Merge pull request #2285 from maxsmythe/dynamic-informer-cache

🌱 Proposal for dynamic informer cache

Author: k8s-ci-robot

pkg/cache/cache.go

@@ -83,6 +83,9 @@ type Informers interface {
 	// of the underlying object.
 	GetInformerForKind(ctx context.Context, gvk schema.GroupVersionKind, opts ...InformerGetOption) (Informer, error)
 
+	// RemoveInformer removes an informer entry and stops it if it was running.
+	RemoveInformer(ctx context.Context, obj client.Object) error
+
 	// Start runs all the informers known to this cache until the context is closed.
 	// It blocks.
 	Start(ctx context.Context) error
@@ -121,6 +124,8 @@ type Informer interface {
 
 	// HasSynced return true if the informers underlying store has synced.
 	HasSynced() bool
+	// IsStopped returns true if the informer has been stopped.
+	IsStopped() bool
 }
 
 // AllNamespaces should be used as the map key to deliminate namespace settings

pkg/cache/cache_test.go

@@ -1887,6 +1887,42 @@ func CacheTest(createCacheFunc func(config *rest.Config, opts cache.Options) (ca
 					By("verifying the object is received on the channel")
 					Eventually(out).Should(Receive(Equal(pod)))
 				})
+				It("should be able to stop and restart informers", func() {
+					By("getting a shared index informer for a pod")
+					pod := &corev1.Pod{
+						ObjectMeta: metav1.ObjectMeta{
+							Name:      "informer-obj",
+							Namespace: "default",
+						},
+						Spec: corev1.PodSpec{
+							Containers: []corev1.Container{
+								{
+									Name:  "nginx",
+									Image: "nginx",
+								},
+							},
+						},
+					}
+					sii, err := informerCache.GetInformer(context.TODO(), pod)
+					Expect(err).NotTo(HaveOccurred())
+					Expect(sii).NotTo(BeNil())
+					Expect(sii.HasSynced()).To(BeTrue())
+
+					By("removing the existing informer")
+					Expect(informerCache.RemoveInformer(context.TODO(), pod)).To(Succeed())
+					Eventually(sii.IsStopped).WithTimeout(5 * time.Second).Should(BeTrue())
+
+					By("recreating the informer")
+
+					sii2, err := informerCache.GetInformer(context.TODO(), pod)
+					Expect(err).NotTo(HaveOccurred())
+					Expect(sii2).NotTo(BeNil())
+					Expect(sii2.HasSynced()).To(BeTrue())
+
+					By("validating the two informers are in different states")
+					Expect(sii.IsStopped()).To(BeTrue())
+					Expect(sii2.IsStopped()).To(BeFalse())
+				})
 				It("should be able to get an informer by group/version/kind", func() {
 					By("getting an shared index informer for gvk = core/v1/pod")
 					gvk := schema.GroupVersionKind{Group: "", Version: "v1", Kind: "Pod"}
@@ -2116,6 +2152,48 @@ func CacheTest(createCacheFunc func(config *rest.Config, opts cache.Options) (ca
 					Eventually(out).Should(Receive(Equal(pod)))
 				})
 
+				It("should be able to stop and restart informers", func() {
+					By("getting a shared index informer for a pod")
+					pod := &unstructured.Unstructured{
+						Object: map[string]interface{}{
+							"spec": map[string]interface{}{
+								"containers": []map[string]interface{}{
+									{
+										"name":  "nginx",
+										"image": "nginx",
+									},
+								},
+							},
+						},
+					}
+					pod.SetName("informer-obj2")
+					pod.SetNamespace("default")
+					pod.SetGroupVersionKind(schema.GroupVersionKind{
+						Group:   "",
+						Version: "v1",
+						Kind:    "Pod",
+					})
+					sii, err := informerCache.GetInformer(context.TODO(), pod)
+					Expect(err).NotTo(HaveOccurred())
+					Expect(sii).NotTo(BeNil())
+					Expect(sii.HasSynced()).To(BeTrue())
+
+					By("removing the existing informer")
+					Expect(informerCache.RemoveInformer(context.TODO(), pod)).To(Succeed())
+					Eventually(sii.IsStopped).WithTimeout(5 * time.Second).Should(BeTrue())
+
+					By("recreating the informer")
+
+					sii2, err := informerCache.GetInformer(context.TODO(), pod)
+					Expect(err).NotTo(HaveOccurred())
+					Expect(sii2).NotTo(BeNil())
+					Expect(sii2.HasSynced()).To(BeTrue())
+
+					By("validating the two informers are in different states")
+					Expect(sii.IsStopped()).To(BeTrue())
+					Expect(sii2.IsStopped()).To(BeFalse())
+				})
+
 				It("should be able to index an object field then retrieve objects by that field", func() {
 					By("creating the cache")
 					informer, err := cache.New(cfg, cache.Options{})

pkg/cache/delegating_by_gvk_cache.go

@@ -52,6 +52,14 @@ func (dbt *delegatingByGVKCache) List(ctx context.Context, list client.ObjectLis
 	return cache.List(ctx, list, opts...)
 }
 
+func (dbt *delegatingByGVKCache) RemoveInformer(ctx context.Context, obj client.Object) error {
+	cache, err := dbt.cacheForObject(obj)
+	if err != nil {
+		return err
+	}
+	return cache.RemoveInformer(ctx, obj)
+}
+
 func (dbt *delegatingByGVKCache) GetInformer(ctx context.Context, obj client.Object, opts ...InformerGetOption) (Informer, error) {
 	cache, err := dbt.cacheForObject(obj)
 	if err != nil {

pkg/cache/informer_cache.go

@@ -190,6 +190,17 @@ func (ic *informerCache) getInformerForKind(ctx context.Context, gvk schema.Grou
 	return ic.Informers.Get(ctx, gvk, obj, &internal.GetOptions{})
 }
 
+// RemoveInformer deactivates and removes the informer from the cache.
+func (ic *informerCache) RemoveInformer(_ context.Context, obj client.Object) error {
+	gvk, err := apiutil.GVKForObject(obj, ic.scheme)
+	if err != nil {
+		return err
+	}
+
+	ic.Informers.Remove(gvk, obj)
+	return nil
+}
+
 // NeedLeaderElection implements the LeaderElectionRunnable interface
 // to indicate that this can be started without requiring the leader lock.
 func (ic *informerCache) NeedLeaderElection() bool {

pkg/cache/informertest/fake_cache.go

@@ -73,6 +73,20 @@ func (c *FakeInformers) GetInformer(ctx context.Context, obj client.Object, opts
 	return c.informerFor(gvk, obj)
 }
 
+// RemoveInformer implements Informers.
+func (c *FakeInformers) RemoveInformer(ctx context.Context, obj client.Object) error {
+	if c.Scheme == nil {
+		c.Scheme = scheme.Scheme
+	}
+	gvks, _, err := c.Scheme.ObjectKinds(obj)
+	if err != nil {
+		return err
+	}
+	gvk := gvks[0]
+	delete(c.InformersByGVK, gvk)
+	return nil
+}
+
 // WaitForCacheSync implements Informers.
 func (c *FakeInformers) WaitForCacheSync(ctx context.Context) bool {
 	if c.Synced == nil {

pkg/cache/internal/informers.go

@@ -36,6 +36,7 @@ import (
 	"k8s.io/client-go/rest"
 	"k8s.io/client-go/tools/cache"
 	"sigs.k8s.io/controller-runtime/pkg/client/apiutil"
+	"sigs.k8s.io/controller-runtime/pkg/internal/syncs"
 )
 
 // InformersOpts configures an InformerMap.
@@ -88,6 +89,20 @@ type Cache struct {
 
 	// CacheReader wraps Informer and implements the CacheReader interface for a single type
 	Reader CacheReader
+
+	// Stop can be used to stop this individual informer.
+	stop chan struct{}
+}
+
+// Start starts the informer managed by a MapEntry.
+// Blocks until the informer stops. The informer can be stopped
+// either individually (via the entry's stop channel) or globally
+// via the provided stop argument.
+func (c *Cache) Start(stop <-chan struct{}) {
+	// Stop on either the whole map stopping or just this informer being removed.
+	internalStop, cancel := syncs.MergeChans(stop, c.stop)
+	defer cancel()
+	c.Informer.Run(internalStop)
 }
 
 type tracker struct {
@@ -180,13 +195,13 @@ func (ip *Informers) Start(ctx context.Context) error {
 
 		// Start each informer
 		for _, i := range ip.tracker.Structured {
-			ip.startInformerLocked(i.Informer)
+			ip.startInformerLocked(i)
 		}
 		for _, i := range ip.tracker.Unstructured {
-			ip.startInformerLocked(i.Informer)
+			ip.startInformerLocked(i)
 		}
 		for _, i := range ip.tracker.Metadata {
-			ip.startInformerLocked(i.Informer)
+			ip.startInformerLocked(i)
 		}
 
 		// Set started to true so we immediately start any informers added later.
@@ -201,7 +216,7 @@ func (ip *Informers) Start(ctx context.Context) error {
 	return nil
 }
 
-func (ip *Informers) startInformerLocked(informer cache.SharedIndexInformer) {
+func (ip *Informers) startInformerLocked(cacheEntry *Cache) {
 	// Don't start the informer in case we are already waiting for the items in
 	// the waitGroup to finish, since waitGroups don't support waiting and adding
 	// at the same time.
@@ -212,7 +227,7 @@ func (ip *Informers) startInformerLocked(informer cache.SharedIndexInformer) {
 	ip.waitGroup.Add(1)
 	go func() {
 		defer ip.waitGroup.Done()
-		informer.Run(ip.ctx.Done())
+		cacheEntry.Start(ip.ctx.Done())
 	}()
 }
 
@@ -288,6 +303,21 @@ func (ip *Informers) Get(ctx context.Context, gvk schema.GroupVersionKind, obj r
 	return started, i, nil
 }
 
+// Remove removes an informer entry and stops it if it was running.
+func (ip *Informers) Remove(gvk schema.GroupVersionKind, obj runtime.Object) {
+	ip.mu.Lock()
+	defer ip.mu.Unlock()
+
+	informerMap := ip.informersByType(obj)
+
+	entry, ok := informerMap[gvk]
+	if !ok {
+		return
+	}
+	close(entry.stop)
+	delete(informerMap, gvk)
+}
+
 func (ip *Informers) informersByType(obj runtime.Object) map[schema.GroupVersionKind]*Cache {
 	switch obj.(type) {
 	case runtime.Unstructured:
@@ -356,13 +386,14 @@ func (ip *Informers) addInformerToMap(gvk schema.GroupVersionKind, obj runtime.O
 			scopeName:        mapping.Scope.Name(),
 			disableDeepCopy:  ip.unsafeDisableDeepCopy,
 		},
+		stop: make(chan struct{}),
 	}
 	ip.informersByType(obj)[gvk] = i
 
 	// Start the informer in case the InformersMap has started, otherwise it will be
 	// started when the InformersMap starts.
 	if ip.started {
-		ip.startInformerLocked(i.Informer)
+		ip.startInformerLocked(i)
 	}
 	return i, ip.started, nil
 }

pkg/cache/multi_namespace_cache.go

@@ -109,6 +109,27 @@ func (c *multiNamespaceCache) GetInformer(ctx context.Context, obj client.Object
 	return &multiNamespaceInformer{namespaceToInformer: namespaceToInformer}, nil
 }
 
+func (c *multiNamespaceCache) RemoveInformer(ctx context.Context, obj client.Object) error {
+	// If the object is clusterscoped, get the informer from clusterCache,
+	// if not use the namespaced caches.
+	isNamespaced, err := apiutil.IsObjectNamespaced(obj, c.Scheme, c.RESTMapper)
+	if err != nil {
+		return err
+	}
+	if !isNamespaced {
+		return c.clusterCache.RemoveInformer(ctx, obj)
+	}
+
+	for _, cache := range c.namespaceToCache {
+		err := cache.RemoveInformer(ctx, obj)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
 func (c *multiNamespaceCache) GetInformerForKind(ctx context.Context, gvk schema.GroupVersionKind, opts ...InformerGetOption) (Informer, error) {
 	// If the object is cluster scoped, get the informer from clusterCache,
 	// if not use the namespaced caches.
@@ -391,3 +412,13 @@ func (i *multiNamespaceInformer) HasSynced() bool {
 	}
 	return true
 }
+
+// IsStopped checks if each namespaced informer has stopped, returns false if any are still running.
+func (i *multiNamespaceInformer) IsStopped() bool {
+	for _, informer := range i.namespaceToInformer {
+		if stopped := informer.IsStopped(); !stopped {
+			return false
+		}
+	}
+	return true
+}

pkg/internal/syncs/syncs.go

@@ -0,0 +1,38 @@
+package syncs
+
+import (
+	"context"
+	"reflect"
+	"sync"
+)
+
+// MergeChans returns a channel that is closed when any of the input channels are signaled.
+// The caller must call the returned CancelFunc to ensure no resources are leaked.
+func MergeChans[T any](chans ...<-chan T) (<-chan T, context.CancelFunc) {
+	var once sync.Once
+	out := make(chan T)
+	cancel := make(chan T)
+	cancelFunc := func() {
+		once.Do(func() {
+			close(cancel)
+		})
+		<-out
+	}
+	cases := make([]reflect.SelectCase, len(chans)+1)
+	for i := range chans {
+		cases[i] = reflect.SelectCase{
+			Dir:  reflect.SelectRecv,
+			Chan: reflect.ValueOf(chans[i]),
+		}
+	}
+	cases[len(cases)-1] = reflect.SelectCase{
+		Dir:  reflect.SelectRecv,
+		Chan: reflect.ValueOf(cancel),
+	}
+	go func() {
+		defer close(out)
+		_, _, _ = reflect.Select(cases)
+	}()
+
+	return out, cancelFunc
+}