Workaround confluent-kafka-go bug #463

neli.go

@@ -46,6 +46,7 @@ type neli struct {
 	barrier      Barrier
 	state        concurrent.AtomicReference
 	stateMutex   sync.Mutex
+	deliveryHandlerDone chan int
 }
 
 // State of the Neli instance.
@@ -83,6 +84,7 @@ func New(config Config, barrier ...Barrier) (Neli, error) {
 		barrier:      barrierArg,
 		pollDeadline: concurrent.NewDeadline(*config.MinPollInterval),
 		state:        concurrent.NewAtomicReference(Live),
+		deliveryHandlerDone: make(chan int),
 	}
 
 	consumerConfigs := copyKafkaConfig(n.config.KafkaConfig)
@@ -134,6 +136,8 @@ func New(config Config, barrier ...Barrier) (Neli, error) {
 	}
 	n.producer = p
 	go func() {
+		defer close(n.deliveryHandlerDone)
+
 		for event := range p.Events() {
 			switch e := event.(type) {
 			case *kafka.Message:
@@ -360,7 +364,17 @@ func (n *neli) Close() error {
 	defer n.state.Set(Closed)
 
 	n.state.Set(Closing)
-	n.producer.Close()
+	defer func() {
+		<-n.deliveryHandlerDone
+	}()
+	defer func() {
+		go func() {
+			// A bug in confluent-kafka-go (#463) occasionally causes an indefinite syscall hang in Close(), after it closes
+			// the Events channel. So we delegate this to a separate goroutine — better an orphaned goroutine than a
+			// frozen harvester. (The rest of the battery will still unwind normally.)
+			n.producer.Close()
+		}()
+	}()
 	return n.consumer.Close()
 }