#celery

Is it safe to use 'self' in a task?

self.obj = 12

or will that self variable be shared between workers?

Such as this: https://gist.github.com/monokrome/0dcc8677d90f9...

monokrome If the task instance is only created once that will _could_ cause issues depending on what type of worker pool you are using. In general it is not very good design. Why not just pass items into each of the methods?

Has anyone had issues getting celery worker to start?

I’m running it on a debian box

I upgraded to 3.1.13

now the worker fails when I try to start it

any help would be most appreciated \

If anyone has an idea please hit me up on twitter @ realkenzic

boris317: It's a bit less DRY to do that, but if celery doesn't create a new instance for each run then I need to do that

That's why I was wondering whether or not celery creates a new instance for each worker

Id argue that does not break dry

so what the difference between `celery worker -A app -Q priority,queue` and `python -m celery.__main__ worker -n app@localhost.localdoman --workdir=/data/myapp/root`

the first one works, celery starts up and process request. the second one.. it starts.. but its looking for RabbitMQ, not Redis like it does when i run celery worker -A