cacher: add support for consistent streaming

design details https://github.com/kubernetes/enhancements/tree/master/keps/sig-api-machinery/3157-watch-list#design-details

staging/src/k8s.io/apiserver/pkg/storage/cacher/cacher.go

@@ -47,7 +47,7 @@ import (
 )
 
 var (
-	emptyFunc = func() {}
+	emptyFunc = func(bool) {}
 )
 
 const (
@@ -147,6 +147,10 @@ func (i *indexedWatchers) deleteWatcher(number int, value string, supported bool
 }
 
 func (i *indexedWatchers) terminateAll(objectType reflect.Type, done func(*cacheWatcher)) {
+	// note that we don't have to call setDrainInputBufferLocked method on the watchers
+	// because we take advantage of the default value - stop immediately
+	// also watchers that have had already its draining strategy set
+	// are no longer available (they were removed from the allWatchers and the valueWatchers maps)
 	if len(i.allWatchers) > 0 || len(i.valueWatchers) > 0 {
 		klog.Warningf("Terminating all watchers from cacher %v", objectType)
 	}
@@ -183,6 +187,10 @@ func newTimeBucketWatchers(clock clock.Clock, bookmarkFrequency time.Duration) *
 // adds a watcher to the bucket, if the deadline is before the start, it will be
 // added to the first one.
 func (t *watcherBookmarkTimeBuckets) addWatcher(w *cacheWatcher) bool {
+	// note that the returned time can be before t.createTime,
+	// especially in cases when the nextBookmarkTime method
+	// give us the zero value of type Time
+	// so buckedID can hold a negative value
 	nextTime, ok := w.nextBookmarkTime(t.clock.Now(), t.bookmarkFrequency)
 	if !ok {
 		return false
@@ -517,7 +525,7 @@ func (c *Cacher) Watch(ctx context.Context, key string, opts storage.ListOptions
 		c.Lock()
 		defer c.Unlock()
 		// Update watcher.forget function once we can compute it.
-		watcher.forget = forgetWatcher(c, c.watcherIdx, triggerValue, triggerSupported)
+		watcher.forget = forgetWatcher(c, watcher, c.watcherIdx, triggerValue, triggerSupported)
 		c.watchers.addWatcher(watcher, c.watcherIdx, triggerValue, triggerSupported)
 
 		// Add it to the queue only when the client support watch bookmarks.
@@ -1028,11 +1036,13 @@ func (c *Cacher) Stop() {
 	c.stopWg.Wait()
 }
 
-func forgetWatcher(c *Cacher, index int, triggerValue string, triggerSupported bool) func() {
-	return func() {
+func forgetWatcher(c *Cacher, w *cacheWatcher, index int, triggerValue string, triggerSupported bool) func(bool) {
+	return func(drainWatcher bool) {
 		c.Lock()
 		defer c.Unlock()
 
+		w.setDrainInputBufferLocked(drainWatcher)
+
 		// It's possible that the watcher is already not in the structure (e.g. in case of
 		// simultaneous Stop() and terminateAllWatchers(), but it is safe to call stopLocked()
 		// on a watcher multiple times.
@@ -1156,7 +1166,7 @@ type cacheWatcher struct {
 	done      chan struct{}
 	filter    filterWithAttrsFunc
 	stopped   bool
-	forget    func()
+	forget    func(bool)
 	versioner storage.Versioner
 	// The watcher will be closed by server after the deadline,
 	// save it here to send bookmark events before that.
@@ -1168,9 +1178,13 @@ type cacheWatcher struct {
 	// human readable identifier that helps assigning cacheWatcher
 	// instance with request
 	identifier string
+
+	// drainInputBuffer indicates whether we should delay closing this watcher
+	// and send all event in the input buffer.
+	drainInputBuffer bool
 }
 
-func newCacheWatcher(chanSize int, filter filterWithAttrsFunc, forget func(), versioner storage.Versioner, deadline time.Time, allowWatchBookmarks bool, objectType reflect.Type, identifier string) *cacheWatcher {
+func newCacheWatcher(chanSize int, filter filterWithAttrsFunc, forget func(bool), versioner storage.Versioner, deadline time.Time, allowWatchBookmarks bool, objectType reflect.Type, identifier string) *cacheWatcher {
 	return &cacheWatcher{
 		input:               make(chan *watchCacheEvent, chanSize),
 		result:              make(chan watch.Event, chanSize),
@@ -1193,16 +1207,29 @@ func (c *cacheWatcher) ResultChan() <-chan watch.Event {
 
 // Implements watch.Interface.
 func (c *cacheWatcher) Stop() {
-	c.forget()
+	c.forget(false)
 }
 
 // we rely on the fact that stopLocked is actually protected by Cacher.Lock()
 func (c *cacheWatcher) stopLocked() {
 	if !c.stopped {
 		c.stopped = true
-		close(c.done)
+		// stop without draining the input channel was requested.
+		if !c.drainInputBuffer {
+			close(c.done)
+		}
 		close(c.input)
 	}
+
+	// Even if the watcher was already stopped, if it previously was
+	// using draining mode and it's not using it now we need to
+	// close the done channel now. Otherwise we could leak the
+	// processing goroutine if it will be trying to put more objects
+	// into result channel, the channel will be full and there will
+	// already be noone on the processing the events on the receiving end.
+	if !c.drainInputBuffer && !c.isDoneChannelClosedLocked() {
+		close(c.done)
+	}
 }
 
 func (c *cacheWatcher) nonblockingAdd(event *watchCacheEvent) bool {
@@ -1227,7 +1254,7 @@ func (c *cacheWatcher) add(event *watchCacheEvent, timer *time.Timer) bool {
 		// we simply terminate it.
 		klog.V(1).Infof("Forcing %v watcher close due to unresponsiveness: %v. len(c.input) = %v, len(c.result) = %v", c.objectType.String(), c.identifier, len(c.input), len(c.result))
 		metrics.TerminatedWatchersCounter.WithLabelValues(c.objectType.String()).Inc()
-		c.forget()
+		c.forget(false)
 	}
 
 	if timer == nil {
@@ -1273,6 +1300,22 @@ func (c *cacheWatcher) nextBookmarkTime(now time.Time, bookmarkFrequency time.Du
 	return heartbeatTime, true
 }
 
+// setDrainInputBufferLocked if set to true indicates that we should delay closing this watcher
+// until we send all events residing in the input buffer.
+func (c *cacheWatcher) setDrainInputBufferLocked(drain bool) {
+	c.drainInputBuffer = drain
+}
+
+// isDoneChannelClosed checks if c.done channel is closed
+func (c *cacheWatcher) isDoneChannelClosedLocked() bool {
+	select {
+	case <-c.done:
+		return true
+	default:
+	}
+	return false
+}
+
 func getMutableObject(object runtime.Object) runtime.Object {
 	if _, ok := object.(*cachingObject); ok {
 		// It is safe to return without deep-copy, because the underlying

staging/src/k8s.io/apiserver/pkg/storage/cacher/cacher_whitebox_test.go

@@ -58,13 +58,14 @@ func TestCacheWatcherCleanupNotBlockedByResult(t *testing.T) {
 	var w *cacheWatcher
 	count := 0
 	filter := func(string, labels.Set, fields.Set) bool { return true }
-	forget := func() {
+	forget := func(drainWatcher bool) {
 		lock.Lock()
 		defer lock.Unlock()
 		count++
 		// forget() has to stop the watcher, as only stopping the watcher
 		// triggers stopping the process() goroutine which we are in the
 		// end waiting for in this test.
+		w.setDrainInputBufferLocked(drainWatcher)
 		w.stopLocked()
 	}
 	initEvents := []*watchCacheEvent{
@@ -89,7 +90,7 @@ func TestCacheWatcherHandlesFiltering(t *testing.T) {
 	filter := func(_ string, _ labels.Set, field fields.Set) bool {
 		return field["spec.nodeName"] == "host"
 	}
-	forget := func() {}
+	forget := func(bool) {}
 
 	testCases := []struct {
 		events   []*watchCacheEvent
@@ -210,6 +211,7 @@ TestCase:
 			break TestCase
 		default:
 		}
+		w.setDrainInputBufferLocked(false)
 		w.stopLocked()
 	}
 }
@@ -524,7 +526,8 @@ func TestCacheWatcherStoppedInAnotherGoroutine(t *testing.T) {
 	var w *cacheWatcher
 	done := make(chan struct{})
 	filter := func(string, labels.Set, fields.Set) bool { return true }
-	forget := func() {
+	forget := func(drainWatcher bool) {
+		w.setDrainInputBufferLocked(drainWatcher)
 		w.stopLocked()
 		done <- struct{}{}
 	}
@@ -556,6 +559,7 @@ func TestCacheWatcherStoppedInAnotherGoroutine(t *testing.T) {
 		case <-time.After(time.Second):
 			t.Fatal("expected received a event on ResultChan")
 		}
+		w.setDrainInputBufferLocked(false)
 		w.stopLocked()
 	}
 }
@@ -667,7 +671,7 @@ func TestTimeBucketWatchersBasic(t *testing.T) {
 	filter := func(_ string, _ labels.Set, _ fields.Set) bool {
 		return true
 	}
-	forget := func() {}
+	forget := func(bool) {}
 
 	newWatcher := func(deadline time.Time) *cacheWatcher {
 		return newCacheWatcher(0, filter, forget, testVersioner{}, deadline, true, objectType, "")
@@ -1581,3 +1585,72 @@ func TestCacheIntervalInvalidationStopsWatch(t *testing.T) {
 		t.Errorf("unexpected number of events received, expected: %d, got: %d", bufferSize+1, received)
 	}
 }
+
+// TestCacheWatcherDraining verifies the cacheWatcher.process goroutine is properly cleaned up when draining was requested
+func TestCacheWatcherDraining(t *testing.T) {
+	var lock sync.RWMutex
+	var w *cacheWatcher
+	count := 0
+	filter := func(string, labels.Set, fields.Set) bool { return true }
+	forget := func(drainWatcher bool) {
+		lock.Lock()
+		defer lock.Unlock()
+		count++
+		w.setDrainInputBufferLocked(drainWatcher)
+		w.stopLocked()
+	}
+	initEvents := []*watchCacheEvent{
+		{Object: &v1.Pod{}},
+		{Object: &v1.Pod{}},
+	}
+	w = newCacheWatcher(1, filter, forget, testVersioner{}, time.Now(), true, objectType, "")
+	go w.processInterval(context.Background(), intervalFromEvents(initEvents), 0)
+	if !w.add(&watchCacheEvent{Object: &v1.Pod{}}, time.NewTimer(1*time.Second)) {
+		t.Fatal("failed adding an even to the watcher")
+	}
+	forget(true) // drain the watcher
+	<-w.ResultChan()
+	<-w.ResultChan()
+	<-w.ResultChan()
+	if err := wait.PollImmediate(1*time.Second, 5*time.Second, func() (bool, error) {
+		lock.RLock()
+		defer lock.RUnlock()
+		return count == 2, nil
+	}); err != nil {
+		t.Fatalf("expected forget() to be called twice, because processInterval should call Stop(): %v", err)
+	}
+}
+
+// TestCacheWatcherDrainingRequestedButNotDrained verifies the cacheWatcher.process goroutine is properly cleaned up when draining was requested
+// but the client never actually get any data
+func TestCacheWatcherDrainingRequestedButNotDrained(t *testing.T) {
+	var lock sync.RWMutex
+	var w *cacheWatcher
+	count := 0
+	filter := func(string, labels.Set, fields.Set) bool { return true }
+	forget := func(drainWatcher bool) {
+		lock.Lock()
+		defer lock.Unlock()
+		count++
+		w.setDrainInputBufferLocked(drainWatcher)
+		w.stopLocked()
+	}
+	initEvents := []*watchCacheEvent{
+		{Object: &v1.Pod{}},
+		{Object: &v1.Pod{}},
+	}
+	w = newCacheWatcher(1, filter, forget, testVersioner{}, time.Now(), true, objectType, "")
+	go w.processInterval(context.Background(), intervalFromEvents(initEvents), 0)
+	if !w.add(&watchCacheEvent{Object: &v1.Pod{}}, time.NewTimer(1*time.Second)) {
+		t.Fatal("failed adding an even to the watcher")
+	}
+	forget(true) // drain the watcher
+	w.Stop()     // client disconnected, timeout expired or ctx was actually closed
+	if err := wait.PollImmediate(1*time.Second, 5*time.Second, func() (bool, error) {
+		lock.RLock()
+		defer lock.RUnlock()
+		return count == 3, nil
+	}); err != nil {
+		t.Fatalf("expected forget() to be called three times, because processInterval should call Stop(): %v", err)
+	}
+}