Skip to content

[release-4.22] OCPBUGS-92011: e2e: fix: clear hugepages before switching kernelPageSize to 4k #1554

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
openshift-cherrypick-robot wants to merge 1 commit into
base: release-4.22
Choose a base branch
from
+26 −51
Open

[release-4.22] OCPBUGS-92011: e2e: fix: clear hugepages before switching kernelPageSize to 4k #1554

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
63 changes: 12 additions & 51 deletions test/e2e/performanceprofile/functests/14_arm/kernelpagesize.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,6 @@ package __arm

import (
"context"
"encoding/json"
"fmt"

corev1 "k8s.io/api/core/v1"
Expand All @@ -18,7 +17,6 @@ import (
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/infrastructure"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/label"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/nodes"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/poolname"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/profiles"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/profilesupdate"
)
Expand All @@ -34,7 +32,6 @@ var _ = Describe("[rfe_id:80342] kernelPageSize configuration validation on aarc
var (
workerRTNodes []corev1.Node
perfProfile, initialProfile *performancev2.PerformanceProfile
poolName string
err error
ctx context.Context = context.Background()
isArm bool
Expand All @@ -50,44 +47,24 @@ var _ = Describe("[rfe_id:80342] kernelPageSize configuration validation on aarc
workerRTNode = workerRTNodes[0]

perfProfile, err = profiles.GetByNodeLabels(testutils.NodeSelectorLabels)
initialProfile = perfProfile.DeepCopy()

poolName = poolname.GetByProfile(ctx, perfProfile)
Expect(err).ToNot(HaveOccurred())
initialProfile = perfProfile.DeepCopy()

isArm, err = infrastructure.IsARM(ctx, &workerRTNode)
Expect(err).ToNot(HaveOccurred())
if !isArm {
Skip("This test requires an aarch64 arm CPU")
}

By("Make sure that the performance profile starts the test suite with 4k kernelPageSize and real time kernel disabled")
currentKernelPageSize, err := getKernelPageSizeFromNode(ctx, workerRTNode)
Expect(err).ToNot(HaveOccurred())

var realTimeKernelEnabled bool
var validKernelPageSize bool

if perfProfile.Spec.RealTimeKernel != nil {
realTimeKernelEnabled = *perfProfile.Spec.RealTimeKernel.Enabled
By("Ensuring profile baseline: kernelPageSize=4k, no hugepages, RT disabled")
perfProfile.Spec.HugePages = nil
perfProfile.Spec.KernelPageSize = ptr.To(performancev2.KernelPageSize(kernelPageSize4k))
perfProfile.Spec.RealTimeKernel = &performancev2.RealTimeKernel{
Enabled: ptr.To(false),
}

if perfProfile.Spec.KernelPageSize != nil {
validKernelPageSize = currentKernelPageSize != kernelPageSizeBytes4k
}

if realTimeKernelEnabled || validKernelPageSize {
perfProfile.Spec.RealTimeKernel = &performancev2.RealTimeKernel{
Enabled: ptr.To(false),
}
*perfProfile.Spec.KernelPageSize = kernelPageSize4k
profiles.UpdateWithRetry(perfProfile)

By(fmt.Sprintf("Applying changes in performance profile and waiting until %s will start updating", poolName))
profilesupdate.WaitForTuningUpdating(ctx, perfProfile)

By(fmt.Sprintf("Waiting when %s finishes updates", poolName))
profilesupdate.WaitForTuningUpdated(ctx, perfProfile)
if !equality.Semantic.DeepEqual(perfProfile.Spec, initialProfile.Spec) {
profilesupdate.ApplyProfileAndWait(ctx, perfProfile)
}
})

Expand All @@ -96,18 +73,8 @@ var _ = Describe("[rfe_id:80342] kernelPageSize configuration validation on aarc
profile, err := profiles.GetByNodeLabels(testutils.NodeSelectorLabels)
Expect(err).ToNot(HaveOccurred())

currentSpec, _ := json.Marshal(profile.Spec)
spec, _ := json.Marshal(initialProfile.Spec)

// revert only if the profile changes
if !equality.Semantic.DeepEqual(currentSpec, spec) {
profiles.UpdateWithRetry(initialProfile)

By(fmt.Sprintf("Applying changes in performance profile and waiting until %s will start updating", poolName))
profilesupdate.WaitForTuningUpdating(ctx, initialProfile)

By(fmt.Sprintf("Waiting when %s finishes updates", poolName))
profilesupdate.WaitForTuningUpdated(ctx, initialProfile)
if !equality.Semantic.DeepEqual(profile.Spec, initialProfile.Spec) {
profilesupdate.ApplyProfileAndWait(ctx, initialProfile)
}
})

Expand All @@ -118,18 +85,12 @@ var _ = Describe("[rfe_id:80342] kernelPageSize configuration validation on aarc
By(fmt.Sprintf("Modifying the profile to use %s", newKernelPageSize))
klog.Infof("Changing kernelPageSize to: %s from: %s", newKernelPageSize, *perfProfile.Spec.KernelPageSize)
perfProfile.Spec.KernelPageSize = ptr.To(performancev2.KernelPageSize(newKernelPageSize))
profiles.UpdateWithRetry(perfProfile)

By(fmt.Sprintf("Applying changes in performance profile and waiting until %s will start updating", poolName))
profilesupdate.WaitForTuningUpdating(ctx, perfProfile)

By(fmt.Sprintf("Waiting when %s finishes updates", poolName))
profilesupdate.WaitForTuningUpdated(ctx, perfProfile)
profilesupdate.ApplyProfileAndWait(ctx, perfProfile)

By(fmt.Sprintf("Verifying the kernelPageSize has changed to %s on the affected node", expectedKernelPageSize))
kernelPageSize, err := getKernelPageSizeFromNode(ctx, workerRTNode)
Expect(err).ToNot(HaveOccurred())
Expect(kernelPageSize).To(Equal(expectedKernelPageSize))
Expect(kernelPageSize).To(Equal(expectedKernelPageSize), "kernel page size mismatch on node")
},
Entry("[test_id:80459] should accept 64k kernelPageSize", kernelPageSize64k, kernelPageSizeBytes64k),
Entry("[test_id:80461] should accept 4k kernelPageSize", kernelPageSize4k, kernelPageSizeBytes4k),
Expand Down
14 changes: 14 additions & 0 deletions test/e2e/performanceprofile/functests/utils/profilesupdate/profile_update.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,7 @@ import (
testlog "github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/log"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/mcps"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/nodepools"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/poolname"
"github.com/openshift/cluster-node-tuning-operator/test/e2e/performanceprofile/functests/utils/profiles"
)

Expand Down Expand Up @@ -128,3 +129,16 @@ func WaitForTuningUpdated(ctx context.Context, profile *performancev2.Performanc
err = nodepools.WaitForConfigToBeReady(ctx, testclient.ControlPlaneClient, np.Name, np.Namespace)
Expect(err).ToNot(HaveOccurred())
}

// ApplyProfileAndWait updates the profile and waits for the MCP rollout (or nodepool on hypershift) to complete.
// Replaces the common 3-step pattern: UpdateWithRetry + WaitForTuningUpdating + WaitForTuningUpdated.
func ApplyProfileAndWait(ctx context.Context, profile *performancev2.PerformanceProfile) {
GinkgoHelper()
poolName := poolname.GetByProfile(ctx, profile)
By("Updating the performance profile")
profiles.UpdateWithRetry(profile)
By(fmt.Sprintf("Applying changes in performance profile and waiting until %s will start updating", poolName))
WaitForTuningUpdating(ctx, profile)
By(fmt.Sprintf("Waiting when %s finishes updates", poolName))
WaitForTuningUpdated(ctx, profile)
}