Monkstar
Monkstar
> It's hard to say anything without at least some skeleton code that shows a rough sequence/flow of API calls that results in this behavior. I feared as much. I'm...
I just found a possibly related problem. Investigating into why one node apparently lost its peering connection with another node while the other node didn't notice this, I checked `netstat`...
> However, if the TCP were to close at the unpeer step, we'd maybe get the desired/expected behavior. I believe so, too. If you could manage to implement that, it...
> I think detecting and dropping duplicated peering requests is the better solution. This is already being done and part of the problem - to improve this, it would mean...
Well, I am not a Broker developer. I merely use it and I would actually prefer not to interfere with it below its API. But thanks for the hint.
I have the same issue. However, it only seems to affect some of the children; not all of them. And apparently it can start working again after a growth moment.