前言
如何让k8s集群某台node暂停调度?
如何恢复调度?
如果某台node需要维护,如果将pod从这台node中驱逐?
cordon暂停调度
暂停node调度，使node不可用，使node不接收新的pod。
命令示例
kubectl cordon node2
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready master 4d6h v1.17.4
node1 Ready <none> 4d6h v1.17.4
node2 Ready <none> 4d6h v1.17.4
[root@master ~]# kubectl cordon node2
node/node2 cordoned
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready master 4d6h v1.17.4
node1 Ready <none> 4d6h v1.17.4
node2 Ready,SchedulingDisabled <none> 4d6h v1.17.4
[root@master ~]#
可以看到node2节点的状态变成了SchedulingDisabled
解释 :SchedulingDisabled 调度不可用
uncordon恢复调度
取消node暂停调度状态，使node可以接收调度
命令示例
kubectl uncordon node2
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready master 4d6h v1.17.4
node1 Ready <none> 4d6h v1.17.4
node2 Ready,SchedulingDisabled <none> 4d6h v1.17.4
[root@master ~]# kubectl uncordon node2
node/node2 uncordoned
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready master 4d6h v1.17.4
node1 Ready <none> 4d6h v1.17.4
node2 Ready <none> 4d6h v1.17.4
drain驱逐
驱逐，把pod赶走
#把pod从node1赶走
kubectl drain node1
[root@master ~]# kubectl get node
NAME STATUS ROLES AGE VERSION
master Ready master 4d6h v1.17.4
node1 Ready <none> 4d6h v1.17.4
node2 Ready <none> 4d6h v1.17.4
[root@master ~]# kubectl get pod -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-64777cd554-dsvzn 1/1 Running 0 5m6s 10.244.1.35 node1 <none> <none>
nginx2-56684bd57d-kgnd4 1/1 Running 0 4m45s 10.244.1.36 node1 <none> <none>
nginx3-74c945c4c4-96dkf 1/1 Running 0 4m36s 10.244.1.37 node1 <none> <none>
nginx4-d989c76f4-6dwkd 1/1 Running 0 2m35s 10.244.2.32 node2 <none> <none>
nginx5-5794c9cd74-n8rl2 1/1 Running 0 2m28s 10.244.2.33 node2 <none> <none>
[root@master ~]# kubectl drain node1
node/node1 cordoned
evicting pod "nginx-6867cdf567-jknt5"
evicting pod "nginx-64777cd554-dsvzn"
evicting pod "nginx2-56684bd57d-kgnd4"
evicting pod "nginx3-74c945c4c4-96dkf"
pod/nginx-6867cdf567-jknt5 evicted
pod/nginx2-56684bd57d-kgnd4 evicted
pod/nginx-64777cd554-dsvzn evicted
pod/nginx3-74c945c4c4-96dkf evicted
node/node1 evicted
[root@master ~]# kubectl get pod -n dev -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-64777cd554-fgrf9 1/1 Running 0 7s 10.244.2.37 node2 <none> <none>
nginx2-56684bd57d-mkcpt 1/1 Running 0 7s 10.244.2.36 node2 <none> <none>
nginx3-74c945c4c4-5clvj 1/1 Running 0 7s 10.244.2.35 node2 <none> <none>
nginx4-d989c76f4-6dwkd 1/1 Running 0 2m51s 10.244.2.32 node2 <none> <none>
nginx5-5794c9cd74-n8rl2 1/1 Running 0 2m44s 10.244.2.33 node2 <none> <none>
可以看到node1上的pod被驱逐到了其它node上(这里是node2)
