flume+kalka 搭建一

flume+kafka搭建

一． kafka集群搭建启动

详细参见文档《kafka集群搭建文档.docx》

安装启动kafka集群

二． 下载安装flume

flume官方下载地址：https://flume.apache.org/download.html

建议下载最新的1.6.0版本的，因为1.6.0版本的集成了整合kafka的插件包可以直接配置使用

下载apache-flume-1.6.0-bin.tar.gz包

通过tar –zxvf apache-flume-1.6.0-bin.tar.gz命令解压

Flume解压既安装成功，配置conf/ flume-conf.properties文件启动完成相应的功能

三． 配置flume连接kafka

前面kafka集群已经成功，这里只需要配置好conf/ flume-conf.properties文件，配置如下

a1.sources = r1

a1.sinks = k1

a1.channels = c1

 

# Describe/configure the source

a1.sources.r1.type=avro

a1.sources.r1.bind=master

a1.sources.r1.port=41414

 

# Describe the sink

a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink  

a1.sinks.k1.topic = testflume  

a1.sinks.k1.brokerList = 192.168.57.4:9092,192.168.57.5:9092,192.168.57.6:9092  

a1.sinks.k1.requiredAcks = 1  

a1.sinks.k1.batchSize = 20  

a1.sinks.k1.channel = c1

 

 

 

# Use a channel which buffers events in memory

a1.channels.c1.type = memory

a1.channels.c1.capacity = 1000000

a1.channels.c1.transactionCapacity = 10000

 

# Bind the source and sink to the channel

a1.sources.r1.channels = c1

a1.sinks.k1.channel = c1

 

注意这里需要根据实际需求来配置sources，这里是用avro的方式配置监听master本机的41414端口注意这里是启动agent的配置 后续的flume client也需要用到这个配置，下面的sink配置a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink是固定的不需要修改，而a1.sinks.k1.topic = testflume是创建的话题名，需要根据自己需要来改，而a1.sinks.k1.brokerList = 192.168.57.4:9092,192.168.57.5:9092,192.168.57.6:9092是根据实际的kafka集群情况配置的

 

四．测试

上面配置完成以后启动flume连接到kafka

bin/flume-ng agent -c ./conf/ -f conf/flume-conf.properties -Dflume.root.logger=DEBUG,console -na1

注意这里的a1指的是配置文件中的agent名字a1不是随意写的

这里的flume对接kafka实际已经完成，接下来就是测试

测试代码：

import org.apache.flume.Event;
import org.apache.flume.EventDeliveryException;
import org.apache.flume.api.RpcClient;
import org.apache.flume.api.RpcClientFactory;
import org.apache.flume.event.EventBuilder;
import java.nio.charset.Charset;

public class MyApp {
  public static void main(String[] args) {
    MyRpcClientFacade client = new MyRpcClientFacade();
    // Initialize client with the remote Flume agent's host and port
    client.init("master",41414);

    // Send 10 events to the remote Flume agent. That agent should be
    // configured to listen with an AvroSource.
    String sampleData = "Hello Flume!";
    for (inti = 0; i < 10; i++) {
      client.sendDataToFlume(sampleData);
    }

    client.cleanUp();
  }
}

class MyRpcClientFacade {
  private RpcClient client;
  private String hostname;
  private int port;

  public void init(String hostname,int port) {
    // Setup the RPC connection
    this.hostname= hostname;
    this.port= port;
    this.client= RpcClientFactory.getDefaultInstance(hostname, port);
    // Use the following method to create a thrift client (instead of the above line):
    // this.client = RpcClientFactory.getThriftInstance(hostname, port);
  }

  public void sendDataToFlume(String data) {
    // Create a Flume Event object that encapsulates the sample data
    Event event = EventBuilder.withBody(data, Charset.forName("UTF-8"));

    // Send the event
    try{
      client.append(event);
    } catch (EventDeliveryException e) {
      // clean up and recreate the client
      client.close();
      client = null;
      client = RpcClientFactory.getDefaultInstance(hostname,port);
      // Use the following method to create a thrift client (instead of the above line):
      // this.client = RpcClientFactory.getThriftInstance(hostname, port);
    }
  }

  public void cleanUp() {
    // Close the RPC connection
    client.close();
  }

}

注意这里的client.init("master",41414);中的master和41414是和flume配置文件里保持一致的

先启动一个consumer来监听

./bin/kafka-console-consumer.sh --zookeeper master:2181,slave1:2181,slave2:2181 --from-beginning --topic testflume

注意这里的topic的名字要和配置文件中一致

执行上面的main方法作为flume的client端来产生数据，可以在上面的consumer监听里面看到结果