nayihz

https://github.com/k8snetworkplumbingwg/network-attachment-definition-client/blob/6ace5615d26a4d00d0a24bbbfa5d5f400ce7593e/pkg/utils/net-attach-def.go#L86-L96 hi @dougbtv , I'm curious about why we use `UpdateStatus(context.TODO(), pod, metav1.UpdateOptions{})` method to patch annotations?

The `active workloads` are the workloads which have been admitted by scheduler. Do I understand it correctly? https://github.com/kubernetes-sigs/kueue/blob/afabae912eb3da09f810ca6bf0fd5fe5bf49874a/pkg/scheduler/scheduler.go#L304-L308

可以启动一个单独的线程去resync 实际资源跟你的后端资源的不同

I don't disable `CGO_ENABLED` manually. `CGO_ENABLED=0` doesn't seem to take effect actually. I have open a pr to fix it(#129). PTAL.

/assign @sanposhiho I'm not very sure if it's the right way to do so. I'll add unit test after we confirm the logic.

/test pull-kubernetes-node-e2e-containerd

/test pull-kubernetes-integration

``` --- FAIL: TestCoreResourceEnqueue/Pod_rejected_by_the_PodAffinity_plugin_is_requeued_when_a_new_Node_is_created_and_turned_to_ready_[SchedulerQueueingHints_enabled:_true] (35.00s) ``` Can't reproduce it using `make test-integration WHAT=./test/integration/scheduler/`. Does @sanposhiho have time to take a look?

/sig apps

IIUC, this issue may be related to `enqueueSSAfter` twice when we reconcile the StatefulSet. First, https://github.com/kubernetes/kubernetes/blob/22243bcfc657122f91db4a4547608d942e08ac12/pkg/controller/statefulset/stateful_set.go#L223-L238 and then https://github.com/kubernetes/kubernetes/blob/22243bcfc657122f91db4a4547608d942e08ac12/pkg/controller/statefulset/stateful_set.go#L487-L489 It will update the workqueue event (double the delay time) after...