系列文章目录
https://zhuanlan.zhihu.com/p/367683572
一. 使用方式
show the code.
public class KafkaProducerDemo {
public static void main(String[] args) {
// step 1: 设置必要参数
Properties config = new Properties();
config.setProperty(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG,
"127.0.0.1:9092,127.0.0.1:9093");
config.setProperty(ProducerConfig.ACKS_CONFIG, "-1");
config.setProperty(ProducerConfig.RETRIES_CONFIG, "3");
// step 2: 创建KafkaProducer
KafkaProducer<String, String> producer = new KafkaProducer<>(config);
// step 3: 构造要发送的消息
String topic = "kafka-source-code-demo-topic";
String key = "demo-key";
String value = "村口老张头: This is a demo message.";
ProducerRecord<String, String> record =
new ProducerRecord<>(topic, key, value);
// step 4: 发送消息
Future<RecordMetadata> future = producer.send(record);
}
}
step 1: 设置必要参数
代码中涉及的几个配置:
- bootstrap.servers:指定Kafka集群节点列表(全部 or 部分均可),用于KafkaProducer初始获取Server端元数据(如完整节点列表、Partition分布等等);
- acks:指定服务端有多少个副本完成同步,才算该Producer发出的消息写入成功(后面讲副本的文章会深入分析,这里按下不表);
- retries:失败重试次数;
更多参数可以参考ProducerConfig类中的常量列表。
step 2: 创建KafkaProducer
KafkaProducer两个模板参数指定了消息的key和value的类型
step 3:构造要发送的消息
- 确定目标topic;
String topic = "kafka-source-code-demo-topic";
- 确定消息的key
key用来决定目标Partition,这个下文细聊。String key = "demo-key";
- 确定消息体
这是待发送的消息内容,传递业务数据。String value = "村口老张头: This is a demo message.";
step 4:发送消息
Future<RecordMetadata> future = producer.send(record);
KafkaProducer中各类send方法均返回Future,并不会直接返回发送结果,其原因便是线程模型设计。
二. 线程模型
这里主要存在两个线程:主线程 和 Sender线程。主线程即调用KafkaProducer.send方法的线程。当send方法被调用时,消息并没有真正被发送,而是暂存到RecordAccumulator。Sender线程在满足一定条件后,会去RecordAccumulator中取消息并发送到Kafka Server端。
那么为啥不直接在主线程就把消息发送出去,非得搞个暂存呢?为了Kafka的目标之一——高吞吐。具体而言有两点好处:
- 可以将多条消息通过一个ProduceRequest批量发送出去;
- 提高数据压缩效率(一般压缩算法都是数据量越大越能接近预期的压缩效果);
三. 源码分析
先给个整体流程图,然后咱们再逐步分析。
1. 主线程
1.1 KafkaProducer属性分析
这里列出KafkaProducer的核心属性。至于全部属性说明,可参考我的"注释版Kafka源码":https://github.com/Hao1296/kafka
1.2 ProducerInterceptors
ProducerInterceptors,消息拦截器集合,维护了多个ProducerInterceptor对象。用于在消息发送前对消息做额外的业务操作。使用时可按如下方式设置:
Properties config = new Properties();
// interceptor.classes
config.setProperty(ProducerConfig.INTERCEPTOR_CLASSES_CONFIG,
"com.kafka.demo.YourProducerInterceptor,com.kafka.demo.InterceptorImpl2");
KafkaProducer<String, String> producer = new KafkaProducer<>(config);
ProducerInterceptor本身是一个接口:
public interface ProducerInterceptor<K, V> extends Configurable {
ProducerRecord<K, V> onSend(ProducerRecord<K, V> record);
void onAcknowledgement(RecordMetadata metadata, Exception exception);
void close();
}
其中,onAcknowledgement是得到Server端正确响应时的回调,后面再细说。onSend是消息在发送前的回调,可在这里做一些消息变更逻辑(如加减字段等)。输入原始消息,输出变更后的消息。KafkaProducer的send方法第一步就是执行ProducerInterceptor:
@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
// 关注这里
ProducerRecord<K, V> interceptedRecord = this.interceptors.onSend(record);
return doSend(interceptedRecord, callback);
}
// 该send方法重载核心逻辑仍是上面的send方法
@Override
public Future<RecordMetadata> send(ProducerRecord<K, V> record) {
return send(record, null);
}
1.3 元数据获取
接上文,ProducerInterceptors执行完毕后会直接调用doSend方法执行发送相关的逻辑。到这为止有个问题,我们并不知道目标Topic下有几个Partition,分别分布在哪些Broker上;故,我们也不知道消息该发给谁。所以,doSend方法第一步就是搞清楚消息集群结构,即获取集群元数据:
private Future<RecordMetadata> doSend(ProducerRecord<K, V> record, Callback callback) {
TopicPartition tp = null;
try {
throwIfProducerClosed();
ClusterAndWaitTime clusterAndWaitTime;
try {
// 获取集群元数据
clusterAndWaitTime = waitOnMetadata(record.topic(), record.partition(), maxBlockTimeMs);
} catch (KafkaException e) {
if (metadata.isClosed())
throw new KafkaException("Producer closed while send in progress", e);
throw e;
}
... ...
}
waiteOnMetadata方法内部大体分为2步:
private ClusterAndWaitTime waitOnMetadata(String topic, Integer partition, long maxWaitMs) throws InterruptedException {
// 第1步, 判断是否已经有了对应topic&partition的元数据
Cluster cluster = metadata.fetch();
Integer partitionsCount = cluster.partitionCountForTopic(topic);
if (partitionsCount != null && (partition == null || partition < partitionsCount))
// 若已存在, 则直接返回
return new ClusterAndWaitTime(cluster, 0);
// 第2步, 获取元数据
do {
... ...
// 2.1 将目标topic加入元数据对象
metadata.add(topic);
// 2.3 将元数据needUpdate字段置为true, 并返回当前元数据版本
int version = metadata.requestUpdate();
// 2.4 唤醒Sender线程
sender.wakeup();
// 2.5 等待已获取的元数据版本大于version时返回, 等待时间超过remainingWaitMs时抛异常
try {
metadata.awaitUpdate(version, remainingWaitMs);
} catch (TimeoutException ex) {
throw new TimeoutException(
String.format("Topic %s not present in metadata after %d ms.",
topic, maxWaitMs));
}
// 2.6 检查新版本元数据是否包含目标partition;
// 若包含, 则结束循环; 若不包含, 则进入下一个迭代, 获取更新版本的元数据
cluster = metadata.fetch();
......
partitionsCount = cluster.partitionCountForTopic(topic);
} while (partitionsCount == null || (partition != null && partition >= partitionsCount));
return new ClusterAndWaitTime(cluster, elapsed);
}
我们看到,waitOnMetadata的思想也和简单,即:唤醒Sender线程来更新元数据,然后等待元数据更新完毕。至于Sender线程是如何更新元数据的,放到下文详解。
1.4 Serialize
这一步是用通过"key.serializer"和"value.serializer"两个配置指定的序列化器分别来序列化key和value
private Future<RecordMetadata> doSend(ProducerRecord<K, V> record, Callback callback) {
.....
// key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);
}
// value序列化
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内置了几个Serializer,如果需要的话,诸君也可以自定义:
org.apache.kafka.common.serialization.StringSerializer;
org.apache.kafka.common.serialization.LongSerializer;
org.apache.kafka.common.serialization.IntegerSerializer;
org.apache.kafka.common.serialization.ShortSerializer;
org.apache.kafka.common.serialization.FloatSerializer;
org.apache.kafka.common.serialization.DoubleSerializer;
org.apache.kafka.common.serialization.BytesSerializer;
org.apache.kafka.common.serialization.ByteBufferSerializer;
org.apache.kafka.common.serialization.ByteArraySerializer;
1.5 Partition选择
到这里,我们已经有了Topic相关的元数据,但也很快遇到了一个问题:Topic下可能有多个Partition,作为生产者,该将待发消息发给哪个Partition?这就用到了上文提到过的KafkaProducer的一个属性——partitioner。
private Future<RecordMetadata> doSend(ProducerRecord<K, V> record, Callback callback) {
......
// 确定目标Partition
int partition = partition(record, serializedKey, serializedValue, cluster);
......
}
private int partition(ProducerRecord<K, V> record, byte[] serializedKey, byte[] serializedValue, Cluster cluster) {
// 若ProducerRecord中强制指定了partition, 则以该值为准
Integer partition = record.partition();
// 否则调用Partitioner动态计算对应的partition
return partition != null ?
partition :
partitioner.partition(
record.topic(), record.key(), serializedKey, record.value(), serializedValue, cluster);
}
在创建KafkaProducer时,可以通过"partitioner.class"配置来指定Partitioner的实现类。若未指定,则使用Kafka内置实现类——DefaultPartitioner。DefaultPartitioner的策略也很简单:若未指定key,则在Topic下多个Partition间Round-Robin;若指定了key,则通过key来hash到一个partition。
public int partition(String topic, Object key, byte[] keyBytes, Object value, byte[] valueBytes, Cluster cluster) {
List<PartitionInfo> partitions = cluster.partitionsForTopic(topic);
int numPartitions = partitions.size();
if (keyBytes == null) {
// 若未指定key
int nextValue = nextValue(topic);
List<PartitionInfo> availablePartitions = cluster.availablePartitionsForTopic(topic);
if (availablePartitions.size() > 0) {
int part = Utils.toPositive(nextValue) % availablePartitions.size();
return availablePartitions.get(part).partition();
} else {
// no partitions are available, give a non-available partition
return Utils.toPositive(nextValue) % numPartitions;
}
} else {
// hash the keyBytes to choose a partition
return Utils.toPositive(Utils.murmur2(keyBytes)) % numPartitions;
}
}
2. RecordAccumulator
RecordAccumulator作为消息暂存者,其思想是将目的地Partition相同的消息放到一起,并按一定的"规格"(由"batch.size"配置指定)划分成多个"批次"(ProducerBatch),然后以批次为单位进行数据压缩&发送。示意图如下:
RecordAccumulator核心属性如下:
RecordAccumulator有两个核心方法,分别对应"存"和"取":
/**
* 主线程会调用此方法追加消息
*/
public RecordAppendResult append(TopicPartition tp,
long timestamp,
byte[] key,
byte[] value,
Header[] headers,
Callback callback,
long maxTimeToBlock) throws InterruptedException;
/**
* Sender线程会调用此方法提取消息
*/
public Map<Integer, List<ProducerBatch>> drain(Cluster cluster,
Set<Node> nodes,
int maxSize,
long now);
3. Sender线程
3.1 NetworkClient
在分析Sender线程业务逻辑前,先来说说通信基础类。
NetworkClient有两个核心方法:
public void send(ClientRequest request, long now);
public List<ClientResponse> poll(long timeout, long now);
其中,send方法很有迷惑性。乍一看,觉得其业务逻辑是将request同步发送出去。然而,send方法其实并不实际执行向网络端口写数据的动作,只是将请求"暂存"起来。poll方法才是实际执行读写动作的地方(NIO)。当请求的目标channel可写时,poll方法会实际执行发送动作;当channel有数据可读时,poll方法读取响应,并做对应处理。
NetworkClient有一个核心属性:
/* 实际实现类为 org.apache.kafka.common.network.Selector */
private final Selectable selector;
send和poll方法都是通过selector来完成的:
public void send(ClientRequest request, long now) {
doSend(request, false, now);
}
private void doSend(ClientRequest clientRequest, boolean isInternalRequest, long now) {
... ...
doSend(clientRequest, isInternalRequest, now, builder.build(version));
}
private void doSend(ClientRequest clientRequest, boolean isInternalRequest, long now, AbstractRequest request) {
... ...
selector.send(send);
}
public List<ClientResponse> poll(long timeout, long now) {
... ...
this.selector.poll(Utils.min(timeout, metadataTimeout, defaultRequestTimeoutMs));
... ...
}
org.apache.kafka.common.network.Selector 内部则通过 java.nio.channels.Selector 来实现。
值得关注的一点是,NetworkClient的poll方法在调用Selector的poll方法前还有段业务逻辑:
// 在selector.poll前有此行逻辑
long metadataTimeout = metadataUpdater.maybeUpdate(now);
try {
this.selector.poll(Utils.min(timeout, metadataTimeout, defaultRequestTimeoutMs));
} catch (IOException e) {
log.error("Unexpected error during I/O", e);
}
metadataUpdater.maybeUpdate可以看出是为元数据更新服务的。其业务逻辑是:判断是否需要更新元数据;若需要,则通过NetworkClient.send方法将MetadataRequest也加入"暂存",等待selector.poll中被实际发送出去。
3.2 Sender线程业务逻辑
KafkaProducer中,和Sender线程相关的有两个属性:
在KafkaProducer的构造函数中被创建:
KafkaProducer(ProducerConfig config,
Serializer<K> keySerializer,
Serializer<V> valueSerializer,
Metadata metadata,
KafkaClient kafkaClient) {
... ...
this.sender = new Sender(logContext,
client,
this.metadata,
this.accumulator,
maxInflightRequests == 1,
config.getInt(ProducerConfig.MAX_REQUEST_SIZE_CONFIG),
acks,
retries,
metricsRegistry.senderMetrics,
Time.SYSTEM,
this.requestTimeoutMs,
config.getLong(ProducerConfig.RETRY_BACKOFF_MS_CONFIG),
this.transactionManager,
apiVersions);
String ioThreadName = NETWORK_THREAD_PREFIX + " | " + clientId;
this.ioThread = new KafkaThread(ioThreadName, this.sender, true);
this.ioThread.start();
... ...
}
Sender线程的业务逻辑也很清晰:
public void run() {
log.debug("Starting Kafka producer I/O thread.");
// 主循环
while (running) {
try {
run(time.milliseconds());
} catch (Exception e) {
log.error("Uncaught error in kafka producer I/O thread: ", e);
}
}
log.debug("Beginning shutdown of Kafka producer I/O thread, sending remaining records.");
// 下面是关闭流程
// okay we stopped accepting requests but there may still be
// requests in the accumulator or waiting for acknowledgment,
// wait until these are completed.
while (!forceClose && (this.accumulator.hasUndrained() || this.client.inFlightRequestCount() > 0)) {
try {
run(time.milliseconds());
} catch (Exception e) {
log.error("Uncaught error in kafka producer I/O thread: ", e);
}
}
if (forceClose) {
// We need to fail all the incomplete batches and wake up the threads waiting on
// the futures.
log.debug("Aborting incomplete batches due to forced shutdown");
this.accumulator.abortIncompleteBatches();
}
try {
this.client.close();
} catch (Exception e) {
log.error("Failed to close network client", e);
}
log.debug("Shutdown of Kafka producer I/O thread has completed.");
}
主循环中仅仅是不断调用另一个run重载,该重载的核心业务逻辑如下:
void run(long now) {
... ...
// 1. 发送请求,并确定下一步的阻塞超时时间
long pollTimeout = sendProducerData(now);
// 2. 处理端口事件,poll的timeout为上一步计算结果
client.poll(pollTimeout, now);
}
其中,sendProducerData会调用RecordAccumulator.drain方法获取待发送消息,然后构造ProduceRequest对象,并调用NetworkClient.send方法"暂存"。sendProducerData方法之后便是调用NetworkClient.poll来执行实际的读写操作。
四. 总结
本文分析了KafkaProducer的业务模型及核心源码实现。才疏学浅,不一定很全面,欢迎诸君随时讨论交流。后续还会有其他模块的分析文章,具体可见系列文章目录: https://zhuanlan.zhihu.com/p/367683572