if (replicas.nonEmpty) {
      try {
        controllerBrokerRequestBatch.newBatch()
        replicas.groupBy(_.replica).map { case (replicaId, replicas) =>
          val partitions = replicas.map(_.topicPartition)
          doHandleStateChanges(replicaId, partitions, targetState, callbacks)
        }
        controllerBrokerRequestBatch.sendRequestsToBrokers(controllerContext.epoch)
      } catch {
        case e: Throwable => error(s"Error while moving some replicas to $targetState state", e)
      }
    }
  }