RickyMa
RickyMa
I think we already support this. Closed.
Can we close this? @xianjingfeng
The root cause of this issue is https://github.com/apache/incubator-uniffle/issues/1626, which will be fixed by https://github.com/apache/incubator-uniffle/pull/1627. And I think maybe we can improve this: when there are flush events being dropped(We need...
@maobaolong You can go ahead if you want. I think @qijiale76 might not have time on this. We now mostly use Netty for sending/getting shuffle data. We only use gRPC...
@EnricoMi It's been a while. Could you please continue working on this?
Do you know the real reason behind the failure of unregistering shuffles? I'm actually quite curious about this. It certainly isn't because of `ShuffleWriteClientImpl.close() / RegisterHeartBeat.shutdown()` being called. Because this...
> I think it is because the executor JVM is shutting down. That might interrupt the waiting threads (`InterruptedException`). Yeah, what improvement measures can we take to avoid this situation...
Normally, the executor should shutdown gracefully. `InterruptedException` is not a desired behavior.
> @rickyma what about #1736? The idea looks good to me.