k8s节点变为NotReady状态

记录一个小教训。切记： k8s安装完成后不要轻易修改hostname。

问题：

k8s的master节点变为notReady，另外一个worker节点正常。

症状：master上的flannel pod显示为pending状态，但是describe pod没有有用信息，最后运行journalctl -f -u kubelet.service 日志中，发现节点名称的问题， 想起来刚才修改过主机名，然后问题就清楚了。
journalctl -f -u kubelet.service显示的日志信息

Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.636853    5724 kubelet.go:2248] node "localhost.localdomain" not found
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.693945    5724 reflector.go:125] k8s.io/kubernetes/pkg/kubelet/config/apiserve
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.696794    5724 reflector.go:125] k8s.io/kubernetes/pkg/kubelet/kubelet.go:453:
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.699268    5724 reflector.go:125] k8s.io/kubernetes/pkg/kubelet/kubelet.go:444:
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: I0309 05:04:49.732622    5724 kubelet_node_status.go:286] Setting node annotation to enable v
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.736983    5724 kubelet.go:2248] node "localhost.localdomain" not found
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: I0309 05:04:49.782223    5724 kubelet_node_status.go:72] Attempting to register node localhos
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.782581    5724 kubelet_node_status.go:94] Unable to register node "localhost.l
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: I0309 05:04:49.806613    5724 kubelet.go:1822] skipping pod synchronization - container runti
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: I0309 05:04:49.807165    5724 plugin_manager.go:116] Starting Kubelet Plugin Manager
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.807373    5724 eviction_manager.go:247] eviction manager: failed to get summar
Mar 09 05:04:49 localhost.localdomain kubelet[5724]: E0309 05:04:49.844272    5724 kubelet.go:2248] node "localhost.localdomain" not found

关键的一句就是node “localhost.localdomain” not found

解决办法：

最后将master机器的主机名修改为原先的名称， 重启master机器，

 hostnamectl set-hostname localhost.localdomain

重启master机器后，一切正常了

[root@localhostxxxxx]# kubectl get node
NAME                    STATUS   ROLES    AGE     VERSION
localhost.localdomain   Ready    master   7d21h   v1.15.0
node2                   Ready       7d20h   v1.15.0