:-D Thanks Niphold, using repeats = 2 and no stop_time is the best solution. I did not realize that I could do it that way.
Thanks again. Best regards. 2013/8/3 Niphlod <[email protected]> > ok, explanation is clearer. Although scheduler was not meant to be > executing task on a precise-timing (its a queue processor, not cron), but > instead it executes the tasks that are available, all that is needed is for > you to understand how it works. > > It's not as precise as cron, because scheduler polls every x seconds the > scheduler_task table to see if there is anything to do, and then spawns a > process to execute the task. > If you need your mails to start exactly at 12:00:00 and at 14:00:00, I'm > assuming you understood how to make it work in the wrong way. > > Scheduling a start_time of 12:00:00 and a stop_time of 14:00:00 assures > you that no tasks will be picked before 12:00:00 and later than 14:00:00. > If you need that task to fire two times only, then setting a period of 4 > hours will never get you 2 executions, because if the first task is picked > around 12:00:00 it will be "rescheduled" for execution 4 hours later, at > 16:00:00. The point is, you set the task to never execute past 14:00:00, so > the second execution will never take place at 16:00:00. > > Be careful though: the scheduler can "spend" some seconds of "sleep" > before picking the tasks, so calculating those times so precisely can get > you in the situation where you set 12:00:00 as start_time, 14:00:00 at > stop_time and 7200 as period, but never get to the second execution. > > That's because scheduler uses the time the task was picked up (that may > very well be 12:00:05) to calculate the next time a task will be > effectively be able to be picked up, and in that case, it will be 14:00:05. > > If so, you have a few choices (based on your loads - meaning how many > concurrent task you may have at 12:00:00), the simplest of them all (given > your current mindset) is to set stop_time at 14:01:00 instead of 14:00:00, > assuming you can "forgive" that one minute ahead. > > But, if I understood your requirements, why don't you just use repeats=2, > start_time=12:00:00, period=7200 and no stop_time at all ? > This will guarantee that the first execution will take place after > 12:00:00 (again, assuming a normal load, it could be picked up a few > seconds later, e.g. 12:00:15) and the second one at , e.g., 14:00:15 , and > then the task will result as COMPLETED after exactly 2 executions. > > > > -- > > --- > 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.
