more than celery we'd need a generalized web2py "create the context" recipe .... running tasks defined in modules is easy..... running tasks outside web2py that needs the usual environment is a PITA ....
On Tuesday, March 12, 2013 6:17:47 PM UTC+1, Eric S wrote: > > > I'm interested in a robust, widely-adopted scheduler. The current web2py > Scheduler is clearly changing very rapidly, which is great, for now I want > a scheduler that is mature. > > Can anyone answer my original question -- how have you gotten Celery > workers to run with web2py? > > > On Thursday, March 7, 2013 9:47:25 AM UTC-8, Massimo Di Pierro wrote: >> >> What I want to know is what do you think celery buys you that the >> built-in scheduler does not provide? >> >> Celery is faster at transferring messages from the application t the >> workers and vice versa but normally when you want to run background tasks >> you have different bottle-necks: computation cycles of background tasks and >> database access from the tasks. In this respect I do not see celery being >> any better than the built-in scheduler. Actually the built-in scheduler >> makes your life easier by exposing the web2py environment to workers which >> is something celery would not be able to do. >> >> On Wednesday, 6 March 2013 21:11:59 UTC-6, rochacbruno wrote: >>> >>> I also would like to see Celery, Solr, Elastic Search and other >>> fantastic tools working with web2py! >>> >>> I think this is an important issue and I am sure it is completely easy >>> and possible to make it. >>> >>> I personally do not like to use the built-in scheduler, so I am using >>> python-rq (Redis Queue) for some production sites and it works very well >>> and offer almost all celery functionalities. >>> >>> Maybe someone can follow my python-rq[1] example and create a >>> wen2py-celery tutorial >>> >>> [1]http://rochacbruno.com.br/web2py-and-redis-queue/ >>> >>> I dont think web2py needs to always reinvent the wheel so I would like >>> to see more integrations. >>> >>> wish list: >>> >>> Whoosh (WIP) >>> Solr (maybe a haystack clone for web2py) >>> ElasticSearch >>> Celery >>> Thumbor >>> Neo4J >>> >>> >>> On Wed, Mar 6, 2013 at 10:47 PM, Eric S <[email protected]> wrote: >>> >>>> >>>> I would like to use Celery in my web2py application, but I'm having >>>> trouble with how to start a Celery worker (I know there is a web2py >>>> scheduler but I would like to use Celery). >>>> >>>> To start a custom scheduler in web2py I would use: >>>> python web2py.py -S appName -M -R worker.py >>>> >>>> Celery workers, however, are launched from the command line such as >>>> with the following command, which won't easily substitute into 'worker.py' >>>> above: >>>> celery -A tasks worker --loglevel=info >>>> >>>> Has anyone had success integrating web2py and Celery? How do you >>>> (robustly) get around this problem? >>>> >>>> Thanks, >>>> Eric >>>> >>>> -- >>>> >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "web2py-users" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> For more options, visit https://groups.google.com/groups/opt_out. >>>> >>>> >>>> >>> >>> -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
