（4.2）kafka生产者源码——RecordAccumulator类

1：RecordAccumulator

位置：org.apache.kafka.clients.producer.internals
作用：内存缓冲区，用于存储数据用来发送给server,作为一个消息累加器

1.1：属性

 private final int batchSize;
    private final CompressionType compression;//压缩方式
    private final long lingerMs;/*batch创建多长时间后必须被发送出去*/
    private final long retryBackoffMs;/*batch发送失败重试的时间间隔*/
    private final BufferPool free;/*相当于内存的缓冲池默认大小16M，用于创建批次存放数据，每次批次发送后就释放放回pool池中*/
    private final Time time;
    private final ApiVersions apiVersions;
    private final ConcurrentMap<TopicPartition, Deque<ProducerBatch>> batches; /*其是线程安全的。存放topicPartition-[batch1,batch2...]批次*/
    private final IncompleteBatches incomplete;
    // The following variables are only accessed by the sender thread, so we don't need to protect them.以下变量仅由发送者线程访问，因此我们不需要保护它们。
    /*存储那些分区需要保证幂等性，分区中消息的有序。正在进行批处理发送中的分区*/
    private final Set<TopicPartition> muted;

1.2：方法

主要的追加数据的方法

   public RecordAppendResult append(TopicPartition tp,
                                     long timestamp,
                                     byte[] key,
                                     byte[] value,
                                     Header[] headers,
                                     Callback callback,
                                     long maxTimeToBlock) throws InterruptedException {
        // We keep track of the number of appending thread to make sure we do not miss batches in
        // abortIncompleteBatches().
        appendsInProgress.incrementAndGet();
        ByteBuffer buffer = null;/*生产者是多线程的，所以下面采用加锁synchronized和双重追加数据检查机制tryAppend处理dq*/
        if (headers == null) headers = Record.EMPTY_HEADERS;
        try {
            // check if we have an in-progress batch检查该topic分区是否有正进行的批次，获取该分区对应的Deque队列，每个队列中可能存在着多批次未发送的数据
            Deque<ProducerBatch> dq = getOrCreateDeque(tp);
            /*只有一个线程进入*/
            synchronized (dq)
            {/*通过加锁给dq批次追加数据，防止多线程不安全*/
                if (closed)
                    throw new IllegalStateException("Cannot send after the producer is closed.");

                /*情况1：如果写入分区对应的没有任何一个批次存在，说明需要先创建批次，添加失败appendResult=null*/
                RecordAppendResult appendResult = tryAppend(timestamp, key, value, headers, callback, dq);
                if (appendResult != null)
                    return appendResult;
            }

            //情况2：写入的分区没有正在进行的记录批处理，尝试创建的批处理追加数据

            byte maxUsableMagic = apiVersions.maxUsableProduceMagic();
            /*取批次（默认16k）设置的值和消息大小（默认1M）取最大值，所以两者一般需要调优*/
            int size = Math.max(this.batchSize, AbstractRecords.estimateSizeInBytesUpperBound(maxUsableMagic, compression, key, value, headers));
            log.trace("Allocating a new {} byte message buffer for topic {} partition {}", size, tp.topic(), tp.partition());
            /*分配内存：取16k和消息大小的最大值*/
            buffer = free.allocate(size, maxTimeToBlock);
            synchronized (dq) {
                // Need to check if producer is closed again after grabbing the dequeue lock.
                if (closed)
                    throw new IllegalStateException("Cannot send after the producer is closed.");

                RecordAppendResult appendResult = tryAppend(timestamp, key, value, headers, callback, dq);
                if (appendResult != null) {
                    // Somebody else found us a batch, return the one we waited for! Hopefully this doesn't happen often...
                    return appendResult;
                }

                /*下面三层封装进行了一些属性的分层解析，没有一次定义多次属性，解耦了*/

                /*将分配的内存空间buffer封装为MemoryRecordsBuilder，用于写入record记录*/
                MemoryRecordsBuilder recordsBuilder = recordsBuilder(buffer, maxUsableMagic);
                /*初始化ProducerBatch，初始化内存空间大小=。时对MemoryRecordsBuilder封装*/
                ProducerBatch batch = new ProducerBatch(tp, recordsBuilder, time.milliseconds());
                /*batch.tryAppend:消息如何按照二进制的规范写入MemoryRecordsBuilder，再写到ProducerBatch*/
                FutureRecordMetadata future = Utils.notNull(batch.tryAppend(timestamp, key, value, headers, callback, time.milliseconds()));

                /*封装后的batch，写入队列*/
                dq.addLast(batch);
                incomplete.add(batch);

                // Don't deallocate this buffer in the finally block as it's being used in the record batch
                buffer = null;
                    //更新是否批的状态，检查是否需要发送
                return new RecordAppendResult(future, dq.size() > 1 || batch.isFull(), true);
            }
        } finally {
            if (buffer != null)
                free.deallocate(buffer);
            appendsInProgress.decrementAndGet();
        }
    }
    
private RecordAppendResult tryAppend(long timestamp, byte[] key, byte[] value, Header[] headers,
                                         Callback callback, Deque<ProducerBatch> deque) {
        /*取出未发送批次中的最新批次的数据*/
        ProducerBatch last = deque.peekLast();
        if (last != null) {
            FutureRecordMetadata future = last.tryAppend(timestamp, key, value, headers, callback, time.milliseconds());
            if (future == null)
                last.closeForRecordAppends();
            else
                return new RecordAppendResult(future, deque.size() > 1 || last.isFull(), false);
        }
        return null;
    }

判断batch是否需要被发送的算法实现，获取已经可以发送分区数据对应的leader所在的node节点。

 public ReadyCheckResult ready(Cluster cluster, long nowMs) {
        Set<Node> readyNodes = new HashSet<>();
        long nextReadyCheckDelayMs = Long.MAX_VALUE;
        Set<String> unknownLeaderTopics = new HashSet<>();

        boolean exhausted = this.free.queued() > 0;
        /*遍历所有的batch，获取可以发送batch对应的partition leader*/
        for (Map.Entry<TopicPartition, Deque<ProducerBatch>> entry : this.batches.entrySet()) {
            TopicPartition part = entry.getKey();
            Deque<ProducerBatch> deque = entry.getValue();

            /*找出分区所在的leader*/
            Node leader = cluster.leaderFor(part);
            /*并发性设计：取数据同样加锁，和写入数据一样，保证线程安全性*/
            synchronized (deque) {
                if (leader == null && !deque.isEmpty()) {
                    // This is a partition for which leader is not known, but messages are available to send.
                    // Note that entries are currently not removed from batches when deque is empty.
                    unknownLeaderTopics.add(part.topic());
                } else if (!readyNodes.contains(leader) && !muted.contains(part)) {
                    /*每轮算法只取每个分区的第一个batch进行发送，而不是一个分区的所有batch都判断*/
                    ProducerBatch batch = deque.peekFirst();
                    if (batch != null) {
                        long waitedTimeMs = batch.waitedTimeMs(nowMs);
                        //判断是否进行过失败重试，第一次肯定是0 ，默认值，进行过重试后判断时间间隔
                        boolean backingOff = batch.attempts() > 0 && waitedTimeMs < retryBackoffMs;
                        /*判断retryBackoffMs超时时间和lingerMs batch允许最长创建时间，两个必须都满足。
                        * lingerMs默认是0，也就是说来了就发送，不会等待batch是否写满
                        * */
                        long timeToWaitMs = backingOff ? retryBackoffMs : lingerMs;
                        boolean full = deque.size() > 1 || batch.isFull();
                        /*通过等待时间进行判断，batch是否过期，默认是0时肯定是超时的，expired=true*/
                        boolean expired = waitedTimeMs >= timeToWaitMs;
                        /*判断batch是否需要被发送，满了或者过期或者客户端被关闭或强制flush时*/
                        boolean sendable = full || expired || exhausted || closed || flushInProgress();
                        if (sendable && !backingOff) {
                            readyNodes.add(leader);
                        } else {//如果没有符合发送的batch
                            /*lingerMs-已经等待时间=该批次下次被发送的最长时间*/
                            long timeLeftMs = Math.max(timeToWaitMs - waitedTimeMs, 0);
                            // Note that this results in a conservative estimate since an un-sendable partition may have
                            // a leader that will later be found to have sendable data. However, this is good enough
                            // since we'll just wake up and then sleep again for the remaining time.
                            /*取所有分区最小的timeLeftMs更新为下次检查dequeue时间*/
                            nextReadyCheckDelayMs = Math.min(timeLeftMs, nextReadyCheckDelayMs);
                        }
                    }
                }
            }
        }

        return new ReadyCheckResult(readyNodes, nextReadyCheckDelayMs, unknownLeaderTopics);
    }

创建或者得到分区对应的Deque队列

    private Deque<ProducerBatch> getOrCreateDeque(TopicPartition tp) {
        Deque<ProducerBatch> d = this.batches.get(tp);
        if (d != null)
            return d;
        d = new ArrayDeque<>();
        /*putIfAbsent线程安全的，保证一个分区只有一个队列*/
        Deque<ProducerBatch> previous = this.batches.putIfAbsent(tp, d);
        if (previous == null)
            return d;
        else
            return previous;
    }