koord-scheduler: add a parameter to mark whether scheduling is allowe…

…d on node with expired nodemetric

Signed-off-by: lucming <2876757716@qq.com>

pkg/scheduler/apis/config/types.go

@@ -37,6 +37,8 @@ type LoadAwareSchedulingArgs struct {
 	// When NodeMetrics expired, the node is considered abnormal.
 	// Default is 180 seconds.
 	NodeMetricExpirationSeconds *int64
+	// EnableScheduleWhenNodeMetricsExpired Indicates whether nodes with expired nodeMetrics are allowed to schedule pods.
+	EnableScheduleWhenNodeMetricsExpired *bool
 	// ResourceWeights indicates the weights of resources.
 	// The weights of CPU and Memory are both 1 by default.
 	ResourceWeights map[corev1.ResourceName]int64

pkg/scheduler/apis/config/v1beta3/types.go

@@ -36,6 +36,8 @@ type LoadAwareSchedulingArgs struct {
 	// When NodeMetrics expired, the node is considered abnormal.
 	// Default is 180 seconds.
 	NodeMetricExpirationSeconds *int64 `json:"nodeMetricExpirationSeconds,omitempty"`
+	// EnableScheduleWhenNodeMetricsExpired Indicates whether nodes with expired nodeMetrics are allowed to schedule pods.
+	EnableScheduleWhenNodeMetricsExpired *bool `json:"enableScheduleWhenNodeMetricsExpired,omitempty"`
 	// ResourceWeights indicates the weights of resources.
 	// The weights of CPU and Memory are both 1 by default.
 	ResourceWeights map[corev1.ResourceName]int64 `json:"resourceWeights,omitempty"`

pkg/scheduler/plugins/loadaware/load_aware.go

@@ -42,6 +42,7 @@ import (
 
 const (
 	Name                                    = "LoadAwareScheduling"
+	ErrReasonNodeMetricExpired              = "node(s) nodeMetric expired"
 	ErrReasonUsageExceedThreshold           = "node(s) %s usage exceed threshold"
 	ErrReasonAggregatedUsageExceedThreshold = "node(s) %s aggregated usage exceed threshold"
 	ErrReasonFailedEstimatePod
@@ -143,6 +144,9 @@ func (p *Plugin) Filter(ctx context.Context, state *framework.CycleState, pod *c
 
 	if p.args.FilterExpiredNodeMetrics != nil && *p.args.FilterExpiredNodeMetrics &&
 		p.args.NodeMetricExpirationSeconds != nil && isNodeMetricExpired(nodeMetric, *p.args.NodeMetricExpirationSeconds) {
+		if p.args.EnableScheduleWhenNodeMetricsExpired != nil && !*p.args.EnableScheduleWhenNodeMetricsExpired {
+			return framework.NewStatus(framework.Unschedulable, ErrReasonNodeMetricExpired)
+		}
 		return nil
 	}