Golang algorithm.NodeLister类代码示例

// CalculateNodeLabelPriority checks whether a particular label exists on a node or not, regardless of its value.
// If presence is true, prioritizes nodes that have the specified label, regardless of value.
// If presence is false, prioritizes nodes that do not have the specified label.
func (n *NodeLabelPrioritizer) CalculateNodeLabelPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	var score int
	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	labeledNodes := map[string]bool{}
	for _, node := range nodes.Items {
		exists := labels.Set(node.Labels).Has(n.label)
		labeledNodes[node.Name] = (exists && n.presence) || (!exists && !n.presence)
	}

	result := []schedulerapi.HostPriority{}
	//score int - scale of 0-10
	// 0 being the lowest priority and 10 being the highest
	for nodeName, success := range labeledNodes {
		if success {
			score = 10
		} else {
			score = 0
		}
		result = append(result, schedulerapi.HostPriority{Host: nodeName, Score: score})
	}
	return result, nil
}

// ImageLocalityPriority is a priority function that favors nodes that already have requested pod container's images.
// It will detect whether the requested images are present on a node, and then calculate a score ranging from 0 to 10
// based on the total size of those images.
// - If none of the images are present, this node will be given the lowest priority.
// - If some of the images are present on a node, the larger their sizes' sum, the higher the node's priority.
func ImageLocalityPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	sumSizeMap := make(map[string]int64)

	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	for _, container := range pod.Spec.Containers {
		for _, node := range nodes.Items {
			// Check if this container's image is present and get its size.
			imageSize := checkContainerImageOnNode(node, container)
			// Add this size to the total result of this node.
			sumSizeMap[node.Name] += imageSize
		}
	}

	result := []schedulerapi.HostPriority{}
	// score int - scale of 0-10
	// 0 being the lowest priority and 10 being the highest.
	for nodeName, sumSize := range sumSizeMap {
		result = append(result, schedulerapi.HostPriority{Host: nodeName,
			Score: calculateScoreFromSize(sumSize)})
	}
	return result, nil
}

// Schedule tries to schedule the given pod to one of node in the node list.
// If it succeeds, it will return the name of the node.
// If it fails, it will return a Fiterror error with reasons.
func (g *genericScheduler) Schedule(pod *api.Pod, nodeLister algorithm.NodeLister) (string, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return "", err
	}
	if len(nodes.Items) == 0 {
		return "", ErrNoNodesAvailable
	}

	// Used for all fit and priority funcs.
	nodeNameToInfo, err := g.cache.GetNodeNameToInfoMap()
	if err != nil {
		return "", err
	}

	filteredNodes, failedPredicateMap, err := findNodesThatFit(pod, nodeNameToInfo, g.predicates, nodes, g.extenders)
	if err != nil {
		return "", err
	}

	if len(filteredNodes.Items) == 0 {
		return "", &FitError{
			Pod:              pod,
			FailedPredicates: failedPredicateMap,
		}
	}

	priorityList, err := PrioritizeNodes(pod, nodeNameToInfo, g.prioritizers, algorithm.FakeNodeLister(filteredNodes), g.extenders)
	if err != nil {
		return "", err
	}

	return g.selectHost(priorityList)
}

func (g *genericScheduler) Schedule(pod *api.Pod, nodeLister algorithm.NodeLister) (string, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return "", err
	}
	if len(nodes.Items) == 0 {
		return "", ErrNoNodesAvailable
	}

	filteredNodes, failedPredicateMap, err := findNodesThatFit(pod, g.pods, g.predicates, nodes)
	if err != nil {
		return "", err
	}

	priorityList, err := PrioritizeNodes(pod, g.pods, g.prioritizers, algorithm.FakeNodeLister(filteredNodes))
	if err != nil {
		return "", err
	}
	if len(priorityList) == 0 {
		return "", &FitError{
			Pod:              pod,
			FailedPredicates: failedPredicateMap,
		}
	}

	return g.selectHost(priorityList)
}

func (g *genericScheduler) Schedule(pod *api.Pod, nodeLister algorithm.NodeLister) (string, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return "", err
	}
	if len(nodes.Items) == 0 {
		return "", ErrNoNodesAvailable
	}

	// TODO: we should compute this once and dynamically update it using Watch, not constantly re-compute.
	// But at least we're now only doing it in one place
	machinesToPods, err := predicates.MapPodsToMachines(g.pods)
	if err != nil {
		return "", err
	}

	filteredNodes, failedPredicateMap, err := findNodesThatFit(pod, machinesToPods, g.predicates, nodes, g.extenders)
	if err != nil {
		return "", err
	}

	priorityList, err := PrioritizeNodes(pod, machinesToPods, g.pods, g.prioritizers, algorithm.FakeNodeLister(filteredNodes), g.extenders)
	if err != nil {
		return "", err
	}
	if len(priorityList) == 0 {
		return "", &FitError{
			Pod:              pod,
			FailedPredicates: failedPredicateMap,
		}
	}

	return g.selectHost(priorityList)
}

// CalculateNodeAffinityPriority prioritizes nodes according to node affinity scheduling preferences
// indicated in PreferredDuringSchedulingIgnoredDuringExecution. Each time a node match a preferredSchedulingTerm,
// it will a get an add of preferredSchedulingTerm.Weight. Thus, the more preferredSchedulingTerms
// the node satisfies and the more the preferredSchedulingTerm that is satisfied weights, the higher
// score the node gets.
func CalculateNodeAffinityPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	var maxCount float64
	counts := make(map[string]float64, len(nodes.Items))

	affinity, err := api.GetAffinityFromPodAnnotations(pod.Annotations)
	if err != nil {
		return nil, err
	}

	// A nil element of PreferredDuringSchedulingIgnoredDuringExecution matches no objects.
	// An element of PreferredDuringSchedulingIgnoredDuringExecution that refers to an
	// empty PreferredSchedulingTerm matches all objects.
	if affinity.NodeAffinity != nil && affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution != nil {
		// Match PreferredDuringSchedulingIgnoredDuringExecution term by term.
		for _, preferredSchedulingTerm := range affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution {
			if preferredSchedulingTerm.Weight == 0 {
				continue
			}

			nodeSelector, err := api.NodeSelectorRequirementsAsSelector(preferredSchedulingTerm.Preference.MatchExpressions)
			if err != nil {
				return nil, err
			}

			for _, node := range nodes.Items {
				if nodeSelector.Matches(labels.Set(node.Labels)) {
					counts[node.Name] += float64(preferredSchedulingTerm.Weight)
				}

				if counts[node.Name] > maxCount {
					maxCount = counts[node.Name]
				}
			}
		}
	}

	result := make(schedulerapi.HostPriorityList, 0, len(nodes.Items))
	for i := range nodes.Items {
		node := &nodes.Items[i]
		if maxCount > 0 {
			fScore := 10 * (counts[node.Name] / maxCount)
			result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: int(fScore)})
			if glog.V(10) {
				// We explicitly don't do glog.V(10).Infof() to avoid computing all the parameters if this is
				// not logged. There is visible performance gain from it.
				glog.Infof("%v -> %v: NodeAffinityPriority, Score: (%d)", pod.Name, node.Name, int(fScore))
			}
		} else {
			result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: 0})
		}
	}
	return result, nil
}

// ComputeTaintTolerationPriority prepares the priority list for all the nodes based on the number of intolerable taints on the node
func ComputeTaintTolerationPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	// the max value of counts
	var maxCount float64
	// counts hold the count of intolerable taints of a pod for a given node
	counts := make(map[string]float64, len(nodes))

	tolerations, err := api.GetTolerationsFromPodAnnotations(pod.Annotations)
	if err != nil {
		return nil, err
	}
	// Fetch a list of all toleration with effect PreferNoSchedule
	tolerationList := getAllTolerationPreferNoSchedule(tolerations)

	// calculate the intolerable taints for all the nodes
	for _, node := range nodes {
		taints, err := api.GetTaintsFromNodeAnnotations(node.Annotations)
		if err != nil {
			return nil, err
		}

		count := countIntolerableTaintsPreferNoSchedule(taints, tolerationList)
		if count > 0 {
			// 0 is default value, so avoid unnecessary map operations.
			counts[node.Name] = count
			if count > maxCount {
				maxCount = count
			}
		}
	}

	// The maximum priority value to give to a node
	// Priority values range from 0 - maxPriority
	const maxPriority = float64(10)
	result := make(schedulerapi.HostPriorityList, 0, len(nodes))
	for _, node := range nodes {
		fScore := maxPriority
		if maxCount > 0 {
			fScore = (1.0 - counts[node.Name]/maxCount) * 10
		}
		if glog.V(10) {
			// We explicitly don't do glog.V(10).Infof() to avoid computing all the parameters if this is
			// not logged. There is visible performance gain from it.
			glog.Infof("%v -> %v: Taint Toleration Priority, Score: (%d)", pod.Name, node.Name, int(fScore))
		}

		result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: int(fScore)})
	}
	return result, nil
}

// CalculateNodeAffinityPriority prioritizes nodes according to node affinity scheduling preferences
// indicated in PreferredDuringSchedulingIgnoredDuringExecution. Each time a node match a preferredSchedulingTerm,
// it will a get an add of preferredSchedulingTerm.Weight. Thus, the more preferredSchedulingTerms
// the node satisfies and the more the preferredSchedulingTerm that is satisfied weights, the higher
// score the node gets.
func (s *NodeAffinity) CalculateNodeAffinityPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {

	var maxCount int
	counts := map[string]int{}

	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	affinity, err := api.GetAffinityFromPodAnnotations(pod.Annotations)
	if err != nil {
		return nil, err
	}

	// A nil element of PreferredDuringSchedulingIgnoredDuringExecution matches no objects.
	// An element of PreferredDuringSchedulingIgnoredDuringExecution that refers to an
	// empty PreferredSchedulingTerm matches all objects.
	if affinity.NodeAffinity != nil && affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution != nil {
		// Match PreferredDuringSchedulingIgnoredDuringExecution term by term.
		for _, preferredSchedulingTerm := range affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution {
			if preferredSchedulingTerm.Weight == 0 {
				continue
			}

			nodeSelector, err := api.NodeSelectorRequirementsAsSelector(preferredSchedulingTerm.Preference.MatchExpressions)
			if err != nil {
				return nil, err
			}

			for _, node := range nodes.Items {
				if nodeSelector.Matches(labels.Set(node.Labels)) {
					counts[node.Name] += preferredSchedulingTerm.Weight
				}

				if counts[node.Name] > maxCount {
					maxCount = counts[node.Name]
				}
			}
		}
	}

	result := []schedulerapi.HostPriority{}
	for _, node := range nodes.Items {
		fScore := float64(0)
		if maxCount > 0 {
			fScore = 10 * (float64(counts[node.Name]) / float64(maxCount))
		}
		result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: int(fScore)})
		glog.V(10).Infof("%v -> %v: NodeAffinityPriority, Score: (%d)", pod.Name, node.Name, int(fScore))
	}
	return result, nil
}

// BalancedResourceAllocation favors nodes with balanced resource usage rate.
// BalancedResourceAllocation should **NOT** be used alone, and **MUST** be used together with LeastRequestedPriority.
// It calculates the difference between the cpu and memory fracion of capacity, and prioritizes the host based on how
// close the two metrics are to each other.
// Detail: score = 10 - abs(cpuFraction-memoryFraction)*10. The algorithm is partly inspired by:
// "Wei Huang et al. An Energy Efficient Virtual Machine Placement Algorithm with Balanced Resource Utilization"
func BalancedResourceAllocation(pod *api.Pod, machinesToPods map[string][]*api.Pod, podLister algorithm.PodLister, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return schedulerapi.HostPriorityList{}, err
	}

	list := schedulerapi.HostPriorityList{}
	for _, node := range nodes.Items {
		list = append(list, calculateBalancedResourceAllocation(pod, node, machinesToPods[node.Name]))
	}
	return list, nil
}

// BalancedResourceAllocation favors nodes with balanced resource usage rate.
// BalancedResourceAllocation should **NOT** be used alone, and **MUST** be used together with LeastRequestedPriority.
// It calculates the difference between the cpu and memory fracion of capacity, and prioritizes the host based on how
// close the two metrics are to each other.
// Detail: score = 10 - abs(cpuFraction-memoryFraction)*10. The algorithm is partly inspired by:
// "Wei Huang et al. An Energy Efficient Virtual Machine Placement Algorithm with Balanced Resource Utilization"
func BalancedResourceAllocation(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return schedulerapi.HostPriorityList{}, err
	}

	list := schedulerapi.HostPriorityList{}
	for _, node := range nodes.Items {
		list = append(list, calculateBalancedResourceAllocation(pod, node, nodeNameToInfo[node.Name].Pods()))
	}
	return list, nil
}

// LeastRequestedPriority is a priority function that favors nodes with fewer requested resources.
// It calculates the percentage of memory and CPU requested by pods scheduled on the node, and prioritizes
// based on the minimum of the average of the fraction of requested to capacity.
// Details: cpu((capacity - sum(requested)) * 10 / capacity) + memory((capacity - sum(requested)) * 10 / capacity) / 2
func LeastRequestedPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return schedulerapi.HostPriorityList{}, err
	}

	list := schedulerapi.HostPriorityList{}
	for _, node := range nodes.Items {
		list = append(list, calculateResourceOccupancy(pod, node, nodeNameToInfo[node.Name]))
	}
	return list, nil
}

// LeastRequestedPriority is a priority function that favors nodes with fewer requested resources.
// It calculates the percentage of memory and CPU requested by pods scheduled on the node, and prioritizes
// based on the minimum of the average of the fraction of requested to capacity.
// Details: cpu((capacity - sum(requested)) * 10 / capacity) + memory((capacity - sum(requested)) * 10 / capacity) / 2
func LeastRequestedPriority(pod *api.Pod, podLister algorithm.PodLister, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return schedulerapi.HostPriorityList{}, err
	}
	podsToMachines, err := predicates.MapPodsToMachines(podLister)

	list := schedulerapi.HostPriorityList{}
	for _, node := range nodes.Items {
		list = append(list, calculateResourceOccupancy(pod, node, podsToMachines[node.Name]))
	}
	return list, nil
}

// BalancedResourceAllocation favors nodes with balanced resource usage rate.
// BalancedResourceAllocation should **NOT** be used alone, and **MUST** be used together with LeastRequestedPriority.
// It calculates the difference between the cpu and memory fracion of capacity, and prioritizes the host based on how
// close the two metrics are to each other.
// Detail: score = 10 - abs(cpuFraction-memoryFraction)*10. The algorithm is partly inspired by:
// "Wei Huang et al. An Energy Efficient Virtual Machine Placement Algorithm with Balanced Resource Utilization"
func BalancedResourceAllocation(pod *api.Pod, podLister algorithm.PodLister, nodeLister algorithm.NodeLister) (algorithm.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return algorithm.HostPriorityList{}, err
	}
	podsToMachines, err := predicates.MapPodsToMachines(podLister)

	list := algorithm.HostPriorityList{}
	for _, node := range nodes.Items {
		list = append(list, calculateBalancedResourceAllocation(pod, node, podsToMachines[node.Name]))
	}
	return list, nil
}

// LeastRequestedPriority is a priority function that favors nodes with fewer requested resources.
// It calculates the percentage of memory and CPU requested by pods scheduled on the node, and prioritizes
// based on the minimum of the average of the fraction of requested to capacity.
// Details: cpu((capacity - sum(requested)) * 10 / capacity) + memory((capacity - sum(requested)) * 10 / capacity) / 2
func LeastRequestedPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	nodes, err := nodeLister.List()
	if err != nil {
		return schedulerapi.HostPriorityList{}, err
	}

	podResources := getNonZeroRequests(pod)
	list := make(schedulerapi.HostPriorityList, 0, len(nodes))
	for _, node := range nodes {
		list = append(list, calculateResourceOccupancy(pod, podResources, node, nodeNameToInfo[node.Name]))
	}
	return list, nil
}

// ComputeTaintTolerationPriority prepares the priority list for all the nodes based on the number of intolerable taints on the node
func (s *TaintToleration) ComputeTaintTolerationPriority(pod *api.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodeLister algorithm.NodeLister) (schedulerapi.HostPriorityList, error) {
	// counts hold the count of intolerable taints of a pod for a given node
	counts := make(map[string]int)

	// the max value of counts
	var maxCount int

	nodes, err := nodeLister.List()
	if err != nil {
		return nil, err
	}

	tolerations, err := api.GetTolerationsFromPodAnnotations(pod.Annotations)
	if err != nil {
		return nil, err
	}
	// Fetch a list of all toleration with effect PreferNoSchedule
	tolerationList := getAllTolerationPreferNoSchedule(tolerations)

	// calculate the intolerable taints for all the nodes
	for i := range nodes.Items {
		node := &nodes.Items[i]
		taints, err := api.GetTaintsFromNodeAnnotations(node.Annotations)
		if err != nil {
			return nil, err
		}

		count := countIntolerableTaintsPreferNoSchedule(taints, tolerationList)
		counts[node.Name] = count
		if count > maxCount {
			maxCount = count
		}
	}

	// The maximum priority value to give to a node
	// Priority values range from 0 - maxPriority
	const maxPriority = 10
	result := make(schedulerapi.HostPriorityList, 0, len(nodes.Items))
	for _, node := range nodes.Items {
		fScore := float64(maxPriority)
		if maxCount > 0 {
			fScore = (1.0 - float64(counts[node.Name])/float64(maxCount)) * 10
		}
		glog.V(10).Infof("%v -> %v: Taint Toleration Priority, Score: (%d)", pod.Name, node.Name, int(fScore))

		result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: int(fScore)})
	}
	return result, nil
}