Kafka 生产者与可靠性保证ACK（2）

生产者消息发送流程

消息发送的整体流程，生产端主要由两个线程协调运行。分别是main线程和sender线程（发送线程）。

在Kafka（2.6.0版本）源码中，可以看到。

源码地址： 
kafka\clients\src\main\java\org.apache.kafka.clients.producer.KafkaProducer.java

测试入口：
KafkaProducerTest.testInvalidGenerationIdAndMemberIdCombinedInSendOffsets()

在创建KafkaProducer时，在430创建了一个Sender对象，并且启动了一个IO线程。

Kafka 生产者与可靠性保证ACK（2）

this.errors = this.metrics.sensor("errors");
this.sender = newSender(logContext, kafkaClient, this.metadata);
String ioThreadName = NETWORK_THREAD_PREFIX + " | " + clientId;
this.ioThread = new KafkaThread(ioThreadName, this.sender, true);
this.ioThread.start();

interceptor

interceptor的作用是实现消息的定制化，类似：spring Interceptor 、MyBatis的插件、Quartz的监听器。

@Override
public Future<RecordMetadata> send(ProducerRecord<K, V> record, Callback callback) {
     // intercept the record, which can be potentially modified; this method does not throw exceptions
     ProducerRecord<K, V> interceptedRecord = this.interceptors.onSend(record);
    return doSend(interceptedRecord, callback);
}

可通过实现org.apache.kafka.clients.producer.ProducerInterceptor接口开发自定义器。

简单自定义例子：

public class CustomInterceptor implements ProducerInterceptor<String, String> {
    // 发送消息时触发
    @Override
    public ProducerRecord<String, String> onSend(ProducerRecord<String, String> record) {
        System.out.println("发送消息时触发");
        return record;
    }

    // 收到服务端的ACK时触发
    @Override
    public void onAcknowledgement(RecordMetadata metadata, Exception exception) {
        System.out.println("消息被服务端接收");
    }

    @Override
    public void close() {
        System.out.println("生产者关闭");
    }

    // 用键值对配置时触发
    @Override
    public void configure(Map<String, ?> configs) {
        System.out.println("configure...");
    }
}

// 生产者中添加
List<String> interceptors = new ArrayList<>();
interceptors.add("com.freecloud.plug.kafka.interceptor.CustomInterceptor");
props.put(ProducerConfig.INTERCEPTOR_CLASSES_CONFIG, interceptors);

序列化

byte[] serializedKey;
try {
    serializedKey = keySerializer.serialize(record.topic(), record.headers(), record.key());
} catch (ClassCastException cce) {
    throw new SerializationException("Can't convert key of class " + record.key().getClass().getName() +
            " to class " + producerConfig.getClass(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG).getName() +
            " specified in key.serializer", cce);
}
byte[] serializedValue;
try {
    serializedValue = valueSerializer.serialize(record.topic(), record.headers(), record.value());
} catch (ClassCastException cce) {
    throw new SerializationException("Can't convert value of class " + record.value().getClass().getName() +
            " to class " + producerConfig.getClass(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG).getName() +
            " specified in value.serializer", cce);
}

Kafka 生产者与可靠性保证ACK（2）

在kafka针对不同的数据类型做了相应的序列化工具。如需自定义实现org.apache.kafka.common.serialization.Serializer接口。

路由器（分区器）

int partition = partition(record, serializedKey, serializedValue, cluster);

消息累加器

RecordAccumulator.RecordAppendResult result = accumulator.append(tp, timestamp, serializedKey,
                    serializedValue, headers, interceptCallback, remainingWaitMs, true, nowMs);

// RecordAccumulator本质是一个ConcurrentMap：

private final ConcurrentMap<TopicPartition, Deque<ProducerBatch>> batches;

一个partition一个Batch。batch满了之后，会唤醒Sender线程发送消息。

if (result.batchIsFull || result.newBatchCreated) {
    log.trace("Waking up the sender since topic {} partition {} is either full or getting a new batch", record.topic(), partition);
    this.sender.wakeup();
}