1、查询pod状态

[root@k8s_master prom]# kubectl -n kube-system get pods | grep Evicted
coredns-67cb7b5f46-2crxg                0/1     Evicted   0          4h56m
coredns-67cb7b5f46-2ds7n                0/1     Evicted   0          4h56m
coredns-67cb7b5f46-2pzw5                0/1     Evicted   0          4h56m
coredns-67cb7b5f46-4f8nm                0/1     Evicted   0          166m
coredns-67cb7b5f46-4gncz                0/1     Evicted   0          10h
coredns-67cb7b5f46-4lnw2                0/1     Evicted   0          166m
coredns-67cb7b5f46-4smp6                0/1     Evicted   0          166m
coredns-67cb7b5f46-5hx49                0/1     Evicted   0          166m
coredns-67cb7b5f46-6jxqn                0/1     Evicted   0          166m
coredns-67cb7b5f46-6qh9z                0/1     Evicted   0          166m
coredns-67cb7b5f46-6rn4c                0/1     Evicted   0          10h
coredns-67cb7b5f46-7jsth                0/1     Evicted   0          10h
coredns-67cb7b5f46-7mzvb                0/1     Evicted   0          166m
coredns-67cb7b5f46-92t2k                0/1     Evicted   0          4h56m
coredns-67cb7b5f46-9jqdm                0/1     Evicted   0          10h
coredns-67cb7b5f46-9rxsc                0/1     Evicted   0          4h56m
coredns-67cb7b5f46-9xmkk                0/1     Evicted   0          166m

2、发现有很多 “Evicted” 状态的pod
问题原因：这是由于node节点的磁盘或内存资源不够后将现有的pod进行了驱赶。
3、批量删除kube-system下所有“Evicted”状态的pod即可

[root@k8s_master ]# kubectl -n kube-system  get pods | grep Evicted |awk '{print$1}'|xargs kubectl -n kube-system delete pods

去掉-n kube-system是删除所有的ｐｏｄ