Linux Kafka源码环境搭建

本文主要讲述的是如何搭建Kafka的源码环境，主要针对的Linux操作系统下IntelliJ IDEA编译器，其余操作系统或者IDE可以类推。

1.安装和配置JDK
确认JDK版本至少为1.7，最好是1.8及以上。使用java -version命令来查看当前JDK的版本，示例如下：

lenmom@M1701:~/workspace/software/hadoop-2.7./bin$ java -version
java version "1.8.0_191"
Java(TM) SE Runtime Environment (build 1.8.0_191-b12)
Java HotSpot(TM) -Bit Server VM (build 25.191-b12, mixed mode)

2.下载并安装配置Gradle
下载地址为：https://gradle.org/releases/，本人使用的版本是5.2.1（官方在kafka2.1.0的源码中使用的是4.10.2版本，本人在linux上使用5.2.1版本是成功的，但是在windows上是失败的,如果在windows上进行搭建，建议采用4.10.2版本)。一般只需要将下载的包解压，然后再将$GRADLE_HOME/bin的路径添加到环境变量Path中即可，其中$GRADLE_HOME指的是Gradle的根目录。可以使用gradle -v命令来验证Gradle是否已经配置完成，示例如下：

wget https://downloads.gradle.org/distributions/gradle-5.2.1-bin.zip
unzip gradle-5.2.-bin.zip -C /home/lenmom/software/

配置环境变量

vim /etc/profile

添加以下内容

export  GRADLE_HOME=/home/lenmom/software/grandle-5.2.
export PATH=$GRADLE_HOME/bin:$PATH

保存退出，source /etc/profile使环境变量生效

3.下载并安装配置Scala
下载地址为：http://www.scala-lang.org/download/all.html，目前最新的版本是2.12.8，不过笔者这里使用的版本是2.11.8。如Gradle一样，只需要解压并将$SCALA_HOME/bin的路径添加到环境变量Path即可，其中$SCALA_HOME指的是Scala的根目录。可以使用scala -version命令来验证scala是否已经配置完成，示例如下：

lenmom@M1701:~/workspace/software/hadoop-2.7./bin$ scala -version
Scala code runner version 2.11. -- Copyright -, LAMP/EPFL

4. 构建Kafka源码环境
Kafka下载地址为：http://kafka.apache.org/downloads，目前最新的版本是2.10。将下载的压缩包解压，并在Kafka的根目录执行gradle idea命令进行构建，如果你使用的是Eclipse，则只需采用gradle eclipse命令构建即可。构建细节如下所示：

lenmom@M1701:~/workspace/open-source/kafka-2.1.-src$ gradle idea
Starting a Gradle Daemon (subsequent builds will be faster)

> Configure project :
Building project 'core' with Scala version 2.11.
Building project 'streams-scala' with Scala version 2.11.

> Task :idea
Generated IDEA project at file:///home/lenmom/workspace/open-source/kafka-2.1.0-src/kafka-2.1.0-src.ipr

Deprecated Gradle features were used in this build, making it incompatible with Gradle 6.0.
Use '--warning-mode all' to show the individual deprecation warnings.
See https://docs.gradle.org/5.2.1/userguide/command_line_interface.html#sec:command_line_warnings

BUILD SUCCESSFUL in 19m 51s
 actionable tasks:  executed

之后将Kafka导入到IDEA中即可。不过这样还没有结束，对于IDEA而言，还需要安装Scala插件，在Setting->Plugin中搜索scala并安装，可以参考下图，笔者这里是已经安装好的状态：

5. 修改kafka_source_home/ gradle.properties中的scala版本

vim  /home/lenmom//workspace/open-source/kafka-2.1.0-src/gradle.properties

将其中的scalaVersion改为2.11.8，原来是2.11.11

修改 /home/lenmom/workspace/open-source/kafka-2.1.0-src/gradle/dependencies.gradle scala版本

/home/lenmom/workspace/open-source/kafka-2.1.-src/gradle/dependencies.gradle 

将其中的def defaultScala211Version = '2.11.11'改为def defaultScala211Version = '2.11.8'

如果更改了scalaVersion，需要重新执行gradle idea命令来重新构建。虽然很多时候在操作系统中安装其他版本的Scala也并没有什么问题，比如安装2.12.12版本。但是有些情况下运行Kafka时会出现一些异常，而这些异常却又是由于Scala版本不一致而引起的，比如会出现下面示例中的报错：

[2019-02-10 17:09:21,119] FATAL (kafka.Kafka$)
java.lang.NoSuchMethodError: scala.collection.TraversableOnce.$init$(Lscala/collection/TraversableOnce;)V
at kafka.message.MessageSet.<init>(MessageSet.scala:72)
at kafka.message.ByteBufferMessageSet.<init>(ByteBufferMessageSet.scala:129)
at kafka.message.MessageSet$.<init>(MessageSet.scala:32)
at kafka.message.MessageSet$.<clinit>(MessageSet.scala)
at kafka.server.Defaults$.<init>(KafkaConfig.scala:52)
at kafka.server.Defaults$.<clinit>(KafkaConfig.scala)
at kafka.server.KafkaConfig$.<init>(KafkaConfig.scala:686)
at kafka.server.KafkaConfig$.<clinit>(KafkaConfig.scala)
at kafka.server.KafkaServerStartable$.fromProps(KafkaServerStartable.scala:28)
at kafka.Kafka$.main(Kafka.scala:82)
at kafka.Kafka.main(Kafka.scala)

所以为了省去一些不必要的麻烦，还是建议读者在安装Scala版本之前先查看下Kafka源码中gradle.properties文件中配置的scalaVersion。

6. 配置Kafka源码环境

6.1 在确保了scalaVersion之后，需要将config目录下的log4j.properties文件拷贝到core/src/main/scala目录下，这样可以让Kafka在运行时能够输出日志信息

cp /home/lenmom/workspace/open-source/kafka-2.1.-src/config/log4j.properties  /home/lenmom/workspace/open-source/kafka-2.1.-src/core/src/main/scala/

6.2 配置server.properties文件，一般只需要修改以下一些配置项

# 是否允许topic被删除，设置为true则topic可以被删除，
# 开启这个功能方便Kafka在运行一段时间之后，能够删除一些不需要的临时topic
delete.topic.enable=true
# 禁用自动创建topic的功能
auto.create.topics.enable=false
# 存储log文件的目录，默认值为/tmp/kafka-logs
# 示例是在Windows环境下运行，所以需要修改这个配置，注意这里的双反斜杠。
log.dir=/home/lenmom/workspace/open-source/kafka-2.1.-src/kafka-logs

# 配置kafka依赖的zookeeper路径地址，这里的前提是在本地开启了一个zookeeper的服务
zookeeper.connect=localhost:

6.3 创建kafka日志保存目录

mkdir /home/lenmom/workspace/open-source/kafka-2.1.-src/kafka-logs

6.4 确保zookeeper已经启动

zookeeper的安装与启动请参见本人的前期博文

6.5 启动kafka

配置参数：

Main class:
kafka.Kafka

VM options:
-Dkafka.logs.dir=/home/lenmom/workspace/open-source/kafka-2.1.-src/logs
-Dlog4j.configuration=file:/home/lenmom/workspace/open-source/kafka-2.1.-src/config/log4j.properties

Program arguments:
/home/lenmom/workspace/open-source/kafka-2.1.-src/config/server.properties

Enviroment variables:
JMX_PORT=

Use classpath of module:
Core_main 

这里配置Main class为kafka.Kafka，并制定启动时所需要的配置文件地址，即：config/server.properties。配置JMX_PORT是为了方便搜集Kafka自身的Metrics数据。

如此便可以顺利的运行Kafka服务了（第一次启动时会有一个耗时较长的编译过程），部分启动日志如下：

    zookeeper.session.timeout.ms =
    zookeeper.set.acl = false
    zookeeper.sync.time.ms =
 (kafka.server.KafkaConfig)
[-- ::,] INFO [ThrottledChannelReaper-Fetch]: Starting (kafka.server.ClientQuotaManager$ThrottledChannelReaper)
[-- ::,] INFO [ThrottledChannelReaper-Produce]: Starting (kafka.server.ClientQuotaManager$ThrottledChannelReaper)
[-- ::,] INFO [ThrottledChannelReaper-Request]: Starting (kafka.server.ClientQuotaManager$ThrottledChannelReaper)
[-- ::,] INFO Loading logs. (kafka.log.LogManager)
[-- ::,] INFO Logs loading complete in  ms. (kafka.log.LogManager)
[-- ::,] INFO Starting log cleanup with a period of  ms. (kafka.log.LogManager)
[-- ::,] INFO Starting log flusher with a default period of  ms. (kafka.log.LogManager)
[-- ::,] INFO Awaiting socket connections on 0.0.0.0:. (kafka.network.Acceptor)
[-- ::,] INFO [SocketServer brokerId=] Started  acceptor threads (kafka.network.SocketServer)
[-- ::,] INFO [ExpirationReaper--Produce]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [ExpirationReaper--Fetch]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [ExpirationReaper--DeleteRecords]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [LogDirFailureHandler]: Starting (kafka.server.ReplicaManager$LogDirFailureHandler)
[-- ::,] INFO Creating /brokers/ids/ (is it secure? false) (kafka.zk.KafkaZkClient)
[-- ::,] INFO Result of znode creation at /brokers/ids/ is: OK (kafka.zk.KafkaZkClient)
[-- ::,] INFO Registered broker  at path /brokers/ids/ with addresses: ArrayBuffer(EndPoint(M1701,,ListenerName(PLAINTEXT),PLAINTEXT)) (kafka.zk.KafkaZkClient)
[-- ::,] INFO [ExpirationReaper--topic]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [ExpirationReaper--Heartbeat]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [ExpirationReaper--Rebalance]: Starting (kafka.server.DelayedOperationPurgatory$ExpiredOperationReaper)
[-- ::,] INFO [GroupCoordinator ]: Starting up. (kafka.coordinator.group.GroupCoordinator)
[-- ::,] INFO [GroupMetadataManager brokerId=] Removed  expired offsets in  milliseconds. (kafka.coordinator.group.GroupMetadataManager)
[-- ::,] INFO [GroupCoordinator ]: Startup complete. (kafka.coordinator.group.GroupCoordinator)
[-- ::,] INFO [ProducerId Manager ]: Acquired new producerId block (brokerId:,blockStartProducerId:,blockEndProducerId:) by writing to Zk with path version  (kafka.coordinator.transaction.ProducerIdManager)
[-- ::,] INFO [TransactionCoordinator id=] Starting up. (kafka.coordinator.transaction.TransactionCoordinator)
[-- ::,] INFO [Transaction Marker Channel Manager ]: Starting (kafka.coordinator.transaction.TransactionMarkerChannelManager)
[-- ::,] INFO [TransactionCoordinator id=] Startup complete. (kafka.coordinator.transaction.TransactionCoordinator)
[-- ::,] INFO [/config/changes-event-process-thread]: Starting (kafka.common.ZkNodeChangeNotificationListener$ChangeEventProcessThread)
[-- ::,] INFO [SocketServer brokerId=] Started processors for  acceptors (kafka.network.SocketServer)
[-- ::,] WARN Error while loading kafka-version.properties :null (org.apache.kafka.common.utils.AppInfoParser)
[-- ::,] INFO Kafka version : unknown (org.apache.kafka.common.utils.AppInfoParser)
[-- ::,] INFO Kafka commitId : unknown (org.apache.kafka.common.utils.AppInfoParser)
[-- ::,] INFO [KafkaServer id=] started (kafka.server.KafkaServer)
[-- ::,] WARN Client session timed out, have not heard from server in 4179ms for sessionid 0x1000144f4a50007 (org.apache.zookeeper.ClientCnxn)
[-- ::,] INFO Client session timed out, have not heard from server in 4179ms for sessionid 0x1000144f4a50007, closing socket connection and attempting reconnect (org.apache.zookeeper.ClientCnxn)
[-- ::,] INFO Opening socket connection to server localhost/127.0.0.1:. Will not attempt to authenticate using SASL (unknown error) (org.apache.zookeeper.ClientCnxn)
[-- ::,] INFO Socket connection established to localhost/127.0.0.1:, initiating session (org.apache.zookeeper.ClientCnxn)
[-- ::,] INFO Session establishment complete on server localhost/127.0.0.1:, sessionid = 0x1000144f4a50007, negotiated timeout =  (org.apache.zookeeper.ClientCnxn)

如果出现：

Failed to load class "org.slf4j.impl.StaticLoggerBinder" 错误消息，解决办法为：

下载

log4j-1.2.17.jar
slf4j-api-1.7.25.jar
slf4j-log4j12-1.7.25.jar

并在File->Project Structure->Project Settings>Modules->core->core_main->Dependencies，添加这几个jar包为依赖项。

本人在windows上按照这个操作亦成功搭建起了调试环境

or using command line, which skip runing tests and checkstyle validation:

gradle build -x test -x checkstyleMain -x checkstyleTest -x checkstyleScoverage -x spotbugsMain