wayne worker run into Infinite loop if rabbitmq is down

[silenceshell]

Describe the bug

If we shutdown the rabbitmq, the wayne worker will print quite a lot of logs such as:

2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:66] delivery not initialized
2019/01/25 14:57:35.718 [C] [workers.go:87] delivery not initialized

To Reproduce
Steps to reproduce the behavior:

1. deploy wayne on k8s
2. scale replicas of rabbitmq to zeor
3. display logs of wayne worker, will get a lot of logs as above.

Expected behavior

As rabbitmq is down, wayne worker should retry to reconnect, and should break this loop:

func processMessage(d *amqp.Delivery, w *BaseMessageWorker) {
	defer func() {
		if r := recover(); r != nil {
			logs.Critical(r)
			d.Reject(false)
		}
	}()

	var m message.Message
	err := json.Unmarshal(d.Body, &m)
	if err != nil {
		ackOrDie(d, false)
		return
	}
	if err = w.Process(&m); err != nil {
		logs.Error(err)
		ackOrDie(d, false) // 认为再次重试也无法处理
	} else {
		ackOrDie(d, false)
	}
}

Screenshots

NA

Desktop (please complete the following information):

NA

Server Version

- Kubernetes version: 1.11.1
- Wayne version: wayne v1.3.1-36-gc5157ca0c 

Additional context