kubeadm upgrade node panics if config has KubeletConfiguration

[nadiamoe]

[root@moniserver ~]# kubeadm --config /etc/kubernetes/kubeadm.yaml upgrade node
W0215 12:30:36.826324  766880 upgradeconfiguration.go:135] [config] WARNING: YAML document with Component Configs kubelet.config.k8s.io is deprecated for upgrade and will be ignored
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x5f2c6db5b612]

goroutine 1 [running]:
k8s.io/kubernetes/cmd/kubeadm/app/util/config.prepareStaticVariables({0x5f2c6e75d560, 0x0})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/util/config/common.go:398 +0x152
k8s.io/kubernetes/cmd/kubeadm/app/util/config.LoadOrDefaultUpgradeConfiguration({0x7ffd1cac1d91, 0x1c}, 0xc00074fba0, {0x0, 0x0})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/util/config/upgradeconfiguration.go:160 +0x1a5
k8s.io/kubernetes/cmd/kubeadm/app/cmd/upgrade.newNodeData(0xc000776908, 0xc000458910, {0x5f2c6e890ae8, 0xc00007c030})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/cmd/upgrade/node.go:184 +0x3ba
k8s.io/kubernetes/cmd/kubeadm/app/cmd/upgrade.newCmdNode.func2(0xc000776908, {0xc00006d720, 0x0, 0x2})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/cmd/upgrade/node.go:127 +0x9b
k8s.io/kubernetes/cmd/kubeadm/app/cmd/phases/workflow.(*Runner).InitData(0xc00016e240, {0xc00006d720, 0x0, 0x2})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/cmd/phases/workflow/runner.go:185 +0xd8
k8s.io/kubernetes/cmd/kubeadm/app/cmd/upgrade.newCmdNode.func1(0xc000776908, {0xc00006d720, 0x0, 0x2})
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/cmd/upgrade/node.go:90 +0xce
github.com/spf13/cobra.(*Command).execute(0xc000776908, {0xc00006d700, 0x2, 0x2})
	/build/kubernetes/src/kubernetes-1.32.1/vendor/github.com/spf13/cobra/command.go:985 +0xfca
github.com/spf13/cobra.(*Command).ExecuteC(0xc000452608)
	/build/kubernetes/src/kubernetes-1.32.1/vendor/github.com/spf13/cobra/command.go:1117 +0x9d0
github.com/spf13/cobra.(*Command).Execute(0xc000452608)
	/build/kubernetes/src/kubernetes-1.32.1/vendor/github.com/spf13/cobra/command.go:1041 +0x32
k8s.io/kubernetes/cmd/kubeadm/app.Run()
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/app/kubeadm.go:47 +0xd1
main.main()
	/build/kubernetes/src/kubernetes-1.32.1/cmd/kubeadm/kubeadm.go:25 +0x13

[root@moniserver ~]# cat /etc/kubernetes/kubeadm.yaml
apiVersion: kubelet.config.k8s.io/v1beta1
kind: KubeletConfiguration
# Use a resolv.conf file without the curry domain/search to prevent ndots spam.
resolvConf: /etc/resolv-k8s.conf
# Wait this duration for pods to gracefully terminate.
shutdownGracePeriod: 120s
# Of the duration above, the following duration will be reserved for terminating critical pods after non-critical ones.
shutdownGracePeriodCriticalPods: 20s

Not specifying a --config with KubeletConfiguration for this operation works around the issue, but I thin it would be best if the config was gracefully ignored as the log message indicates.

[neolit123]

kubeletconfig on upgrade is not supported but the panic should be fixed. we should also backport the fix for >= 1.31.