cserwen comments

Results 25 comments of


                                            cserwen

[ISSUE #4463] fix broker is normal exit, but indexFile not flush disk

> > > Is it possible to add the logic of regular flushing of `IndexFile`, just like refreshing `ConsumeQueue`, `StoreCheckpoint`, and `CommitLog` regularly. > > > > > > i...

[ISSUE#4552] Add asySendBlockMode to implement BackPressure

> > > > The new API uses a new protocol model. it is an accomplishment, not a replacement. > > BTW, the new API needs a long time to...

the interface `rocketmq.Producer` can't be implemented outside

In what case do we need to implement this interface ourselves?

After mq message consumption fails, the number of the message in the retry queue continues to grow

有一种可能是消费端存在重复的clientID，导致rebalance错误，即重试队列被多个实例同时消费，举例：假设有三个重复的clientID，都在消费重试队列，一条消息消费失败后会变成三条，三条会被每个客户端消费，然后消费失败变成9条。。。就会呈现指数值增加 @sologgfun

After mq message consumption fails, the number of the message in the retry queue continues to grow

没有用处，可以删掉 @sologgfun

fix WithRecover 无法recover问题

@snail-plus Maybe it would be better to have a simple unit test.

primitive WithRecover do not work

是的，目前是这样的。。。感兴趣可以修复下 @snail-plus

[Panic] init trace client, the same intstance

也许通过traceConfig 设置不同的groupName 可以解决这个问题 @iGoogle-ink

[Panic] init trace client, the same intstance

> > 也许通过traceConfig 设置不同的groupName 可以解决这个问题 @iGoogle-ink > > 试过，不行，trace用的是 default的，默认写死的一个 group 你能发下相关的代码在哪里吗，我看的是根据传入的groupName 设置的 @iGoogle-ink

[Panic] init trace client, the same intstance

@iGoogle-ink 我仔细看下了你的报错截图，你应该是两个客户端使用了不同的namesrv地址导致的，具体报错在这一行 https://github.com/apache/rocketmq-client-go/blob/a15c7771e84b00a33ba6f054d8c31d85654e2dac/internal/client.go#L209 可以在tarcer的` instanceName` 上默认加上 namesrv地址，具体可以参考 Java SDK 的实现，https://github.com/apache/rocketmq/blob/develop/client/src/main/java/org/apache/rocketmq/client/trace/AsyncTraceDispatcher.java#L108