Paco Xu
There is a feature gate named GracefulNodeShutdownBasedOnPodPriority. It can influence the drain order during node graceful shutdown.
I mean you can follow the design of the feature to use PodPriority to decide the order during the drain of pods.
/sig cli /transfer kubectl
真的很nice 👍
> 请教一下，这里pr的明细列表，有什么办法快速获取到，自己也参与过不少，然后很多都忘记了，想做个汇总统计 @eryajf 参考 https://github.com/yihong0618/github-readme-stats
/retest
/test pull-kubernetes-integration
https://storage.googleapis.com/k8s-triage/index.html?test=%20Pods%20should%20run%20through%20the%20lifecycle%20of%20Pods%20and%20PodStatus It still flakes but not that much. I checked the history: https://github.com/kubernetes/kubernetes/commit/079764ebd5477c77acb11a55587d2499f9c25138 adjusted the timeout from 30s to 1m. https://github.com/kubernetes/kubernetes/pull/108892 changed it not to fail for watching Delete Event,...
> Sep 19 13:13:47 kinder-external-etcd-worker-2 containerd[220]: time="2023-09-19T13:13:47.313835166Z" level=info msg="StopContainer for \"2dd4d0a67942263b4dbc37d677fd3c53e3607f21487480c77dfb2a9c2f37d2ee\" with timeout 1 (s)" > Sep 19 13:13:47 kinder-external-etcd-worker-2 containerd[220]: time="2023-09-19T13:13:47.323305023Z" level=info msg="Stop container \"2dd4d0a67942263b4dbc37d677fd3c53e3607f21487480c77dfb2a9c2f37d2ee\" with signal terminated" >...