github
/
k3s
mirror of https://github.com/k3s-io/k3s
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge pull request #64919 from losipiuk/lo/e2e-tests-nap 

Automatic merge from submit-queue (batch tested with PRs 65256, 64236, 64919, 64879, 57932). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

 Add cluster autoscaler w/NAP test involving GPUs

This is very trivial e2e tests verifying interoperability of Cluster Autoscaler with Node Auto Provisioning and GPUs.
More specific tests to come.

```release-note
NONE
```
pull/8/head
Kubernetes Submit Queue 2018-06-20 17:22:15 -07:00 committed by GitHub
parent 14268aef75 e16b1fdc41
commit 03413cc126
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 26 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
test/e2e/autoscaling/cluster_size_autoscaling.go
View File

@ -227,6 +227,7 @@ var _ = SIGDescribe("Cluster size autoscaling [Slow]", func() {
By("Schedule a pod which requires GPU")
framework.ExpectNoError(scheduleGpuPod(f, "gpu-pod-rc"))
defer framework.DeleteRCAndWaitForGC(f.ClientSet, f.Namespace.Name, "gpu-pod-rc")
framework.ExpectNoError(WaitForClusterSizeFunc(f.ClientSet,
func(size int) bool { return size == nodeCount+1 }, scaleUpTimeout))
@ -244,6 +245,7 @@ var _ = SIGDescribe("Cluster size autoscaling [Slow]", func() {
By("Schedule a single pod which requires GPU")
framework.ExpectNoError(scheduleGpuPod(f, "gpu-pod-rc"))
defer framework.DeleteRCAndWaitForGC(f.ClientSet, f.Namespace.Name, "gpu-pod-rc")
By("Enable autoscaler")
framework.ExpectNoError(enableAutoscaler(gpuPoolName, 0, 2))
@ -294,6 +296,7 @@ var _ = SIGDescribe("Cluster size autoscaling [Slow]", func() {
By("Schedule a single pod which requires GPU")
framework.ExpectNoError(scheduleGpuPod(f, "gpu-pod-rc"))
defer framework.DeleteRCAndWaitForGC(f.ClientSet, f.Namespace.Name, "gpu-pod-rc")
By("Enable autoscaler")
framework.ExpectNoError(enableAutoscaler(gpuPoolName, 0, 1))
@ -309,6 +312,29 @@ var _ = SIGDescribe("Cluster size autoscaling [Slow]", func() {
})
}
// TODO consider moving to [Feature:ClusterSizeAutoscalingGpu] as soon as NAP goes out of beta. Currently
// project needed to run the NAP tests require whitelisting for NAP alpha
It("NAP should add a pool with GPUs if unschedulable POD which require GPU exists [Feature:ClusterSizeAutoscalingScaleWithNAP]", func() {
framework.SkipUnlessProviderIs("gke")
installNvidiaDriversDaemonSet()
framework.ExpectNoError(enableAutoprovisioning(`
"resource_limits":{"name":"nvidia-tesla-k80", "minimum":0, "maximum": 3},
"resource_limits":{"name":"cpu", "minimum":0, "maximum":64},
"resource_limits":{"name":"memory", "minimum":0, "maximum":1000000000000}`))
By("Schedule a pod which requires GPU and wait until it is started")
framework.ExpectNoError(scheduleGpuPod(f, "gpu-pod-rc"))
defer framework.DeleteRCAndWaitForGC(f.ClientSet, f.Namespace.Name, "gpu-pod-rc")
By("Verify cluster size increased")
framework.ExpectNoError(WaitForClusterSizeFunc(f.ClientSet,
func(size int) bool { return size == nodeCount+1 }, scaleUpTimeout))
By("Check if NAP group was created")
Expect(getNAPNodePoolsNumber()).Should(Equal(1))
})
It("should increase cluster size if pending pods are small and one node is broken [Feature:ClusterSizeAutoscalingScaleUp]",
func() {
framework.TestUnderTemporaryNetworkFailure(c, "default", getAnyNode(c), func() { simpleScaleUpTest(1) })