IBolt IRichBolt IBasicBolt IBatchBolt 区分

一 ：IBolt : 定义了Bolt的功能集合，主要有

void prepare(Map stormConf,TopologyContext context, OutputCollector collector);

void execut(e Tuple input);

void cleanup();

三个方法

二： IRichBolt ：需要实现IBolt和IComponent接口

public interface IRichBolt extends IBolt, IComponent {

}

三： IBasicBolt: 接口定义与IBolt基本一致，具体的实现要求也与IBolt相同，

区别在于以下两点：

1：他的收集器用的是 BasicOutputCollector，并且该参数放在execute的方法中二不是perpare中

2：他实现了 IComponent 接口

源码：

public interface IBasicBolt extends IComponent {

void prepare(Map stormConf, TopologyContext context);

/**

* Process the input tuple and optionally emit new tuples based on the input tuple.

*

* All acking is managed for you. Throw a FailedException if you want to fail the tuple.

*/

void execute(Tuple input, BasicOutputCollector collector);

void cleanup();

}

为什么会有这个接口，作用是提供一种更简单的Bolt编写方式，基于 IBasicBolt编写的好处是storm框架帮你处理发出消息的Ack，Fail和Anchor操作，这是由执行器 BasicBoltExecutor

实现的。

BasicBoltExecutor 已经帮你实现了IRichBolt

源码：

public class BasicBoltExecutor implements IRichBolt {

public static final Logger LOG = LoggerFactory. getLogger (BasicBoltExecutor. class );

private IBasicBolt _bolt ;

private transient BasicOutputCollector _collector ;

public BasicBoltExecutor(IBasicBolt bolt) {

_bolt = bolt;

}

public void declareOutputFields(OutputFieldsDeclarer declarer) {

_bolt .declareOutputFields(declarer);

}

public void prepare(Map stormConf, TopologyContext context, OutputCollector collector) {

_bolt .prepare(stormConf, context);

_collector = new BasicOutputCollector(collector);

}

public void execute(Tuple input) {

_collector .setContext(input);

try {

_bolt .execute(input, _collector );

_collector .getOutputter().ack(input);

} catch (FailedException e) {

if (e instanceof ReportedFailedException) {

_collector .reportError(e);

}

_collector .getOutputter().fail(input);

}

}

public void cleanup() {

_bolt .cleanup();

}

public Map<String, Object> getComponentConfiguration() {

return _bolt .getComponentConfiguration();

}

}

四 ：IBatchBolt

IBatchBolt主要用于storm中的批处理，目前storm主要用该接口来实现消息的可靠传输，storm的事务Topology以及Trident主要是基于IBatchBolt的，相比以上三个，它多了一个 finishBatch 方法，它在一个批处理结束时候被调用。此外， IBatchBolt还除去了cleanup方法。

public class BatchBoltExecutor implements IRichBolt, FinishedCallback, TimeoutCallback {

public static final Logger LOG = LoggerFactory. getLogger (BatchBoltExecutor. class );

byte [] _boltSer ;

Map<Object, IBatchBolt> _openTransactions ;

Map _conf ;

TopologyContext _context ;

BatchOutputCollectorImpl _collector ;

public BatchBoltExecutor(IBatchBolt bolt) {

_boltSer = Utils. javaSerialize (bolt);

}

@Override

public void prepare(Map conf, TopologyContext context, OutputCollector collector) {

_conf = conf;

_context = context;

_collector = new BatchOutputCollectorImpl(collector);

_openTransactions = new HashMap<>();

}

@Override

public void execute(Tuple input) {

Object id = input.getValue( 0 );

IBatchBolt bolt = getBatchBolt(id);

try {

bolt.execute(input);

_collector .ack(input);

} catch (FailedException e) {

LOG .error( "Failed to process tuple in batch" , e);

_collector .fail(input);

}

}

@Override

public void cleanup() {

}

@Override

public void finishedId(Object id) {

IBatchBolt bolt = getBatchBolt(id);

_openTransactions .remove(id);

bolt.finishBatch();

}

@Override

public void timeoutId(Object attempt) {

_openTransactions .remove(attempt);

}

@Override

public void declareOutputFields(OutputFieldsDeclarer declarer) {

newTransactionalBolt().declareOutputFields(declarer);

}

@Override

public Map<String, Object> getComponentConfiguration() {

return newTransactionalBolt().getComponentConfiguration();

}

private IBatchBolt getBatchBolt(Object id) {

IBatchBolt bolt = _openTransactions .get(id);

if (bolt== null ) {

bolt = newTransactionalBolt();

bolt.prepare( _conf , _context , _collector , id);

_openTransactions .put(id, bolt);

}

return bolt;

}

private IBatchBolt newTransactionalBolt() {

return Utils. javaDeserialize ( _boltSer , IBatchBolt. class );

}

}