github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2025-09-20 17:38:50 +00:00
Code Issues Projects Releases Wiki Activity

sched: retry unschedule pods immediately after a waiting pod's deletion

Browse Source
This commit is contained in:
Wei Huang
2021-07-19 15:46:55 -07:00
parent ebc87c39d3
commit dc079acc2b
8 changed files with 312 additions and 101 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
pkg/scheduler/scheduler.go
View File

@@ -639,6 +639,16 @@ func (sched *Scheduler) scheduleOne(ctx context.Context) {
fwk.RunReservePluginsUnreserve(bindingCycleCtx, state, assumedPod, scheduleResult.SuggestedHost)
if forgetErr := sched.SchedulerCache.ForgetPod(assumedPod); forgetErr != nil {
klog.ErrorS(forgetErr, "scheduler cache ForgetPod failed")
} else {
// "Forget"ing an assumed Pod in binding cycle should be treated as a PodDelete event,
// as the assumed Pod had occupied a certain amount of resources in scheduler cache.
// TODO(#103853): de-duplicate the logic.
// Avoid moving the assumed Pod itself as it's always Unschedulable.
// It's intentional to "defer" this operation; otherwise MoveAllToActiveOrBackoffQueue() would
// update `q.moveRequest` and thus move the assumed pod to backoffQ anyways.
defer sched.SchedulingQueue.MoveAllToActiveOrBackoffQueue(internalqueue.AssignedPodDelete, func(pod *v1.Pod) bool {
return assumedPod.UID != pod.UID
})
}
sched.recordSchedulingFailure(fwk, assumedPodInfo, waitOnPermitStatus.AsError(), reason, "")
return
@@ -652,6 +662,11 @@ func (sched *Scheduler) scheduleOne(ctx context.Context) {
fwk.RunReservePluginsUnreserve(bindingCycleCtx, state, assumedPod, scheduleResult.SuggestedHost)
if forgetErr := sched.SchedulerCache.ForgetPod(assumedPod); forgetErr != nil {
klog.ErrorS(forgetErr, "scheduler cache ForgetPod failed")
} else {
// "Forget"ing an assumed Pod in binding cycle should be treated as a PodDelete event,
// as the assumed Pod had occupied a certain amount of resources in scheduler cache.
// TODO(#103853): de-duplicate the logic.
sched.SchedulingQueue.MoveAllToActiveOrBackoffQueue(internalqueue.AssignedPodDelete, nil)
}
sched.recordSchedulingFailure(fwk, assumedPodInfo, preBindStatus.AsError(), SchedulerError, "")
return
@@ -664,6 +679,11 @@ func (sched *Scheduler) scheduleOne(ctx context.Context) {
fwk.RunReservePluginsUnreserve(bindingCycleCtx, state, assumedPod, scheduleResult.SuggestedHost)
if err := sched.SchedulerCache.ForgetPod(assumedPod); err != nil {
klog.ErrorS(err, "scheduler cache ForgetPod failed")
} else {
// "Forget"ing an assumed Pod in binding cycle should be treated as a PodDelete event,
// as the assumed Pod had occupied a certain amount of resources in scheduler cache.
// TODO(#103853): de-duplicate the logic.
sched.SchedulingQueue.MoveAllToActiveOrBackoffQueue(internalqueue.AssignedPodDelete, nil)
}
sched.recordSchedulingFailure(fwk, assumedPodInfo, fmt.Errorf("binding rejected: %w", err), SchedulerError, "")
} else {