问题描述
在AKS集群中部署calico时候,遇见 ImagePullBackOff 问题。 在创建POD calico-typha-horizontal-autoscale 时候遇见拉取镜像失败问题。错误消息为:
Failed to pull image "k8s.gcr.io/cluster-proportional-autoscaler-amd64:1.1.2-r2": rpc error: code = Unknown desc = Error response from daemon: Get https://k8s.gcr.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
问题解答
这个问题的原因是从中国区访问全球的镜像库时,遇见了无法连接的问题。所以需要修改镜像的地址,在中国区有专门为一些基础镜像做的代理库。
比如:k8s.gcr.io 修改成中国区专用的 gcr.azk8s.cn/google_containers/
通常情况下,AKS组件的yaml文件存储在主节点中的 /etc/kubernetes/addons 目录中,以下示例时修改主节点中的 /etc/kubernetes/addons/calico-daemonset.yaml 文件。
修改前:
metadata: name: calico-typha-horizontal-autoscaler namespace: kube-system … … containers: - image: k8s.gcr.io/cluster-proportional-autoscaler-amd64:1.1.2-r2 name: autoscaler
修改后:
metadata: name: calico-typha-horizontal-autoscaler namespace: kube-system … … containers: - image: gcr.azk8s.cn/google_containers/cluster-proportional-autoscaler-amd64:1.1.2-r2 name: autoscaler
然后,用一下的命令apply这次的改动:
kubectl apply -f /etc/kubernetes/addons/calico-daemonset.yaml
这样,Kubernetes 旧可以拉取镜像并创建POD。
Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 3m11s default-scheduler Successfully assigned kube-system/calico-typha-horizontal-autoscaler-5b8f8b8479-rj6ff to k8s-xxx-111-0 Normal Pulling 3m8s kubelet, k8s-xxx-111-0 pulling image "gcr.azk8s.cn/google_containers/cluster-proportional-autoscaler-amd64:1.1.2-r2" Normal Pulled 3m4s kubelet, k8s-xxx-111-0 Successfully pulled image "gcr.azk8s.cn/google_containers/cluster-proportional-autoscaler-amd64:1.1.2-r2" Normal Created 3m4s kubelet, k8s-xxx-111-0 Created container Normal Started 3m2s kubelet, k8s-xxx-111-0 Started container
参考资料
Container Registry Proxy : https://github.com/Azure/container-service-for-azure-china/blob/master/aks/README.md#22-container-registry-proxy