Liu Shengzhong

> 2023-07-31T09:13:52.19左右出现过mq服务器重启，mq重启之后，就出现了 brokerName=rocketmq-broker-b, queueId=1的消息堆积，直到2023-07-31T13:13:52.195Z才恢复，offset从529直接跳到了553 > > 恢复正常时候的mq日志如下： 2023-07-31T13:13:52.195Z WARN mq/producer.go:100 fetch offset of mq from broker success {"consumerGroup": "op_controller_cluster_ph", "MessageQueue": "MessageQueue [topic=ydocs_op_uplink_porder, brokerName=rocketmq-broker-b, queueId=1]", "offset": 529} > > 2023-07-31T13:13:52.195Z DEBUG...

@ltamber PTAL

> Maybe we should refactor it to like ` DefaultMQPushConsumer`, initialize the `TraceDispatcher` on the constructor. It was before the [commit](https://github.com/apache/rocketmq/commit/2dc75d32099cf12e3caff0b32c7fa5fd1ce66a4c). And bug in this issue is also introduced by...

If confirmed, I can try to make a fix.

> CompletableFuture future = getBizMessage(popCheckPoint.getTopic(), msgOffset, popCheckPoint.getQueueId(), popCheckPoint.getBrokerName()) > > 这里建议加个注释，Boolean 表示消费结果，false 时需要 rePutCk() ok，加了

@drpmma @lizhimins I think that the custom NettyServerHandler#channelWritabilityChanged can achieve the purpose of back-pressure. Once the high watermark of write buffer is reached, reading data from this channel will be...

Configure your consumer group as order with "-o" option. ./mqadmin updateSubGroup -c DefaultCluster -n 192.168.10.233:9876 -g OrderConsumerGroupV3 -o true

This exception indicates the pollingTime (awaitDuration) set by client is too small. There are two configs related to pollingtime in proxy, grpcClientConsumerMinLongPollingTimeoutMillis (default=5s) and grpcClientConsumerMaxLongPollingTimeoutMillis (default=20s). Set the pollingTime within...

Transaction log indicates that the broker cannot find producer channel when trying to call client CheckTransactionState. In this case, the producer send message before its first heartbeat, which causes offline...

> @redlsz need rebase develop @lizhimins rebase done