Avoid unnecessary GET request when updating pod status

When sending out an pod status update, kubelet GETs the pod from the apiserver Terminates if the apiserver returns an not found error; otherwise, proceed to to update. Even after a pod has been deleted, there might still be queued up updates for the pod. This leads to expensive, unncessary GET operations. The situation is worse when there are batch creation/deletion of a significant number of pods (e.g., E2E tests), leaving many updates in the queue. This change checks whether a pod exists before GET the pod from the apiserver to avoid redundant GETs.

Avoid unnecessary GET request when updating pod status
3072754c · Yu-Ju Hong · 5516b868 · 3072754c
Commit 3072754c authored Feb 12, 2016 by Yu-Ju Hong
Hide whitespace changes
Inline Side-by-side

Showing with 5 additions and 4 deletions

manager.go pkg/kubelet/status/manager.go +5 -4

No files found.
--- a/pkg/kubelet/status/manager.go
+++ b/pkg/kubelet/status/manager.go
@@ -347,6 +347,11 @@ func (m *manager) syncBatch() {

 // syncPod syncs the given status with the API server. The caller must not hold the lock.
 func (m *manager) syncPod(uid types.UID, status versionedPodStatus) {
+	if !m.needsUpdate(uid, status) {
+		glog.V(1).Infof("Status for pod %q is up-to-date; skipping", uid)
+		return
+	}
+
 	// TODO: make me easier to express from client code
 	pod, err := m.kubeClient.Core().Pods(status.podNamespace).Get(status.podName)
 	if errors.IsNotFound(err) {
@@ -362,10 +367,6 @@ func (m *manager) syncPod(uid types.UID, status versionedPodStatus) {
 			m.deletePodStatus(uid)
 			return
 		}
-		if !m.needsUpdate(pod.UID, status) {
-			glog.V(1).Infof("Status for pod %q is up-to-date; skipping", format.Pod(pod))
-			return
-		}
 		pod.Status = status.status
 		// TODO: handle conflict as a retry, make that easier too.
 		pod, err = m.kubeClient.Core().Pods(pod.Namespace).UpdateStatus(pod)