10.应用例子.docxVIP

应用例子本文帮助读者快速实现一个JStorm 例子Example 源码最简单的JStorm例子分为4个步骤：生成TopologyMap conf =newHashMp();//topology所有自定义的配置均放入这个MapTopologyBuilder builder =newTopologyBuilder();//创建topology的生成器int spoutParal = get(spout.parallel, 1);//获取spout的并发设置SpoutDeclarer spout = builder.setSpout(SequenceTopologyDef.SEQUENCE_SPOUT_NAME,newSequenceSpout(), spoutParal);//创建Spout， 其中new SequenceSpout() 为真正spout对象，SequenceTopologyDef.SEQUENCE_SPOUT_NAME 为spout的名字，注意名字中不要含有空格int boltParal = get(bolt.parallel, 1);//获取bolt的并发设置BoltDeclarer totalBolt = builder.setBolt(SequenceTopologyDef.TOTAL_BOLT_NAME, newTotalCount(), boltParal).shuffleGrouping(SequenceTopologyDef.SEQUENCE_SPOUT_NAME);//创建bolt， SequenceTopologyDef.TOTAL_BOLT_NAME 为bolt名字，TotalCount 为bolt对象，boltParal为bolt并发数，//shuffleGrouping（SequenceTopologyDef.SEQUENCE_SPOUT_NAME）， //表示接收SequenceTopologyDef.SEQUENCE_SPOUT_NAME的数据，并且以shuffle方式，//即每个spout随机轮询发送tuple到下一级bolt中int ackerParal = get(acker.parallel, 1);Config.setNumAckers(conf, ackerParal);//设置表示acker的并发数int workerNum = get(worker.num, 10);conf.put(Config.TOPOLOGY_WORKERS, workerNum);//表示整个topology将使用几个workerconf.put(Config.STORM_CLUSTER_MODE, distributed);//设置topolog模式为分布式，这样topology就可以放到JStorm集群上运行StormSubmitter.submitTopology(streamName, conf, builder.createTopology());//提交topologyIRichSpoutIRichSpout 为最简单的Spout接口IRichSpout{ @Overridepublicvoid open(Map conf, TopologyContext context, SpoutOutputCollector collector) { } @Overridepublicvoid close() { } @Overridepublicvoid activate() { } @Overridepublicvoid deactivate() { } @Overridepublicvoid nextTuple() { } @Overridepublicvoid ack(Object msgId) { } @Overridepublicvoid fail(Object msgId) { } @Overridepublicvoid declareOutputFields(OutputFieldsDeclarer declarer) { } @OverridepublicMapString, Object getComponentConfiguration() {returnnull; }其中注意：spout对象必须是继承Serializable，因此要求spout内所有数据结构必须是可序列化的spout可以有构造函数，但构造函数只执行一次，是在提交任务时，创建spout对象，因此在task分配到具体worker之前的初始化工