K8s1.13 cannot schedule pods from unreachable node

Some of k8s1.13 worker nodes down due to power outage, but I found the pods on them stuck in “Ternimating” status. It was good in 1.11 time frame. Any bugs here?