#celery

Ok, I'm logging off; if you'd like to contact me: gnrhxni@outlook.com

Hi, when I do sudo /etc/init.d/celeryd status I get Error: No nodes replied within time constraint.

Everything else is working fine. -B beat status is working. I see nothing in the log file

any idea what could cause this?

if someone has an idea how to solve the status error, please tweet #ashavan thanks

hey celery ppl... have a problem, hopefully someone here can help: I'm trying to call a group directly, ie: "res = group(taskname.subtask(...) for i in arr)()" ...

the code calling the group is executing on server A, the tasks within the group are executing in a queue being processed on server B

I see, in my celery logs (using celery-django), that the leaf tasks making up the group complete successfully... but a call to "res.get()" blocks forever/never completes, printing status while waiting for ready() on the group result shows that nothing looks like it's completing as far as the group() is concerned

code works 100% afaict in my dev environment (celery 3.0.9 w/ python 2.6.5)... staging/prod where I'm having issues is 3.0.12 with python 2.6.6)

only other difference aside from minor versions is in my dev env it's all running on one machine instead of the subtasks in the group being run in a queue on another machine (ie: in dev this queue is also local)

I tried running the grouping task via delay(queue='other queue on other machine') so that it'd run on the same machine to emulate the dev setup but no diff

any hints/ideas/suggestions would be most welcome

oh backend is rabbitmq

and rabbitmqctl list_queues shows me that the subtasks are indeed completing successfully... they just never his the GroupResult

*hit

well... that's embarassing... FWIW this was the result of server B pointing to localhost for mongo (as the result backend)... doh

Hi, when I do "sudo /etc/init.d/celeryd status" I get the Error: "No nodes replied within time constraint."

Everything else is working fine. I see nothing in the log file. any idea what would cause this please?

houman: did you start a celery worker?

mher: Thanks for reply. This is my celeryd configuration: http://pastebin.com/eQDMHsHU

I have run it as "/etc/init.d/celeryd start"

And it beats and sends the task as expected. But I cant see the status

Some claim it is a permission problem on the folders.

houman: what is the output of the 'celery status' command?

mher: $ celery status --> No command 'celery' found, did you mean:

Command 'celeryd' from package 'python-celery' (universe)

celery: command not found

I have installed: pip install -U celery | pip install django-celery | pip install eventlet

which version of celery are you using?

mhre: celery==3.0.13

houman: i'd try to run a worker from console and then "/etc/init.d/celeryd status"

mhre: You mean like this? ./manage.py celery worker -E -B --loglevel=info --pool=eventlet

But if the daemon is running, why do I have to start this manually?

how can you set rate limit for a task at runtime?

celery.control doesn't seem to exist as suggested in http://docs.celeryproject.org/en/latest/usergui...

mher: sorry I mistyped your name. :) see my answer above

found it, celery control rate_limit api.tasks.photos.fetch_photo 5/s

but now it says its disabled. so I guess its off by default

sorry back.

any idea why celery control rate_limit api.tasks.photos.fetch_photo 1/s first gives OK, and on second run says rate limits are disabled?

hi guys... i have a question: Can i use Redis to store results persistently? and how would i retrieve a list of results? i can't really find something in celeryctl commands

ok, i see now in redis-cli that there are results.. will they stay there indefinitely?

bstrr: see http://docs.celeryproject.org/en/latest/configu... and http://docs.celeryproject.org/en/latest/configu... settings

To recreate a result object you can do result = YourTaskFunction.AsyncResult(task_id)

thx runeh, so i set those values to very high... can i set max-cached-results to 0 and it will keep them?

bstrr: not sure actually. Check the code perhaps. I haven't used it