Separate backOff policy for static pod

llhhbc · llhhbc · commit db0f65c55a74 · 2024-07-18T10:27:14.000+08:00
Most static pods run as critical components. When an exception occurs and a restart is required, the sooner the better, so a separate backoff policy is set for static pods.
diff --git a/pkg/kubelet/kubelet.go b/pkg/kubelet/kubelet.go
@@ -873,6 +873,7 @@ func NewMainKubelet(kubeCfg *kubeletconfiginternal.KubeletConfiguration,
 		volumepathhandler.NewBlockVolumePathHandler())
 
 	klet.backOff = flowcontrol.NewBackOff(backOffPeriod, MaxContainerBackOff)
+	klet.staticBackOff = flowcontrol.NewBackOff(time.Second*3, time.Second*10)
 
 	// setup eviction manager
 	evictionManager, evictionAdmitHandler := eviction.NewManager(klet.resourceAnalyzer, evictionConfig,
@@ -1259,6 +1260,8 @@ type Kubelet struct {
 
 	// Container restart Backoff
 	backOff *flowcontrol.Backoff
+	// static pod Backoff
+	staticBackOff *flowcontrol.Backoff
 
 	// Information about the ports which are opened by daemons on Node running this Kubelet server.
 	daemonEndpoints *v1.NodeDaemonEndpoints
@@ -1940,7 +1943,14 @@ func (kl *Kubelet) SyncPod(ctx context.Context, updateType kubetypes.SyncPodType
 	// Use WithoutCancel instead of a new context.TODO() to propagate trace context
 	// Call the container runtime's SyncPod callback
 	sctx := context.WithoutCancel(ctx)
-	result := kl.containerRuntime.SyncPod(sctx, pod, podStatus, pullSecrets, kl.backOff)
+	backOff := kl.backOff
+	if pod.Annotations != nil && pod.Annotations[kubetypes.ConfigSourceAnnotationKey] == "file" {
+		klog.V(5).InfoS("use static backOff for pod. ", "pod", klog.KObj(pod))
+		backOff = kl.staticBackOff
+	} else {
+		klog.V(5).InfoS("use normal backOff for pod. ", "pod", klog.KObj(pod))
+	}
+	result := kl.containerRuntime.SyncPod(sctx, pod, podStatus, pullSecrets, backOff)
 	kl.reasonCache.Update(pod.UID, result)
 	if err := result.Error(); err != nil {
 		// Do not return error if the only failures were pods in backoff
diff --git a/pkg/kubelet/kubelet_pods.go b/pkg/kubelet/kubelet_pods.go
@@ -1370,6 +1370,7 @@ func (kl *Kubelet) HandlePodCleanups(ctx context.Context) error {
 
 	// Cleanup any backoff entries.
 	kl.backOff.GC()
+	kl.staticBackOff.GC()
 	return nil
 }
 
diff --git a/pkg/kubelet/kubelet_test.go b/pkg/kubelet/kubelet_test.go
@@ -35,14 +35,12 @@ import (
 	"go.opentelemetry.io/otel/sdk/trace/tracetest"
 	oteltrace "go.opentelemetry.io/otel/trace"
 	noopoteltrace "go.opentelemetry.io/otel/trace/noop"
+	"k8s.io/utils/mount"
 
 	cadvisorapi "github.com/google/cadvisor/info/v1"
 	cadvisorapiv2 "github.com/google/cadvisor/info/v2"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
-	core "k8s.io/client-go/testing"
-	"k8s.io/mount-utils"
-
 	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
@@ -52,6 +50,7 @@ import (
 	"k8s.io/apimachinery/pkg/util/wait"
 	utilfeature "k8s.io/apiserver/pkg/util/feature"
 	"k8s.io/client-go/kubernetes/fake"
+	core "k8s.io/client-go/testing"
 	"k8s.io/client-go/tools/record"
 	"k8s.io/client-go/util/flowcontrol"
 	featuregatetesting "k8s.io/component-base/featuregate/testing"
@@ -330,7 +329,9 @@ func newTestKubeletWithImageList(
 
 	fakeClock := testingclock.NewFakeClock(time.Now())
 	kubelet.backOff = flowcontrol.NewBackOff(time.Second, time.Minute)
+	kubelet.staticBackOff = flowcontrol.NewBackOff(time.Second, time.Minute)
 	kubelet.backOff.Clock = fakeClock
+	kubelet.staticBackOff.Clock = fakeClock
 	kubelet.resyncInterval = 10 * time.Second
 	kubelet.workQueue = queue.NewBasicWorkQueue(fakeClock)
 	// Relist period does not affect the tests.

Original file line number	Diff line number	Diff line change
`@@ -1370,6 +1370,7 @@ func (kl *Kubelet) HandlePodCleanups(ctx context.Context) error {`
`1370`	`1370`
`1371`	`1371`	`// Cleanup any backoff entries.`
`1372`	`1372`	`kl.backOff.GC()`
	`1373`	`+ kl.staticBackOff.GC()`
`1373`	`1374`	`return nil`
`1374`	`1375`	`}`
`1375`	`1376`