Update go dependencies

This commit is contained in:
Manuel de Brito Fontes 2018-07-12 13:19:04 -04:00 committed by Manuel Alejandro de Brito Fontes
parent d5cf22c129
commit 063cc68d1c
No known key found for this signature in database
GPG key ID: 786136016A8BA02A
1321 changed files with 52830 additions and 31081 deletions

58
vendor/k8s.io/kubernetes/pkg/scheduler/algorithm/BUILD generated vendored Normal file
View file

@ -0,0 +1,58 @@
package(default_visibility = ["//visibility:public"])
load(
"@io_bazel_rules_go//go:def.bzl",
"go_library",
"go_test",
)
go_library(
name = "go_default_library",
srcs = [
"doc.go",
"scheduler_interface.go",
"types.go",
"well_known_labels.go",
],
importpath = "k8s.io/kubernetes/pkg/scheduler/algorithm",
deps = [
"//pkg/apis/core:go_default_library",
"//pkg/scheduler/api:go_default_library",
"//pkg/scheduler/cache:go_default_library",
"//vendor/k8s.io/api/apps/v1beta1:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/api/extensions/v1beta1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
],
)
go_test(
name = "go_default_test",
srcs = [
"scheduler_interface_test.go",
"types_test.go",
],
embed = [":go_default_library"],
deps = [
"//pkg/scheduler/cache:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
],
)
filegroup(
name = "package-srcs",
srcs = glob(["**"]),
tags = ["automanaged"],
visibility = ["//visibility:private"],
)
filegroup(
name = "all-srcs",
srcs = [
":package-srcs",
"//pkg/scheduler/algorithm/predicates:all-srcs",
"//pkg/scheduler/algorithm/priorities:all-srcs",
],
tags = ["automanaged"],
)

View file

@ -0,0 +1,19 @@
/*
Copyright 2014 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 algorithm contains a generic Scheduler interface and several
// implementations.
package algorithm // import "k8s.io/kubernetes/pkg/scheduler/algorithm"

View file

@ -0,0 +1,101 @@
package(default_visibility = ["//visibility:public"])
load(
"@io_bazel_rules_go//go:def.bzl",
"go_library",
"go_test",
)
go_library(
name = "go_default_library",
srcs = [
"balanced_resource_allocation.go",
"image_locality.go",
"interpod_affinity.go",
"least_requested.go",
"metadata.go",
"most_requested.go",
"node_affinity.go",
"node_label.go",
"node_prefer_avoid_pods.go",
"reduce.go",
"requested_to_capacity_ratio.go",
"resource_allocation.go",
"resource_limits.go",
"selector_spreading.go",
"taint_toleration.go",
"test_util.go",
],
importpath = "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities",
deps = [
"//pkg/apis/core/v1/helper:go_default_library",
"//pkg/features:go_default_library",
"//pkg/scheduler/algorithm:go_default_library",
"//pkg/scheduler/algorithm/predicates:go_default_library",
"//pkg/scheduler/algorithm/priorities/util:go_default_library",
"//pkg/scheduler/api:go_default_library",
"//pkg/scheduler/cache:go_default_library",
"//pkg/util/node:go_default_library",
"//pkg/util/parsers:go_default_library",
"//vendor/github.com/golang/glog:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/api/errors:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/api/resource:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
"//vendor/k8s.io/apiserver/pkg/util/feature:go_default_library",
"//vendor/k8s.io/client-go/util/workqueue:go_default_library",
],
)
go_test(
name = "go_default_test",
srcs = [
"balanced_resource_allocation_test.go",
"image_locality_test.go",
"interpod_affinity_test.go",
"least_requested_test.go",
"metadata_test.go",
"most_requested_test.go",
"node_affinity_test.go",
"node_label_test.go",
"node_prefer_avoid_pods_test.go",
"requested_to_capacity_ratio_test.go",
"resource_limits_test.go",
"selector_spreading_test.go",
"taint_toleration_test.go",
],
embed = [":go_default_library"],
deps = [
"//pkg/features:go_default_library",
"//pkg/kubelet/apis:go_default_library",
"//pkg/scheduler/algorithm/priorities/util:go_default_library",
"//pkg/scheduler/api:go_default_library",
"//pkg/scheduler/cache:go_default_library",
"//pkg/scheduler/testing:go_default_library",
"//pkg/util/parsers:go_default_library",
"//vendor/github.com/stretchr/testify/assert:go_default_library",
"//vendor/k8s.io/api/apps/v1beta1:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/api/extensions/v1beta1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/api/resource:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/apiserver/pkg/util/feature:go_default_library",
],
)
filegroup(
name = "package-srcs",
srcs = glob(["**"]),
tags = ["automanaged"],
visibility = ["//visibility:private"],
)
filegroup(
name = "all-srcs",
srcs = [
":package-srcs",
"//pkg/scheduler/algorithm/priorities/util:all-srcs",
],
tags = ["automanaged"],
)

View file

@ -0,0 +1,77 @@
/*
Copyright 2016 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 priorities
import (
"math"
utilfeature "k8s.io/apiserver/pkg/util/feature"
"k8s.io/kubernetes/pkg/features"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
var (
balancedResourcePriority = &ResourceAllocationPriority{"BalancedResourceAllocation", balancedResourceScorer}
// BalancedResourceAllocationMap favors nodes with balanced resource usage rate.
// BalancedResourceAllocationMap should **NOT** be used alone, and **MUST** be used together
// with LeastRequestedPriority. It calculates the difference between the cpu and memory fraction
// 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"
BalancedResourceAllocationMap = balancedResourcePriority.PriorityMap
)
func balancedResourceScorer(requested, allocable *schedulercache.Resource, includeVolumes bool, requestedVolumes int, allocatableVolumes int) int64 {
cpuFraction := fractionOfCapacity(requested.MilliCPU, allocable.MilliCPU)
memoryFraction := fractionOfCapacity(requested.Memory, allocable.Memory)
// This to find a node which has most balanced CPU, memory and volume usage.
if includeVolumes && utilfeature.DefaultFeatureGate.Enabled(features.BalanceAttachedNodeVolumes) && allocatableVolumes > 0 {
volumeFraction := float64(requestedVolumes) / float64(allocatableVolumes)
if cpuFraction >= 1 || memoryFraction >= 1 || volumeFraction >= 1 {
// if requested >= capacity, the corresponding host should never be preferred.
return 0
}
// Compute variance for all the three fractions.
mean := (cpuFraction + memoryFraction + volumeFraction) / float64(3)
variance := float64((((cpuFraction - mean) * (cpuFraction - mean)) + ((memoryFraction - mean) * (memoryFraction - mean)) + ((volumeFraction - mean) * (volumeFraction - mean))) / float64(3))
// Since the variance is between positive fractions, it will be positive fraction. 1-variance lets the
// score to be higher for node which has least variance and multiplying it with 10 provides the scaling
// factor needed.
return int64((1 - variance) * float64(schedulerapi.MaxPriority))
}
if cpuFraction >= 1 || memoryFraction >= 1 {
// if requested >= capacity, the corresponding host should never be preferred.
return 0
}
// Upper and lower boundary of difference between cpuFraction and memoryFraction are -1 and 1
// respectively. Multiplying the absolute value of the difference by 10 scales the value to
// 0-10 with 0 representing well balanced allocation and 10 poorly balanced. Subtracting it from
// 10 leads to the score which also scales from 0 to 10 while 10 representing well balanced.
diff := math.Abs(cpuFraction - memoryFraction)
return int64((1 - diff) * float64(schedulerapi.MaxPriority))
}
func fractionOfCapacity(requested, capacity int64) float64 {
if capacity == 0 {
return 1
}
return float64(requested) / float64(capacity)
}

View file

@ -0,0 +1,97 @@
/*
Copyright 2016 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 priorities
import (
"fmt"
"strings"
"k8s.io/api/core/v1"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
"k8s.io/kubernetes/pkg/util/parsers"
)
// This is a reasonable size range of all container images. 90%ile of images on dockerhub drops into this range.
const (
mb int64 = 1024 * 1024
minImgSize int64 = 23 * mb
maxImgSize int64 = 1000 * mb
)
// ImageLocalityPriorityMap 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 ImageLocalityPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
sumSize := totalImageSize(nodeInfo, pod.Spec.Containers)
return schedulerapi.HostPriority{
Host: node.Name,
Score: calculateScoreFromSize(sumSize),
}, nil
}
// calculateScoreFromSize calculates the priority of a node. sumSize is sum size of requested images on this node.
// 1. Split image size range into 10 buckets.
// 2. Decide the priority of a given sumSize based on which bucket it belongs to.
func calculateScoreFromSize(sumSize int64) int {
switch {
case sumSize == 0 || sumSize < minImgSize:
// 0 means none of the images required by this pod are present on this
// node or the total size of the images present is too small to be taken into further consideration.
return 0
case sumSize >= maxImgSize:
// If existing images' total size is larger than max, just make it highest priority.
return schedulerapi.MaxPriority
}
return int((int64(schedulerapi.MaxPriority) * (sumSize - minImgSize) / (maxImgSize - minImgSize)) + 1)
}
// totalImageSize returns the total image size of all the containers that are already on the node.
func totalImageSize(nodeInfo *schedulercache.NodeInfo, containers []v1.Container) int64 {
var total int64
imageSizes := nodeInfo.ImageSizes()
for _, container := range containers {
if size, ok := imageSizes[normalizedImageName(container.Image)]; ok {
total += size
}
}
return total
}
// normalizedImageName returns the CRI compliant name for a given image.
// TODO: cover the corner cases of missed matches, e.g,
// 1. Using Docker as runtime and docker.io/library/test:tag in pod spec, but only test:tag will present in node status
// 2. Using the implicit registry, i.e., test:tag or library/test:tag in pod spec but only docker.io/library/test:tag
// in node status; note that if users consistently use one registry format, this should not happen.
func normalizedImageName(name string) string {
if strings.LastIndex(name, ":") <= strings.LastIndex(name, "/") {
name = name + ":" + parsers.DefaultImageTag
}
return name
}

View file

@ -0,0 +1,240 @@
/*
Copyright 2016 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 priorities
import (
"sync"
"k8s.io/api/core/v1"
apierrors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/client-go/util/workqueue"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
"k8s.io/kubernetes/pkg/scheduler/algorithm/predicates"
priorityutil "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities/util"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
"github.com/golang/glog"
)
// InterPodAffinity contains information to calculate inter pod affinity.
type InterPodAffinity struct {
info predicates.NodeInfo
nodeLister algorithm.NodeLister
podLister algorithm.PodLister
hardPodAffinityWeight int32
}
// NewInterPodAffinityPriority creates an InterPodAffinity.
func NewInterPodAffinityPriority(
info predicates.NodeInfo,
nodeLister algorithm.NodeLister,
podLister algorithm.PodLister,
hardPodAffinityWeight int32) algorithm.PriorityFunction {
interPodAffinity := &InterPodAffinity{
info: info,
nodeLister: nodeLister,
podLister: podLister,
hardPodAffinityWeight: hardPodAffinityWeight,
}
return interPodAffinity.CalculateInterPodAffinityPriority
}
type podAffinityPriorityMap struct {
sync.Mutex
// nodes contain all nodes that should be considered
nodes []*v1.Node
// counts store the mapping from node name to so-far computed score of
// the node.
counts map[string]float64
// The first error that we faced.
firstError error
}
func newPodAffinityPriorityMap(nodes []*v1.Node) *podAffinityPriorityMap {
return &podAffinityPriorityMap{
nodes: nodes,
counts: make(map[string]float64, len(nodes)),
}
}
func (p *podAffinityPriorityMap) setError(err error) {
p.Lock()
defer p.Unlock()
if p.firstError == nil {
p.firstError = err
}
}
func (p *podAffinityPriorityMap) processTerm(term *v1.PodAffinityTerm, podDefiningAffinityTerm, podToCheck *v1.Pod, fixedNode *v1.Node, weight float64) {
namespaces := priorityutil.GetNamespacesFromPodAffinityTerm(podDefiningAffinityTerm, term)
selector, err := metav1.LabelSelectorAsSelector(term.LabelSelector)
if err != nil {
p.setError(err)
return
}
match := priorityutil.PodMatchesTermsNamespaceAndSelector(podToCheck, namespaces, selector)
if match {
func() {
p.Lock()
defer p.Unlock()
for _, node := range p.nodes {
if priorityutil.NodesHaveSameTopologyKey(node, fixedNode, term.TopologyKey) {
p.counts[node.Name] += weight
}
}
}()
}
}
func (p *podAffinityPriorityMap) processTerms(terms []v1.WeightedPodAffinityTerm, podDefiningAffinityTerm, podToCheck *v1.Pod, fixedNode *v1.Node, multiplier int) {
for i := range terms {
term := &terms[i]
p.processTerm(&term.PodAffinityTerm, podDefiningAffinityTerm, podToCheck, fixedNode, float64(term.Weight*int32(multiplier)))
}
}
// CalculateInterPodAffinityPriority compute a sum by iterating through the elements of weightedPodAffinityTerm and adding
// "weight" to the sum if the corresponding PodAffinityTerm is satisfied for
// that node; the node(s) with the highest sum are the most preferred.
// Symmetry need to be considered for preferredDuringSchedulingIgnoredDuringExecution from podAffinity & podAntiAffinity,
// symmetry need to be considered for hard requirements from podAffinity
func (ipa *InterPodAffinity) CalculateInterPodAffinityPriority(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodes []*v1.Node) (schedulerapi.HostPriorityList, error) {
affinity := pod.Spec.Affinity
hasAffinityConstraints := affinity != nil && affinity.PodAffinity != nil
hasAntiAffinityConstraints := affinity != nil && affinity.PodAntiAffinity != nil
allNodeNames := make([]string, 0, len(nodeNameToInfo))
for name := range nodeNameToInfo {
allNodeNames = append(allNodeNames, name)
}
// convert the topology key based weights to the node name based weights
var maxCount float64
var minCount float64
// priorityMap stores the mapping from node name to so-far computed score of
// the node.
pm := newPodAffinityPriorityMap(nodes)
processPod := func(existingPod *v1.Pod) error {
existingPodNode, err := ipa.info.GetNodeInfo(existingPod.Spec.NodeName)
if err != nil {
if apierrors.IsNotFound(err) {
glog.Errorf("Node not found, %v", existingPod.Spec.NodeName)
return nil
}
return err
}
existingPodAffinity := existingPod.Spec.Affinity
existingHasAffinityConstraints := existingPodAffinity != nil && existingPodAffinity.PodAffinity != nil
existingHasAntiAffinityConstraints := existingPodAffinity != nil && existingPodAffinity.PodAntiAffinity != nil
if hasAffinityConstraints {
// For every soft pod affinity term of <pod>, if <existingPod> matches the term,
// increment <pm.counts> for every node in the cluster with the same <term.TopologyKey>
// value as that of <existingPods>`s node by the term`s weight.
terms := affinity.PodAffinity.PreferredDuringSchedulingIgnoredDuringExecution
pm.processTerms(terms, pod, existingPod, existingPodNode, 1)
}
if hasAntiAffinityConstraints {
// For every soft pod anti-affinity term of <pod>, if <existingPod> matches the term,
// decrement <pm.counts> for every node in the cluster with the same <term.TopologyKey>
// value as that of <existingPod>`s node by the term`s weight.
terms := affinity.PodAntiAffinity.PreferredDuringSchedulingIgnoredDuringExecution
pm.processTerms(terms, pod, existingPod, existingPodNode, -1)
}
if existingHasAffinityConstraints {
// For every hard pod affinity term of <existingPod>, if <pod> matches the term,
// increment <pm.counts> for every node in the cluster with the same <term.TopologyKey>
// value as that of <existingPod>'s node by the constant <ipa.hardPodAffinityWeight>
if ipa.hardPodAffinityWeight > 0 {
terms := existingPodAffinity.PodAffinity.RequiredDuringSchedulingIgnoredDuringExecution
// TODO: Uncomment this block when implement RequiredDuringSchedulingRequiredDuringExecution.
//if len(existingPodAffinity.PodAffinity.RequiredDuringSchedulingRequiredDuringExecution) != 0 {
// terms = append(terms, existingPodAffinity.PodAffinity.RequiredDuringSchedulingRequiredDuringExecution...)
//}
for _, term := range terms {
pm.processTerm(&term, existingPod, pod, existingPodNode, float64(ipa.hardPodAffinityWeight))
}
}
// For every soft pod affinity term of <existingPod>, if <pod> matches the term,
// increment <pm.counts> for every node in the cluster with the same <term.TopologyKey>
// value as that of <existingPod>'s node by the term's weight.
terms := existingPodAffinity.PodAffinity.PreferredDuringSchedulingIgnoredDuringExecution
pm.processTerms(terms, existingPod, pod, existingPodNode, 1)
}
if existingHasAntiAffinityConstraints {
// For every soft pod anti-affinity term of <existingPod>, if <pod> matches the term,
// decrement <pm.counts> for every node in the cluster with the same <term.TopologyKey>
// value as that of <existingPod>'s node by the term's weight.
terms := existingPodAffinity.PodAntiAffinity.PreferredDuringSchedulingIgnoredDuringExecution
pm.processTerms(terms, existingPod, pod, existingPodNode, -1)
}
return nil
}
processNode := func(i int) {
nodeInfo := nodeNameToInfo[allNodeNames[i]]
if nodeInfo.Node() != nil {
if hasAffinityConstraints || hasAntiAffinityConstraints {
// We need to process all the nodes.
for _, existingPod := range nodeInfo.Pods() {
if err := processPod(existingPod); err != nil {
pm.setError(err)
}
}
} else {
// The pod doesn't have any constraints - we need to check only existing
// ones that have some.
for _, existingPod := range nodeInfo.PodsWithAffinity() {
if err := processPod(existingPod); err != nil {
pm.setError(err)
}
}
}
}
}
workqueue.Parallelize(16, len(allNodeNames), processNode)
if pm.firstError != nil {
return nil, pm.firstError
}
for _, node := range nodes {
if pm.counts[node.Name] > maxCount {
maxCount = pm.counts[node.Name]
}
if pm.counts[node.Name] < minCount {
minCount = pm.counts[node.Name]
}
}
// calculate final priority score for each node
result := make(schedulerapi.HostPriorityList, 0, len(nodes))
for _, node := range nodes {
fScore := float64(0)
if (maxCount - minCount) > 0 {
fScore = float64(schedulerapi.MaxPriority) * ((pm.counts[node.Name] - minCount) / (maxCount - minCount))
}
result = append(result, schedulerapi.HostPriority{Host: node.Name, Score: int(fScore)})
if glog.V(10) {
glog.Infof("%v -> %v: InterPodAffinityPriority, Score: (%d)", pod.Name, node.Name, int(fScore))
}
}
return result, nil
}

View file

@ -0,0 +1,53 @@
/*
Copyright 2016 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 priorities
import (
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
var (
leastResourcePriority = &ResourceAllocationPriority{"LeastResourceAllocation", leastResourceScorer}
// LeastRequestedPriorityMap 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
LeastRequestedPriorityMap = leastResourcePriority.PriorityMap
)
func leastResourceScorer(requested, allocable *schedulercache.Resource, includeVolumes bool, requestedVolumes int, allocatableVolumes int) int64 {
return (leastRequestedScore(requested.MilliCPU, allocable.MilliCPU) +
leastRequestedScore(requested.Memory, allocable.Memory)) / 2
}
// The unused capacity is calculated on a scale of 0-10
// 0 being the lowest priority and 10 being the highest.
// The more unused resources the higher the score is.
func leastRequestedScore(requested, capacity int64) int64 {
if capacity == 0 {
return 0
}
if requested > capacity {
return 0
}
return ((capacity - requested) * int64(schedulerapi.MaxPriority)) / capacity
}

View file

@ -0,0 +1,114 @@
/*
Copyright 2016 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 priorities
import (
"k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
priorityutil "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities/util"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// PriorityMetadataFactory is a factory to produce PriorityMetadata.
type PriorityMetadataFactory struct {
serviceLister algorithm.ServiceLister
controllerLister algorithm.ControllerLister
replicaSetLister algorithm.ReplicaSetLister
statefulSetLister algorithm.StatefulSetLister
}
// NewPriorityMetadataFactory creates a PriorityMetadataFactory.
func NewPriorityMetadataFactory(serviceLister algorithm.ServiceLister, controllerLister algorithm.ControllerLister, replicaSetLister algorithm.ReplicaSetLister, statefulSetLister algorithm.StatefulSetLister) algorithm.PriorityMetadataProducer {
factory := &PriorityMetadataFactory{
serviceLister: serviceLister,
controllerLister: controllerLister,
replicaSetLister: replicaSetLister,
statefulSetLister: statefulSetLister,
}
return factory.PriorityMetadata
}
// priorityMetadata is a type that is passed as metadata for priority functions
type priorityMetadata struct {
nonZeroRequest *schedulercache.Resource
podTolerations []v1.Toleration
affinity *v1.Affinity
podSelectors []labels.Selector
controllerRef *metav1.OwnerReference
podFirstServiceSelector labels.Selector
}
// PriorityMetadata is a PriorityMetadataProducer. Node info can be nil.
func (pmf *PriorityMetadataFactory) PriorityMetadata(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo) interface{} {
// If we cannot compute metadata, just return nil
if pod == nil {
return nil
}
return &priorityMetadata{
nonZeroRequest: getNonZeroRequests(pod),
podTolerations: getAllTolerationPreferNoSchedule(pod.Spec.Tolerations),
affinity: pod.Spec.Affinity,
podSelectors: getSelectors(pod, pmf.serviceLister, pmf.controllerLister, pmf.replicaSetLister, pmf.statefulSetLister),
controllerRef: priorityutil.GetControllerRef(pod),
podFirstServiceSelector: getFirstServiceSelector(pod, pmf.serviceLister),
}
}
// getFirstServiceSelector returns one selector of services the given pod.
func getFirstServiceSelector(pod *v1.Pod, sl algorithm.ServiceLister) (firstServiceSelector labels.Selector) {
if services, err := sl.GetPodServices(pod); err == nil && len(services) > 0 {
return labels.SelectorFromSet(services[0].Spec.Selector)
}
return nil
}
// getSelectors returns selectors of services, RCs and RSs matching the given pod.
func getSelectors(pod *v1.Pod, sl algorithm.ServiceLister, cl algorithm.ControllerLister, rsl algorithm.ReplicaSetLister, ssl algorithm.StatefulSetLister) []labels.Selector {
var selectors []labels.Selector
if services, err := sl.GetPodServices(pod); err == nil {
for _, service := range services {
selectors = append(selectors, labels.SelectorFromSet(service.Spec.Selector))
}
}
if rcs, err := cl.GetPodControllers(pod); err == nil {
for _, rc := range rcs {
selectors = append(selectors, labels.SelectorFromSet(rc.Spec.Selector))
}
}
if rss, err := rsl.GetPodReplicaSets(pod); err == nil {
for _, rs := range rss {
if selector, err := metav1.LabelSelectorAsSelector(rs.Spec.Selector); err == nil {
selectors = append(selectors, selector)
}
}
}
if sss, err := ssl.GetPodStatefulSets(pod); err == nil {
for _, ss := range sss {
if selector, err := metav1.LabelSelectorAsSelector(ss.Spec.Selector); err == nil {
selectors = append(selectors, selector)
}
}
}
return selectors
}

View file

@ -0,0 +1,55 @@
/*
Copyright 2016 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 priorities
import (
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
var (
mostResourcePriority = &ResourceAllocationPriority{"MostResourceAllocation", mostResourceScorer}
// MostRequestedPriorityMap is a priority function that favors nodes with most requested resources.
// It calculates the percentage of memory and CPU requested by pods scheduled on the node, and prioritizes
// based on the maximum of the average of the fraction of requested to capacity.
// Details: (cpu(10 * sum(requested) / capacity) + memory(10 * sum(requested) / capacity)) / 2
MostRequestedPriorityMap = mostResourcePriority.PriorityMap
)
func mostResourceScorer(requested, allocable *schedulercache.Resource, includeVolumes bool, requestedVolumes int, allocatableVolumes int) int64 {
return (mostRequestedScore(requested.MilliCPU, allocable.MilliCPU) +
mostRequestedScore(requested.Memory, allocable.Memory)) / 2
}
// The used capacity is calculated on a scale of 0-10
// 0 being the lowest priority and 10 being the highest.
// The more resources are used the higher the score is. This function
// is almost a reversed version of least_requested_priority.calculatUnusedScore
// (10 - calculateUnusedScore). The main difference is in rounding. It was added to
// keep the final formula clean and not to modify the widely used (by users
// in their default scheduling policies) calculateUSedScore.
func mostRequestedScore(requested, capacity int64) int64 {
if capacity == 0 {
return 0
}
if requested > capacity {
return 0
}
return (requested * schedulerapi.MaxPriority) / capacity
}

View file

@ -0,0 +1,77 @@
/*
Copyright 2015 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/labels"
v1helper "k8s.io/kubernetes/pkg/apis/core/v1/helper"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// CalculateNodeAffinityPriorityMap 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 CalculateNodeAffinityPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
// default is the podspec.
affinity := pod.Spec.Affinity
if priorityMeta, ok := meta.(*priorityMetadata); ok {
// We were able to parse metadata, use affinity from there.
affinity = priorityMeta.affinity
}
var count int32
// A nil element of PreferredDuringSchedulingIgnoredDuringExecution matches no objects.
// An element of PreferredDuringSchedulingIgnoredDuringExecution that refers to an
// empty PreferredSchedulingTerm matches all objects.
if affinity != nil && affinity.NodeAffinity != nil && affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution != nil {
// Match PreferredDuringSchedulingIgnoredDuringExecution term by term.
for i := range affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution {
preferredSchedulingTerm := &affinity.NodeAffinity.PreferredDuringSchedulingIgnoredDuringExecution[i]
if preferredSchedulingTerm.Weight == 0 {
continue
}
// TODO: Avoid computing it for all nodes if this becomes a performance problem.
nodeSelector, err := v1helper.NodeSelectorRequirementsAsSelector(preferredSchedulingTerm.Preference.MatchExpressions)
if err != nil {
return schedulerapi.HostPriority{}, err
}
if nodeSelector.Matches(labels.Set(node.Labels)) {
count += preferredSchedulingTerm.Weight
}
}
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: int(count),
}, nil
}
// CalculateNodeAffinityPriorityReduce is a reduce function for node affinity priority calculation.
var CalculateNodeAffinityPriorityReduce = NormalizeReduce(schedulerapi.MaxPriority, false)

View file

@ -0,0 +1,62 @@
/*
Copyright 2016 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// NodeLabelPrioritizer contains information to calculate node label priority.
type NodeLabelPrioritizer struct {
label string
presence bool
}
// NewNodeLabelPriority creates a NodeLabelPrioritizer.
func NewNodeLabelPriority(label string, presence bool) (algorithm.PriorityMapFunction, algorithm.PriorityReduceFunction) {
labelPrioritizer := &NodeLabelPrioritizer{
label: label,
presence: presence,
}
return labelPrioritizer.CalculateNodeLabelPriorityMap, nil
}
// CalculateNodeLabelPriorityMap 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) CalculateNodeLabelPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
exists := labels.Set(node.Labels).Has(n.label)
score := 0
if (exists && n.presence) || (!exists && !n.presence) {
score = schedulerapi.MaxPriority
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: score,
}, nil
}

View file

@ -0,0 +1,68 @@
/*
Copyright 2015 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
v1helper "k8s.io/kubernetes/pkg/apis/core/v1/helper"
priorityutil "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities/util"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// CalculateNodePreferAvoidPodsPriorityMap priorities nodes according to the node annotation
// "scheduler.alpha.kubernetes.io/preferAvoidPods".
func CalculateNodePreferAvoidPodsPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
var controllerRef *metav1.OwnerReference
if priorityMeta, ok := meta.(*priorityMetadata); ok {
controllerRef = priorityMeta.controllerRef
} else {
// We couldn't parse metadata - fallback to the podspec.
controllerRef = priorityutil.GetControllerRef(pod)
}
if controllerRef != nil {
// Ignore pods that are owned by other controller than ReplicationController
// or ReplicaSet.
if controllerRef.Kind != "ReplicationController" && controllerRef.Kind != "ReplicaSet" {
controllerRef = nil
}
}
if controllerRef == nil {
return schedulerapi.HostPriority{Host: node.Name, Score: schedulerapi.MaxPriority}, nil
}
avoids, err := v1helper.GetAvoidPodsFromNodeAnnotations(node.Annotations)
if err != nil {
// If we cannot get annotation, assume it's schedulable there.
return schedulerapi.HostPriority{Host: node.Name, Score: schedulerapi.MaxPriority}, nil
}
for i := range avoids.PreferAvoidPods {
avoid := &avoids.PreferAvoidPods[i]
if avoid.PodSignature.PodController.Kind == controllerRef.Kind && avoid.PodSignature.PodController.UID == controllerRef.UID {
return schedulerapi.HostPriority{Host: node.Name, Score: 0}, nil
}
}
return schedulerapi.HostPriority{Host: node.Name, Score: schedulerapi.MaxPriority}, nil
}

View file

@ -0,0 +1,64 @@
/*
Copyright 2017 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 priorities
import (
"k8s.io/api/core/v1"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// NormalizeReduce generates a PriorityReduceFunction that can normalize the result
// scores to [0, maxPriority]. If reverse is set to true, it reverses the scores by
// subtracting it from maxPriority.
func NormalizeReduce(maxPriority int, reverse bool) algorithm.PriorityReduceFunction {
return func(
_ *v1.Pod,
_ interface{},
_ map[string]*schedulercache.NodeInfo,
result schedulerapi.HostPriorityList) error {
var maxCount int
for i := range result {
if result[i].Score > maxCount {
maxCount = result[i].Score
}
}
if maxCount == 0 {
if reverse {
for i := range result {
result[i].Score = maxPriority
}
}
return nil
}
for i := range result {
score := result[i].Score
score = maxPriority * score / maxCount
if reverse {
score = maxPriority - score
}
result[i].Score = score
}
return nil
}
}

View file

@ -0,0 +1,141 @@
/*
Copyright 2018 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 priorities
import (
"fmt"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// FunctionShape represents shape of scoring function.
// For safety use NewFunctionShape which performs precondition checks for struct creation.
type FunctionShape []FunctionShapePoint
// FunctionShapePoint represents single point in scoring function shape.
type FunctionShapePoint struct {
// Utilization is function argument.
Utilization int64
// Score is function value.
Score int64
}
var (
// give priority to least utilized nodes by default
defaultFunctionShape, _ = NewFunctionShape([]FunctionShapePoint{{0, 10}, {100, 0}})
)
const (
minUtilization = 0
maxUtilization = 100
minScore = 0
maxScore = schedulerapi.MaxPriority
)
// NewFunctionShape creates instance of FunctionShape in a safe way performing all
// necessary sanity checks.
func NewFunctionShape(points []FunctionShapePoint) (FunctionShape, error) {
n := len(points)
if n == 0 {
return nil, fmt.Errorf("at least one point must be specified")
}
for i := 1; i < n; i++ {
if points[i-1].Utilization >= points[i].Utilization {
return nil, fmt.Errorf("utilization values must be sorted. Utilization[%d]==%d >= Utilization[%d]==%d", i-1, points[i-1].Utilization, i, points[i].Utilization)
}
}
for i, point := range points {
if point.Utilization < minUtilization {
return nil, fmt.Errorf("utilization values must not be less than %d. Utilization[%d]==%d", minUtilization, i, point.Utilization)
}
if point.Utilization > maxUtilization {
return nil, fmt.Errorf("utilization values must not be greater than %d. Utilization[%d]==%d", maxUtilization, i, point.Utilization)
}
if point.Score < minScore {
return nil, fmt.Errorf("score values must not be less than %d. Score[%d]==%d", minScore, i, point.Score)
}
if point.Score > maxScore {
return nil, fmt.Errorf("score valuses not be greater than %d. Score[%d]==%d", maxScore, i, point.Score)
}
}
// We make defensive copy so we make no assumption if array passed as argument is not changed afterwards
pointsCopy := make(FunctionShape, n)
copy(pointsCopy, points)
return pointsCopy, nil
}
// RequestedToCapacityRatioResourceAllocationPriorityDefault creates a requestedToCapacity based
// ResourceAllocationPriority using default resource scoring function shape.
// The default function assigns 1.0 to resource when all capacity is available
// and 0.0 when requested amount is equal to capacity.
func RequestedToCapacityRatioResourceAllocationPriorityDefault() *ResourceAllocationPriority {
return RequestedToCapacityRatioResourceAllocationPriority(defaultFunctionShape)
}
// RequestedToCapacityRatioResourceAllocationPriority creates a requestedToCapacity based
// ResourceAllocationPriority using provided resource scoring function shape.
func RequestedToCapacityRatioResourceAllocationPriority(scoringFunctionShape FunctionShape) *ResourceAllocationPriority {
return &ResourceAllocationPriority{"RequestedToCapacityRatioResourceAllocationPriority", buildRequestedToCapacityRatioScorerFunction(scoringFunctionShape)}
}
func buildRequestedToCapacityRatioScorerFunction(scoringFunctionShape FunctionShape) func(*schedulercache.Resource, *schedulercache.Resource, bool, int, int) int64 {
rawScoringFunction := buildBrokenLinearFunction(scoringFunctionShape)
resourceScoringFunction := func(requested, capacity int64) int64 {
if capacity == 0 || requested > capacity {
return rawScoringFunction(maxUtilization)
}
return rawScoringFunction(maxUtilization - (capacity-requested)*maxUtilization/capacity)
}
return func(requested, allocable *schedulercache.Resource, includeVolumes bool, requestedVolumes int, allocatableVolumes int) int64 {
cpuScore := resourceScoringFunction(requested.MilliCPU, allocable.MilliCPU)
memoryScore := resourceScoringFunction(requested.Memory, allocable.Memory)
return (cpuScore + memoryScore) / 2
}
}
// Creates a function which is built using linear segments. Segments are defined via shape array.
// Shape[i].Utilization slice represents points on "utilization" axis where different segments meet.
// Shape[i].Score represents function values at meeting points.
//
// function f(p) is defined as:
// shape[0].Score for p < f[0].Utilization
// shape[i].Score for p == shape[i].Utilization
// shape[n-1].Score for p > shape[n-1].Utilization
// and linear between points (p < shape[i].Utilization)
func buildBrokenLinearFunction(shape FunctionShape) func(int64) int64 {
n := len(shape)
return func(p int64) int64 {
for i := 0; i < n; i++ {
if p <= shape[i].Utilization {
if i == 0 {
return shape[0].Score
}
return shape[i-1].Score + (shape[i].Score-shape[i-1].Score)*(p-shape[i-1].Utilization)/(shape[i].Utilization-shape[i-1].Utilization)
}
}
return shape[n-1].Score
}
}

View file

@ -0,0 +1,90 @@
/*
Copyright 2017 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 priorities
import (
"fmt"
"github.com/golang/glog"
"k8s.io/api/core/v1"
priorityutil "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities/util"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// ResourceAllocationPriority contains information to calculate resource allocation priority.
type ResourceAllocationPriority struct {
Name string
scorer func(requested, allocable *schedulercache.Resource, includeVolumes bool, requestedVolumes int, allocatableVolumes int) int64
}
// PriorityMap priorities nodes according to the resource allocations on the node.
// It will use `scorer` function to calculate the score.
func (r *ResourceAllocationPriority) PriorityMap(
pod *v1.Pod,
meta interface{},
nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
allocatable := nodeInfo.AllocatableResource()
var requested schedulercache.Resource
if priorityMeta, ok := meta.(*priorityMetadata); ok {
requested = *priorityMeta.nonZeroRequest
} else {
// We couldn't parse metadata - fallback to computing it.
requested = *getNonZeroRequests(pod)
}
requested.MilliCPU += nodeInfo.NonZeroRequest().MilliCPU
requested.Memory += nodeInfo.NonZeroRequest().Memory
var score int64
// Check if the pod has volumes and this could be added to scorer function for balanced resource allocation.
if len(pod.Spec.Volumes) >= 0 && nodeInfo.TransientInfo != nil {
score = r.scorer(&requested, &allocatable, true, nodeInfo.TransientInfo.TransNodeInfo.RequestedVolumes, nodeInfo.TransientInfo.TransNodeInfo.AllocatableVolumesCount)
} else {
score = r.scorer(&requested, &allocatable, false, 0, 0)
}
if glog.V(10) {
glog.Infof(
"%v -> %v: %v, capacity %d millicores %d memory bytes, total request %d millicores %d memory bytes, score %d",
pod.Name, node.Name, r.Name,
allocatable.MilliCPU, allocatable.Memory,
requested.MilliCPU+allocatable.MilliCPU, requested.Memory+allocatable.Memory,
score,
)
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: int(score),
}, nil
}
func getNonZeroRequests(pod *v1.Pod) *schedulercache.Resource {
result := &schedulercache.Resource{}
for i := range pod.Spec.Containers {
container := &pod.Spec.Containers[i]
cpu, memory := priorityutil.GetNonzeroRequests(&container.Resources.Requests)
result.MilliCPU += cpu
result.Memory += memory
}
return result
}

View file

@ -0,0 +1,99 @@
/*
Copyright 2017 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
"github.com/golang/glog"
)
// ResourceLimitsPriorityMap is a priority function that increases score of input node by 1 if the node satisfies
// input pod's resource limits. In detail, this priority function works as follows: If a node does not publish its
// allocatable resources (cpu and memory both), the node score is not affected. If a pod does not specify
// its cpu and memory limits both, the node score is not affected. If one or both of cpu and memory limits
// of the pod are satisfied, the node is assigned a score of 1.
// Rationale of choosing the lowest score of 1 is that this is mainly selected to break ties between nodes that have
// same scores assigned by one of least and most requested priority functions.
func ResourceLimitsPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
allocatableResources := nodeInfo.AllocatableResource()
// compute pod limits
podLimits := getResourceLimits(pod)
cpuScore := computeScore(podLimits.MilliCPU, allocatableResources.MilliCPU)
memScore := computeScore(podLimits.Memory, allocatableResources.Memory)
score := int(0)
if cpuScore == 1 || memScore == 1 {
score = 1
}
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: Resource Limits Priority, allocatable %d millicores %d memory bytes, pod limits %d millicores %d memory bytes, score %d",
pod.Name, node.Name,
allocatableResources.MilliCPU, allocatableResources.Memory,
podLimits.MilliCPU, podLimits.Memory,
score,
)
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: score,
}, nil
}
// computeScore return 1 if limit value is less than or equal to allocable
// value, otherwise it returns 0.
func computeScore(limit, allocatable int64) int64 {
if limit != 0 && allocatable != 0 && limit <= allocatable {
return 1
}
return 0
}
// getResourceLimits computes resource limits for input pod.
// The reason to create this new function is to be consistent with other
// priority functions because most or perhaps all priority functions work
// with schedulercache.Resource.
// TODO: cache it as part of metadata passed to priority functions.
func getResourceLimits(pod *v1.Pod) *schedulercache.Resource {
result := &schedulercache.Resource{}
for _, container := range pod.Spec.Containers {
result.Add(container.Resources.Limits)
}
// take max_resource(sum_pod, any_init_container)
for _, container := range pod.Spec.InitContainers {
result.SetMaxResource(container.Resources.Limits)
}
return result
}

View file

@ -0,0 +1,285 @@
/*
Copyright 2014 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
utilnode "k8s.io/kubernetes/pkg/util/node"
"github.com/golang/glog"
)
// When zone information is present, give 2/3 of the weighting to zone spreading, 1/3 to node spreading
// TODO: Any way to justify this weighting?
const zoneWeighting float64 = 2.0 / 3.0
// SelectorSpread contains information to calculate selector spread priority.
type SelectorSpread struct {
serviceLister algorithm.ServiceLister
controllerLister algorithm.ControllerLister
replicaSetLister algorithm.ReplicaSetLister
statefulSetLister algorithm.StatefulSetLister
}
// NewSelectorSpreadPriority creates a SelectorSpread.
func NewSelectorSpreadPriority(
serviceLister algorithm.ServiceLister,
controllerLister algorithm.ControllerLister,
replicaSetLister algorithm.ReplicaSetLister,
statefulSetLister algorithm.StatefulSetLister) (algorithm.PriorityMapFunction, algorithm.PriorityReduceFunction) {
selectorSpread := &SelectorSpread{
serviceLister: serviceLister,
controllerLister: controllerLister,
replicaSetLister: replicaSetLister,
statefulSetLister: statefulSetLister,
}
return selectorSpread.CalculateSpreadPriorityMap, selectorSpread.CalculateSpreadPriorityReduce
}
// CalculateSpreadPriorityMap spreads pods across hosts, considering pods
// belonging to the same service,RC,RS or StatefulSet.
// When a pod is scheduled, it looks for services, RCs,RSs and StatefulSets that match the pod,
// then finds existing pods that match those selectors.
// It favors nodes that have fewer existing matching pods.
// i.e. it pushes the scheduler towards a node where there's the smallest number of
// pods which match the same service, RC,RSs or StatefulSets selectors as the pod being scheduled.
func (s *SelectorSpread) CalculateSpreadPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
var selectors []labels.Selector
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
priorityMeta, ok := meta.(*priorityMetadata)
if ok {
selectors = priorityMeta.podSelectors
} else {
selectors = getSelectors(pod, s.serviceLister, s.controllerLister, s.replicaSetLister, s.statefulSetLister)
}
if len(selectors) == 0 {
return schedulerapi.HostPriority{
Host: node.Name,
Score: int(0),
}, nil
}
count := int(0)
for _, nodePod := range nodeInfo.Pods() {
if pod.Namespace != nodePod.Namespace {
continue
}
// When we are replacing a failed pod, we often see the previous
// deleted version while scheduling the replacement.
// Ignore the previous deleted version for spreading purposes
// (it can still be considered for resource restrictions etc.)
if nodePod.DeletionTimestamp != nil {
glog.V(4).Infof("skipping pending-deleted pod: %s/%s", nodePod.Namespace, nodePod.Name)
continue
}
matches := false
for _, selector := range selectors {
if selector.Matches(labels.Set(nodePod.ObjectMeta.Labels)) {
matches = true
break
}
}
if matches {
count++
}
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: int(count),
}, nil
}
// CalculateSpreadPriorityReduce calculates the source of each node
// based on the number of existing matching pods on the node
// where zone information is included on the nodes, it favors nodes
// in zones with fewer existing matching pods.
func (s *SelectorSpread) CalculateSpreadPriorityReduce(pod *v1.Pod, meta interface{}, nodeNameToInfo map[string]*schedulercache.NodeInfo, result schedulerapi.HostPriorityList) error {
countsByZone := make(map[string]int, 10)
maxCountByZone := int(0)
maxCountByNodeName := int(0)
for i := range result {
if result[i].Score > maxCountByNodeName {
maxCountByNodeName = result[i].Score
}
zoneID := utilnode.GetZoneKey(nodeNameToInfo[result[i].Host].Node())
if zoneID == "" {
continue
}
countsByZone[zoneID] += result[i].Score
}
for zoneID := range countsByZone {
if countsByZone[zoneID] > maxCountByZone {
maxCountByZone = countsByZone[zoneID]
}
}
haveZones := len(countsByZone) != 0
maxCountByNodeNameFloat64 := float64(maxCountByNodeName)
maxCountByZoneFloat64 := float64(maxCountByZone)
MaxPriorityFloat64 := float64(schedulerapi.MaxPriority)
for i := range result {
// initializing to the default/max node score of maxPriority
fScore := MaxPriorityFloat64
if maxCountByNodeName > 0 {
fScore = MaxPriorityFloat64 * (float64(maxCountByNodeName-result[i].Score) / maxCountByNodeNameFloat64)
}
// If there is zone information present, incorporate it
if haveZones {
zoneID := utilnode.GetZoneKey(nodeNameToInfo[result[i].Host].Node())
if zoneID != "" {
zoneScore := MaxPriorityFloat64
if maxCountByZone > 0 {
zoneScore = MaxPriorityFloat64 * (float64(maxCountByZone-countsByZone[zoneID]) / maxCountByZoneFloat64)
}
fScore = (fScore * (1.0 - zoneWeighting)) + (zoneWeighting * zoneScore)
}
}
result[i].Score = int(fScore)
if glog.V(10) {
glog.Infof(
"%v -> %v: SelectorSpreadPriority, Score: (%d)", pod.Name, result[i].Host, int(fScore),
)
}
}
return nil
}
// ServiceAntiAffinity contains information to calculate service anti-affinity priority.
type ServiceAntiAffinity struct {
podLister algorithm.PodLister
serviceLister algorithm.ServiceLister
label string
}
// NewServiceAntiAffinityPriority creates a ServiceAntiAffinity.
func NewServiceAntiAffinityPriority(podLister algorithm.PodLister, serviceLister algorithm.ServiceLister, label string) (algorithm.PriorityMapFunction, algorithm.PriorityReduceFunction) {
antiAffinity := &ServiceAntiAffinity{
podLister: podLister,
serviceLister: serviceLister,
label: label,
}
return antiAffinity.CalculateAntiAffinityPriorityMap, antiAffinity.CalculateAntiAffinityPriorityReduce
}
// Classifies nodes into ones with labels and without labels.
func (s *ServiceAntiAffinity) getNodeClassificationByLabels(nodes []*v1.Node) (map[string]string, []string) {
labeledNodes := map[string]string{}
nonLabeledNodes := []string{}
for _, node := range nodes {
if labels.Set(node.Labels).Has(s.label) {
label := labels.Set(node.Labels).Get(s.label)
labeledNodes[node.Name] = label
} else {
nonLabeledNodes = append(nonLabeledNodes, node.Name)
}
}
return labeledNodes, nonLabeledNodes
}
// filteredPod get pods based on namespace and selector
func filteredPod(namespace string, selector labels.Selector, nodeInfo *schedulercache.NodeInfo) (pods []*v1.Pod) {
if nodeInfo.Pods() == nil || len(nodeInfo.Pods()) == 0 || selector == nil {
return []*v1.Pod{}
}
for _, pod := range nodeInfo.Pods() {
// Ignore pods being deleted for spreading purposes
// Similar to how it is done for SelectorSpreadPriority
if namespace == pod.Namespace && pod.DeletionTimestamp == nil && selector.Matches(labels.Set(pod.Labels)) {
pods = append(pods, pod)
}
}
return
}
// CalculateAntiAffinityPriorityMap spreads pods by minimizing the number of pods belonging to the same service
// on given machine
func (s *ServiceAntiAffinity) CalculateAntiAffinityPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
var firstServiceSelector labels.Selector
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
priorityMeta, ok := meta.(*priorityMetadata)
if ok {
firstServiceSelector = priorityMeta.podFirstServiceSelector
} else {
firstServiceSelector = getFirstServiceSelector(pod, s.serviceLister)
}
//pods matched namespace,selector on current node
matchedPodsOfNode := filteredPod(pod.Namespace, firstServiceSelector, nodeInfo)
return schedulerapi.HostPriority{
Host: node.Name,
Score: int(len(matchedPodsOfNode)),
}, nil
}
// CalculateAntiAffinityPriorityReduce computes each node score with the same value for a particular label.
// The label to be considered is provided to the struct (ServiceAntiAffinity).
func (s *ServiceAntiAffinity) CalculateAntiAffinityPriorityReduce(pod *v1.Pod, meta interface{}, nodeNameToInfo map[string]*schedulercache.NodeInfo, result schedulerapi.HostPriorityList) error {
var numServicePods int
var label string
podCounts := map[string]int{}
labelNodesStatus := map[string]string{}
maxPriorityFloat64 := float64(schedulerapi.MaxPriority)
for _, hostPriority := range result {
numServicePods += hostPriority.Score
if !labels.Set(nodeNameToInfo[hostPriority.Host].Node().Labels).Has(s.label) {
continue
}
label = labels.Set(nodeNameToInfo[hostPriority.Host].Node().Labels).Get(s.label)
labelNodesStatus[hostPriority.Host] = label
podCounts[label] += hostPriority.Score
}
//score int - scale of 0-maxPriority
// 0 being the lowest priority and maxPriority being the highest
for i, hostPriority := range result {
label, ok := labelNodesStatus[hostPriority.Host]
if !ok {
result[i].Host = hostPriority.Host
result[i].Score = int(0)
continue
}
// initializing to the default/max node score of maxPriority
fScore := maxPriorityFloat64
if numServicePods > 0 {
fScore = maxPriorityFloat64 * (float64(numServicePods-podCounts[label]) / float64(numServicePods))
}
result[i].Host = hostPriority.Host
result[i].Score = int(fScore)
}
return nil
}

View file

@ -0,0 +1,76 @@
/*
Copyright 2016 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 priorities
import (
"fmt"
"k8s.io/api/core/v1"
v1helper "k8s.io/kubernetes/pkg/apis/core/v1/helper"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// CountIntolerableTaintsPreferNoSchedule gives the count of intolerable taints of a pod with effect PreferNoSchedule
func countIntolerableTaintsPreferNoSchedule(taints []v1.Taint, tolerations []v1.Toleration) (intolerableTaints int) {
for _, taint := range taints {
// check only on taints that have effect PreferNoSchedule
if taint.Effect != v1.TaintEffectPreferNoSchedule {
continue
}
if !v1helper.TolerationsTolerateTaint(tolerations, &taint) {
intolerableTaints++
}
}
return
}
// getAllTolerationEffectPreferNoSchedule gets the list of all Tolerations with Effect PreferNoSchedule or with no effect.
func getAllTolerationPreferNoSchedule(tolerations []v1.Toleration) (tolerationList []v1.Toleration) {
for _, toleration := range tolerations {
// Empty effect means all effects which includes PreferNoSchedule, so we need to collect it as well.
if len(toleration.Effect) == 0 || toleration.Effect == v1.TaintEffectPreferNoSchedule {
tolerationList = append(tolerationList, toleration)
}
}
return
}
// ComputeTaintTolerationPriorityMap prepares the priority list for all the nodes based on the number of intolerable taints on the node
func ComputeTaintTolerationPriorityMap(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error) {
node := nodeInfo.Node()
if node == nil {
return schedulerapi.HostPriority{}, fmt.Errorf("node not found")
}
// To hold all the tolerations with Effect PreferNoSchedule
var tolerationsPreferNoSchedule []v1.Toleration
if priorityMeta, ok := meta.(*priorityMetadata); ok {
tolerationsPreferNoSchedule = priorityMeta.podTolerations
} else {
tolerationsPreferNoSchedule = getAllTolerationPreferNoSchedule(pod.Spec.Tolerations)
}
return schedulerapi.HostPriority{
Host: node.Name,
Score: countIntolerableTaintsPreferNoSchedule(node.Spec.Taints, tolerationsPreferNoSchedule),
}, nil
}
// ComputeTaintTolerationPriorityReduce calculates the source of each node based on the number of intolerable taints on the node
var ComputeTaintTolerationPriorityReduce = NormalizeReduce(schedulerapi.MaxPriority, true)

View file

@ -0,0 +1,61 @@
/*
Copyright 2016 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 priorities
import (
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/api/resource"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/kubernetes/pkg/scheduler/algorithm"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
func makeNode(node string, milliCPU, memory int64) *v1.Node {
return &v1.Node{
ObjectMeta: metav1.ObjectMeta{Name: node},
Status: v1.NodeStatus{
Capacity: v1.ResourceList{
v1.ResourceCPU: *resource.NewMilliQuantity(milliCPU, resource.DecimalSI),
v1.ResourceMemory: *resource.NewQuantity(memory, resource.BinarySI),
},
Allocatable: v1.ResourceList{
v1.ResourceCPU: *resource.NewMilliQuantity(milliCPU, resource.DecimalSI),
v1.ResourceMemory: *resource.NewQuantity(memory, resource.BinarySI),
},
},
}
}
func priorityFunction(mapFn algorithm.PriorityMapFunction, reduceFn algorithm.PriorityReduceFunction, mataData interface{}) algorithm.PriorityFunction {
return func(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodes []*v1.Node) (schedulerapi.HostPriorityList, error) {
result := make(schedulerapi.HostPriorityList, 0, len(nodes))
for i := range nodes {
hostResult, err := mapFn(pod, mataData, nodeNameToInfo[nodes[i].Name])
if err != nil {
return nil, err
}
result = append(result, hostResult)
}
if reduceFn != nil {
if err := reduceFn(pod, mataData, nodeNameToInfo, result); err != nil {
return nil, err
}
}
return result, nil
}
}

View file

@ -0,0 +1,55 @@
package(default_visibility = ["//visibility:public"])
load(
"@io_bazel_rules_go//go:def.bzl",
"go_library",
"go_test",
)
go_test(
name = "go_default_test",
srcs = [
"non_zero_test.go",
"topologies_test.go",
"util_test.go",
],
embed = [":go_default_library"],
deps = [
"//vendor/github.com/stretchr/testify/assert:go_default_library",
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/api/resource:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/selection:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/util/sets:go_default_library",
],
)
go_library(
name = "go_default_library",
srcs = [
"non_zero.go",
"topologies.go",
"util.go",
],
importpath = "k8s.io/kubernetes/pkg/scheduler/algorithm/priorities/util",
deps = [
"//vendor/k8s.io/api/core/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/apis/meta/v1:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/labels:go_default_library",
"//vendor/k8s.io/apimachinery/pkg/util/sets:go_default_library",
],
)
filegroup(
name = "package-srcs",
srcs = glob(["**"]),
tags = ["automanaged"],
visibility = ["//visibility:private"],
)
filegroup(
name = "all-srcs",
srcs = [":package-srcs"],
tags = ["automanaged"],
)

View file

@ -0,0 +1,52 @@
/*
Copyright 2016 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 util
import "k8s.io/api/core/v1"
// For each of these resources, a pod that doesn't request the resource explicitly
// will be treated as having requested the amount indicated below, for the purpose
// of computing priority only. This ensures that when scheduling zero-request pods, such
// pods will not all be scheduled to the machine with the smallest in-use request,
// and that when scheduling regular pods, such pods will not see zero-request pods as
// consuming no resources whatsoever. We chose these values to be similar to the
// resources that we give to cluster addon pods (#10653). But they are pretty arbitrary.
// As described in #11713, we use request instead of limit to deal with resource requirements.
// DefaultMilliCPURequest defines default milli cpu request number.
const DefaultMilliCPURequest int64 = 100 // 0.1 core
// DefaultMemoryRequest defines default memory request size.
const DefaultMemoryRequest int64 = 200 * 1024 * 1024 // 200 MB
// GetNonzeroRequests returns the default resource request if none is found or
// what is provided on the request.
func GetNonzeroRequests(requests *v1.ResourceList) (int64, int64) {
var outMilliCPU, outMemory int64
// Override if un-set, but not if explicitly set to zero
if _, found := (*requests)[v1.ResourceCPU]; !found {
outMilliCPU = DefaultMilliCPURequest
} else {
outMilliCPU = requests.Cpu().MilliValue()
}
// Override if un-set, but not if explicitly set to zero
if _, found := (*requests)[v1.ResourceMemory]; !found {
outMemory = DefaultMemoryRequest
} else {
outMemory = requests.Memory().Value()
}
return outMilliCPU, outMemory
}

View file

@ -0,0 +1,81 @@
/*
Copyright 2016 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 util
import (
"k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/apimachinery/pkg/util/sets"
)
// GetNamespacesFromPodAffinityTerm returns a set of names
// according to the namespaces indicated in podAffinityTerm.
// If namespaces is empty it considers the given pod's namespace.
func GetNamespacesFromPodAffinityTerm(pod *v1.Pod, podAffinityTerm *v1.PodAffinityTerm) sets.String {
names := sets.String{}
if len(podAffinityTerm.Namespaces) == 0 {
names.Insert(pod.Namespace)
} else {
names.Insert(podAffinityTerm.Namespaces...)
}
return names
}
// PodMatchesTermsNamespaceAndSelector returns true if the given <pod>
// matches the namespace and selector defined by <affinityPod>`s <term>.
func PodMatchesTermsNamespaceAndSelector(pod *v1.Pod, namespaces sets.String, selector labels.Selector) bool {
if !namespaces.Has(pod.Namespace) {
return false
}
if !selector.Matches(labels.Set(pod.Labels)) {
return false
}
return true
}
// NodesHaveSameTopologyKey checks if nodeA and nodeB have same label value with given topologyKey as label key.
// Returns false if topologyKey is empty.
func NodesHaveSameTopologyKey(nodeA, nodeB *v1.Node, topologyKey string) bool {
if len(topologyKey) == 0 {
return false
}
if nodeA.Labels == nil || nodeB.Labels == nil {
return false
}
nodeALabel, okA := nodeA.Labels[topologyKey]
nodeBLabel, okB := nodeB.Labels[topologyKey]
// If found label in both nodes, check the label
if okB && okA {
return nodeALabel == nodeBLabel
}
return false
}
// Topologies contains topologies information of nodes.
type Topologies struct {
DefaultKeys []string
}
// NodesHaveSameTopologyKey checks if nodeA and nodeB have same label value with given topologyKey as label key.
func (tps *Topologies) NodesHaveSameTopologyKey(nodeA, nodeB *v1.Node, topologyKey string) bool {
return NodesHaveSameTopologyKey(nodeA, nodeB, topologyKey)
}

View file

@ -0,0 +1,36 @@
/*
Copyright 2016 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 util
import (
"k8s.io/api/core/v1"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
)
// GetControllerRef gets pod's owner controller reference from a pod object.
func GetControllerRef(pod *v1.Pod) *metav1.OwnerReference {
if len(pod.OwnerReferences) == 0 {
return nil
}
for i := range pod.OwnerReferences {
ref := &pod.OwnerReferences[i]
if ref.Controller != nil && *ref.Controller {
return ref
}
}
return nil
}

View file

@ -0,0 +1,88 @@
/*
Copyright 2014 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 algorithm
import (
"k8s.io/api/core/v1"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// SchedulerExtender is an interface for external processes to influence scheduling
// decisions made by Kubernetes. This is typically needed for resources not directly
// managed by Kubernetes.
type SchedulerExtender interface {
// Filter based on extender-implemented predicate functions. The filtered list is
// expected to be a subset of the supplied list. failedNodesMap optionally contains
// the list of failed nodes and failure reasons.
Filter(pod *v1.Pod,
nodes []*v1.Node, nodeNameToInfo map[string]*schedulercache.NodeInfo,
) (filteredNodes []*v1.Node, failedNodesMap schedulerapi.FailedNodesMap, err error)
// Prioritize based on extender-implemented priority functions. The returned scores & weight
// are used to compute the weighted score for an extender. The weighted scores are added to
// the scores computed by Kubernetes scheduler. The total scores are used to do the host selection.
Prioritize(pod *v1.Pod, nodes []*v1.Node) (hostPriorities *schedulerapi.HostPriorityList, weight int, err error)
// Bind delegates the action of binding a pod to a node to the extender.
Bind(binding *v1.Binding) error
// IsBinder returns whether this extender is configured for the Bind method.
IsBinder() bool
// IsInterested returns true if at least one extended resource requested by
// this pod is managed by this extender.
IsInterested(pod *v1.Pod) bool
// ProcessPreemption returns nodes with their victim pods processed by extender based on
// given:
// 1. Pod to schedule
// 2. Candidate nodes and victim pods (nodeToVictims) generated by previous scheduling process.
// 3. nodeNameToInfo to restore v1.Node from node name if extender cache is enabled.
// The possible changes made by extender may include:
// 1. Subset of given candidate nodes after preemption phase of extender.
// 2. A different set of victim pod for every given candidate node after preemption phase of extender.
ProcessPreemption(
pod *v1.Pod,
nodeToVictims map[*v1.Node]*schedulerapi.Victims,
nodeNameToInfo map[string]*schedulercache.NodeInfo,
) (map[*v1.Node]*schedulerapi.Victims, error)
// SupportsPreemption returns if the scheduler extender support preemption or not.
SupportsPreemption() bool
// IsIgnorable returns true indicates scheduling should not fail when this extender
// is unavailable. This gives scheduler ability to fail fast and tolerate non-critical extenders as well.
IsIgnorable() bool
}
// ScheduleAlgorithm is an interface implemented by things that know how to schedule pods
// onto machines.
type ScheduleAlgorithm interface {
Schedule(*v1.Pod, NodeLister) (selectedMachine string, err error)
// Preempt receives scheduling errors for a pod and tries to create room for
// the pod by preempting lower priority pods if possible.
// It returns the node where preemption happened, a list of preempted pods, a
// list of pods whose nominated node name should be removed, and error if any.
Preempt(*v1.Pod, NodeLister, error) (selectedNode *v1.Node, preemptedPods []*v1.Pod, cleanupNominatedPods []*v1.Pod, err error)
// Predicates() returns a pointer to a map of predicate functions. This is
// exposed for testing.
Predicates() map[string]FitPredicate
// Prioritizers returns a slice of priority config. This is exposed for
// testing.
Prioritizers() []PriorityConfig
}

View file

@ -0,0 +1,172 @@
/*
Copyright 2014 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 algorithm
import (
apps "k8s.io/api/apps/v1beta1"
"k8s.io/api/core/v1"
extensions "k8s.io/api/extensions/v1beta1"
"k8s.io/apimachinery/pkg/labels"
schedulerapi "k8s.io/kubernetes/pkg/scheduler/api"
schedulercache "k8s.io/kubernetes/pkg/scheduler/cache"
)
// NodeFieldSelectorKeys is a map that: the key are node field selector keys; the values are
// the functions to get the value of the node field.
var NodeFieldSelectorKeys = map[string]func(*v1.Node) string{
NodeFieldSelectorKeyNodeName: func(n *v1.Node) string { return n.Name },
}
// FitPredicate is a function that indicates if a pod fits into an existing node.
// The failure information is given by the error.
type FitPredicate func(pod *v1.Pod, meta PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []PredicateFailureReason, error)
// PriorityMapFunction is a function that computes per-node results for a given node.
// TODO: Figure out the exact API of this method.
// TODO: Change interface{} to a specific type.
type PriorityMapFunction func(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (schedulerapi.HostPriority, error)
// PriorityReduceFunction is a function that aggregated per-node results and computes
// final scores for all nodes.
// TODO: Figure out the exact API of this method.
// TODO: Change interface{} to a specific type.
type PriorityReduceFunction func(pod *v1.Pod, meta interface{}, nodeNameToInfo map[string]*schedulercache.NodeInfo, result schedulerapi.HostPriorityList) error
// PredicateMetadataProducer is a function that computes predicate metadata for a given pod.
type PredicateMetadataProducer func(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo) PredicateMetadata
// PriorityMetadataProducer is a function that computes metadata for a given pod. This
// is now used for only for priority functions. For predicates please use PredicateMetadataProducer.
type PriorityMetadataProducer func(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo) interface{}
// PriorityFunction is a function that computes scores for all nodes.
// DEPRECATED
// Use Map-Reduce pattern for priority functions.
type PriorityFunction func(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo, nodes []*v1.Node) (schedulerapi.HostPriorityList, error)
// PriorityConfig is a config used for a priority function.
type PriorityConfig struct {
Name string
Map PriorityMapFunction
Reduce PriorityReduceFunction
// TODO: Remove it after migrating all functions to
// Map-Reduce pattern.
Function PriorityFunction
Weight int
}
// EmptyPredicateMetadataProducer returns a no-op MetadataProducer type.
func EmptyPredicateMetadataProducer(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo) PredicateMetadata {
return nil
}
// EmptyPriorityMetadataProducer returns a no-op PriorityMetadataProducer type.
func EmptyPriorityMetadataProducer(pod *v1.Pod, nodeNameToInfo map[string]*schedulercache.NodeInfo) interface{} {
return nil
}
// PredicateFailureReason interface represents the failure reason of a predicate.
type PredicateFailureReason interface {
GetReason() string
}
// NodeLister interface represents anything that can list nodes for a scheduler.
type NodeLister interface {
// We explicitly return []*v1.Node, instead of v1.NodeList, to avoid
// performing expensive copies that are unneeded.
List() ([]*v1.Node, error)
}
// PodLister interface represents anything that can list pods for a scheduler.
type PodLister interface {
// We explicitly return []*v1.Pod, instead of v1.PodList, to avoid
// performing expensive copies that are unneeded.
List(labels.Selector) ([]*v1.Pod, error)
// This is similar to "List()", but the returned slice does not
// contain pods that don't pass `podFilter`.
FilteredList(podFilter schedulercache.PodFilter, selector labels.Selector) ([]*v1.Pod, error)
}
// ServiceLister interface represents anything that can produce a list of services; the list is consumed by a scheduler.
type ServiceLister interface {
// Lists all the services
List(labels.Selector) ([]*v1.Service, error)
// Gets the services for the given pod
GetPodServices(*v1.Pod) ([]*v1.Service, error)
}
// ControllerLister interface represents anything that can produce a list of ReplicationController; the list is consumed by a scheduler.
type ControllerLister interface {
// Lists all the replication controllers
List(labels.Selector) ([]*v1.ReplicationController, error)
// Gets the services for the given pod
GetPodControllers(*v1.Pod) ([]*v1.ReplicationController, error)
}
// ReplicaSetLister interface represents anything that can produce a list of ReplicaSet; the list is consumed by a scheduler.
type ReplicaSetLister interface {
// Gets the replicasets for the given pod
GetPodReplicaSets(*v1.Pod) ([]*extensions.ReplicaSet, error)
}
var _ ControllerLister = &EmptyControllerLister{}
// EmptyControllerLister implements ControllerLister on []v1.ReplicationController returning empty data
type EmptyControllerLister struct{}
// List returns nil
func (f EmptyControllerLister) List(labels.Selector) ([]*v1.ReplicationController, error) {
return nil, nil
}
// GetPodControllers returns nil
func (f EmptyControllerLister) GetPodControllers(pod *v1.Pod) (controllers []*v1.ReplicationController, err error) {
return nil, nil
}
var _ ReplicaSetLister = &EmptyReplicaSetLister{}
// EmptyReplicaSetLister implements ReplicaSetLister on []extensions.ReplicaSet returning empty data
type EmptyReplicaSetLister struct{}
// GetPodReplicaSets returns nil
func (f EmptyReplicaSetLister) GetPodReplicaSets(pod *v1.Pod) (rss []*extensions.ReplicaSet, err error) {
return nil, nil
}
// StatefulSetLister interface represents anything that can produce a list of StatefulSet; the list is consumed by a scheduler.
type StatefulSetLister interface {
// Gets the StatefulSet for the given pod.
GetPodStatefulSets(*v1.Pod) ([]*apps.StatefulSet, error)
}
var _ StatefulSetLister = &EmptyStatefulSetLister{}
// EmptyStatefulSetLister implements StatefulSetLister on []apps.StatefulSet returning empty data.
type EmptyStatefulSetLister struct{}
// GetPodStatefulSets of EmptyStatefulSetLister returns nil.
func (f EmptyStatefulSetLister) GetPodStatefulSets(pod *v1.Pod) (sss []*apps.StatefulSet, err error) {
return nil, nil
}
// PredicateMetadata interface represents anything that can access a predicate metadata.
type PredicateMetadata interface {
ShallowCopy() PredicateMetadata
AddPod(addedPod *v1.Pod, nodeInfo *schedulercache.NodeInfo) error
RemovePod(deletedPod *v1.Pod) error
}

View file

@ -0,0 +1,85 @@
/*
Copyright 2015 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 algorithm
import (
api "k8s.io/kubernetes/pkg/apis/core"
)
const (
// TaintNodeNotReady will be added when node is not ready
// and feature-gate for TaintBasedEvictions flag is enabled,
// and removed when node becomes ready.
TaintNodeNotReady = "node.kubernetes.io/not-ready"
// DeprecatedTaintNodeNotReady is the deprecated version of TaintNodeNotReady.
// It is deprecated since 1.9
DeprecatedTaintNodeNotReady = "node.alpha.kubernetes.io/notReady"
// TaintNodeUnreachable will be added when node becomes unreachable
// (corresponding to NodeReady status ConditionUnknown)
// and feature-gate for TaintBasedEvictions flag is enabled,
// and removed when node becomes reachable (NodeReady status ConditionTrue).
TaintNodeUnreachable = "node.kubernetes.io/unreachable"
// DeprecatedTaintNodeUnreachable is the deprecated version of TaintNodeUnreachable.
// It is deprecated since 1.9
DeprecatedTaintNodeUnreachable = "node.alpha.kubernetes.io/unreachable"
// TaintNodeUnschedulable will be added when node becomes unschedulable
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when node becomes scheduable.
TaintNodeUnschedulable = "node.kubernetes.io/unschedulable"
// TaintNodeOutOfDisk will be added when node becomes out of disk
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when node has enough disk.
TaintNodeOutOfDisk = "node.kubernetes.io/out-of-disk"
// TaintNodeMemoryPressure will be added when node has memory pressure
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when node has enough memory.
TaintNodeMemoryPressure = "node.kubernetes.io/memory-pressure"
// TaintNodeDiskPressure will be added when node has disk pressure
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when node has enough disk.
TaintNodeDiskPressure = "node.kubernetes.io/disk-pressure"
// TaintNodeNetworkUnavailable will be added when node's network is unavailable
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when network becomes ready.
TaintNodeNetworkUnavailable = "node.kubernetes.io/network-unavailable"
// TaintNodePIDPressure will be added when node has pid pressure
// and feature-gate for TaintNodesByCondition flag is enabled,
// and removed when node has enough disk.
TaintNodePIDPressure = "node.kubernetes.io/pid-pressure"
// TaintExternalCloudProvider sets this taint on a node to mark it as unusable,
// when kubelet is started with the "external" cloud provider, until a controller
// from the cloud-controller-manager intitializes this node, and then removes
// the taint
TaintExternalCloudProvider = "node.cloudprovider.kubernetes.io/uninitialized"
// TaintNodeShutdown when node is shutdown in external cloud provider
TaintNodeShutdown = "node.cloudprovider.kubernetes.io/shutdown"
// NodeFieldSelectorKeyNodeName ('metadata.name') uses this as node field selector key
// when selecting node by node's name.
NodeFieldSelectorKeyNodeName = api.ObjectNameField
)