sealos部署K8s，安装docker时master节点突然NotReady

1、集群正常运行中，在集群master-1上安装了docker+harbor，却发现master-1节点NotReady，使用的网络插件为 Cilium

bash 复制代码

#安装docker和harbor（docker运行正常）
root@master-1:/etc/apt# apt install docker-ce=5:19.03.15~3-0~ubuntu-focal docker-ce-cli=5:19.03.15~3-0~ubuntu-focal docker-compose

#安装harbor（harbor运行正常）
root@master-1:/usr/local/harbor# ./install.sh --with-trivy

#开始出现问题
root@master-1:~# crictl images

FATA[0000] validate service connection: validate CRI v1 image API for endpoint "unix:///var/run/image-cri-shim.sock": rpc error: code = Unimplemented desc = unknown service runtime.v1.ImageService

bash 复制代码

#验证sock存在
root@master-1:~# ll /var/run/image-cri-shim.sock
srw-rw---- 1 root root 0 Nov  8 13:26 /var/run/image-cri-shim.sock=

#服务状态没有明显异常
root@master-1:~# systemctl status containerd.service image-cri-shim.service

bash 复制代码

#查看kubelet服务启动失败
root@master-1:~# systemctl status kubelet.service

bash 复制代码

#检查服务日志看到明显报错（基本来看，这是gRPC 的错误，由于kubelet无法找到 CRI v1 运行时服务而引起的。）
root@master-1:~# journalctl -u kubelet.service

查看节点信息master-1失败

bash 复制代码

root@master-2:~# kubectl get nodes
NAME       STATUS     ROLES           AGE   VERSION
node-2     Ready      <none>          9d    v1.27.7
master-1   NotReady   control-plane   9d    v1.27.7
master-2   Ready      control-plane   9d    v1.27.7
master-3   Ready      control-plane   9d    v1.27.7
node-1     Ready      <none>          9d    v1.27.7

看了一下节点的iptables规则

bash 复制代码

root@master-1:~# iptables -L -v -n

对比了一下master-2节点的（这看起来明显是docker把所有的配置都篡改了）

bash 复制代码

root@master-2:/etc/containerd# iptables -L -v -n

这里我停了docker和harbor

bash 复制代码

docker-compose down
systemctl disable --now docker

检查sock文件是存在的

bash 复制代码

root@master-2:~# ll /run/containerd/containerd.sock
srw-rw---- 1 root root 0 Nov  8 03:20 /run/containerd/containerd.sock=

查看containerd配置文件（这里发现了错误，和master-2节点对比发现文件不同，感觉是安装docker时有被篡改的感觉配置完全变了）

bash 复制代码

root@master-1:~# vim /etc/containerd/config.toml
root@master-2:~# vim /etc/containerd/config.toml

2、解决

将master-2节点上的拷贝过来

bash 复制代码

#拷贝文件
root@master-2:~# scp /etc/containerd/config.toml root@172.16.100.241://etc/containerd/config.toml

#重启服务
root@master-1:~# systemctl daemon-reload && systemctl restart containerd kubelet

#查看状态都正常
root@master-1:~# systemctl status containerd kubelet

#查看集群状态
root@master-1:~# kubectl get nodes
NAME       STATUS   ROLES           AGE   VERSION
node-2     Ready    <none>          10d   v1.27.7
master-1   Ready    control-plane   10d   v1.27.7
master-2   Ready    control-plane   10d   v1.27.7
master-3   Ready    control-plane   10d   v1.27.7
node-1     Ready    <none>          10d   v1.27.7

查看一下etcd的状态

bash 复制代码

#查看了etcd的状态和集群的选举都是OK的，自查完成
root@master-1:~#  kubectl exec -it etcd-master-1 -n kube-system \
> -- etcdctl --write-out=table member list --endpoints=https://172.16.100.241:2379 \
> --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt \
> --key=/etc/kubernetes/pki/etcd/server.key endpoint health