#pyramid

Looking into implementing a calendar where people can add appointments. Is celery and scheduling a task with eta=... a good, scalable solution here, or are there better ones? E.g. using a daily midnight celerybeat that schedules the tasks for the next 24h?

i'm about to set up my first public demo site with a database, and i want it to get wiped and reinstalled every hour. is there a preferred way of going about it? any pitfalls to avoid? i haven't thought much more than use cron to execute a DROP and execute an SQL statement to rebuild the dropped table

steverpiercy: i wipe redis and restart celery as well. just to ensure a clean slate

good point. no caching going on here

stevepiercy: and make sure no data lingers in sqlalchemy ;)

celery is already part of my server setup, which is why i thought into that direction. apscheduler looks similar, so i don't see a benefit from adding that one...

celery eta isn't quite precise though, if i understand the documentation correctly.

stevepiercy: drop all and create all should do it

_savage: stevepiercy: apscheduler is more of a scheduling within a single process

moo-_-: so Celery task with eta is a scalable solution?

[waitress] bertjwregeer pushed 1 new commit to master: https://git.io/vDGBc

waitress/master 094d1cd Bert JW Regeer: Merge pull request #153 from Pylons/bugfix/nameinfo...

_savage: celery can handle billions of events, so ye

moo-_-: the comment that eta is not exact makes me nervous though. if I schedule at time X then i'd like the task to start at time X.

_savage: I don't understand your use case

_savage: you are talking about appointments

Pylons/waitress#207 (master - 094d1cd : Bert JW Regeer): The build passed.

Change view : https://github.com/Pylons/waitress/compare/9ecb...

Build details : https://travis-ci.org/Pylons/waitress/builds/19...

do you want to send out a notification to the user on a certain millisecond or a certain nansecond?

moo-_-: oh... the X+delta is not in the minutes? if it's a second off that's just fine

moo-_-: i interpreted the delta as seconds, even perhaps minutes

_savage: it doesn't matter what scheduling system you use

the failure mode is always the same

it cannot process all the tasks within the time frame

and it means your code runs too slowly that all tasks could be processed within th deadline

moo-_-: right

_savage: furthermore you can instruct celery to drop the tasks if the deadline expections are not met

moo-_-: ah ok, nice.

moo-_-: one thing i didn't quite understand is where celery stores the schedule data... redis? (that's what i've set it up with)

moo-_-: so if celery workers restart, redis will hold the schedules? just need to make sure the redis instance is backed up ok?

moo-_-: and yes, i'm talking about appointments

morning!

hello

encolpe: disko yo

morning

morning