Kafka生产者介绍（七）：创建发送请求_kafka producerequest-优快云博客

本文链接：https://blog.youkuaiyun.com/lianggx3/article/details/90616044

博客介绍了Sender.createProduceRequests()将待发送消息封装成ClientRequest的过程。每个Node最多生成一个ClientRequest对象，具体步骤包括整理RecordBatch集合、创建RequestSend和ProduceRequest、创建回调对象，最后创建并返回ClientRequest，后续会发送RequestSend并缓存ClientRequest。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

createProducerRequest
Sender.createProduceRequests()把待发送的消息封装成ClientRequest。无论一个Node对应多少个RecordBatch，或是发到多少个分区的，每个Node最多生成一个ClientRequest对象：

public class Sender implements Runnable {
    private List<ClientRequest> createProduceRequests(Map<Integer, List<RecordBatch>> collated, long now) {
        List<ClientRequest> requests = new ArrayList<ClientRequest>(collated.size());
        // 调用produceRequest方法，把发往同一个node的RecordBatch封装成一个ClientRequset对象。
        for (Map.Entry<Integer, List<RecordBatch>> entry : collated.entrySet())
            requests.add(produceRequest(now, entry.getKey(), acks, requestTimeout, entry.getValue()));
        return requests;
    }
    /**
     * Create a produce request from the given record batches
     */
    private ClientRequest produceRequest(long now, int destination, short acks, int timeout, List<RecordBatch> batches) {
        //produceRecordsByPartition，value是ByteBuffer
        Map<TopicPartition, ByteBuffer> produceRecordsByPartition = new HashMap<TopicPartition, ByteBuffer>(batches.size());
        // recordsByPartition，value是RecordBatch
        final Map<TopicPartition, RecordBatch> recordsByPartition = new HashMap<TopicPartition, RecordBatch>(batches.size());
        // 把batches里面的RecordBatch按照partiton进行分类，放入上列两个集合。
        for (RecordBatch batch : batches) {
            TopicPartition tp = batch.topicPartition;
            produceRecordsByPartition.put(tp, batch.records.buffer());
            recordsByPartition.put(tp, batch);
        // 创建ProduceRequest 和 RequestSend
        ProduceRequest request = new ProduceRequest(acks, timeout, produceRecordsByPartition);
        RequestSend send = new RequestSend(Integer.toString(destination),
                                           this.client.nextRequestHeader(ApiKeys.PRODUCE),
                                           request.toStruct());
        // RequestCompletionHandler为回调对象
        RequestCompletionHandler callback = new RequestCompletionHandler() {
            public void onComplete(ClientResponse response) {
                handleProduceResponse(response, recordsByPartition, time.milliseconds());
            }
        };
        // 创建ClientRequest，根据ack判断是否需要响应。
        return new ClientRequest(now, acks != 0, send, callback);
    }
}

1、将一个NodeId对应的RecordBatch集合，重新整理为produceRecordsByPartition和recordsByPartition两个集合
2、创建RequestSend和ProduceRequest
3、创建回调对象
4、创建并返回ClientRequest

后面会发送RequestSend对象，把ClientRequest放入InFlightRequest中缓存，当收到响应或者出现异常时候，通过缓存的ClientRequset调用其RequestCompletionHandler对象。