rocketmq 发送顺序消息和普通消息的主流程区别大部分一致的,区别在于:普通消息发送时,从所有broker的队列集合中 轮询选择一个队列,而顺序队列可以提供用户自定义消息队列选择器,从NameServer 分配的顺序 broker集合中选择一个队列。
源码版本:4.9.3
源码架构图
发送普通消息源码在另外一篇文章https://blog.csdn.net/hzwangmr/article/details/135411495,这里主要阅读和普通消息有差异的部分。
org.apache.rocketmq.client.producer.DefaultMQProducer#send(org.apache.rocketmq.common.message.Message, org.apache.rocketmq.client.producer.MessageQueueSelector, java.lang.Object)
可以看到系统提供了一个 MessageQueueSelector 消息队列选择器,用于自定义选择队列的逻辑。
/**
* Same to {@link #send(Message)} with message queue selector specified.
*
* @param msg Message to send.
* @param selector Message queue selector, through which we get target message queue to deliver message to.
* @param arg Argument to work along with message queue selector.
* @return {@link SendResult} instance to inform senders details of the deliverable, say Message ID of the message,
* {@link SendStatus} indicating broker storage/replication status, message queue sent to, etc.
* @throws MQClientException if there is any client error.
* @throws RemotingException if there is any network-tier error.
* @throws MQBrokerException if there is any error with broker.
* @throws InterruptedException if the sending thread is interrupted.
*/
@Override
public SendResult send(Message msg, MessageQueueSelector selector, Object arg)
throws MQClientException, RemotingException, MQBrokerException, InterruptedException {
msg.setTopic(withNamespace(msg.getTopic()));
return this.defaultMQProducerImpl.send(msg, selector, arg);
}
org.apache.rocketmq.client.impl.producer.DefaultMQProducerImpl#sendSelectImpl
这里可以看到在调用发送消息核心 sendKernelImpl() 方法之前,会调用 selector.select() 函数,执行我们自定的选择逻辑。那么自定义的逻辑具体是什么呢?接着往下看
private SendResult sendSelectImpl(
Message msg,
MessageQueueSelector selector,
Object arg,
final CommunicationMode communicationMode,
final SendCallback sendCallback, final long timeout
) throws MQClientException, RemotingException, MQBrokerException, InterruptedException {
long beginStartTime = System.currentTimeMillis();
this.makeSureStateOK();
Validators.checkMessage(msg, this.defaultMQProducer);
// 获取topic路由数据
TopicPublishInfo topicPublishInfo = this.tryToFindTopicPublishInfo(msg.getTopic());
if (topicPublishInfo != null && topicPublishInfo.ok()) {
MessageQueue mq = null;
try {
// 解析发布消息队列
List messageQueueList =
mQClientFactory.getMQAdminImpl().parsePublishMessageQueues(topicPublishInfo.getMessageQueueList());
// 克隆消息
Message userMessage = MessageAccessor.cloneMessage(msg);
String userTopic = NamespaceUtil.withoutNamespace(userMessage.getTopic(), mQClientFactory.getClientConfig().getNamespace());
userMessage.setTopic(userTopic);
// 利用消息队列选择器,选择一个队列
mq = mQClientFactory.getClientConfig().queueWithNamespace(
// 自定义选择队列
selector.select(messageQueueList, userMessage, arg)
);
} catch (Throwable e) {
throw new MQClientException("select message queue threw exception.", e);
}
long costTime = System.currentTimeMillis() - beginStartTime;
if (timeout < costTime) {
throw new RemotingTooMuchRequestException("sendSelectImpl call timeout");
}
if (mq != null) {
// 如果选择出来的 MessageQueue 存在,这调用核心发送消息函数,发送消息
return this.sendKernelImpl(msg, mq, communicationMode, sendCallback, null, timeout - costTime);
} else {
throw new MQClientException("select message queue return null.", null);
}
}
validateNameServerSetting();
throw new MQClientException("No route info for this topic, " + msg.getTopic(), null);
}
可以看到,工程自带的发送顺序消息的example实例,针对 id相同的数据,选择了相同的消息队列,这样对于同一个实体的数据变化一定是有顺序的。
public class Producer {
public static void main(String[] args) throws UnsupportedEncodingException {
try {
DefaultMQProducer producer = new DefaultMQProducer("please_rename_unique_group_name");
producer.start();
String[] tags = new String[] {"TagA", "TagB", "TagC", "TagD", "TagE"};
for (int i = 0; i < 100; i++) {
int orderId = i % 10;
Message msg =
new Message("TopicTestjjj", tags[i % tags.length], "KEY" + i,
("Hello RocketMQ " + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
SendResult sendResult = producer.send(msg, new MessageQueueSelector() {
@Override
public MessageQueue select(List mqs, Message msg, Object arg) {
// 自定义选择逻辑,可以理解为将相同的 orderId订单id,投递到相同的队列中
Integer id = (Integer) arg;
int index = id % mqs.size();
return mqs.get(index);
}
}, orderId);
System.out.printf("%s%n", sendResult);
}
producer.shutdown();
} catch (MQClientException | RemotingException | MQBrokerException | InterruptedException e) {
e.printStackTrace();
}
}
}
mq生产者客户端,在发送消息前,会从 NameServer中读取消息队里指定 topic对应的 topic路由信息,然后写到转换和缓存在内存多个数据结构里。其中,有一处就是下方的将 topic路由数据转换为topic 已发布信息,这里写入了messageQueueList (顺序消息待选择队列)。
转换逻辑是:
public static TopicPublishInfo topicRouteData2TopicPublishInfo(final String topic, final TopicRouteData route) {
TopicPublishInfo info = new TopicPublishInfo();
// 写入路由数据
info.setTopicRouteData(route);
// 顺序消息配置,顺序消息分配了多少个broker
if (route.getOrderTopicConf() != null && route.getOrderTopicConf().length() > 0) {
// 是顺序消息
// 解析顺序消息topic配置
String[] brokers = route.getOrderTopicConf().split(";");
// 遍历顺序 broker,从namersrv拿到的broker数量
for (String broker : brokers) {
String[] item = broker.split(":");
// 特定broker中的队列数量
int nums = Integer.parseInt(item[1]);
for (int i = 0; i < nums; i++) {
// 分装队列数据,topic--brokerName--队列id
MessageQueue mq = new MessageQueue(topic, item[0], i);
info.getMessageQueueList().add(mq);
}
}
// 是顺序topic,打标
info.setOrderTopic(true);
}
}