Merge pull request #1399 from SargunNarula/refactor_getsmtlevel

OCPBUGS-62605: e2e: refactor GetSMTLevel to remove Gomega assertions

Author: openshift-merge-bot[bot]

test/e2e/performanceprofile/functests/1_performance/cpu_management.go

@@ -109,7 +109,8 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 		onlineCPUSet, err = nodes.GetOnlineCPUsSet(ctx, workerRTNode)
 		Expect(err).ToNot(HaveOccurred())
 		cpuID := onlineCPUSet.UnsortedList()[0]
-		smtLevel = nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+		smtLevel, err = nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+		Expect(err).ToNot(HaveOccurred(), "Unable to fetch SMT level on node %s, Error: %v", workerRTNode.Name, err)
 		getter, err = cgroup.BuildGetter(ctx, testclient.DataPlaneClient, testclient.K8sClient)
 		Expect(err).ToNot(HaveOccurred())
 		cgroupV2, err = cgroup.IsVersion2(ctx, testclient.DataPlaneClient)
@@ -251,7 +252,8 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 
 		DescribeTable("Verify CPU usage by stress PODs", func(ctx context.Context, guaranteed bool) {
 			cpuID := onlineCPUSet.UnsortedList()[0]
-			smtLevel := nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+			smtLevel, err := nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+			Expect(err).ToNot(HaveOccurred(), "Unable to fetch SMT level on node %s, Error: %v", workerRTNode.Name, err)
 			if smtLevel < 2 {
 				Skip(fmt.Sprintf("designated worker node %q has SMT level %d - minimum required 2", workerRTNode.Name, smtLevel))
 			}
@@ -281,7 +283,6 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 			}
 
 			By(fmt.Sprintf("create a %s QoS stress pod requesting %d cpus", expectedQos, cpuRequest))
-			var err error
 			err = testclient.DataPlaneClient.Create(ctx, testpod)
 			Expect(err).ToNot(HaveOccurred())
 
@@ -414,7 +415,8 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 			}
 
 			cpuID := onlineCPUSet.UnsortedList()[0]
-			smtLevel = nodes.GetSMTLevel(context.TODO(), cpuID, workerRTNode)
+			smtLevel, err = nodes.GetSMTLevel(context.TODO(), cpuID, workerRTNode)
+			Expect(err).ToNot(HaveOccurred(), "Unable to fetch SMT level on node %s, Error: %v", workerRTNode.Name, err)
 		})
 
 		AfterEach(func() {
@@ -584,7 +586,8 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 			// also covers Hyper-thread aware sheduling [test_id:46545] Odd number of isolated CPU threads
 			// any random existing cpu is fine
 			cpuID := onlineCPUSet.UnsortedList()[0]
-			smtLevel := nodes.GetSMTLevel(context.TODO(), cpuID, workerRTNode)
+			smtLevel, err := nodes.GetSMTLevel(context.TODO(), cpuID, workerRTNode)
+			Expect(err).ToNot(HaveOccurred(), "Unable to fetch SMT level on node %s, Error: %v", workerRTNode.Name, err)
 			if smtLevel < 2 {
 				Skip(fmt.Sprintf("designated worker node %q has SMT level %d - minimum required 2", workerRTNode.Name, smtLevel))
 			}
@@ -593,7 +596,7 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 			testpod = promotePodToGuaranteed(getStressPod(workerRTNode.Name, cpuCount))
 			testpod.Namespace = testutils.NamespaceTesting
 
-			err := testclient.DataPlaneClient.Create(context.TODO(), testpod)
+			err = testclient.DataPlaneClient.Create(context.TODO(), testpod)
 			Expect(err).ToNot(HaveOccurred())
 
 			currentPod, err := pods.WaitForPredicate(context.TODO(), client.ObjectKeyFromObject(testpod), 10*time.Minute, func(pod *corev1.Pod) (bool, error) {
@@ -643,7 +646,8 @@ var _ = Describe("[rfe_id:27363][performance] CPU Management", Ordered, func() {
 				// Check for SMT enabled
 				// any random existing cpu is fine
 				cpuID := onlineCPUSet.UnsortedList()[0]
-				smtLevel := nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+				smtLevel, err := nodes.GetSMTLevel(ctx, cpuID, workerRTNode)
+				Expect(err).ToNot(HaveOccurred(), "Unable to fetch SMT level on node %s, Error: %v", workerRTNode.Name, err)
 				hasWP := checkForWorkloadPartitioning(ctx)
 
 				// Following checks are required to map test_id scenario correctly to the type of node under test

test/e2e/performanceprofile/functests/utils/nodes/nodes.go

@@ -262,16 +262,20 @@ func GetOnlineCPUsSet(ctx context.Context, node *corev1.Node) (cpuset.CPUSet, er
 
 // GetSMTLevel returns the SMT level on the node using the given cpuID as target
 // Use a random cpuID from the return value of GetOnlineCPUsSet if not sure
-func GetSMTLevel(ctx context.Context, cpuID int, node *corev1.Node) int {
+func GetSMTLevel(ctx context.Context, cpuID int, node *corev1.Node) (int, error) {
 	cmd := []string{"/bin/sh", "-c", fmt.Sprintf("cat /sys/devices/system/cpu/cpu%d/topology/thread_siblings_list | tr -d \"\n\r\"", cpuID)}
 	out, err := ExecCommand(ctx, node, cmd)
-	ExpectWithOffset(1, err).ToNot(HaveOccurred())
+	if err != nil {
+		return 0, err
+	}
 	threadSiblingsList := testutils.ToString(out)
 	// how many thread sibling you have = SMT level
 	// example: 2-way SMT means 2 threads sibling for each thread
 	cpus, err := cpuset.Parse(strings.TrimSpace(threadSiblingsList))
-	ExpectWithOffset(1, err).ToNot(HaveOccurred())
-	return cpus.Size()
+	if err != nil {
+		return 0, err
+	}
+	return cpus.Size(), nil
 }
 
 // GetNumaNodes returns the number of numa nodes and the associated cpus as list on the node