fix: 解决 redis mq 消息丢失问题

yudao-framework/yudao-spring-boot-starter-mq/src/main/java/cn/iocoder/yudao/framework/mq/job/RedisPendingMessageResendJob.java

@@ -3,6 +3,7 @@ package cn.iocoder.yudao.framework.mq.job;
 import cn.hutool.core.collection.CollUtil;
 import cn.iocoder.yudao.framework.mq.core.RedisMQTemplate;
 import cn.iocoder.yudao.framework.mq.core.stream.AbstractStreamMessageListener;
+import lombok.AllArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.redisson.api.RLock;
 import org.redisson.api.RedissonClient;
@@ -22,7 +23,9 @@ import java.util.Map;
  * 这个任务用于处理，crash 之后的消费者未消费完的消息
  */
 @Slf4j
+@AllArgsConstructor
 public class RedisPendingMessageResendJob {
+
     private static final String LOCK_KEY = "redis:pending:msg:lock";
 
     private final List<AbstractStreamMessageListener<?>> listeners;
@@ -30,13 +33,6 @@ public class RedisPendingMessageResendJob {
     private final String groupName;
     private final RedissonClient redissonClient;
 
-    public RedisPendingMessageResendJob(List<AbstractStreamMessageListener<?>> listeners, RedisMQTemplate redisTemplate, String groupName, RedissonClient redissonClient) {
-        this.listeners = listeners;
-        this.redisTemplate = redisTemplate;
-        this.groupName = groupName;
-        this.redissonClient = redissonClient;
-    }
-
     /**
      * 一分钟执行一次,这里选择每分钟的35秒执行，是为了避免整点任务过多的问题
      */
@@ -57,28 +53,26 @@ public class RedisPendingMessageResendJob {
 
     private void execute() {
         StreamOperations<String, Object, Object> ops = redisTemplate.getRedisTemplate().opsForStream();
-
         listeners.forEach(listener -> {
             PendingMessagesSummary pendingMessagesSummary = ops.pending(listener.getStreamKey(), groupName);
             // 每个消费者的 pending 队列消息数量
             Map<String, Long> pendingMessagesPerConsumer = pendingMessagesSummary.getPendingMessagesPerConsumer();
-            pendingMessagesPerConsumer.entrySet().forEach(entry -> {
-                String consumerName = entry.getKey();
-                Long pendingMessageCount = entry.getValue();
+            pendingMessagesPerConsumer.forEach((consumerName, pendingMessageCount) -> {
                 log.info("[processPendingMessage][消费者({}) 消息数量({})]", consumerName, pendingMessageCount);
 
                 // 从消费者的 pending 队列中读取消息
                 List<MapRecord<String, Object, Object>> records = ops.read(Consumer.from(groupName, consumerName), StreamOffset.create(listener.getStreamKey(), ReadOffset.from("0")));
-                if (CollUtil.isNotEmpty(records)) {
-                    for (MapRecord<String, Object, Object> record : records) {
-                        // 重新投递消息
-                        redisTemplate.getRedisTemplate().opsForStream().add(StreamRecords.newRecord()
-                                .ofObject(record.getValue()) // 设置内容
-                                .withStreamKey(listener.getStreamKey()));
+                if (CollUtil.isEmpty(records)) {
+                    return;
+                }
+                for (MapRecord<String, Object, Object> record : records) {
+                    // 重新投递消息
+                    redisTemplate.getRedisTemplate().opsForStream().add(StreamRecords.newRecord()
+                            .ofObject(record.getValue()) // 设置内容
+                            .withStreamKey(listener.getStreamKey()));
 
-                        // ack 消息消费完成
-                        redisTemplate.getRedisTemplate().opsForStream().acknowledge(groupName, record);
-                    }
+                    // ack 消息消费完成
+                    redisTemplate.getRedisTemplate().opsForStream().acknowledge(groupName, record);
                 }
             });
         });