253df78f1b
As per mdame, we can't ensure that the cluster is actually balanced if other tests are adding or deleting pods in parallel.
250 lines
8.5 KiB
Go
250 lines
8.5 KiB
Go
/*
|
|
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 scheduling
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
"math"
|
|
"time"
|
|
|
|
"github.com/onsi/ginkgo"
|
|
"github.com/onsi/gomega"
|
|
v1 "k8s.io/api/core/v1"
|
|
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
|
|
"k8s.io/apimachinery/pkg/labels"
|
|
"k8s.io/apimachinery/pkg/util/intstr"
|
|
"k8s.io/apimachinery/pkg/util/uuid"
|
|
clientset "k8s.io/client-go/kubernetes"
|
|
"k8s.io/kubernetes/test/e2e/framework"
|
|
e2enode "k8s.io/kubernetes/test/e2e/framework/node"
|
|
e2epod "k8s.io/kubernetes/test/e2e/framework/pod"
|
|
e2erc "k8s.io/kubernetes/test/e2e/framework/rc"
|
|
e2eskipper "k8s.io/kubernetes/test/e2e/framework/skipper"
|
|
testutils "k8s.io/kubernetes/test/utils"
|
|
imageutils "k8s.io/kubernetes/test/utils/image"
|
|
)
|
|
|
|
var _ = SIGDescribe("Multi-AZ Clusters", func() {
|
|
f := framework.NewDefaultFramework("multi-az")
|
|
var zoneCount int
|
|
var err error
|
|
var cleanUp func()
|
|
ginkgo.BeforeEach(func() {
|
|
e2eskipper.SkipUnlessProviderIs("gce", "gke", "aws")
|
|
if zoneCount <= 0 {
|
|
zoneCount, err = getZoneCount(f.ClientSet)
|
|
framework.ExpectNoError(err)
|
|
}
|
|
ginkgo.By(fmt.Sprintf("Checking for multi-zone cluster. Zone count = %d", zoneCount))
|
|
msg := fmt.Sprintf("Zone count is %d, only run for multi-zone clusters, skipping test", zoneCount)
|
|
e2eskipper.SkipUnlessAtLeast(zoneCount, 2, msg)
|
|
// TODO: SkipUnlessDefaultScheduler() // Non-default schedulers might not spread
|
|
|
|
cs := f.ClientSet
|
|
e2enode.WaitForTotalHealthy(cs, time.Minute)
|
|
nodeList, err := e2enode.GetReadySchedulableNodes(cs)
|
|
framework.ExpectNoError(err)
|
|
|
|
// make the nodes have balanced cpu,mem usage
|
|
cleanUp, err = createBalancedPodForNodes(f, cs, f.Namespace.Name, nodeList.Items, podRequestedResource, 0.0)
|
|
framework.ExpectNoError(err)
|
|
})
|
|
ginkgo.AfterEach(func() {
|
|
if cleanUp != nil {
|
|
cleanUp()
|
|
}
|
|
})
|
|
ginkgo.It("should spread the pods of a service across zones [Serial]", func() {
|
|
SpreadServiceOrFail(f, 5*zoneCount, imageutils.GetPauseImageName())
|
|
})
|
|
|
|
ginkgo.It("should spread the pods of a replication controller across zones [Serial]", func() {
|
|
SpreadRCOrFail(f, int32(5*zoneCount), framework.ServeHostnameImage, []string{"serve-hostname"})
|
|
})
|
|
})
|
|
|
|
// SpreadServiceOrFail check that the pods comprising a service
|
|
// get spread evenly across available zones
|
|
func SpreadServiceOrFail(f *framework.Framework, replicaCount int, image string) {
|
|
// First create the service
|
|
serviceName := "test-service"
|
|
serviceSpec := &v1.Service{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Name: serviceName,
|
|
Namespace: f.Namespace.Name,
|
|
},
|
|
Spec: v1.ServiceSpec{
|
|
Selector: map[string]string{
|
|
"service": serviceName,
|
|
},
|
|
Ports: []v1.ServicePort{{
|
|
Port: 80,
|
|
TargetPort: intstr.FromInt(80),
|
|
}},
|
|
},
|
|
}
|
|
_, err := f.ClientSet.CoreV1().Services(f.Namespace.Name).Create(context.TODO(), serviceSpec, metav1.CreateOptions{})
|
|
framework.ExpectNoError(err)
|
|
|
|
// Now create some pods behind the service
|
|
podSpec := &v1.Pod{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Name: serviceName,
|
|
Labels: map[string]string{"service": serviceName},
|
|
},
|
|
Spec: v1.PodSpec{
|
|
Containers: []v1.Container{
|
|
{
|
|
Name: "test",
|
|
Image: image,
|
|
},
|
|
},
|
|
},
|
|
}
|
|
|
|
// Caution: StartPods requires at least one pod to replicate.
|
|
// Based on the callers, replicas is always positive number: zoneCount >= 0 implies (2*zoneCount)+1 > 0.
|
|
// Thus, no need to test for it. Once the precondition changes to zero number of replicas,
|
|
// test for replicaCount > 0. Otherwise, StartPods panics.
|
|
framework.ExpectNoError(testutils.StartPods(f.ClientSet, replicaCount, f.Namespace.Name, serviceName, *podSpec, false, framework.Logf))
|
|
|
|
// Wait for all of them to be scheduled
|
|
selector := labels.SelectorFromSet(labels.Set(map[string]string{"service": serviceName}))
|
|
pods, err := e2epod.WaitForPodsWithLabelScheduled(f.ClientSet, f.Namespace.Name, selector)
|
|
framework.ExpectNoError(err)
|
|
|
|
// Now make sure they're spread across zones
|
|
zoneNames, err := e2enode.GetClusterZones(f.ClientSet)
|
|
framework.ExpectNoError(err)
|
|
checkZoneSpreading(f.ClientSet, pods, zoneNames.List())
|
|
}
|
|
|
|
// Find the name of the zone in which a Node is running
|
|
func getZoneNameForNode(node v1.Node) (string, error) {
|
|
if z, ok := node.Labels[v1.LabelFailureDomainBetaZone]; ok {
|
|
return z, nil
|
|
} else if z, ok := node.Labels[v1.LabelTopologyZone]; ok {
|
|
return z, nil
|
|
}
|
|
return "", fmt.Errorf("node %s doesn't have zone label %s or %s",
|
|
node.Name, v1.LabelFailureDomainBetaZone, v1.LabelTopologyZone)
|
|
}
|
|
|
|
// Return the number of zones in which we have nodes in this cluster.
|
|
func getZoneCount(c clientset.Interface) (int, error) {
|
|
zoneNames, err := e2enode.GetClusterZones(c)
|
|
if err != nil {
|
|
return -1, err
|
|
}
|
|
return len(zoneNames), nil
|
|
}
|
|
|
|
// Find the name of the zone in which the pod is scheduled
|
|
func getZoneNameForPod(c clientset.Interface, pod v1.Pod) (string, error) {
|
|
ginkgo.By(fmt.Sprintf("Getting zone name for pod %s, on node %s", pod.Name, pod.Spec.NodeName))
|
|
node, err := c.CoreV1().Nodes().Get(context.TODO(), pod.Spec.NodeName, metav1.GetOptions{})
|
|
framework.ExpectNoError(err)
|
|
return getZoneNameForNode(*node)
|
|
}
|
|
|
|
// Determine whether a set of pods are approximately evenly spread
|
|
// across a given set of zones
|
|
func checkZoneSpreading(c clientset.Interface, pods *v1.PodList, zoneNames []string) {
|
|
podsPerZone := make(map[string]int)
|
|
for _, zoneName := range zoneNames {
|
|
podsPerZone[zoneName] = 0
|
|
}
|
|
for _, pod := range pods.Items {
|
|
if pod.DeletionTimestamp != nil {
|
|
continue
|
|
}
|
|
zoneName, err := getZoneNameForPod(c, pod)
|
|
framework.ExpectNoError(err)
|
|
podsPerZone[zoneName] = podsPerZone[zoneName] + 1
|
|
}
|
|
minPodsPerZone := math.MaxInt32
|
|
maxPodsPerZone := 0
|
|
for _, podCount := range podsPerZone {
|
|
if podCount < minPodsPerZone {
|
|
minPodsPerZone = podCount
|
|
}
|
|
if podCount > maxPodsPerZone {
|
|
maxPodsPerZone = podCount
|
|
}
|
|
}
|
|
gomega.Expect(maxPodsPerZone-minPodsPerZone).To(gomega.BeNumerically("~", 0, 2),
|
|
"Pods were not evenly spread across zones. %d in one zone and %d in another zone",
|
|
minPodsPerZone, maxPodsPerZone)
|
|
}
|
|
|
|
// SpreadRCOrFail Check that the pods comprising a replication
|
|
// controller get spread evenly across available zones
|
|
func SpreadRCOrFail(f *framework.Framework, replicaCount int32, image string, args []string) {
|
|
name := "ubelite-spread-rc-" + string(uuid.NewUUID())
|
|
ginkgo.By(fmt.Sprintf("Creating replication controller %s", name))
|
|
controller, err := f.ClientSet.CoreV1().ReplicationControllers(f.Namespace.Name).Create(context.TODO(), &v1.ReplicationController{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Namespace: f.Namespace.Name,
|
|
Name: name,
|
|
},
|
|
Spec: v1.ReplicationControllerSpec{
|
|
Replicas: &replicaCount,
|
|
Selector: map[string]string{
|
|
"name": name,
|
|
},
|
|
Template: &v1.PodTemplateSpec{
|
|
ObjectMeta: metav1.ObjectMeta{
|
|
Labels: map[string]string{"name": name},
|
|
},
|
|
Spec: v1.PodSpec{
|
|
Containers: []v1.Container{
|
|
{
|
|
Name: name,
|
|
Image: image,
|
|
Args: args,
|
|
Ports: []v1.ContainerPort{{ContainerPort: 9376}},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
}, metav1.CreateOptions{})
|
|
framework.ExpectNoError(err)
|
|
// Cleanup the replication controller when we are done.
|
|
defer func() {
|
|
// Resize the replication controller to zero to get rid of pods.
|
|
if err := e2erc.DeleteRCAndWaitForGC(f.ClientSet, f.Namespace.Name, controller.Name); err != nil {
|
|
framework.Logf("Failed to cleanup replication controller %v: %v.", controller.Name, err)
|
|
}
|
|
}()
|
|
// List the pods, making sure we observe all the replicas.
|
|
selector := labels.SelectorFromSet(labels.Set(map[string]string{"name": name}))
|
|
_, err = e2epod.PodsCreated(f.ClientSet, f.Namespace.Name, name, replicaCount)
|
|
framework.ExpectNoError(err)
|
|
|
|
// Wait for all of them to be scheduled
|
|
ginkgo.By(fmt.Sprintf("Waiting for %d replicas of %s to be scheduled. Selector: %v", replicaCount, name, selector))
|
|
pods, err := e2epod.WaitForPodsWithLabelScheduled(f.ClientSet, f.Namespace.Name, selector)
|
|
framework.ExpectNoError(err)
|
|
|
|
// Now make sure they're spread across zones
|
|
zoneNames, err := e2enode.GetClusterZones(f.ClientSet)
|
|
framework.ExpectNoError(err)
|
|
checkZoneSpreading(f.ClientSet, pods, zoneNames.List())
|
|
}
|