When kubelet tries to start on my Kubernetes worker nodes, I'm getting messages like this in the system log:

May 25 19:43:57 ip-10-240-0-223 kubelet[4882]: I0525 19:43:57.627389    4882 kubelet_node_status.go:82] Attempting to register node worker-1
May 25 19:43:57 ip-10-240-0-223 kubelet[4882]: E0525 19:43:57.628967    4882 kubelet_node_status.go:106] Unable to register node "worker-1" with API server: nodes is forbidden: User "system:node:" cannot create nodes at the cluster scope: unknown node for user "system:node:"
May 25 19:43:58 ip-10-240-0-223 kubelet[4882]: E0525 19:43:58.256557    4882 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/kubelet.go:451: Failed to list *v1.Service: services is forbidden: User "system:node:" cannot list services at the cluster scope: unknown node for user "system:node:"
May 25 19:43:58 ip-10-240-0-223 kubelet[4882]: E0525 19:43:58.257381    4882 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/config/apiserver.go:47: Failed to list *v1.Pod: pods is forbidden: User "system:node:" cannot list pods at the cluster scope: unknown node for user "system:node:"

If I'm reading these correctly, the problem is that the node is using the username system:node: when connecting to the API server rather than system:node:worker-1. But as far as I can tell, it should be using a worker-specific one. Here's my kubeconfig (with private stuff elided):

apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: [elided]
    server: https://[elided]:6443
  name: kubernetes-the-hard-way
contexts:
- context:
    cluster: kubernetes-the-hard-way
    user: system:node:worker-1
  name: default
current-context: default
kind: Config
preferences: {}
users:
- name: system:node:worker-1
  user:
    client-certificate-data:  [elided]
    client-key-data:  [elided]

I was under the impression that the users specified there were the ones used when contacting the API, but clearly I'm wrong. Is there somewhere else I've missed out a reference to worker-1?

I'm following the Kubernetes the Hard Way tutorial, but adjusting it for AWS as I go, so this problem is almost certainly a mistake I made when adjusting the config files. If there are any other config files that I should provide to make this easier/possible to debug, please do let me know.