消息队列-RockMQ-重试参数设置

最大重试消费实践

生产者和消费者都会遇到失败的问题。生产者比如网络失败，系统原因。消费者比如消费消息的过程中发生了异常，希望再次消费。

生产者重试

1 ResponseCode 通过ResponseCode 判断是否需要重试

producer.addRetryResponseCode(ResponseCode.FLUSH_SLAVE_TIMEOUT);
// 默认已经有了这些情况
private final Set<Integer> retryResponseCodes = new CopyOnWriteArraySet<Integer>(Arrays.asList(
          ResponseCode.TOPIC_NOT_EXIST,
          ResponseCode.SERVICE_NOT_AVAILABLE,
          ResponseCode.SYSTEM_ERROR,
          ResponseCode.NO_PERMISSION,
          ResponseCode.NO_BUYER_ID,
          ResponseCode.NOT_IN_CURRENT_UNIT
  ));
 // 更多的重试情况可以参考
 public class ResponseCode extends RemotingSysResponseCode {

    public static final int FLUSH_DISK_TIMEOUT = 10;

    public static final int SLAVE_NOT_AVAILABLE = 11;

    public static final int FLUSH_SLAVE_TIMEOUT = 12;

    public static final int MESSAGE_ILLEGAL = 13;

    public static final int SERVICE_NOT_AVAILABLE = 14;

    public static final int VERSION_NOT_SUPPORTED = 15;

    public static final int NO_PERMISSION = 16;

    public static final int TOPIC_NOT_EXIST = 17;
    public static final int TOPIC_EXIST_ALREADY = 18;
    public static final int PULL_NOT_FOUND = 19;

    public static final int PULL_RETRY_IMMEDIATELY = 20;

    public static final int PULL_OFFSET_MOVED = 21;

    public static final int QUERY_NOT_FOUND = 22;

    public static final int SUBSCRIPTION_PARSE_FAILED = 23;

    public static final int SUBSCRIPTION_NOT_EXIST = 24;

    public static final int SUBSCRIPTION_NOT_LATEST = 25;

    public static final int SUBSCRIPTION_GROUP_NOT_EXIST = 26;

    public static final int FILTER_DATA_NOT_EXIST = 27;

    public static final int FILTER_DATA_NOT_LATEST = 28;

    public static final int TRANSACTION_SHOULD_COMMIT = 200;

    public static final int TRANSACTION_SHOULD_ROLLBACK = 201;

    public static final int TRANSACTION_STATE_UNKNOW = 202;

    public static final int TRANSACTION_STATE_GROUP_WRONG = 203;
    public static final int NO_BUYER_ID = 204;

    public static final int NOT_IN_CURRENT_UNIT = 205;

    public static final int CONSUMER_NOT_ONLINE = 206;

    public static final int CONSUME_MSG_TIMEOUT = 207;

    public static final int NO_MESSAGE = 208;

    public static final int UPDATE_AND_CREATE_ACL_CONFIG_FAILED = 209;

    public static final int DELETE_ACL_CONFIG_FAILED = 210;

    public static final int UPDATE_GLOBAL_WHITE_ADDRS_CONFIG_FAILED = 211;

}

2 setRetryAnotherBrokerWhenNotStoreOK
如果发送失败，是否需要尝试发送到其他的 Broker 节点，就是没有特意的关注，到底是同步发送失败、还是异步发送失败，总之，只要是发送失败了后，看一看该变量，如果是 true 的话，那么就自动尝试将消息发送到其他的 Broker 节点
3 setRetryTimesWhenSendFailed：同步，参数默认值是2
4 setRetryTimesWhenSendAsyncFailed：异步，参数默认值是2

    /**
     * Maximum number of retry to perform internally before claiming sending failure in synchronous mode. </p>
     *
     * This may potentially cause message duplication which is up to application developers to resolve.
     */
    private int retryTimesWhenSendFailed = 2;

    /**
     * Maximum number of retry to perform internally before claiming sending failure in asynchronous mode. </p>
     *
     * This may potentially cause message duplication which is up to application developers to resolve.
     */
    private int retryTimesWhenSendAsyncFailed = 2;

异步的时候会设置，异步线程池ExecutorService：
producer.setAsyncSenderExecutor();

this.asyncSenderThreadPoolQueue = new LinkedBlockingQueue<Runnable>(50000);
        this.defaultAsyncSenderExecutor = new ThreadPoolExecutor(
            Runtime.getRuntime().availableProcessors(),
            Runtime.getRuntime().availableProcessors(),
            1000 * 60,
            TimeUnit.MILLISECONDS,
            this.asyncSenderThreadPoolQueue,
            new ThreadFactory() {
                private AtomicInteger threadIndex = new AtomicInteger(0);

                @Override
                public Thread newThread(Runnable r) {
                    return new Thread(r, "AsyncSenderExecutor_" + this.threadIndex.incrementAndGet());
                }
            });

消费者重试

1 设置最大消费次数
setMaxReconsumeTimes

consumer.setMaxReconsumeTimes();

/**
    * Max re-consume times. 
    * In concurrently mode, -1 means 16;
    * In orderly mode, -1 means Integer.MAX_VALUE.
    *
    * If messages are re-consumed more than {@link #maxReconsumeTimes} before success.
    */
   private int maxReconsumeTimes = -1;

2 通过消费状态，设置过一会儿再重新消费
ConsumeOrderlyStatus#SUSPEND_CURRENT_QUEUE_A_MOMENT

ConsumeConcurrentlyStatus#RECONSUME_LATER