Closed t-io closed 7 years ago
After setting 'exchange': 'parser', 'routing_key': 'parser'
in CELERY_TASK_ROUTES
settings and adding my celery app to the proc everything runs fine
It's very interesting that you seem to have found a workaround for this issue. AFAICT, this is a bug in Celery. celery/celery#3767. The workaround isn't a general solution, but it may work for some people that need to have a workaround, thank you.
Since it's a bug in Celery, it's not something that I can reasonably fix. However, The latest master includes a new feature, called simple_queues
, that allows you to tell HireFire that your queue names are the same as the exchange and the routing_key. This handily side-steps the issue for simple cases, but it doesn't fix it.
Hi, I'm using celery 4 with the latest hirefire. Scaling the worker up works fine but after this the communication with hirefire seems broken. I get the same results with the latest master and release.