Move from glog to klog

- Move from the old github.com/golang/glog to k8s.io/klog
- klog as explicit InitFlags() so we add them as necessary
- we update the other repositories that we vendor that made a similar
change from glog to klog
  * github.com/kubernetes/repo-infra
  * k8s.io/gengo/
  * k8s.io/kube-openapi/
  * github.com/google/cadvisor
- Entirely remove all references to glog
- Fix some tests by explicit InitFlags in their init() methods

Change-Id: I92db545ff36fcec83afe98f550c9e630098b3135

test/integration/scheduler_perf/BUILD

@@ -42,7 +42,7 @@ go_test(
         "//staging/src/k8s.io/apimachinery/pkg/labels:go_default_library",
         "//test/integration/framework:go_default_library",
         "//test/utils:go_default_library",
-        "//vendor/github.com/golang/glog:go_default_library",
+        "//vendor/k8s.io/klog:go_default_library",
     ],
 )
 

test/integration/scheduler_perf/scheduler_bench_test.go

@@ -28,7 +28,7 @@ import (
 	"k8s.io/kubernetes/test/integration/framework"
 	testutils "k8s.io/kubernetes/test/utils"
 
-	"github.com/golang/glog"
+	"k8s.io/klog"
 )
 
 var (
@@ -227,7 +227,7 @@ func benchmarkScheduling(numNodes, numExistingPods, minPods int,
 		"scheduler-perf-",
 	)
 	if err := nodePreparer.PrepareNodes(); err != nil {
-		glog.Fatalf("%v", err)
+		klog.Fatalf("%v", err)
 	}
 	defer nodePreparer.CleanupNodes()
 
@@ -239,7 +239,7 @@ func benchmarkScheduling(numNodes, numExistingPods, minPods int,
 	for {
 		scheduled, err := schedulerConfigFactory.GetScheduledPodLister().List(labels.Everything())
 		if err != nil {
-			glog.Fatalf("%v", err)
+			klog.Fatalf("%v", err)
 		}
 		if len(scheduled) >= numExistingPods {
 			break
@@ -257,7 +257,7 @@ func benchmarkScheduling(numNodes, numExistingPods, minPods int,
 		// TODO: Setup watch on apiserver and wait until all pods scheduled.
 		scheduled, err := schedulerConfigFactory.GetScheduledPodLister().List(labels.Everything())
 		if err != nil {
-			glog.Fatalf("%v", err)
+			klog.Fatalf("%v", err)
 		}
 		if len(scheduled) >= numExistingPods+b.N {
 			break

test/integration/scheduler_perf/scheduler_test.go

@@ -18,11 +18,11 @@ package benchmark
 
 import (
 	"fmt"
-	"github.com/golang/glog"
 	"k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/labels"
+	"k8s.io/klog"
 	"k8s.io/kubernetes/pkg/scheduler/factory"
 	testutils "k8s.io/kubernetes/test/utils"
 	"math"
@@ -137,7 +137,7 @@ func schedulePods(config *testConfig) int32 {
 		time.Sleep(50 * time.Millisecond)
 		scheduled, err := config.schedulerSupportFunctions.GetScheduledPodLister().List(labels.Everything())
 		if err != nil {
-			glog.Fatalf("%v", err)
+			klog.Fatalf("%v", err)
 		}
 		// 30,000 pods -> wait till @ least 300 are scheduled to start measuring.
 		// TODO Find out why sometimes there may be scheduling blips in the beginning.
@@ -155,7 +155,7 @@ func schedulePods(config *testConfig) int32 {
 		// TODO: Setup watch on apiserver and wait until all pods scheduled.
 		scheduled, err := config.schedulerSupportFunctions.GetScheduledPodLister().List(labels.Everything())
 		if err != nil {
-			glog.Fatalf("%v", err)
+			klog.Fatalf("%v", err)
 		}
 
 		// We will be completed when all pods are done being scheduled.