Way to run tasks

[hirokiky]

Hi, I don't know how to run tasks by using this Dockerfile.

• Where should I place tasks.py file to register tasks to run?
• How can I notice new tasks for this container?

If you know, please tell me. 😃

[kjr]

You can mount your tasks.py into the container in the users home directory (/home/user) and I think use the CELERY_APP environment variable (this worked for me, but I was browsing the issues here hoping to find a list of available environment variables...)

docker run --volume /path/to/tasks.py:/home/user/tasks.py --env CELERY_APP=tasks celery

Not sure how you would notice new tasks.. any change to the file on the host would be seen by the container, but I don't think it will reload. Still, it's quick to restart the container...

[AdrianRibao]

This needs to be documented.

How can this be used with django for example?

[tianon]

+1 for some documentation 😇 (what we've got is impossibly sparse right now: https://github.com/docker-library/docs/blob/master/celery/content.md)

[esseti]

that's tricky. i tried with docker-compose but i can't find a way. has anyone been able to run the celery in docker and use the integration with django??

[bittner]

Using this image seems ridiculous. If you have an application container, as you usually have with Django, you need all dependencies (things you import in tasks.py) installed in this container again.

That's why other projects (e.g. cookiecutter-django) reuse the application container for Celery, and only run a different command (command: celery ... worker) against it with docker-compose.

[tianon]

Indeed, Sentry includes a way to run Celery directly as well. Pondering whether this image actually makes sense to simply deprecate, similar to the deprecation of the django image (and for similar reasons).

[tianon]

I've opened #12 to discuss this image's deprecation. In the meantime, I'm going to close this issue since I think the original question is as solved as it's going to get. 👍

[ArnaudPel]

Agreed, it's not going to be much more difficult to replace this image with a build on a standard Python image with celery added to pip's requirements.txt for example.

Actually, doing so in the first place would have saved me two hours yesterday: This celery docker image ignores the broker url when provided from inside python like so app = Celery('tasks', broker='my url'), and only allows it to be set from the docker env variable CELERY_BROKER_URL. Ok this is kinda hinted in the doc, my bad for not reading carefully. But it's clearly counter-intuitive and time consuming to find this out when having your nose in the code and not expecting some of the parameters to be forced but not the others (the backend argument is not forced like this for example, although it has the same kind of properties that could refer to other docker containers: host, port, ...).

Anyway I'm rambling again, I came here to comment about this parameters forcing but since the whole thing is deprecated I reckon there's no need.