Thanks a lot for your detailed explanation for the difference between Celery and scheduler! I think currently scheduler is enough for our application, and it is easier to implement (thanks to great work of your team). May I have one more question. If I want to allow user to delete one of his projects, which means the task associated with the project will also be deleted, is there a simple command in scheduler to do that. In the data table scheduler_worker and scheduler_run I want to delete both entries. Should I use db(db.scheduler_*.uuid==<an_id>).delete() for both of them, or there is a single command like scheduler.delete_task(uuid) to do that. I am not sure if scheduler.terminate(), scheduler.stop_task() or scheduler.kill can do that.
On Friday, October 3, 2014 2:23:22 AM UTC-4, Niphlod wrote: > > > > On Wednesday, October 1, 2014 4:34:08 PM UTC+2, Pengfei Yu wrote: >> >> So that means all the communications between different machines are from >> database management system, and has little thing to do with the scheduler? >> Sorry, I am new to this field and the question may be stupid. >> > > all the communications between any process sitting on any system just > needs to access the database you point your scheduler instance to. > > >> And I see Celery has features to support distributed computing ( >> http://stackoverflow.com/questions/23916413/celery-parallel-distributed-task-with-multiprocessing), >> >> Are those features also included in Scheduler? >> >> > same exact thing, two different concepts in mind: web2py's scheduler > assumes that the time it takes you to send a task is neglegible in regard > of the actual time the task will take to be processed, while celery use a > software stack that tries hard to minimize that timespan. > Also, consider that the whole Celery team is focused on the creation of a > task processor, while web2py's is not (just sayin') > In real world terms, this means that if your goal is to process 1 trillion > task per minute, Celery (and a big freaky farm of servers) is the right > tool for the job. > Don't even start with web2py's scheduler with those numbers. > web2py's scheduler performances are hardly dependant on the underlying db > (since it's where all of the communication and coordination for all the > activities takes place). Scheduler's usually runs out (on "commodity > hardware") at 2000 tasks per minute. Still, it's usually enough for a vast > majority of web applications. > -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- 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/d/optout.
