github/kubernetes
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #50949 from bsalamat/preemption_eviction

Browse Source 
Automatic merge from submit-queue

Add pod preemption to the scheduler

**What this PR does / why we need it**:
This is the last of a series of PRs to add priority-based preemption to the scheduler. This PR connects the preemption logic to the scheduler workflow.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #48646

**Special notes for your reviewer**:
This PR includes other PRs which are under review (#50805, #50405, #50190). All the new code is located in 43627afdf9.

**Release note**:

```release-note
Add priority-based preemption to the scheduler.
```

ref/ #47604

/assign @davidopp 

@kubernetes/sig-scheduling-pr-reviews
This commit is contained in:
Kubernetes Submit Queue
2017-09-08 14:19:42 -07:00
committed by GitHub
parent ed154988c5 c0b718373b
commit f695a3120a
34 changed files with 1900 additions and 91 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
plugin/pkg/scheduler/algorithm/predicates/predicates.go
View File

@@ -45,6 +45,10 @@ import (
"github.com/golang/glog"
)
const (
MatchInterPodAffinity = "MatchInterPodAffinity"
)
// NodeInfo: Other types for predicate functions...
type NodeInfo interface {
GetNodeInfo(nodeID string) (*v1.Node, error)
@@ -152,7 +156,7 @@ func isVolumeConflict(volume v1.Volume, pod *v1.Pod) bool {
// - Ceph RBD forbids if any two pods share at least same monitor, and match pool and image.
// - ISCSI forbids if any two pods share at least same IQN, LUN and Target
// TODO: migrate this into some per-volume specific code?
func NoDiskConflict(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func NoDiskConflict(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
for _, v := range pod.Spec.Volumes {
for _, ev := range nodeInfo.Pods() {
if isVolumeConflict(v, ev) {
@@ -250,7 +254,7 @@ func (c *MaxPDVolumeCountChecker) filterVolumes(volumes []v1.Volume, namespace s
return nil
}
func (c *MaxPDVolumeCountChecker) predicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (c *MaxPDVolumeCountChecker) predicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
// If a pod doesn't have any volume attached to it, the predicate will always be true.
// Thus we make a fast path for it, to avoid unnecessary computations in this case.
if len(pod.Spec.Volumes) == 0 {
@@ -371,7 +375,7 @@ func NewVolumeZonePredicate(pvInfo PersistentVolumeInfo, pvcInfo PersistentVolum
return c.predicate
}
func (c *VolumeZoneChecker) predicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (c *VolumeZoneChecker) predicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
// If a pod doesn't have any volume attached to it, the predicate will always be true.
// Thus we make a fast path for it, to avoid unnecessary computations in this case.
if len(pod.Spec.Volumes) == 0 {
@@ -529,7 +533,7 @@ func podName(pod *v1.Pod) string {
// PodFitsResources checks if a node has sufficient resources, such as cpu, memory, gpu, opaque int resources etc to run a pod.
// First return value indicates whether a node has sufficient resources to run a pod while the second return value indicates the
// predicate failure reasons if the node has insufficient resources to run the pod.
func PodFitsResources(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodFitsResources(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
node := nodeInfo.Node()
if node == nil {
return false, nil, fmt.Errorf("node not found")
@@ -658,7 +662,7 @@ func podMatchesNodeLabels(pod *v1.Pod, node *v1.Node) bool {
}
// PodMatchNodeSelector checks if a pod node selector matches the node label.
func PodMatchNodeSelector(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodMatchNodeSelector(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
node := nodeInfo.Node()
if node == nil {
return false, nil, fmt.Errorf("node not found")
@@ -670,7 +674,7 @@ func PodMatchNodeSelector(pod *v1.Pod, meta interface{}, nodeInfo *schedulercach
}
// PodFitsHost checks if a pod spec node name matches the current node.
func PodFitsHost(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodFitsHost(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
if len(pod.Spec.NodeName) == 0 {
return true, nil, nil
}
@@ -709,7 +713,7 @@ func NewNodeLabelPredicate(labels []string, presence bool) algorithm.FitPredicat
// Alternately, eliminating nodes that have a certain label, regardless of value, is also useful
// A node may have a label with "retiring" as key and the date as the value
// and it may be desirable to avoid scheduling new pods on this node
func (n *NodeLabelChecker) CheckNodeLabelPresence(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (n *NodeLabelChecker) CheckNodeLabelPresence(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
node := nodeInfo.Node()
if node == nil {
return false, nil, fmt.Errorf("node not found")
@@ -792,7 +796,7 @@ func NewServiceAffinityPredicate(podLister algorithm.PodLister, serviceLister al
//
// WARNING: This Predicate is NOT guaranteed to work if some of the predicateMetadata data isn't precomputed...
// For that reason it is not exported, i.e. it is highly coupled to the implementation of the FitPredicate construction.
func (s *ServiceAffinity) checkServiceAffinity(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (s *ServiceAffinity) checkServiceAffinity(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var services []*v1.Service
var pods []*v1.Pod
if pm, ok := meta.(*predicateMetadata); ok && (pm.serviceAffinityMatchingPodList != nil || pm.serviceAffinityMatchingPodServices != nil) {
@@ -804,6 +808,7 @@ func (s *ServiceAffinity) checkServiceAffinity(pod *v1.Pod, meta interface{}, no
s.serviceAffinityMetadataProducer(pm)
pods, services = pm.serviceAffinityMatchingPodList, pm.serviceAffinityMatchingPodServices
}
filteredPods := nodeInfo.FilterOutPods(pods)
node := nodeInfo.Node()
if node == nil {
return false, nil, fmt.Errorf("node not found")
@@ -813,8 +818,8 @@ func (s *ServiceAffinity) checkServiceAffinity(pod *v1.Pod, meta interface{}, no
// Step 1: If we don't have all constraints, introspect nodes to find the missing constraints.
if len(s.labels) > len(affinityLabels) {
if len(services) > 0 {
if len(pods) > 0 {
nodeWithAffinityLabels, err := s.nodeInfo.GetNodeInfo(pods[0].Spec.NodeName)
if len(filteredPods) > 0 {
nodeWithAffinityLabels, err := s.nodeInfo.GetNodeInfo(filteredPods[0].Spec.NodeName)
if err != nil {
return false, nil, err
}
@@ -830,7 +835,7 @@ func (s *ServiceAffinity) checkServiceAffinity(pod *v1.Pod, meta interface{}, no
}
// PodFitsHostPorts checks if a node has free ports for the requested pod ports.
func PodFitsHostPorts(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodFitsHostPorts(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var wantPorts map[int]bool
if predicateMeta, ok := meta.(*predicateMetadata); ok {
wantPorts = predicateMeta.podPorts
@@ -871,7 +876,7 @@ func haveSame(a1, a2 []string) bool {
// GeneralPredicates checks whether noncriticalPredicates and EssentialPredicates pass. noncriticalPredicates are the predicates
// that only non-critical pods need and EssentialPredicates are the predicates that all pods, including critical pods, need
func GeneralPredicates(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func GeneralPredicates(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var predicateFails []algorithm.PredicateFailureReason
fit, reasons, err := noncriticalPredicates(pod, meta, nodeInfo)
if err != nil {
@@ -893,7 +898,7 @@ func GeneralPredicates(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.N
}
// noncriticalPredicates are the predicates that only non-critical pods need
func noncriticalPredicates(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func noncriticalPredicates(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var predicateFails []algorithm.PredicateFailureReason
fit, reasons, err := PodFitsResources(pod, meta, nodeInfo)
if err != nil {
@@ -907,7 +912,7 @@ func noncriticalPredicates(pod *v1.Pod, meta interface{}, nodeInfo *schedulercac
}
// EssentialPredicates are the predicates that all pods, including critical pods, need
func EssentialPredicates(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func EssentialPredicates(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var predicateFails []algorithm.PredicateFailureReason
fit, reasons, err := PodFitsHost(pod, meta, nodeInfo)
if err != nil {
@@ -953,7 +958,7 @@ func NewPodAffinityPredicate(info NodeInfo, podLister algorithm.PodLister) algor
// InterPodAffinityMatches checks if a pod can be scheduled on the specified node with pod affinity/anti-affinity configuration.
// First return value indicates whether a pod can be scheduled on the specified node while the second return value indicates the
// predicate failure reasons if the pod cannot be scheduled on the specified node.
func (c *PodAffinityChecker) InterPodAffinityMatches(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (c *PodAffinityChecker) InterPodAffinityMatches(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
node := nodeInfo.Node()
if node == nil {
return false, nil, fmt.Errorf("node not found")
@@ -1138,7 +1143,7 @@ func (c *PodAffinityChecker) getMatchingAntiAffinityTerms(pod *v1.Pod, allPods [
// Checks if scheduling the pod onto this node would break any anti-affinity
// rules indicated by the existing pods.
func (c *PodAffinityChecker) satisfiesExistingPodsAntiAffinity(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) bool {
func (c *PodAffinityChecker) satisfiesExistingPodsAntiAffinity(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) bool {
node := nodeInfo.Node()
if node == nil {
return false
@@ -1246,7 +1251,7 @@ func (c *PodAffinityChecker) satisfiesPodsAffinityAntiAffinity(pod *v1.Pod, node
}
// PodToleratesNodeTaints checks if a pod tolerations can tolerate the node taints
func PodToleratesNodeTaints(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodToleratesNodeTaints(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
return podToleratesNodeTaints(pod, nodeInfo, func(t *v1.Taint) bool {
// PodToleratesNodeTaints is only interested in NoSchedule and NoExecute taints.
return t.Effect == v1.TaintEffectNoSchedule || t.Effect == v1.TaintEffectNoExecute
@@ -1254,7 +1259,7 @@ func PodToleratesNodeTaints(pod *v1.Pod, meta interface{}, nodeInfo *schedulerca
}
// PodToleratesNodeNoExecuteTaints checks if a pod tolerations can tolerate the node's NoExecute taints
func PodToleratesNodeNoExecuteTaints(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func PodToleratesNodeNoExecuteTaints(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
return podToleratesNodeTaints(pod, nodeInfo, func(t *v1.Taint) bool {
return t.Effect == v1.TaintEffectNoExecute
})
@@ -1279,7 +1284,7 @@ func isPodBestEffort(pod *v1.Pod) bool {
// CheckNodeMemoryPressurePredicate checks if a pod can be scheduled on a node
// reporting memory pressure condition.
func CheckNodeMemoryPressurePredicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func CheckNodeMemoryPressurePredicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
var podBestEffort bool
if predicateMeta, ok := meta.(*predicateMetadata); ok {
podBestEffort = predicateMeta.podBestEffort
@@ -1301,7 +1306,7 @@ func CheckNodeMemoryPressurePredicate(pod *v1.Pod, meta interface{}, nodeInfo *s
// CheckNodeDiskPressurePredicate checks if a pod can be scheduled on a node
// reporting disk pressure condition.
func CheckNodeDiskPressurePredicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func CheckNodeDiskPressurePredicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
// check if node is under disk pressure
if nodeInfo.DiskPressureCondition() == v1.ConditionTrue {
return false, []algorithm.PredicateFailureReason{ErrNodeUnderDiskPressure}, nil
@@ -1311,7 +1316,7 @@ func CheckNodeDiskPressurePredicate(pod *v1.Pod, meta interface{}, nodeInfo *sch
// CheckNodeConditionPredicate checks if a pod can be scheduled on a node reporting out of disk,
// network unavailable and not ready condition. Only node conditions are accounted in this predicate.
func CheckNodeConditionPredicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func CheckNodeConditionPredicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
reasons := []algorithm.PredicateFailureReason{}
if nodeInfo == nil || nodeInfo.Node() == nil {
@@ -1359,7 +1364,7 @@ func NewVolumeNodePredicate(pvInfo PersistentVolumeInfo, pvcInfo PersistentVolum
return c.predicate
}
func (c *VolumeNodeChecker) predicate(pod *v1.Pod, meta interface{}, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
func (c *VolumeNodeChecker) predicate(pod *v1.Pod, meta algorithm.PredicateMetadata, nodeInfo *schedulercache.NodeInfo) (bool, []algorithm.PredicateFailureReason, error) {
if !utilfeature.DefaultFeatureGate.Enabled(features.PersistentLocalVolumes) {
return true, nil, nil
}