From 98a3182398c1bb42df3a3d8addcb661b00319597 Mon Sep 17 00:00:00 2001 From: Kensei Nakada Date: Thu, 20 Jun 2024 23:48:42 +0000 Subject: [PATCH] correct comment --- pkg/scheduler/internal/queue/scheduling_queue.go | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/pkg/scheduler/internal/queue/scheduling_queue.go b/pkg/scheduler/internal/queue/scheduling_queue.go index ba3a8b50d5a..9f6a592eccf 100644 --- a/pkg/scheduler/internal/queue/scheduling_queue.go +++ b/pkg/scheduler/internal/queue/scheduling_queue.go @@ -1200,8 +1200,10 @@ func (p *PriorityQueue) movePodsToActiveOrBackoffQueue(logger klog.Logger, podIn // Scheduling-gated Pods never get schedulable with any events, // except the Pods themselves got updated, which isn't handled by movePodsToActiveOrBackoffQueue. // So, we can skip them early here so that they don't go through isPodWorthRequeuing, - // which isn't fast enough when the number of scheduling-gated Pods in unschedulablePods is large. - // This is a hotfix, which might be changed + // which isn't fast enough to keep a sufficient scheduling throughput + // when the number of scheduling-gated Pods in unschedulablePods is large. + // https://github.com/kubernetes/kubernetes/issues/124384 + // This is a hotfix for this issue, which might be changed // once we have a better general solution for the shared lock issue. // // Note that we cannot skip all pInfo.Gated Pods here