Using the following settings and calling my_task_results.get() hangs:
CELERY_RESULT_BACKEND = "cache"
CELERY_CACHE_BACKEND = "memory"
Using other backends seem to work as expected. No errors reported. In fact, celery monitor shows all tasks were successful.
Not sure how this would work exactly, the worker does not have access to the memory of the parent...
If you want to use the in-memory cache backend you would have to write the result before
you can retrieve it:
>>> result = add.delay(2, 2)
>>> add.backend.mark_as_done(result.id, 4)