Timeout DispatchJob if no dispatch return is received (#42)

This is to avoid blocking in case of error.

Author: raphael

pool/node.go

@@ -42,6 +42,7 @@ type (
 		workerTTL         time.Duration     // Worker considered dead if keep-alive not updated after this duration
 		workerShutdownTTL time.Duration     // Worker considered dead if not shutdown after this duration
 		pendingJobTTL     time.Duration     // Job lease expires if not acked after this duration
+		ackGracePeriod    time.Duration     // Wait for return status up to this duration
 		logger            pulse.Logger
 		h                 hasher
 		stop              chan struct{}  // closed when node is stopped
@@ -191,6 +192,7 @@ func AddNode(ctx context.Context, name string, rdb *redis.Client, opts ...NodeOp
 		workerTTL:         o.workerTTL,
 		workerShutdownTTL: o.workerShutdownTTL,
 		pendingJobTTL:     o.pendingJobTTL,
+		ackGracePeriod:    o.ackGracePeriod,
 		h:                 jumpHash{crc64.New(crc64.MakeTable(crc64.ECMA))},
 		stop:              make(chan struct{}),
 		rdb:               rdb,
@@ -313,16 +315,26 @@ func (node *Node) DispatchJob(ctx context.Context, key string, payload []byte) e
 	node.pendingJobs[eventID] = cherr
 	node.lock.Unlock()
 
-	// Wait for return status.
+	// Wait for return status up to ack grace period.
+	timer := time.NewTimer(2 * node.ackGracePeriod)
+	defer timer.Stop()
+
 	select {
 	case err = <-cherr:
+	case <-timer.C:
+		err = fmt.Errorf("DispatchJob: job %q timed out, TTL: %v", key, 2*node.ackGracePeriod)
 	case <-ctx.Done():
 		err = ctx.Err()
 	}
 
+	node.lock.Lock()
+	delete(node.pendingJobs, eventID)
 	close(cherr)
+	node.lock.Unlock()
 	if err == nil {
 		node.logger.Info("dispatched", "key", key)
+	} else {
+		node.logger.Error(fmt.Errorf("DispatchJob: failed to dispatch job: %w", err), "key", key)
 	}
 	return err
 }
@@ -636,9 +648,9 @@ func (node *Node) returnDispatchStatus(_ context.Context, ev *streaming.Event) {
 		return
 	}
 	node.logger.Debug("dispatch return", "event", ev.EventName, "id", ev.ID, "ack-id", ack.EventID)
-	delete(node.pendingJobs, ack.EventID)
 	if cherr == nil {
 		// Event was requeued.
+		delete(node.pendingJobs, ack.EventID)
 		return
 	}
 	var err error

pool/scheduler.go

@@ -134,11 +134,11 @@ func (sched *scheduler) startJobs(ctx context.Context, jobs []*JobParam) error {
 	for _, job := range jobs {
 		err := sched.node.DispatchJob(ctx, job.Key, job.Payload)
 		if err != nil {
-			sched.logger.Error(err, "failed to dispatch job", "job", job.Key)
+			sched.logger.Error(fmt.Errorf("failed to dispatch job: %w", err), "job", job.Key)
 			continue
 		}
 		if _, err := sched.jobMap.Set(ctx, job.Key, time.Now().String()); err != nil {
-			sched.logger.Error(err, "failed to store job", "job", job.Key)
+			sched.logger.Error(fmt.Errorf("failed to store job: %w", err), "job", job.Key)
 			continue
 		}
 	}
@@ -164,11 +164,11 @@ func (sched *scheduler) stopJobs(ctx context.Context, plan *JobPlan) error {
 	for _, key := range toStop {
 		err := sched.node.StopJob(ctx, key)
 		if err != nil {
-			sched.logger.Error(err, "failed to stop job", "job", key)
+			sched.logger.Error(fmt.Errorf("failed to stop job: %w", err), "job", key)
 			continue
 		}
 		if _, err := sched.jobMap.Delete(ctx, key); err != nil {
-			sched.logger.Error(err, "failed to delete job", "job", key)
+			sched.logger.Error(fmt.Errorf("failed to delete job: %w", err), "job", key)
 		}
 	}
 	return nil

streaming/sink.go

@@ -410,12 +410,13 @@ func (s *Sink) periodicKeepAlive() {
 	for {
 		select {
 		case <-ticker.C:
+			s.lock.Lock()
 			now := time.Now().UnixNano()
 			if _, err := s.consumersKeepAliveMap.Set(ctx, s.consumer, strconv.FormatInt(now, 10)); err != nil {
 				s.logger.Error(fmt.Errorf("failed to update sink keep-alive: %v", err))
+				s.lock.Unlock()
 				continue
 			}
-			s.lock.Lock()
 			s.lastKeepAlive = now
 			s.lock.Unlock()