芹菜result.get超时 [英] Celery result.get times out

问题描述

我有两个不同的Django项目，分别是 projA 和 projB ，每个项目都有自己的celery守护程序在不同的队列上运行但同一个虚拟主机，projA的任务为 taskA ，projB的任务为 taskB ，我尝试运行 taskB 从 taskA 内部，例如

I have two different django projects say projA and projB, each have its own celery daemon running on separate queues but same vhost, projA have a task taskA and projB have a task taskB, I try to run taskB from inside taskA e.g.

@task(routing_key='taskA')
def taskA(event_id):
    # do some work , then call taskB and wait for result
    result = send_task('taskB',routing_key='taskB')
    res = result.get(timeout=20)

我可以看到在projB的日志中，taskB在一秒钟内完成，但是taskA继续等待结果，并在20秒后超时。

I can see in logs of projB that taskB finished within a second, but taskA keeps on waiting for result and times out after 20 seconds

对于后端，我有Rabbitmq。

For backend I have rabbitmq.

推荐答案

设置结果后端解决了问题

Setting the result back-end fixed the problem

CELERY_RESULT_BACKEND = "amqp"
CELERY_AMQP_TASK_RESULT_EXPIRES = 1000 

IMO（如果结果）后端未设置result.get应该引发错误或至少记录警告

IMO if result back-end is not set result.get should throw error or at-least log a warning

尽管celery 2.3确实引发了错误，如此处所述 https://github.com/ask/django-celery/issues/66

Though celery 2.3 does throw error as described here https://github.com/ask/django-celery/issues/66

这篇关于芹菜result.get超时的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！