kubernetes集群很久没看了,发现集群挂掉,无法启动,报错:
The connection to the server 172.26.237.196:6443 was refused - did you specify the right host or port?
查看kubelet是stop状态:
systemctl status kubelet
并且重启失败:systemctl restart kubelet
查看系统日志中关于kubelet的部分:
journalctl -fu kubelet
可以看到kubelet在不断重启(因为在daemon-reload的范围之内,stop后会自动重启)。
没有什么有用信息,手动启动kubelet看一下日志:
/usr/bin/kubelet
主要有两个错误:
Observed a panic: "invalid memory address or nil pointer dereference" (runtime error: invalid memory address or nil pointer dereference)
和
orphaned pod "0f660127-21c3-461d-b1a7-fc12f2dcafe7" found, but volume paths are still present on disk : There were a total of 4 errors similar to this. Turn up verbosity to see them
github有一个issue:
kubelet: Observed a panic: "invalid memory address or nil pointer dereference"
https://github.com/kubernetes/ingress-nginx/issues/5219
但是没有提供明确的解决方式。
由于是个人机器,不想花费太多时间,所以采用最直接的方式,直接重置集群后在init。
这也是我强烈不建议生产使用自建kubernetes集群,99%的公司承受不了这个成本。
kubeadm reset
重新初始化集群:
kubeadm init --ignore-preflight-errors='NumCPU' --kubernetes-version v1.19.3 --pod-network-cidr=10.244.0.0/16
记住下述命令,按需执行:
To start using your cluster, you need to run the following as a regular user:
要让集群work,需要执行:
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown (id -u):(id -g)
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
https://kubernetes.io/docs/concepts/cluster-administration/addons/
Then you can join any number of worker nodes by running the following on each as root:
kubeadm join 172.26.237.196:6443 --token loe9wq.rt1w12hyeaantw6z \
--discovery-token-ca-cert-hash sha256:5544027d81f4580cbab5f8b6318c6719470263338e7278b13c3e857cfa36ecf8
但是发现集群是ready,但是coredns无法创建:
describe coredns pod:
可以看到是网络问题。
老集群和新集群的网络配置发生了混乱,kubeadm reset重置后将相关网络配置都删除:
kubeadm reset
systemctl stop kubelet
systemctl stop docker
rm -rf /var/lib/cni/ /var/lib/kubelet/* /etc/cni/
ifconfig cni0 down
ifconfig flannel.1 down
ifconfig docker0 down
ip link delete cni0
ip link delete flannel.1
systemctl start docker
然后顺次执行:
kubeadm init --ignore-preflight-errors='NumCPU' --kubernetes-version v1.19.3 --pod-network-cidr=10.244.0.0/16
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown (id -u):(id -g)
kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml
可能coredns上pending状态:
descirbe pod coredns:
这是因为删除了集群的配置,需要执行下允许master调度pod的命令:
kubectl taint node future node-role.kubernetes.io/master:NoSchedule-
集群恢复。