罗泽轩

IMHO, it would make thing complex if we put the connection management inside the redis client. If we ensure there is only one coroutine that can write to a cosocket,...

I don't have mingw64 to test with. But I guess it's because there is a space in the path of the cmd ps?

> Do we have the implementation status in the different proxy implementations ? I think any proposed API should start with the list of proxies that provide the feature already...

@bzp2010 @shreemaan-abhishek Maybe you can fix the broken CI and kick off a new release? It won't take too much of your time to do this, probably just three or...

> Is it possible the new one gets a new IP allocated since the allocation isn't stable? The allocation only happens in autoAllocateIPs function. Its comment says that the allocation...

> Services don't share the same IP, there is a conflict resolution mechanism to assign it a different one. So it's not a stable function but rather depends on the...

> istio/pilot/pkg/serviceregistry/serviceentry/controller.go Thanks for pointing it out. So it's a bit tricky to solve the problem. Let me think about it a little more.

@howardjohn What do you think about moving the IP allocation from the read part to the write one? We can allocate the IP only for the changed services in these...

> A ServiceEntry change event needs to trigger a full push because LDS or RDS also relies on the ServiceEntry. If only endpoints are changed, istio can only send an...

The unit tests failed because we added an upstream check and they can't satisfy. Could we mock the upstream check in the test?