celery update_state issue

export.py

from tasks import full_export_task

def export_year_zip(user):
  full_export_task.update_state(state="PROGRESS", meta={})

tasks.py

task()
def full_export_task(user, task_id=None):
  from export.views.user_full_export import export_year_zip
  result = export_year_zip(user)
  return result

I think the idea of the custom task state is to notify any applications that are aware of this running of the changing state. For example, suppose you have a video transcoding job that you run in a task. This is done in chunks because the video is very large. Before you start processing the video, you know you have 100 chunks to process. The update_state method will allow you to update the task's state (i.e. progress in the video example) as the task is running. This means that you can say, "If anyone cares, I have just finished the first chunk of 100. I'll let you know when I'm done with the next one."

The reason why there is no reference to task_id is because it refers to the current task.

Does that make sense?

The above code is just an example. export_year_zip is actually much longer and does exactly what you've described: telling the user how much of the export has already been processed with a long–running task.

The problem I'm having is this: You say that it refers to the "current task". But how does celery know what the "current task" is when I'm having 4 tasks running synchronously? I just want to prevent that all 4 tasks suddenly have the same state. And before I push this to the live instance I wanted to make sure that this code is supposed to work (although the tests are running fine).

I think the problem is that you're calling update_state from outside the @task-decorated function. Not really sure, sorry.