#celery

asksol: yeah. all of our values are json compatible. If you're not objecting, that gives me a bit more confidence to proceed.

bmbouter: we should probably remove the __del__

bmbouter: that's probably why they're not collected in the first place

as defining __del__ means the gc cannot collect cyclic references

I removed one __del__ in py-amqp recently, inherited from amqplib. In my experience you never really need them, they're just used to enable sloppy programming

there's no __del__ in kombu master at all afaict

hey is anyone here?

asksol: ping

hi

dis there a way to create a task that returns a group() from an input list ? (like the dmap example) the only solution I have seems to work, but returns an asyncresult containing the groupresult, so I cannot pipe it

the dmap function described here: http://stackoverflow.com/questions/13271056/how...

Any tips on how to start debugging memory usage? Our nodes takes 1-2GB of RAM (per process) on start.

using RabbitMQ as a broker, celery==3.1.23, librabbitmq==1.6.1

Hey guys!

I have a question: does schedule tasks allocate a worker?

Because I have a service which launch a lot of tasks and I expect to have 5 worker at the same time, but I have only 3 running.

(I have 2 scheduled tasks but they run at 2 PM)

Hi . How can celery read static file ('/static/') from django?

gulzar: how can something read a static file from django?

malinoff: there are static files like css, or .db files. I have one such salite.db which I need to pass to a task in celery

gulzar: you didn't answer my question

malinoff:Oh sorry. I only know to read css files using 'load static' in html

malinoff: urls.static have static function , I think that can work

malinoff: thank you

gulzar: i wouldn't suggest to rely on celery being installed on the same server as your django app. Django works via http, and its use-case is to load resources via http. Why don't simply use requests and load these static files via http from django?

Hi. We're investigating an event driven architecture of different webservices.

I was wondering whether we could achieve the following with celery: https://docs.google.com/drawings/d/1jh4gxya7qFj...

malinoff: I am noob. that tech went above my head

malinoff: I am a biologist and not a web developer, coding is my hobby and i have few projects in my mind so working with python

gulzar: well, it would be very difficult to you to write a sane django app if you don't know http so i'd suggest to read about it first

to make a couple of manual requests via telnet, curl etc

cecemel: why HTTP as delivery method/protocol? AMQP is much more robust and easier to use

cecemel: also what you named "celery broker" is actually just "broker", for example, rabbitmq

cecemel: celery can be used as emitters (task.apply_async) and as workers (connection.drain_events) so services 1, 2 would be custom code calling task.apply_async and services 3, 4 would be celery workers

malinoff: http is for convenience/infrastructural constraints. Are Webhooks not ok?

cecemel: not when it comes to task queues. Use rabbitmq/amqp

I have a question: does schedule tasks allocate a worker?

Because I have a service which launch a lot of tasks and I expect to have 5 worker at the same time, but I have only 3 running.

(I have 2 scheduled tasks but they run at 2 PM)

Debnet: no, they don't. Workers are usually pre-allocated

OK. Thanks.

malinof: what are typical use cases for Webhooks then?

cecemel: I honestly don't know. Never used them. But they're removed in incoming 4.0 release so you should definitely not rely on them: https://github.com/celery/celery/blob/master/do...

eofs RE: debugging celery memory usage:: I'm assuming you have already read the http://docs.celeryproject.org/en/latest/usergui...