Merge pull request #99163 from ahg-g/ahg-pod-deletion

Implements pod deletion cost
This commit is contained in:
Kubernetes Prow Robot
2021-03-01 18:19:21 -08:00
committed by GitHub
15 changed files with 292 additions and 86 deletions

View File

@@ -39,13 +39,16 @@ import (
"k8s.io/apimachinery/pkg/util/sets"
"k8s.io/apimachinery/pkg/util/strategicpatch"
"k8s.io/apimachinery/pkg/util/wait"
utilfeature "k8s.io/apiserver/pkg/util/feature"
clientset "k8s.io/client-go/kubernetes"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/tools/record"
clientretry "k8s.io/client-go/util/retry"
podutil "k8s.io/kubernetes/pkg/api/v1/pod"
"k8s.io/kubernetes/pkg/apis/core/helper"
_ "k8s.io/kubernetes/pkg/apis/core/install"
"k8s.io/kubernetes/pkg/apis/core/validation"
"k8s.io/kubernetes/pkg/features"
hashutil "k8s.io/kubernetes/pkg/util/hash"
taintutils "k8s.io/kubernetes/pkg/util/taints"
"k8s.io/utils/integer"
@@ -791,15 +794,17 @@ func (s ActivePods) Less(i, j int) bool {
// is unknown, and a pod whose phase is unknown comes before a running pod.
// 3. If exactly one of the pods is ready, the pod that is not ready comes
// before the ready pod.
// 4. If the pods' ranks differ, the pod with greater rank comes before the pod
// 4. If controller.kubernetes.io/pod-deletion-cost annotation is set, then
// the pod with the lower value will come first.
// 5. If the pods' ranks differ, the pod with greater rank comes before the pod
// with lower rank.
// 5. If both pods are ready but have not been ready for the same amount of
// 6. If both pods are ready but have not been ready for the same amount of
// time, the pod that has been ready for a shorter amount of time comes
// before the pod that has been ready for longer.
// 6. If one pod has a container that has restarted more than any container in
// 7. If one pod has a container that has restarted more than any container in
// the other pod, the pod with the container with more restarts comes
// before the other pod.
// 7. If the pods' creation times differ, the pod that was created more recently
// 8. If the pods' creation times differ, the pod that was created more recently
// comes before the older pod.
//
// If none of these rules matches, the second pod comes before the first pod.
@@ -842,7 +847,17 @@ func (s ActivePodsWithRanks) Less(i, j int) bool {
if podutil.IsPodReady(s.Pods[i]) != podutil.IsPodReady(s.Pods[j]) {
return !podutil.IsPodReady(s.Pods[i])
}
// 4. Doubled up < not doubled up
// 4. higher pod-deletion-cost < lower pod-deletion cost
if utilfeature.DefaultFeatureGate.Enabled(features.PodDeletionCost) {
pi, _ := helper.GetDeletionCostFromPodAnnotations(s.Pods[i].Annotations)
pj, _ := helper.GetDeletionCostFromPodAnnotations(s.Pods[j].Annotations)
if pi != pj {
return pi < pj
}
}
// 5. Doubled up < not doubled up
// If one of the two pods is on the same node as one or more additional
// ready pods that belong to the same replicaset, whichever pod has more
// colocated ready pods is less
@@ -851,7 +866,7 @@ func (s ActivePodsWithRanks) Less(i, j int) bool {
}
// TODO: take availability into account when we push minReadySeconds information from deployment into pods,
// see https://github.com/kubernetes/kubernetes/issues/22065
// 5. Been ready for empty time < less time < more time
// 6. Been ready for empty time < less time < more time
// If both pods are ready, the latest ready one is smaller
if podutil.IsPodReady(s.Pods[i]) && podutil.IsPodReady(s.Pods[j]) {
readyTime1 := podReadyTime(s.Pods[i])
@@ -860,11 +875,11 @@ func (s ActivePodsWithRanks) Less(i, j int) bool {
return afterOrZero(readyTime1, readyTime2)
}
}
// 6. Pods with containers with higher restart counts < lower restart counts
// 7. Pods with containers with higher restart counts < lower restart counts
if maxContainerRestarts(s.Pods[i]) != maxContainerRestarts(s.Pods[j]) {
return maxContainerRestarts(s.Pods[i]) > maxContainerRestarts(s.Pods[j])
}
// 7. Empty creation time pods < newer pods < older pods
// 8. Empty creation time pods < newer pods < older pods
if !s.Pods[i].CreationTimestamp.Equal(&s.Pods[j].CreationTimestamp) {
return afterOrZero(&s.Pods[i].CreationTimestamp, &s.Pods[j].CreationTimestamp)
}

View File

@@ -38,6 +38,7 @@ import (
"k8s.io/apimachinery/pkg/util/clock"
"k8s.io/apimachinery/pkg/util/sets"
"k8s.io/apimachinery/pkg/util/uuid"
utilfeature "k8s.io/apiserver/pkg/util/feature"
clientset "k8s.io/client-go/kubernetes"
"k8s.io/client-go/kubernetes/fake"
clientscheme "k8s.io/client-go/kubernetes/scheme"
@@ -45,8 +46,11 @@ import (
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/tools/record"
utiltesting "k8s.io/client-go/util/testing"
featuregatetesting "k8s.io/component-base/featuregate/testing"
"k8s.io/kubernetes/pkg/apis/core"
_ "k8s.io/kubernetes/pkg/apis/core/install"
"k8s.io/kubernetes/pkg/controller/testutil"
"k8s.io/kubernetes/pkg/features"
"k8s.io/kubernetes/pkg/securitycontext"
"github.com/stretchr/testify/assert"
@@ -435,7 +439,7 @@ func TestSortingActivePodsWithRanks(t *testing.T) {
now := metav1.Now()
then := metav1.Time{Time: now.AddDate(0, -1, 0)}
zeroTime := metav1.Time{}
pod := func(podName, nodeName string, phase v1.PodPhase, ready bool, restarts int32, readySince metav1.Time, created metav1.Time) *v1.Pod {
pod := func(podName, nodeName string, phase v1.PodPhase, ready bool, restarts int32, readySince metav1.Time, created metav1.Time, annotations map[string]string) *v1.Pod {
var conditions []v1.PodCondition
var containerStatuses []v1.ContainerStatus
if ready {
@@ -446,6 +450,7 @@ func TestSortingActivePodsWithRanks(t *testing.T) {
ObjectMeta: metav1.ObjectMeta{
CreationTimestamp: created,
Name: podName,
Annotations: annotations,
},
Spec: v1.PodSpec{NodeName: nodeName},
Status: v1.PodStatus{
@@ -456,15 +461,17 @@ func TestSortingActivePodsWithRanks(t *testing.T) {
}
}
var (
unscheduledPod = pod("unscheduled", "", v1.PodPending, false, 0, zeroTime, zeroTime)
scheduledPendingPod = pod("pending", "node", v1.PodPending, false, 0, zeroTime, zeroTime)
unknownPhasePod = pod("unknown-phase", "node", v1.PodUnknown, false, 0, zeroTime, zeroTime)
runningNotReadyPod = pod("not-ready", "node", v1.PodRunning, false, 0, zeroTime, zeroTime)
runningReadyNoLastTransitionTimePod = pod("ready-no-last-transition-time", "node", v1.PodRunning, true, 0, zeroTime, zeroTime)
runningReadyNow = pod("ready-now", "node", v1.PodRunning, true, 0, now, now)
runningReadyThen = pod("ready-then", "node", v1.PodRunning, true, 0, then, then)
runningReadyNowHighRestarts = pod("ready-high-restarts", "node", v1.PodRunning, true, 9001, now, now)
runningReadyNowCreatedThen = pod("ready-now-created-then", "node", v1.PodRunning, true, 0, now, then)
unscheduledPod = pod("unscheduled", "", v1.PodPending, false, 0, zeroTime, zeroTime, nil)
scheduledPendingPod = pod("pending", "node", v1.PodPending, false, 0, zeroTime, zeroTime, nil)
unknownPhasePod = pod("unknown-phase", "node", v1.PodUnknown, false, 0, zeroTime, zeroTime, nil)
runningNotReadyPod = pod("not-ready", "node", v1.PodRunning, false, 0, zeroTime, zeroTime, nil)
runningReadyNoLastTransitionTimePod = pod("ready-no-last-transition-time", "node", v1.PodRunning, true, 0, zeroTime, zeroTime, nil)
runningReadyNow = pod("ready-now", "node", v1.PodRunning, true, 0, now, now, nil)
runningReadyThen = pod("ready-then", "node", v1.PodRunning, true, 0, then, then, nil)
runningReadyNowHighRestarts = pod("ready-high-restarts", "node", v1.PodRunning, true, 9001, now, now, nil)
runningReadyNowCreatedThen = pod("ready-now-created-then", "node", v1.PodRunning, true, 0, now, then, nil)
lowPodDeletionCost = pod("low-deletion-cost", "node", v1.PodRunning, true, 0, now, then, map[string]string{core.PodDeletionCost: "10"})
highPodDeletionCost = pod("high-deletion-cost", "node", v1.PodRunning, true, 0, now, then, map[string]string{core.PodDeletionCost: "100"})
)
equalityTests := []*v1.Pod{
unscheduledPod,
@@ -491,33 +498,40 @@ func TestSortingActivePodsWithRanks(t *testing.T) {
rank int
}
inequalityTests := []struct {
lesser, greater podWithRank
lesser, greater podWithRank
disablePodDeletioncost bool
}{
{podWithRank{unscheduledPod, 1}, podWithRank{scheduledPendingPod, 2}},
{podWithRank{unscheduledPod, 2}, podWithRank{scheduledPendingPod, 1}},
{podWithRank{scheduledPendingPod, 1}, podWithRank{unknownPhasePod, 2}},
{podWithRank{unknownPhasePod, 1}, podWithRank{runningNotReadyPod, 2}},
{podWithRank{runningNotReadyPod, 1}, podWithRank{runningReadyNoLastTransitionTimePod, 1}},
{podWithRank{runningReadyNoLastTransitionTimePod, 1}, podWithRank{runningReadyNow, 1}},
{podWithRank{runningReadyNow, 2}, podWithRank{runningReadyNoLastTransitionTimePod, 1}},
{podWithRank{runningReadyNow, 1}, podWithRank{runningReadyThen, 1}},
{podWithRank{runningReadyNow, 2}, podWithRank{runningReadyThen, 1}},
{podWithRank{runningReadyNowHighRestarts, 1}, podWithRank{runningReadyNow, 1}},
{podWithRank{runningReadyNow, 2}, podWithRank{runningReadyNowHighRestarts, 1}},
{podWithRank{runningReadyNow, 1}, podWithRank{runningReadyNowCreatedThen, 1}},
{podWithRank{runningReadyNowCreatedThen, 2}, podWithRank{runningReadyNow, 1}},
{lesser: podWithRank{unscheduledPod, 1}, greater: podWithRank{scheduledPendingPod, 2}},
{lesser: podWithRank{unscheduledPod, 2}, greater: podWithRank{scheduledPendingPod, 1}},
{lesser: podWithRank{scheduledPendingPod, 1}, greater: podWithRank{unknownPhasePod, 2}},
{lesser: podWithRank{unknownPhasePod, 1}, greater: podWithRank{runningNotReadyPod, 2}},
{lesser: podWithRank{runningNotReadyPod, 1}, greater: podWithRank{runningReadyNoLastTransitionTimePod, 1}},
{lesser: podWithRank{runningReadyNoLastTransitionTimePod, 1}, greater: podWithRank{runningReadyNow, 1}},
{lesser: podWithRank{runningReadyNow, 2}, greater: podWithRank{runningReadyNoLastTransitionTimePod, 1}},
{lesser: podWithRank{runningReadyNow, 1}, greater: podWithRank{runningReadyThen, 1}},
{lesser: podWithRank{runningReadyNow, 2}, greater: podWithRank{runningReadyThen, 1}},
{lesser: podWithRank{runningReadyNowHighRestarts, 1}, greater: podWithRank{runningReadyNow, 1}},
{lesser: podWithRank{runningReadyNow, 2}, greater: podWithRank{runningReadyNowHighRestarts, 1}},
{lesser: podWithRank{runningReadyNow, 1}, greater: podWithRank{runningReadyNowCreatedThen, 1}},
{lesser: podWithRank{runningReadyNowCreatedThen, 2}, greater: podWithRank{runningReadyNow, 1}},
{lesser: podWithRank{lowPodDeletionCost, 2}, greater: podWithRank{highPodDeletionCost, 1}},
{lesser: podWithRank{highPodDeletionCost, 2}, greater: podWithRank{lowPodDeletionCost, 1}, disablePodDeletioncost: true},
}
for _, test := range inequalityTests {
podsWithRanks := ActivePodsWithRanks{
Pods: []*v1.Pod{test.lesser.pod, test.greater.pod},
Rank: []int{test.lesser.rank, test.greater.rank},
}
if !podsWithRanks.Less(0, 1) {
t.Errorf("expected pod %q with rank %v to be less than %q with rank %v", podsWithRanks.Pods[0].Name, podsWithRanks.Rank[0], podsWithRanks.Pods[1].Name, podsWithRanks.Rank[1])
}
if podsWithRanks.Less(1, 0) {
t.Errorf("expected pod %q with rank %v not to be less than %v with rank %v", podsWithRanks.Pods[1].Name, podsWithRanks.Rank[1], podsWithRanks.Pods[0].Name, podsWithRanks.Rank[0])
}
for i, test := range inequalityTests {
t.Run(fmt.Sprintf("test%d", i), func(t *testing.T) {
defer featuregatetesting.SetFeatureGateDuringTest(t, utilfeature.DefaultFeatureGate, features.PodDeletionCost, !test.disablePodDeletioncost)()
podsWithRanks := ActivePodsWithRanks{
Pods: []*v1.Pod{test.lesser.pod, test.greater.pod},
Rank: []int{test.lesser.rank, test.greater.rank},
}
if !podsWithRanks.Less(0, 1) {
t.Errorf("expected pod %q with rank %v to be less than %q with rank %v", podsWithRanks.Pods[0].Name, podsWithRanks.Rank[0], podsWithRanks.Pods[1].Name, podsWithRanks.Rank[1])
}
if podsWithRanks.Less(1, 0) {
t.Errorf("expected pod %q with rank %v not to be less than %v with rank %v", podsWithRanks.Pods[1].Name, podsWithRanks.Rank[1], podsWithRanks.Pods[0].Name, podsWithRanks.Rank[0])
}
})
}
}