Cliff Burdick

I too saw a connection reset by peer today, which caused the thread to hang. Here was the output: ``` Root task 'poster of events' is failed: [Errno 104] Connection...

Hi @nolar, I'm seeing the same thing with an on.delete handler for pods. Do you need help reproducing it? It seems to happen every time the pod is deleted. Here...

@nolar / @brutus333 I think I see what's happening. Like the OP, I have a controller for a CRD that creates pods. I think what should happen is kopf would...

A bit more information: it looks like the pod did indeed have the finalizer handle to begin with. The on.delete handler is being called when the pod is deleted, and...

@nolar sure! I'll try it out and report back. For what it's worth, when this happens `requires_finalizer` was True, and `has_finalizer` was False

@nolar I can confirm that 0.27rc6 indeed fixes the problem! I did notice a lot more aiohttp traffic to the k8s server while the pod was active compared to 0.25,...

@nolar sure. It was rps -- the bytes didn't increase much. I had some debug print statements in the aio library from trying to debug this issue, and saw those...

> @cliffburdick PS: 0.27rc6 is going to be 0.27 in a few days — I have finally finished testing it in action. But test it carefully before upgrading anyway —...

@nolar I think this is related to #301 I opened. The way I'm imagining it is you create a custom scheduler using kopf that simply looks for a CRD to...

> @cliffburdick Are you talking about CRDs or CRs? > > If it is about CRs (i.e. objects), then it is already doable now when the parent's handler puts the...