Celery keeps creating rabbitmq queues, pilling all over

问题

Using celery (3.1.8 + ) with django 1.6 all tasks are defined to ignore results (is this the correct syntax?)

@shared_task(ignore_result=True)
def somefunc():
   pass

When I look at the rabbitmq queues I see more and more queues created by celery with names like:

19926fa9965e40c19ed9640c2b42ce1e

and contain one message (similar to the following):

correlation_id: 19926fa9-965e-40c1-9ed9-640c2b42ce1e priority: 0 delivery_mode: 2 headers: content_encoding: binary content_type: application/x-python-serialize Payload 118 bytes Encoding: base64 gAJ9cQEoVQZzdGF0dXNxAlUHU1VDQ0VTU3EDVQl0cmFjZWJhY2txBE5VBnJlc3VsdHEFTlUHdGFza19pZHEGVSQxOTkyNmZhOS05NjVlLTQwYzEtOWVkOS02 NDBjMmI0MmNlMWVxB1UIY2hpbGRyZW5xCF11Lg==

If I set celery to "always eager" mode then the problem is solved, but this is obviously not a good solution for a production server.

Any Clues? Is this connected to the ignore_result option? and there is some missing task somewhere? something else?

Thanks for the help

回答1:

I'm not sure how django or your code defines the @shared_task decorator, but have you tried explicitly setting the queue within it?

@shared_task(ignore_results=True, queue="myexamplequeue")

This would apply the message to the myexamplequeue whenever you call .delay() or .apply_async() on somefunc.

来源：https://stackoverflow.com/questions/24910613/celery-keeps-creating-rabbitmq-queues-pilling-all-over