Celery: Executing a task after another task, but return the result of the first task

Question

Sometimes I need to run task B after task A, but return the result of task A. Looking at the docs, I have tried to ( Avoid launching synchronous subtasks )

So I would use a chain, like this:

@celery.task
def A():
    return 5

@celery.task
def B():
    return 2

def do_all():
    chain = A.s() | B.s()
    chain()
    return result_of_A

But this would not work. In my case I need:

• B does not have any parameter. It does not accept the result of A
• B must be executed after A completes
• I do not need to wait for B to complete, and I do not need its result (B must be async?)
• How do I return the result of A in do_all? (A must be sync?)

Is it possible to implement this with chains or any other subtasks primitives?

Answer 1

I can't noodle a way to do this with primitives; there probably is one. But a different approach that would work is, use the on_success or after_return handler in task A to initiate task B. If you also use the ignore_result option, then the invocation of task B would be truly fire-and-forget.