TKE节点管理

原创

何飞良

修改于 2020-06-30 09:59:07

1.6K0

修改于 2020-06-30 09:59:07

文章被收录于专栏：hfl的运维笔记hfl的运维笔记

节点生命周期在集群的控制面板上查看

状态有健康，异常，已封锁，驱逐中

新增节点

一个明显的蓝色按钮，新增节点，也可以灵活选择配置，数量

查看初始化进度

kubectl get nodes
NAME           STATUS                     ROLES    AGE     VERSION
172.27.0.16    Ready                      <none>   48m     v1.16.3-tke.9
172.27.16.13   Ready                      master   30h     v1.16.3-tke.9
172.27.16.14   Ready                      <none>   3m56s   v1.16.3-tke.9
172.27.16.15   Ready,SchedulingDisabled   <none>   29h     v1.16.3-tke.9
172.27.16.3    Ready                      <none>   4m22s   v1.16.3-tke.9
172.27.16.4    Ready                      master   30h     v1.16.3-tke.9
172.27.16.5    Ready                      <none>   29h     v1.16.3-tke.9
172.27.16.7    Ready                      master   30h     v1.16.3-tke.9

移除节点

移除节点会有提示

如果是workload的pod在节点移除了，pod会在新的的节点新建；如果是裸pod，是会删除的，不会新建

驱逐或封锁节点

封锁节点，不调度节点

集群里查看状态

kubectl get nodes
NAME           STATUS                     ROLES    AGE   VERSION
172.27.16.13   Ready                      master   30h   v1.16.3-tke.9
172.27.16.15   Ready                      <none>   30h   v1.16.3-tke.9
172.27.16.3    Ready                      <none>   20m   v1.16.3-tke.9
172.27.16.4    Ready                      master   30h   v1.16.3-tke.9
172.27.16.5    Ready,SchedulingDisabled   <none>   30h   v1.16.3-tke.9
172.27.16.7    Ready                      master   30h   v1.16.3-tke.9

极端情况，创建pod没有node节点可以调度时

kubectl get pods|grep nginx
nginx                    0/1     Pending   0          7s
kubectl get events
<unknown>   Warning   FailedScheduling          pod/nginx                     0/6 nodes are available: 3 node(s) had taints that the pod didn't tolerate, 3 node(s) were unschedulable.

取消封锁