0x01 故障
集群中一个deployment中有很多Evicted的pod
root@iZbp12bkuvg20e1j3y9gtxZ:~# kubectl get pod -n base-ops
NAME READY STATUS RESTARTS AGE
registry-6d7b7c9fb4-22f2x 0/1 Evicted 0 10d
registry-6d7b7c9fb4-24r7j 0/1 Evicted 0 10d
registry-6d7b7c9fb4-25clg 0/1 Evicted 0 10d
...
0x02 原因
原因：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。可以把磁盘资源释放后，批量删除
Evicted Pod
kubectl -n base-ops get pods | grep Evicted | awk '{print$1}' |xargs kubectl -n base-ops delete pods
0x03 解决方案
批量删除所有“Evicted”状态的pod即可
root@iZbp12bkuvg20e1j3y9gtxZ:~# kubectl -n base-ops get pods | grep Evicted |awk '{print$1}'|xargs kubectl -n base-ops delete pods
pod "registry-6d7b7c9fb4-22f2x" deleted
pod "registry-6d7b7c9fb4-24r7j" deleted
pod "registry-6d7b7c9fb4-25clg" deleted
发表评论