O Kubenet não possui o netConfig

Question

Eu configurei um cluster kubernetes com CNI para redes, conexões RBAC e TLS. Estou preso há algumas horas para conseguir meus nós prontos.

root@master ~ # kubectl get nodes
NAME                STATUS     AGE       VERSION
us1-worker-node01   NotReady   3h        v1.6.1
us1-worker-node02   NotReady   3h        v1.6.1

O erro pareceu vir do kubelet

[...]
juin 14 13:54:03 us1-worker-node01 kubelet[19570]: I0614 13:54:03.192345   19570 manager.go:288] Starting recovery of all containers
juin 14 13:54:03 us1-worker-node01 kubelet[19570]: I0614 13:54:03.239022   19570 manager.go:293] Recovery completed
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.754864   19570 kubelet_node_status.go:230] Setting node annotation to enable volume controller attach/detach
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.760381   19570 kubelet_node_status.go:379] Recording NodeHasSufficientDisk event message for node us1-worker-node01
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.760566   19570 kubelet_node_status.go:379] Recording NodeHasSufficientMemory event message for node us1-worker-node01
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.760721   19570 kubelet_node_status.go:379] Recording NodeHasNoDiskPressure event message for node us1-worker-node01
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.760872   19570 kubelet_node_status.go:77] Attempting to register node us1-worker-node01
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.791650   19570 kubelet_node_status.go:128] Node us1-worker-node01 was previously registered
juin 14 13:54:02 us1-worker-node01 kubelet[19570]: I0614 13:54:02.791837   19570 kubelet_node_status.go:80] Successfully registered node us1-worker-node01
juin 14 13:54:08 us1-worker-node01 kubelet[19570]: I0614 13:54:08.171841   19570 kubelet.go:1808] SyncLoop (ADD, "api"): ""
juin 14 13:54:07 us1-worker-node01 kubelet[19570]: E0614 13:54:07.778441   19570 kubelet.go:2067] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: Kubenet does not have netConfig. This is most likely due to lack of PodCIDR
juin 14 13:54:12 us1-worker-node01 kubelet[19570]: E0614 13:54:12.779166   19570 kubelet.go:2067] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: Kubenet does not have netConfig. This is most likely due to lack of PodCIDR
juin 14 13:54:17 us1-worker-node01 kubelet[19570]: E0614 13:54:17.781032   19570 kubelet.go:2067] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: Kubenet does not have netConfig. This is most likely due to lack of PodCIDR
juin 14 13:54:22 us1-worker-node01 kubelet[19570]: E0614 13:54:22.783057   19570 kubelet.go:2067] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: Kubenet does not have netConfig. This is most likely due to lack of PodCIDR
juin 14 13:54:27 us1-worker-node01 kubelet[19570]: E0614 13:54:27.783449   19570 kubelet.go:2067] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: Kubenet does not have netConfig. This is most likely due to lack of PodCIDR

O último erro é repetido várias vezes ...

aqui está a configuração do serviço do kubelet

[Unit]
Description=Kubernetes Kubelet
Documentation=https://github.com/GoogleCloudPlatform/kubernetes
After=docker.service
Requires=docker.service

[Service]
ExecStart=/usr/local/bin/kubelet \
  --allow-privileged=true \
  --cluster-dns=10.250.0.10 \
  --cluster-domain=cluster.local \
  --container-runtime=docker \
  --network-plugin=kubenet \
  --kubeconfig=/var/lib/kubelet/kubelet.kubeconfig \
  --serialize-image-pulls=false \
  --register-node=true \
  --require-kubeconfig \
  --tls-cert-file=/var/lib/kubelet/kubelet-client.pem \
  --tls-private-key-file=/var/lib/kubelet/kubelet-client.key \
  --v=2
Restart=on-failure
RestartSec=5s

[Install]
WantedBy=multi-user.target

Os binários CNI estão instalados em / opt / cni / bin. O certificado de assunto do Kubelet é CN = system: node: us1-worker-node01, O = system: nós, OU = Cluster

Alguma idéia?

    

Answer 1

OK ...

O kube-controller-manager não teve acesso ao CA.key para as opções: - service-account-private-key-file e - arquivo-chave de assinatura de cluster

Com o bom CA.key, não há mais problemas!