为什么 celery 在执行我的任务时返回 KeyError？

Question

I keep getting this keyError. I am sending strings and id (integers) to the task function, so I don't think it is serialization issue. Also it says the keyerror is on the path to the function itself, not the contents. Please help.

Tasks.py

from celery.decorators import task
from notification import models as notification

@task(ignore_result=True)
def notify_match_creation(match, home_team, away_team, home_team_captain, away_team_captain):
    notification.send(User.objects.filter(profile__teams__pk__in=(home_team, away_team)),
                      "tournaments_new_match",
                      {'match': unicode(match),
                       'home_team_captain': home_team_captain,
                       'away_team_captain': away_team_captain,
                       })

Relevant settings

CELERY_RESULT_BACKEND = "database"
CELERY_RESULT_DBURI = "postgresql://user:pass@localhost/ahgl"
BROKER_HOST = "localhost"
BROKER_PORT = 5672
BROKER_USER = "guest"
BROKER_PASSWORD = "guest"
BROKER_VHOST = "/"

Celery output:

[Tasks]

  . apps.tournaments.tasks.notify_match_creation
  . tournaments.tasks.notify_match_creation
[2012-02-25 02:34:06,209: WARNING/MainProcess] celery@NATTOWER has started.
[2012-02-25 02:34:06,477: WARNING/PoolWorker-4] E:\Webdesign\ahgl\ENV\lib\site-packages\djcelery\loaders.py:84: UserWarn
ing: Using settings.DEBUG leads to a memory leak, never use this setting in production environments!
  warnings.warn("Using settings.DEBUG leads to a memory leak, never "
[2012-02-25 02:34:06,479: WARNING/PoolWorker-2] E:\Webdesign\ahgl\ENV\lib\site-packages\djcelery\loaders.py:84: UserWarn
ing: Using settings.DEBUG leads to a memory leak, never use this setting in production environments!
  warnings.warn("Using settings.DEBUG leads to a memory leak, never "
[2012-02-25 02:34:06,523: WARNING/PoolWorker-3] E:\Webdesign\ahgl\ENV\lib\site-packages\djcelery\loaders.py:84: UserWarn
ing: Using settings.DEBUG leads to a memory leak, never use this setting in production environments!
  warnings.warn("Using settings.DEBUG leads to a memory leak, never "
[2012-02-25 02:34:06,566: WARNING/PoolWorker-1] E:\Webdesign\ahgl\ENV\lib\site-packages\djcelery\loaders.py:84: UserWarn
ing: Using settings.DEBUG leads to a memory leak, never use this setting in production environments!
  warnings.warn("Using settings.DEBUG leads to a memory leak, never "
[2012-02-25 02:34:31,520: INFO/MainProcess] Got task from broker: apps.tournaments.tasks.notify_match_creation[4dbd6258-
5cee-49e9-8c8a-2d2105a2d52a]
[2012-02-25 02:34:31,569: ERROR/MainProcess] Task apps.tournaments.tasks.notify_match_creation[4dbd6258-5cee-49e9-8c8a-2
d2105a2d52a] raised exception: KeyError('apps.tournaments.tasks.notify_match_creation',)
Traceback (most recent call last):
  File "E:\Webdesign\ahgl\ENV\lib\site-packages\celery\concurrency\processes\pool.py", line 211, in worker
    result = (True, func(*args, **kwds))
  File "E:\Webdesign\ahgl\ENV\lib\site-packages\celery\worker\job.py", line 50, in execute_and_trace
    task = tasks[name]
KeyError: 'apps.tournaments.tasks.notify_match_creation'
[2012-02-25 02:38:29,773: WARNING/MainProcess] celeryd: Hitting Ctrl+C again will terminate all running tasks!
[2012-02-25 02:38:29,773: WARNING/MainProcess] celeryd: Warm shutdown (MainProcess)
[2012-02-25 02:38:31,779: INFO/MainProcess] process shutting down

Answer 1

In the celery output, you see that the task that gets picked up is

tournaments.tasks.notify_match_creation (#1)

and in your key error it is

KeyError: 'apps.tournaments.tasks.notify_match_creation'

before you call the celery task, make sure that it is imported with the same name (structure) as it is in the celery task that gets picked up (#1). Please refer to the link from celery docs for getting your relative imports rights.

one possible solution, when you launch celery - try

celery worker -A apps.tournaments.tasks.notify_match_creation

this could align your task names

Answer 2

I had the exact same issue, however, the following worked for me,

in my tasks.py I changed from

from celery.task import task

@task(name="generate_and_email_membership_invoice")
def my_task_name():
      ...

to

from ..celeryconf import app # the file where you have app.autodiscover_tasks()

@app.task()
def my_task_name():
     ..

Answer 3

I know it's late but this worked for me When I had the same error

I was using task.nameoftask

It is rather a relative import as

app.task.nameoftask