Question

我正在尝试重新平衡使用KafkaSpout的Storm拓扑。我的代码是：

    TopologyBuilder builder = new TopologyBuilder();
    Properties kafkaProps = new Properties();
    kafkaProps.put("zk.connect", "localhost:2181");
    kafkaProps.put("zk.connectiontimeout.ms", "1000000");
    kafkaProps.put("groupid", "storm");

    builder.setSpout( "kafkaSpout" , new KafkaSpout(kafkaProps, "test"), 3);
    builder.setBolt( "eventBolt", new EventBolt(), 2 ).shuffleGrouping( "kafkaSpout", "eventStream" );
    builder.setBolt( "tableBolt", new TableBolt(), 2 ).shuffleGrouping( "kafkaSpout", "tableStream");

    Map<String, Object> conf = new HashMap<String, Object>();
    conf.put(Config.TOPOLOGY_DEBUG, true);

    LocalCluster cluster = new LocalCluster();
    cluster.submitTopology("test", conf, builder.createTopology());

    Utils.sleep( 1000*5 );

    List<TopologySummary> topologySummaries = cluster.getClusterInfo().get_topologies();
    for ( TopologySummary summary : topologySummaries ) {
        StormTopology topology = cluster.getTopology( summary.get_id() );
        RebalanceOptions options = new RebalanceOptions();
        options.set_wait_secs( 0 );
        options.set_num_workers( 4 );

        for ( String name : topology.get_bolts().keySet() ) {
            System.err.println( name + "   " + topology.get_bolts().get(name).get_common().get_json_conf() );
            options.put_to_num_executors( name , 5);
        }
        for ( String name : topology.get_spouts().keySet() ) {
            System.err.println( name + "   " + topology.get_spouts().get(name).get_common().get_json_conf() );
            options.put_to_num_executors( name , 5);
        }

        cluster.rebalance( summary.get_name() , options);
    }

但是，在重新平衡期间，会显示以下错误跟踪：

10341 [storm_rishabh-1361473654345-95461d10_watcher_executor] INFO  kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-95461d10 begin rebalancing consumer storm_rishabh-1361473654345-95461d10 try #1
10341 [storm_rishabh-1361473654345-3b26ed76_watcher_executor] INFO kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-3b26ed76 begin rebalancing consumer storm_rishabh-1361473654345-3b26ed76 try #1
10342 [storm_rishabh-1361473654345-95461d10_watcher_executor] ERROR kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-95461d10 error during syncedRebalance
java.lang.NullPointerException: null
at kafka.utils.ZkUtils$.getChildrenParentMayNotExist(ZkUtils.scala:181) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.utils.ZkUtils$.getCluster(ZkUtils.scala:202) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener$$anonfun$syncedRebalance$1.apply$mcVI$sp(ZookeeperConsumerConnector.scala:447) ~[kafka_2.9.2-0.7.0.jar:na]
at scala.collection.immutable.Range.foreach$mVc$sp(Range.scala:78) ~[scala-library-2.9.2.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener.syncedRebalance(ZookeeperConsumerConnector.scala:444) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener$$anon$1.run(ZookeeperConsumerConnector.scala:401) ~[kafka_2.9.2-0.7.0.jar:na]
10342 [storm_rishabh-1361473654345-3b26ed76_watcher_executor] ERROR kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-3b26ed76 error during syncedRebalance
java.lang.NullPointerException: null
at kafka.utils.ZkUtils$.getChildrenParentMayNotExist(ZkUtils.scala:181) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.utils.ZkUtils$.getCluster(ZkUtils.scala:202) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener$$anonfun$syncedRebalance$1.apply$mcVI$sp(ZookeeperConsumerConnector.scala:447) ~[kafka_2.9.2-0.7.0.jar:na]
at scala.collection.immutable.Range.foreach$mVc$sp(Range.scala:78) ~[scala-library-2.9.2.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener.syncedRebalance(ZookeeperConsumerConnector.scala:444) ~[kafka_2.9.2-0.7.0.jar:na]
at kafka.consumer.ZookeeperConsumerConnector$ZKRebalancerListener$$anon$1.run(ZookeeperConsumerConnector.scala:401) ~[kafka_2.9.2-0.7.0.jar:na]
10342 [storm_rishabh-1361473654345-95461d10_watcher_executor] INFO  kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-95461d10 stopping watcher executor thread for consumer storm_rishabh-1361473654345-95461d10
10343 [storm_rishabh-1361473654345-3b26ed76_watcher_executor] INFO  kafka.consumer.ZookeeperConsumerConnector - storm_rishabh-1361473654345-3b26ed76 stopping watcher executor thread for consumer storm_rishabh-1361473654345-3b26ed76

有人可以告诉我可能是什么问题吗？我是否需要在kafkaSpout中定义更多内容，以便在重新平衡时正确关闭然后重新开始？

Answer 1

在LocalCluster中运行时（出于开发目的），我遇到了同样的问题。我更改了我的测试配置YAML以将工作人员数量减少到1：

topology.workers: 1

这纠正了这个问题。我还没有尝试在实际的分布式集群上运行它，所以我不知道这是否只是在LocalCluster模式下运行的工件。

（在我的代码中，我从不调用LocalCluster.rebalance。）

Answer 2

使用来自supervisor或nimbus节点的storm rebalance命令。

例如，风暴重新平衡mytopology -n 5 -e blue-spout = 3 -e yellow-bolt = 10。

请参阅本网站。 www.michael-noll.com.

使用Java代码进行风暴拓扑重新平衡

2 个答案: