storm-kafka（storm spout做爲kafka的消費端）

storm是grovvy寫的java

kafka是scala寫的
git

storm-kafka  storm鏈接kafka consumer的插件github

下載地址：apache

https://github.com/wurstmeister/storm-kafka-0.8-plus併發

除了須要storm和kafka相關jar包還須要google-collections-1.0.jarsocket

以及zookeeper相關包 curator-framework-1.3.3.jar和curator-client-1.3.3.jaride

之前由com.netflix.curator組織開發如今歸到org.apache.curator下面ui

1.Kafka Consumer即Storm Spout代碼google

package demo;

import java.util.ArrayList;
import java.util.List;

import backtype.storm.Config;
import backtype.storm.LocalCluster;
import backtype.storm.StormSubmitter;
import backtype.storm.generated.AlreadyAliveException;
import backtype.storm.generated.InvalidTopologyException;
import backtype.storm.spout.SchemeAsMultiScheme;
import backtype.storm.topology.TopologyBuilder;
import storm.kafka.KafkaSpout;
import storm.kafka.SpoutConfig;
import storm.kafka.StringScheme;
import storm.kafka.ZkHosts;

public class MyKafkaSpout {
public static void main(String[] args) {
    
    String topic ="track";
    ZkHosts zkhosts  = new ZkHosts("192.168.1.107:2181,192.168.1.108:2181,192.168.1.109:2181");
    
    SpoutConfig spoutConfig = new SpoutConfig(zkhosts, topic,
            "/MyKafka", //偏移量offset的根目錄
            "MyTrack");//子目錄對應一個應用    
    List<String> zkServers=new ArrayList<String>();
    //zkServers.add("192.168.1.107");
    //zkServers.add("192.168.1.108");
    for(String host:zkhosts.brokerZkStr.split(","))
    {
        zkServers.add(host.split(":")[0]);
    }
    
    spoutConfig.zkServers=zkServers;
    spoutConfig.zkPort=2181;
    spoutConfig.forceFromStart=true;//從頭開始消費，其實是要改爲false的
    spoutConfig.socketTimeoutMs=60;
    spoutConfig.scheme=new SchemeAsMultiScheme(new StringScheme());//定義輸出爲string類型
    
    TopologyBuilder builder=new TopologyBuilder();
    builder.setSpout("spout", new KafkaSpout(spoutConfig),1);//引用spout，併發度設爲1
    builder.setBolt("bolt1", new MyKafkaBolt(),1).shuffleGrouping("spout");
    
    Config config =new Config();
    config.setDebug(true);//上線以前都要改爲false不然日誌會很是多
    if(args.length>0){
        
        try {
            StormSubmitter.submitTopology(args[0], config, builder.createTopology());
        } catch (AlreadyAliveException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        } catch (InvalidTopologyException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
        
    }else{
        
        LocalCluster localCluster=new LocalCluster();
        localCluster.submitTopology("mytopology", config,  builder.createTopology());
        //本地模式在一個進程裏面模擬一個storm集羣的全部功能
    }
    
    
    
}
}

2.Bolt代碼只是簡單打印輸出，覆寫execute方法便可
spa

package demo;

import java.util.Map;

import backtype.storm.task.TopologyContext;
import backtype.storm.topology.BasicOutputCollector;
import backtype.storm.topology.IBasicBolt;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.tuple.Tuple;

public class MyKafkaBolt implements IBasicBolt {

    @Override
    public void declareOutputFields(OutputFieldsDeclarer arg0) {
        // TODO Auto-generated method stub

    }

    @Override
    public Map<String, Object> getComponentConfiguration() {
        // TODO Auto-generated method stub
        return null;
    }

    @Override
    public void cleanup() {
        // TODO Auto-generated method stub

    }

    @Override
    public void execute(Tuple input, BasicOutputCollector arg1) {
    String kafkaMsg =input.getString(0);
    System.err.println("bolt"+kafkaMsg);

    }

    @Override
    public void prepare(Map arg0, TopologyContext arg1) {
        // TODO Auto-generated method stub

    }

}