apache storm, load balance, json

Question

I am using Kafka storm, kafka sends/emits json string to storm, in the storm, I want to distribute the load to a couple of workers based on the key/field in the json. How to do that? In my case, it is groupid field in json string.

For example, I have json like that:

{groupid: 1234, userid: 145, comments:"I want to distribute all this group 1234  to one worker", size:50,type:"group json"}
{groupid: 1235, userid: 134, comments:"I want to distribute all this group 1234 to another worker", size:90,type:"group json"}
{groupid: 1234, userid: 158, comments:"I want to be sent to same worker as group 1234", size:50,type:"group json"}   

I try too use following codes:

      1.  TopologyBuilder builder = new TopologyBuilder();
      2.  builder.setSpout(SPOUTNAME, kafkaSpout, 1);
      3.  builder.setBolt(MYDISTRIBUTEDWORKER, new DistributedBolt()).setFieldsGroup(SPOUTNAME,new Fields("groupid"));  <---???

I am wondering how to put arguments in setFieldsGroup method in line 3. Could someone give me a hint?

Juhani

==Testing using storm 0.9.4 ============

=============source codes==============

import java.util.List;
import java.util.Map;
import java.util.concurrent.atomic.AtomicInteger;

import storm.kafka.KafkaSpout;
import storm.kafka.SpoutConfig;
import storm.kafka.StringScheme;
import storm.kafka.ZkHosts;
import backtype.storm.Config;
import backtype.storm.LocalCluster;
import backtype.storm.spout.SchemeAsMultiScheme;
import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.topology.TopologyBuilder;
import backtype.storm.topology.base.BaseRichBolt;
import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Tuple;
import backtype.storm.tuple.Values;


public class KafkaBoltMain {
   private static final String SPOUTNAME="TopicSpout"; 
   private static final String ANALYSISBOLT = "AnalysisWorker";
   private static final String CLIENTID = "Storm";
   private static final String TOPOLOGYNAME = "LocalTopology";


   private static class AppAnalysisBolt extends BaseRichBolt {
       private static final long serialVersionUID = -6885792881303198646L;
        private OutputCollector _collector;
       private long groupid=-1L;
       private String log="test";

       public void prepare(Map conf, TopologyContext context, OutputCollector collector) {
           _collector = collector;
       }

       public void execute(Tuple tuple) {
           List<Object> objs = tuple.getValues();
           int i=0;
           for(Object obj:objs){
               System.out.println(""+i+"th object's value is:"+obj.toString());
               i++;
           }

//         _collector.emit(new Values(groupid,log));
           _collector.ack(tuple);
       }

        public void declareOutputFields(OutputFieldsDeclarer declarer) {
            declarer.declare(new Fields("groupid","log"));
        }
   } 

   public static void main(String[] args){
       String zookeepers = null;
       String topicName = null;
       if(args.length == 2 ){
           zookeepers = args[0];
           topicName = args[1];
        }else if(args.length == 1 && args[0].equalsIgnoreCase("help")){ 
           System.out.println("xxxx");
           System.exit(0);
        }
       else{
           System.out.println("You need to have two arguments: kafka zookeeper:port and topic name");
           System.out.println("xxxx");
           System.exit(-1);
        }       

        SpoutConfig spoutConfig = new SpoutConfig(new ZkHosts(zookeepers),
                topicName,
                "",// zookeeper root path for offset storing
                CLIENTID);
        spoutConfig.scheme = new SchemeAsMultiScheme(new StringScheme());
        KafkaSpout kafkaSpout = new KafkaSpout(spoutConfig);

        TopologyBuilder builder = new TopologyBuilder();
        builder.setSpout(SPOUTNAME, kafkaSpout, 1);
        builder.setBolt(ANALYSISBOLT, new AppAnalysisBolt(),2)
            .fieldsGrouping(SPOUTNAME,new Fields("groupid"));

        //Configuration
        Config conf = new Config();
        conf.setDebug(false);
        //Topology run
        conf.put(Config.TOPOLOGY_MAX_SPOUT_PENDING, 1);
        LocalCluster cluster = new LocalCluster();
        cluster.submitTopology(TOPOLOGYNAME, conf, builder.createTopology());
    }
}

================================================== when I start to submit topology(local cluster), it gives following error:

11658 [SyncThread:0] INFO  org.apache.storm.zookeeper.server.ZooKeeperServer - Established session 0x14d097d338c0009 with negotiated timeout 20000 for client /127.0.0.1:34656
11658 [main-SendThread(localhost:2000)] INFO  org.apache.storm.zookeeper.ClientCnxn - Session establishment complete on server localhost/127.0.0.1:2000, sessionid = 0x14d097d338c0009, negotiated timeout = 20000
11659 [main-EventThread] INFO  org.apache.storm.curator.framework.state.ConnectionStateManager - State change: CONNECTED
12670 [main] INFO  backtype.storm.daemon.supervisor - Starting supervisor with id ccc57de0-29ff-4cb4-89de-fea1ea9b6e28 at host storm-VirtualBox
12794 [main] WARN  backtype.storm.daemon.nimbus - Topology submission exception. (topology name='LocalTopology') #<InvalidTopologyException InvalidTopologyException(msg:Component: [AnalysisWorker] subscribes from stream: [default] of component [TopicSpout] with non-existent fields: #{"groupid"})>
12800 [main] ERROR org.apache.storm.zookeeper.server.NIOServerCnxnFactory - Thread Thread[main,5,main] died
backtype.storm.generated.InvalidTopologyException: null
        at backtype.storm.daemon.common$validate_structure_BANG_.invoke(common.clj:178) ~[storm-core-0.9.4.jar:0.9.4]
        at backtype.storm.daemon.common$system_topology_BANG_.invoke(common.clj:307) ~[storm-core-0.9.4.jar:0.9.4]
        at backtype.storm.daemon.nimbus$fn__4290$exec_fn__1754__auto__$reify__4303.submitTopologyWithOpts(nimbus.clj:948) ~[storm-core-0.9.4.jar:0.9.4]
        at backtype.storm.daemon.nimbus$fn__4290$exec_fn__1754__auto__$reify__4303.submitTopology(nimbus.clj:966) ~[storm-core-0.9.4.jar:0.9.4]
        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) ~[na:1.7.0_80]
        at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57) ~[na:1.7.0_80]
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) ~[na:1.7.0_80]
        at java.lang.reflect.Method.invoke(Method.java:606) ~[na:1.7.0_80]
        at clojure.lang.Reflector.invokeMatchingMethod(Reflector.java:93) ~[clojure-1.5.1.jar:na]
        at clojure.lang.Reflector.invokeInstanceMethod(Reflector.java:28) ~[clojure-1.5.1.jar:na]
        at backtype.storm.testing$submit_local_topology.invoke(testing.clj:264) ~[storm-core-0.9.4.jar:0.9.4]
        at backtype.storm.LocalCluster$_submitTopology.invoke(LocalCluster.clj:43) ~[storm-core-0.9.4.jar:0.9.4]
        at backtype.storm.LocalCluster.submitTopology(Unknown Source) ~[storm-core-0.9.4.jar:0.9.4]
        at com.callstats.stream.analyzer.KafkaBoltMain.main(KafkaBoltMain.java:94) ~[StreamAnalyzer-1.0-SNAPSHOT-jar-with-dependencies.jar:na]

Answer 1

I'm not sure which version of Storm you are using, as of 0.9.4, your requirement can be implemented as follows.

builder.setBolt(MYDISTRIBUTEDWORKER, new DistributedBolt()).fieldsGrouping(SPOUTNAME, new Fields("groupid"));

In prepare method of DistributedBolt,

public void declareOutputFields(OutputFieldsDeclarer declarer) {
    declarer.declare(new Fields("groupid", "log"));
}

Somewhere in execute method of it, you will call

collector.emit(new Values(groupid, log));

then tuples which have same groupid will be delivered to same instance of next bolt.