centos8安装kubernets，k8s 1.18.6 后，重启系统后,k8s,kubelet没法启动问题

centos8安装kubernets，k8s 1.18.6 后，重启系统后,k8s没有启动。

查看docker ps  也没有容器实例启动。

journalctl -xe 或是cat /var/log/messages |tail  -30

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.166570    8011 topology_manager.go:126] [topologymanager] Cre

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.166608    8011 container_manager_linux.go:301] [topologymanag

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.166633    8011 container_manager_linux.go:306] Creating devic

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.166870    8011 client.go:75] Connecting to docker on unix:///

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.166910    8011 client.go:92] Start docker client with request

Aug 12 00:07:14 centosk8s kubelet[8011]: W0812 00:07:14.170784    8011 docker_service.go:561] Hairpin mode set to "pro "hairpin-veth"

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.170861    8011 docker_service.go:238] Hairpin mode set to "ha

Aug 12 00:07:14 centosk8s kubelet[8011]: W0812 00:07:14.171157    8011 cni.go:237] Unable to update cni config: no ne

Aug 12 00:07:14 centosk8s kubelet[8011]: W0812 00:07:14.179737    8011 cni.go:237] Unable to update cni config: no ne

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.179879    8011 docker_service.go:253] Docker cri networking m

Aug 12 00:07:14 centosk8s kubelet[8011]: W0812 00:07:14.180105    8011 cni.go:237] Unable to update cni config: no ne

Aug 12 00:07:14 centosk8s kubelet[8011]: I0812 00:07:14.207741    8011 docker_service.go:258] Docker Info: &{ID:DDVQ:iners:10 ContainersRunning:0 ContainersPaused:0 ContainersStopped:10 Images:7 Driver:overlay2 DriverStatus:[[Backing true]] SystemStatus:[] Plugins:{Volume:[local] Network:[bridge host macvlan null overlay] Authorization:[] Log:[awslosplunk syslog]} MemoryLimit:true SwapLimit:true KernelMemory:true KernelMemoryTCP:false CPUCfsPeriod:true CPUCfsQuotaing:true BridgeNfIptables:true BridgeNfIP6tables:true Debug:false NFd:24 OomKillDisable:true NGoroutines:37 SystemTim CgroupDriver:systemd NEventsListener:0 KernelVersion:4.18.0-193.14.2.el8_2.x86_64 OperatingSystem:CentOS Linux 8 (Cos://index.docker.io/v1/ RegistryConfig:0xc000242770 NCPU:2 MemTotal:1364574208 GenericResources:[] DockerRootDir:/varabels:[] ExperimentalBuild:false ServerVersion:18.09.1 ClusterStore: ClusterAdvertise: Runtimes:map[runc:{Path:runc AodeState:inactive ControlAvailable:false Error: RemoteManagers:[] Nodes:0 Managers:0 Cluster:<nil> Warnings:[]} LiveRnerdCommit:{ID:7ad184331fa3e55e52b890ea95e65ba581ae3429 Expected:7ad184331fa3e55e52b890ea95e65ba581ae3429} RuncCommit8a3303feef5b3839f4323d9beb36df0a9dd} InitCommit:{ID:fec3683 Expected:fec3683} SecurityOptions:[name=seccomp,profile=d

Aug 12 00:07:14 centosk8s kubelet[8011]: F0812 00:07:14.208006    8011 server.go:274] failed to run Kubelet: misconfim docker cgroup driver: "systemd"

Aug 12 00:07:14 centosk8s systemd[1]: kubelet.service: Main process exited, code=exited, status=255/n/a

Aug 12 00:07:14 centosk8s systemd[1]: kubelet.service: Failed with result 'exit-code'.

1 问题描述

        启动 kubelet 报错，总是报如下错误：

failed to run Kubelet: failed to create kubelet: misconfiguration: kubelet cgroup driver: "systemd" is different from docker cgroup driver: "cgroupfs"

2 解决方案：

        修改 /etc/docker/daemon.json 文件，添加如下内容：

"exec-opts": ["native.cgroupdriver=systemd"]

systemctl restart docker

        保证 kubelet 跟 docker 的 cgroup driver 一致即可。查看 docker 所使用的 cgroup driver:

docker info

 查看 kubelet 启动时的 cgroup-driver 配置，编辑 vim /var/lib/kubelet/kubeadm-flags.env 文件：

最后再重新启动 

systemctl restart kubelet

再次查看docker已经有容器在运行了。