kubernetes fit 源码

kubernetes fit 代码

文件路径：/pkg/scheduler/framework/plugins/noderesources/fit.go

/*
Copyright 2019 The Kubernetes Authors.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/

package noderesources

import (
	"context"
	"fmt"
	"strings"

	v1 "k8s.io/api/core/v1"
	"k8s.io/apimachinery/pkg/runtime"
	"k8s.io/apimachinery/pkg/util/sets"
	v1helper "k8s.io/kubernetes/pkg/apis/core/v1/helper"
	"k8s.io/kubernetes/pkg/scheduler/apis/config"
	"k8s.io/kubernetes/pkg/scheduler/apis/config/validation"
	"k8s.io/kubernetes/pkg/scheduler/framework"
	"k8s.io/kubernetes/pkg/scheduler/framework/plugins/feature"
	"k8s.io/kubernetes/pkg/scheduler/framework/plugins/names"
)

var _ framework.PreFilterPlugin = &Fit{}
var _ framework.FilterPlugin = &Fit{}
var _ framework.EnqueueExtensions = &Fit{}
var _ framework.ScorePlugin = &Fit{}

const (
	// Name is the name of the plugin used in the plugin registry and configurations.
	Name = names.NodeResourcesFit

	// preFilterStateKey is the key in CycleState to NodeResourcesFit pre-computed data.
	// Using the name of the plugin will likely help us avoid collisions with other plugins.
	preFilterStateKey = "PreFilter" + Name
)

// nodeResourceStrategyTypeMap maps strategy to scorer implementation
var nodeResourceStrategyTypeMap = map[config.ScoringStrategyType]scorer{
	config.LeastAllocated: func(args *config.NodeResourcesFitArgs) *resourceAllocationScorer {
		resToWeightMap := resourcesToWeightMap(args.ScoringStrategy.Resources)
		return &resourceAllocationScorer{
			Name:                string(config.LeastAllocated),
			scorer:              leastResourceScorer(resToWeightMap),
			resourceToWeightMap: resToWeightMap,
		}
	},
	config.MostAllocated: func(args *config.NodeResourcesFitArgs) *resourceAllocationScorer {
		resToWeightMap := resourcesToWeightMap(args.ScoringStrategy.Resources)
		return &resourceAllocationScorer{
			Name:                string(config.MostAllocated),
			scorer:              mostResourceScorer(resToWeightMap),
			resourceToWeightMap: resToWeightMap,
		}
	},
	config.RequestedToCapacityRatio: func(args *config.NodeResourcesFitArgs) *resourceAllocationScorer {
		resToWeightMap := resourcesToWeightMap(args.ScoringStrategy.Resources)
		return &resourceAllocationScorer{
			Name:                string(config.RequestedToCapacityRatio),
			scorer:              requestedToCapacityRatioScorer(resToWeightMap, args.ScoringStrategy.RequestedToCapacityRatio.Shape),
			resourceToWeightMap: resToWeightMap,
		}
	},
}

// Fit is a plugin that checks if a node has sufficient resources.
type Fit struct {
	ignoredResources      sets.String
	ignoredResourceGroups sets.String
	handle                framework.Handle
	resourceAllocationScorer
}

// ScoreExtensions of the Score plugin.
func (f *Fit) ScoreExtensions() framework.ScoreExtensions {
	return nil
}

// preFilterState computed at PreFilter and used at Filter.
type preFilterState struct {
	framework.Resource
}

// Clone the prefilter state.
func (s *preFilterState) Clone() framework.StateData {
	return s
}

// Name returns name of the plugin. It is used in logs, etc.
func (f *Fit) Name() string {
	return Name
}

// NewFit initializes a new plugin and returns it.
func NewFit(plArgs runtime.Object, h framework.Handle, fts feature.Features) (framework.Plugin, error) {
	args, ok := plArgs.(*config.NodeResourcesFitArgs)
	if !ok {
		return nil, fmt.Errorf("want args to be of type NodeResourcesFitArgs, got %T", plArgs)
	}
	if err := validation.ValidateNodeResourcesFitArgs(nil, args); err != nil {
		return nil, err
	}

	if args.ScoringStrategy == nil {
		return nil, fmt.Errorf("scoring strategy not specified")
	}

	strategy := args.ScoringStrategy.Type
	scorePlugin, exists := nodeResourceStrategyTypeMap[strategy]
	if !exists {
		return nil, fmt.Errorf("scoring strategy %s is not supported", strategy)
	}

	return &Fit{
		ignoredResources:         sets.NewString(args.IgnoredResources...),
		ignoredResourceGroups:    sets.NewString(args.IgnoredResourceGroups...),
		handle:                   h,
		resourceAllocationScorer: *scorePlugin(args),
	}, nil
}

// computePodResourceRequest returns a framework.Resource that covers the largest
// width in each resource dimension. Because init-containers run sequentially, we collect
// the max in each dimension iteratively. In contrast, we sum the resource vectors for
// regular containers since they run simultaneously.
//
// # The resources defined for Overhead should be added to the calculated Resource request sum
//
// Example:
//
// Pod:
//
//	InitContainers
//	  IC1:
//	    CPU: 2
//	    Memory: 1G
//	  IC2:
//	    CPU: 2
//	    Memory: 3G
//	Containers
//	  C1:
//	    CPU: 2
//	    Memory: 1G
//	  C2:
//	    CPU: 1
//	    Memory: 1G
//
// Result: CPU: 3, Memory: 3G
func computePodResourceRequest(pod *v1.Pod) *preFilterState {
	result := &preFilterState{}
	for _, container := range pod.Spec.Containers {
		result.Add(container.Resources.Requests)
	}

	// take max_resource(sum_pod, any_init_container)
	for _, container := range pod.Spec.InitContainers {
		result.SetMaxResource(container.Resources.Requests)
	}

	// If Overhead is being utilized, add to the total requests for the pod
	if pod.Spec.Overhead != nil {
		result.Add(pod.Spec.Overhead)
	}
	return result
}

// PreFilter invoked at the prefilter extension point.
func (f *Fit) PreFilter(ctx context.Context, cycleState *framework.CycleState, pod *v1.Pod) (*framework.PreFilterResult, *framework.Status) {
	cycleState.Write(preFilterStateKey, computePodResourceRequest(pod))
	return nil, nil
}

// PreFilterExtensions returns prefilter extensions, pod add and remove.
func (f *Fit) PreFilterExtensions() framework.PreFilterExtensions {
	return nil
}

func getPreFilterState(cycleState *framework.CycleState) (*preFilterState, error) {
	c, err := cycleState.Read(preFilterStateKey)
	if err != nil {
		// preFilterState doesn't exist, likely PreFilter wasn't invoked.
		return nil, fmt.Errorf("error reading %q from cycleState: %w", preFilterStateKey, err)
	}

	s, ok := c.(*preFilterState)
	if !ok {
		return nil, fmt.Errorf("%+v  convert to NodeResourcesFit.preFilterState error", c)
	}
	return s, nil
}

// EventsToRegister returns the possible events that may make a Pod
// failed by this plugin schedulable.
// NOTE: if in-place-update (KEP 1287) gets implemented, then PodUpdate event
// should be registered for this plugin since a Pod update may free up resources
// that make other Pods schedulable.
func (f *Fit) EventsToRegister() []framework.ClusterEvent {
	return []framework.ClusterEvent{
		{Resource: framework.Pod, ActionType: framework.Delete},
		{Resource: framework.Node, ActionType: framework.Add | framework.Update},
	}
}

// Filter invoked at the filter extension point.
// Checks if a node has sufficient resources, such as cpu, memory, gpu, opaque int resources etc to run a pod.
// It returns a list of insufficient resources, if empty, then the node has all the resources requested by the pod.
func (f *Fit) Filter(ctx context.Context, cycleState *framework.CycleState, pod *v1.Pod, nodeInfo *framework.NodeInfo) *framework.Status {
	s, err := getPreFilterState(cycleState)
	if err != nil {
		return framework.AsStatus(err)
	}

	insufficientResources := fitsRequest(s, nodeInfo, f.ignoredResources, f.ignoredResourceGroups)

	if len(insufficientResources) != 0 {
		// We will keep all failure reasons.
		failureReasons := make([]string, 0, len(insufficientResources))
		for i := range insufficientResources {
			failureReasons = append(failureReasons, insufficientResources[i].Reason)
		}
		return framework.NewStatus(framework.Unschedulable, failureReasons...)
	}
	return nil
}

// InsufficientResource describes what kind of resource limit is hit and caused the pod to not fit the node.
type InsufficientResource struct {
	ResourceName v1.ResourceName
	// We explicitly have a parameter for reason to avoid formatting a message on the fly
	// for common resources, which is expensive for cluster autoscaler simulations.
	Reason    string
	Requested int64
	Used      int64
	Capacity  int64
}

// Fits checks if node have enough resources to host the pod.
func Fits(pod *v1.Pod, nodeInfo *framework.NodeInfo) []InsufficientResource {
	return fitsRequest(computePodResourceRequest(pod), nodeInfo, nil, nil)
}

func fitsRequest(podRequest *preFilterState, nodeInfo *framework.NodeInfo, ignoredExtendedResources, ignoredResourceGroups sets.String) []InsufficientResource {
	insufficientResources := make([]InsufficientResource, 0, 4)

	allowedPodNumber := nodeInfo.Allocatable.AllowedPodNumber
	if len(nodeInfo.Pods)+1 > allowedPodNumber {
		insufficientResources = append(insufficientResources, InsufficientResource{
			ResourceName: v1.ResourcePods,
			Reason:       "Too many pods",
			Requested:    1,
			Used:         int64(len(nodeInfo.Pods)),
			Capacity:     int64(allowedPodNumber),
		})
	}

	if podRequest.MilliCPU == 0 &&
		podRequest.Memory == 0 &&
		podRequest.EphemeralStorage == 0 &&
		len(podRequest.ScalarResources) == 0 {
		return insufficientResources
	}

	if podRequest.MilliCPU > (nodeInfo.Allocatable.MilliCPU - nodeInfo.Requested.MilliCPU) {
		insufficientResources = append(insufficientResources, InsufficientResource{
			ResourceName: v1.ResourceCPU,
			Reason:       "Insufficient cpu",
			Requested:    podRequest.MilliCPU,
			Used:         nodeInfo.Requested.MilliCPU,
			Capacity:     nodeInfo.Allocatable.MilliCPU,
		})
	}
	if podRequest.Memory > (nodeInfo.Allocatable.Memory - nodeInfo.Requested.Memory) {
		insufficientResources = append(insufficientResources, InsufficientResource{
			ResourceName: v1.ResourceMemory,
			Reason:       "Insufficient memory",
			Requested:    podRequest.Memory,
			Used:         nodeInfo.Requested.Memory,
			Capacity:     nodeInfo.Allocatable.Memory,
		})
	}
	if podRequest.EphemeralStorage > (nodeInfo.Allocatable.EphemeralStorage - nodeInfo.Requested.EphemeralStorage) {
		insufficientResources = append(insufficientResources, InsufficientResource{
			ResourceName: v1.ResourceEphemeralStorage,
			Reason:       "Insufficient ephemeral-storage",
			Requested:    podRequest.EphemeralStorage,
			Used:         nodeInfo.Requested.EphemeralStorage,
			Capacity:     nodeInfo.Allocatable.EphemeralStorage,
		})
	}

	for rName, rQuant := range podRequest.ScalarResources {
		// Skip in case request quantity is zero
		if rQuant == 0 {
			continue
		}

		if v1helper.IsExtendedResourceName(rName) {
			// If this resource is one of the extended resources that should be ignored, we will skip checking it.
			// rName is guaranteed to have a slash due to API validation.
			var rNamePrefix string
			if ignoredResourceGroups.Len() > 0 {
				rNamePrefix = strings.Split(string(rName), "/")[0]
			}
			if ignoredExtendedResources.Has(string(rName)) || ignoredResourceGroups.Has(rNamePrefix) {
				continue
			}
		}

		if rQuant > (nodeInfo.Allocatable.ScalarResources[rName] - nodeInfo.Requested.ScalarResources[rName]) {
			insufficientResources = append(insufficientResources, InsufficientResource{
				ResourceName: rName,
				Reason:       fmt.Sprintf("Insufficient %v", rName),
				Requested:    podRequest.ScalarResources[rName],
				Used:         nodeInfo.Requested.ScalarResources[rName],
				Capacity:     nodeInfo.Allocatable.ScalarResources[rName],
			})
		}
	}

	return insufficientResources
}

// Score invoked at the Score extension point.
func (f *Fit) Score(ctx context.Context, state *framework.CycleState, pod *v1.Pod, nodeName string) (int64, *framework.Status) {
	nodeInfo, err := f.handle.SnapshotSharedLister().NodeInfos().Get(nodeName)
	if err != nil {
		return 0, framework.AsStatus(fmt.Errorf("getting node %q from Snapshot: %w", nodeName, err))
	}

	return f.score(pod, nodeInfo)
}

相关信息

kubernetes 源码目录

相关文章

kubernetes balanced_allocation 源码

kubernetes balanced_allocation_test 源码

kubernetes fit_test 源码

kubernetes least_allocated 源码

kubernetes least_allocated_test 源码

kubernetes most_allocated 源码

kubernetes most_allocated_test 源码

kubernetes requested_to_capacity_ratio 源码

kubernetes requested_to_capacity_ratio_test 源码

kubernetes resource_allocation 源码