nextTuple（）在Storm上使用BaseRichSpout调用无限次

Question

我实现了一个简单的Storm拓扑，它具有一个spout和一个在本地集群模式下运行的bolt。

由于某种原因，spout的nextTuple（）被多次调用。

知道为什么吗？

public class CommitFeedListener extends BaseRichSpout {
    private SpoutOutputCollector outputCollector;
    private List<String> commits;

    @Override
    public void declareOutputFields(OutputFieldsDeclarer declarer) {
        declarer.declare(new Fields("commit"));
    }

    @Override
    public void open(Map configMap,
                     TopologyContext context,
                     SpoutOutputCollector outputCollector) {
        this.outputCollector = outputCollector;
    }

    **//that method is invoked more than once**
    @Override
    public void nextTuple() {

            outputCollector.emit(new Values("testValue"));

    }
}

public class EmailExtractor extends BaseBasicBolt {
    @Override
    public void declareOutputFields(OutputFieldsDeclarer declarer) {
        declarer.declare(new Fields("email"));
    }
    @Override
    public void execute(Tuple tuple,
                        BasicOutputCollector outputCollector) {
        String commit = tuple.getStringByField("commit");
        System.out.println(commit);        
    }  
}

运行配置：

public class LocalTopologyRunner {
    private static final int TEN_MINUTES = 600000;
    public static void main(String[] args) {
        TopologyBuilder builder = new TopologyBuilder();
        builder.setSpout("commit-feed-listener", new CommitFeedListener());
                builder
        .setBolt("email-extractor", new EmailExtractor())
                .shuffleGrouping("commit-feed-listener");
        Config config = new Config();
        config.setDebug(true);
        StormTopology topology = builder.createTopology();
        LocalCluster cluster = new LocalCluster();
        cluster.submitTopology("github-commit-count-topology",
                config,
                topology);
        Utils.sleep(TEN_MINUTES);
        cluster.killTopology("github-commit-count");
        cluster.shutdown();
    }
}

谢谢你，雷。

Answer 1

nextTuple（）由设计在无限循环中调用。这样做是为了使用外部资源（数据库、流、IO等）的脏检查。

如果您在nextTuple（）中无事可做，则应该Hibernate一段时间，以防止使用backtype.storm.utils.utils滥发CPU信息

Utils.sleep(pollIntervalInMilliseconds);

Storm是一个实时处理架构，所以它确实是正确的行为。查看一些示例，看看如何根据您的需要实现一个喷口。