Kubernetes Kube-proxy failed to retrieve node info

2
votes

Trying to understand why I'm seeing this output from my kube-proxy logs

W0328 08:00:53.755379  1 server.go:468] Failed to retrieve node info: nodes "ip-172-31-55-175" not found
W0328 08:00:53.755505  1 proxier.go:249] invalid nodeIP, initialize kube-proxy with 127.0.0.1 as nodeIP

The cluster is working just fine, is that indicating an issue with the cluster configuration?

2
kuberneteskube-dnskubelet
Looks like a warning. Does this occur every time kube-proxy starts on the node? - ahmet alp balkan
How did you setup cluster? If you used kubeadm, you need specify the advertise addr or check your kubelete attributes that you specify your listen addr. It sometime happens when you have multiple eth interfaces. - Michael Nikitochkin

2 Answers

6
votes

Can you please show the output of the command kubectl get node?

Probably the registered name used when kubelet starts is different from the name that kube-proxy is using.

You can force the usage of this very same name starting kube-proxy with the --hostname-override directive, and point to the same name kubelet registered itself.

1
votes

For newcomers start the minions(nodes) services in this order. 1. kubelet 2. kube-proxy 3. docker

Incorrect follow ups can trip up the service registrations.