daemon: close EventsService on shutdown

On daemon shutdown, the HTTP server tries to gracefully shutdown for 5
seconds. If there's an open API connection to the '/events' endpoint, it
fails to do so as nothing interrupts that connection, thus forcing the
daemon to wait until that timeout is reached.

Add a Close method to the EventsService, and call it during daemon
shutdown. It'll close any events channel, signaling to the '/events'
handler to return and close the connection.

It now takes ~1s (or less) to shutdown the daemon when there's an active
'/events' connection, instead of 5.

Signed-off-by: Albin Kerouanton <albin.kerouanton@docker.com>

daemon/daemon.go

@@ -1493,6 +1493,12 @@ func (daemon *Daemon) Shutdown(ctx context.Context) error {
 		daemon.mdDB.Close()
 	}
 
+	// At this point, everything has been shut down and no containers are
+	// running anymore. If there are still some open connections to the
+	// '/events' endpoint, closing the EventsService should tear them down
+	// immediately.
+	daemon.EventsService.Close()
+
 	return daemon.cleanupMounts(cfg)
 }
 

daemon/events/events.go

@@ -151,3 +151,9 @@ func (e *Events) loadBufferedEvents(since, until time.Time, topic func(any) bool
 	}
 	return buffered
 }
+
+// Close all the channels returned to event subscribers.
+func (e *Events) Close() error {
+	e.pub.Close()
+	return nil
+}

daemon/server/router/system/system_routes.go

@@ -353,7 +353,12 @@ func (s *systemRouter) getEvents(ctx context.Context, w http.ResponseWriter, r *
 
 	for {
 		select {
-		case ev := <-l:
+		case ev, ok := <-l:
+			if !ok {
+				log.G(ctx).Debug("event channel closed")
+				return nil
+			}
+
 			jev, ok := ev.(events.Message)
 			if !ok {
 				log.G(ctx).Warnf("unexpected event message: %q", ev)

integration/daemon/daemon_linux_test.go

@@ -6,6 +6,7 @@ import (
 	"net/netip"
 	"slices"
 	"testing"
+	"time"
 
 	"github.com/google/go-cmp/cmp/cmpopts"
 	"github.com/moby/moby/api/types/network"
@@ -518,3 +519,42 @@ func TestSwarmNoNftables(t *testing.T) {
 		assert.Check(t, is.ErrorContains(err, "daemon exited during startup"))
 	})
 }
+
+// TestDaemonShutsDownQuicklyDespiteEventsConnection checks whether the daemon
+// shuts down in less than 5 secs when there's an active API connection to the
+// '/events' endpoint.
+//
+// As this test verifies timing behavior, it may become flaky. Feel free to
+// delete it if it's too annoying.
+//
+// Regression test for https://github.com/moby/moby/issues/32357#issuecomment-3496848492
+func TestDaemonShutsDownQuicklyDespiteEventsConnection(t *testing.T) {
+	skip.If(t, testEnv.IsRemoteDaemon, "cannot start daemon on remote test run")
+	// The Engine is presumably working the same way on Windows and Linux.
+	// Avoid running on Windows as it may be more flaky there.
+	skip.If(t, testEnv.DaemonInfo.OSType == "windows")
+
+	t.Parallel()
+
+	ctx := testutil.StartSpan(baseContext, t)
+
+	d := daemon.New(t)
+	defer d.Cleanup(t)
+
+	// This is a parallel test, disable iptables integration.
+	d.StartWithBusybox(ctx, t, "--iptables=false", "--ip6tables=false")
+	defer d.Stop(t)
+
+	apiClient := d.NewClientT(t)
+	// Open a connection to the '/events' endpoint.
+	apiClient.Events(ctx, client.EventsListOptions{})
+
+	// Kill the daemon
+	t0 := time.Now()
+	d.Stop(t)
+
+	dt := time.Since(t0)
+	if dt.Seconds() > 5 {
+		t.Error("the daemon took more than 5 secs to shutdown")
+	}
+}