e2e: fix CNI Logging level test case

It can happen that clock of the cluster nodes is not aligned
to the clock of test runner. In such cases, asking a pod for logs
since a given time might end in an empty response, as the requested
time can be in the future (from the cluster perspective).

Use the pod's CreationTimestamp timestamp field to be sure
to pass to the API server a meaningful value.

Signed-off-by: Andrea Panattoni <apanatto@redhat.com>

test/conformance/tests/test_sriov_operator.go

@@ -1013,11 +1013,9 @@ var _ = Describe("[sriov] operator", func() {
 				Expect(err).ToNot(HaveOccurred())
 				waitForNetAttachDef(sriovNetworkName, namespaces.Test)
 
-				podDeployTime := time.Now()
-
 				testPod := createTestPod(node, []string{sriovNetworkName})
 
-				recentMultusLogs := getMultusPodLogs(testPod.Spec.NodeName, podDeployTime)
+				recentMultusLogs := getMultusPodLogs(testPod.Spec.NodeName, testPod.ObjectMeta.CreationTimestamp.Time)
 
 				Expect(recentMultusLogs).To(
 					ContainElement(