diff --git a/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager.go b/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager.go
index 7d7176603e4..8f4b5c6437b 100644
--- a/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager.go
+++ b/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager.go
@@ -48,7 +48,7 @@ const (
 
 type activePodsLister interface {
 	// Returns a list of active pods on the node.
-	GetRunningPods() ([]*v1.Pod, error)
+	GetActivePods() []*v1.Pod
 }
 
 // nvidiaGPUManager manages nvidia gpu devices.
@@ -148,9 +148,7 @@ func (ngm *nvidiaGPUManager) AllocateGPU(pod *v1.Pod, container *v1.Container) (
 		ngm.allocated = allocated
 	} else {
 		// update internal list of GPUs in use prior to allocating new GPUs.
-		if err := ngm.updateAllocatedGPUs(); err != nil {
-			return nil, fmt.Errorf("Failed to allocate GPUs because of issues with updating GPUs in use: %v", err)
-		}
+		ngm.updateAllocatedGPUs()
 	}
 	// Check if GPUs have already been allocated. If so return them right away.
 	// This can happen if a container restarts for example.
@@ -179,13 +177,10 @@ func (ngm *nvidiaGPUManager) AllocateGPU(pod *v1.Pod, container *v1.Container) (
 }
 
 // updateAllocatedGPUs updates the list of GPUs in use.
-// It gets a list of running pods and then frees any GPUs that are bound to terminated pods.
+// It gets a list of active pods and then frees any GPUs that are bound to terminated pods.
 // Returns error on failure.
-func (ngm *nvidiaGPUManager) updateAllocatedGPUs() error {
-	activePods, err := ngm.activePodsLister.GetRunningPods()
-	if err != nil {
-		return fmt.Errorf("Failed to list active pods: %v", err)
-	}
+func (ngm *nvidiaGPUManager) updateAllocatedGPUs() {
+	activePods := ngm.activePodsLister.GetActivePods()
 	activePodUids := sets.NewString()
 	for _, pod := range activePods {
 		activePodUids.Insert(string(pod.UID))
@@ -194,7 +189,6 @@ func (ngm *nvidiaGPUManager) updateAllocatedGPUs() error {
 	podsToBeRemoved := allocatedPodUids.Difference(activePodUids)
 	glog.V(5).Infof("pods to be removed: %v", podsToBeRemoved.List())
 	ngm.allocated.delete(podsToBeRemoved.List())
-	return nil
 }
 
 // discoverGPUs identifies allGPUs NVIDIA GPU devices available on the local node by walking `/dev` directory.
@@ -224,10 +218,7 @@ func (ngm *nvidiaGPUManager) discoverGPUs() error {
 
 // gpusInUse returns a list of GPUs in use along with the respective pods that are using it.
 func (ngm *nvidiaGPUManager) gpusInUse() (*podGPUs, error) {
-	pods, err := ngm.activePodsLister.GetRunningPods()
-	if err != nil {
-		return nil, err
-	}
+	pods := ngm.activePodsLister.GetActivePods()
 	type containerIdentifier struct {
 		id   string
 		name string
diff --git a/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager_test.go b/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager_test.go
index e4565525e49..db80f03cf1e 100644
--- a/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager_test.go
+++ b/pkg/kubelet/gpu/nvidia/nvidia_gpu_manager_test.go
@@ -32,8 +32,8 @@ type testActivePodsLister struct {
 	activePods []*v1.Pod
 }
 
-func (tapl *testActivePodsLister) GetRunningPods() ([]*v1.Pod, error) {
-	return tapl.activePods, nil
+func (tapl *testActivePodsLister) GetActivePods() []*v1.Pod {
+	return tapl.activePods
 }
 
 func makeTestPod(numContainers, gpusPerContainer int) *v1.Pod {
diff --git a/pkg/kubelet/kubelet.go b/pkg/kubelet/kubelet.go
index 14bcce91f8b..9584ef5f1b0 100644
--- a/pkg/kubelet/kubelet.go
+++ b/pkg/kubelet/kubelet.go
@@ -792,7 +792,7 @@ func NewMainKubelet(kubeCfg *componentconfig.KubeletConfiguration, kubeDeps *Kub
 	klet.AddPodSyncLoopHandler(activeDeadlineHandler)
 	klet.AddPodSyncHandler(activeDeadlineHandler)
 
-	criticalPodAdmissionHandler := preemption.NewCriticalPodAdmissionHandler(klet.getActivePods, killPodNow(klet.podWorkers, kubeDeps.Recorder), kubeDeps.Recorder)
+	criticalPodAdmissionHandler := preemption.NewCriticalPodAdmissionHandler(klet.GetActivePods, killPodNow(klet.podWorkers, kubeDeps.Recorder), kubeDeps.Recorder)
 	klet.admitHandlers.AddPodAdmitHandler(lifecycle.NewPredicateAdmitHandler(klet.getNodeAnyWay, criticalPodAdmissionHandler))
 	// apply functional Option's
 	for _, opt := range kubeDeps.Options {
@@ -1204,7 +1204,7 @@ func (kl *Kubelet) initializeModules() error {
 		return fmt.Errorf("Kubelet failed to get node info: %v", err)
 	}
 
-	if err := kl.containerManager.Start(node, kl.getActivePods); err != nil {
+	if err := kl.containerManager.Start(node, kl.GetActivePods); err != nil {
 		return fmt.Errorf("Failed to start ContainerManager %v", err)
 	}
 
@@ -1230,7 +1230,7 @@ func (kl *Kubelet) initializeRuntimeDependentModules() {
 		glog.Fatalf("Failed to start cAdvisor %v", err)
 	}
 	// eviction manager must start after cadvisor because it needs to know if the container runtime has a dedicated imagefs
-	kl.evictionManager.Start(kl, kl.getActivePods, kl, evictionMonitoringPeriod)
+	kl.evictionManager.Start(kl, kl.GetActivePods, kl, evictionMonitoringPeriod)
 }
 
 // Run starts the kubelet reacting to config updates
diff --git a/pkg/kubelet/kubelet_pods.go b/pkg/kubelet/kubelet_pods.go
index d96627f9f7f..e1eda16586a 100644
--- a/pkg/kubelet/kubelet_pods.go
+++ b/pkg/kubelet/kubelet_pods.go
@@ -76,8 +76,8 @@ func (kl *Kubelet) listPodsFromDisk() ([]types.UID, error) {
 	return pods, nil
 }
 
-// getActivePods returns non-terminal pods
-func (kl *Kubelet) getActivePods() []*v1.Pod {
+// GetActivePods returns non-terminal pods
+func (kl *Kubelet) GetActivePods() []*v1.Pod {
 	allPods := kl.podManager.GetPods()
 	activePods := kl.filterOutTerminatedPods(allPods)
 	return activePods