fixup! Add logic to account for pods nominated to run on nodes, but are not…

plugin/pkg/scheduler/core/generic_scheduler.go

@@ -362,7 +362,14 @@ func addNominatedPods(podPriority int32, meta algorithm.PredicateMetadata,
 	return true, metaOut, nodeInfoOut
 }
 
-// Checks whether node with a given name and NodeInfo satisfies all predicateFuncs.
+// podFitsOnNode checks whether a node given by NodeInfo satisfies the given predicate functions.
+// This function is called from two different places: Schedule and Preempt.
+// When it is called from Schedule, we want to test whether the pod is schedulable
+// on the node with all the existing pods on the node plus higher and equal priority
+// pods nominated to run on the node.
+// When it is called from Preempt, we should remove the victims of preemption and
+// add the nominated pods. Removal of the victims is done by SelectVictimsOnNode().
+// It removes victims from meta and NodeInfo before calling this function.
 func podFitsOnNode(
 	pod *v1.Pod,
 	meta algorithm.PredicateMetadata,

plugin/pkg/scheduler/factory/factory.go

@@ -1162,7 +1162,7 @@ func (p *podPreemptor) RemoveNominatedNodeAnnotation(pod *v1.Pod) error {
 	if _, exists := podCopy.Annotations[core.NominatedNodeAnnotationKey]; !exists {
 		return nil
 	}
-	// Note: Deleting the entry from the annotations and passing it Patch() will
+	// Note: Deleting the entry from the annotations and passing it to Patch() will
 	// not remove the annotation. That's why we set it to empty string.
 	podCopy.Annotations[core.NominatedNodeAnnotationKey] = ""
 	ret := &unstructured.Unstructured{}

plugin/pkg/scheduler/scheduler.go

@@ -226,6 +226,10 @@ func (sched *Scheduler) preempt(preemptor *v1.Pod, scheduleErr error) (string, e
 			sched.config.Recorder.Eventf(victim, v1.EventTypeNormal, "Preempted", "by %v/%v on node %v", preemptor.Namespace, preemptor.Name, nodeName)
 		}
 	}
+	// Clearing nominated pods should happen outside of "if node != nil". Node could
+	// be nil when a pod with nominated node name is eligible to preempt again,
+	// but preemption logic does not find any node for it. In that case Preempt()
+	// function of generic_scheduler.go returns the pod itself for removal of the annotation.
 	for _, p := range nominatedPodsToClear {
 		rErr := sched.config.PodPreemptor.RemoveNominatedNodeAnnotation(p)
 		if rErr != nil {