Date triggered job continuously gets called

[seyed...@noaa.gov]

Hi I have a date triggered job that I schedule using the following

            scheduler.add_job(MY_PROCEDURE, trigger='date', args=[PARAMETERS....],

                              next_run_time=(datetime.now() + timedelta(days=2)))

So the thought is that MY_PROCEDURE gets called 2 days from "now" however it seems to call MY_PROCEDURE on a daily basis.

Am I using this wrong?

[Alex Grönholm]

Yes, you're supposed to use the run_date parameter. See the documentation of the date trigger: http://apscheduler.readthedocs.io/en/latest/modules/triggers/date.html#module-apscheduler.triggers.date

--
You received this message because you are subscribed to the Google Groups "APScheduler" group.
To unsubscribe from this group and stop receiving emails from it, send an email to apscheduler...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[seyed...@noaa.gov]

Got it...so this should work right? (meaning two days after today and only run once)

       scheduler.add_job(MY_PROCEDURE, 'date', run_date=(datetime.now() + timedelta(days=2)), args=[PARAMETERS....])

[Alex Grönholm]

Yes.

[seyed...@noaa.gov]

Thanks Alex, that did the trick. What's strange is is that once in a while I get this error

TypeError: unsupported operand type(s) for -: 'decimal.Decimal' and 'float'
[2016-10-29 14:52:01,598] ERROR in wsgi: Application exception:
Traceback (most recent call last):
  File "/var/www/wsgi-scripts/BudgetDB/miniconda3/envs/budgetdb-py34-env/lib/pyt                                                                  hon3.4/site-packages/apscheduler/schedulers/base.py", line 876, in _create_plugi                                                                  n_instance
    plugin_cls = class_container[alias]
KeyError: 'date'

I am running the scheduler as described in the previous post

  scheduler.add_job(MY_PROCEDURE, 'date', run_date=(datetime.now() + timedelta(days=2)), args=[PARAMETERS....])

Any thoughts?

[Alex Grönholm]

That error would suggest that you haven't properly installed APScheduler, so the entry point for "date" is missing. Try instantiating the trigger manually and passing it to add_job() instead of 'date'.

[seyed...@noaa.gov]

On server startup, I instantiate APScheduler like this

jobstores = {
    'default': SQLAlchemyJobStore(url='sqlite:///jobs.sqlite')
}
executors = {
    'default': ThreadPoolExecutor(20),
    'processpool': ProcessPoolExecutor(5)
}
job_defaults = {
    'coalesce': False,
    'max_instances': 3
}
scheduler = BackgroundScheduler(jobstores=jobstores, executors=executors, job_defaults=job_defaults)
scheduler.start()

Do I need to make a change to this to make "date" entry point available? Outside of code, I just did a pip install. 

If it's not my initialization above, how do I manually instantiate the trigger? I have checked out https://apscheduler.readthedocs.io/en/latest/modules/triggers/date.html#module-apscheduler.triggers.date

Do I just do trigger = apscheduler.triggers.date?

[Alex Grönholm]

Pip installing should have been enough. But to manually instantiate the trigger, simply import the appropriate class (apscheduler.triggers.date.DateTrigger), create an instance of it and pass it to add_job() in place of 'date'.

[seyed...@noaa.gov]

Awesome, thank you. That worked.