k8s 1.12.1 的坑和解決方法

k8s 1.12.1 的坑和解決方法

pull 鏡像:

gcr.io 被牆，須要 pull 本身的鏡像，而後改 tag。具體須要 pull 哪些鏡像呢，kubeadm config images 可查看
我本身 build 的都放到了 https://github.com/FingerLiu/... ， 須要的話也能夠直接用裏面的腳本：

wget -O - https://raw.githubusercontent.com/FingerLiu/k8s.gcr.io/master/pull.sh | bash

提示沒權限：

kubeadm reset 重複安裝的時候，.kube 文件夾不會清空，但 key 已經從新生成了，全部會key secret 不匹配。解決辦法是清空 .kube 目錄，而後將 /etc/kubernetes/kube-admin.json 拷貝過來

coredns

pending,network not ready:安裝對應版本的 flannel。kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

1 node(s) had taints that the pod didn't tolerate:

默認 k8s 不容許往 master 節點裝東西，強行設置下容許：kubectl taint nodes --all node-role.kubernetes.io/master-

azure xxx:

kubelet 本身的 bug, 無視

The connection to the server localhost:8080 was refused :

sudo cp /etc/kubernetes/admin.conf $HOME/
sudo chown $(id -u):$(id -g) $HOME/admin.conf
export KUBECONFIG=$HOME/admin.conf

helm 沒權限

That's because you don't have permission to deploy tiller, add an account for it:

kubectl create serviceaccount --namespace kube-system tiller
serviceaccount "tiller" created

kubectl create clusterrolebinding tiller-cluster-rule --clusterrole=cluster-admin --serviceaccount=kube-system:tiller
clusterrolebinding "tiller-cluster-rule" created

kubectl patch deploy --namespace kube-system tiller-deploy -p '{"spec":{"template":{"spec":{"serviceAccount":"tiller"}}}}'
deployment "tiller-deploy" patched
Then run below to check it :

helm listhelm repo update