What happened?

LoadBalancer&NodePort service cannot access after upgrade from 1.20 to 1.22.
The NodeName field in v1.endpointslice.endpoint missed.

The convertion in apiserver emit the field Topology in v1beta1.endpointslice and not convert to field NodeName in v1.endpointslice.

And the kube-proxy on 1.22 only use the NodeName field to judge whether the endpoint is Local. So kube-proxy on each node removed the endpoint from NodePort. Then LoadBalancer&NodePort network broken.

What did you expect to happen?

The LoadBalancer&NodePort network should not broken duration upgrade.

How can we reproduce it (as minimally and precisely as possible)?

1. Create 1.20 cluster.
2. Craete LoadBalancer Service with ExterernalTrafficPolicy setting to Local
3. Upgrade the cluster.
4. Observe the cluster's LoadBalancer networking.

Anything else we need to know?

No response

Kubernetes version

$ kubectl version
Client Version: version.Info{Major:"1", Minor:"19", GitVersion:"v1.19.7", GitCommit:"1dd5338295409edcfff11505e7bb246f0d325d15", GitTreeState:"clean", BuildDate:"2021-01-13T13:23:52Z", GoVersion:"go1.15.5", Compiler:"gc", Platform:"darwin/amd64"}
Server Version: version.Info{Major:"1", Minor:"22+", GitVersion:"v1.22.3-aliyun.1", GitCommit:"7a509a37c6f3b8f82175df3fb149974999882a54", GitTreeState:"clean", BuildDate:"2022-05-24T08:59:55Z", GoVersion:"go1.16.9", Compiler:"gc", Platform:"linux/amd64"}

Cloud provider

OS version

# On Linux:
$ cat /etc/os-release
# paste output here
$ uname -a
# paste output here

# On Windows:
C:\> wmic os get Caption, Version, BuildNumber, OSArchitecture
# paste output here

Install tools

Container runtime (CRI) and version (if applicable)

Related plugins (CNI, CSI, ...) and versions (if applicable)