flannel一直处于crashloopbackoff

在部署flannel网络插件时，发现flannel一直处于crashloopbackoff状态：

[root@fv120-100 manifest]# kubectl get pods -n kube-system
NAME                    READY   STATUS             RESTARTS   AGE
kube-flannel-ds-2xsqv   0/1     CrashLoopBackOff   5          5m13s
kube-flannel-ds-p7j2d   0/1     CrashLoopBackOff   5          5m13s

查看相关日志：

 kubectl logs kube-flannel-ds-2xsqv -n kube-system

E0515 02:10:20.838517       1 main.go:293] Error registering network: failed to acquire lease: node "172.16.120.101" pod cidr not assigned
I0515 02:10:20.838711       1 main.go:372] Stopping shutdownHandler...

意思是说注册网络时，pod 的cidr没有分配；

修改kube-controller-manager的配置参数：

增加如下参数解决

--allocate-node-cidrs=true

重启systemctl restart kube-controller-manager.service