scheduler support coming .....
I think you didn't understand how models works / the scheduler works. Sorry
for probably being redundant/naive, but just to check.....
in models belong:
def myfunction():
####whatever ####
from gluon.scheduler import Scheduler
myscheduler = Scheduler(db)
This line basically give you access to the scheduler api and the underlying
tables.
Now, the following line
myscheduler.queue_task(
function='my_function',
task_name = 'my_function',
repeats = 0, # unlimited runs
period = 300 # every 5 minutes )
that can be shortened as well as
myscheduler.queue_task(
my_function,
repeats = 0, # unlimited runs
period = 300 # every 5 minutes )
does an actual insert on the scheduler_task table.
That should be used whenever you need to queue a NEW task.
For repeating task, executing that line *one time only* it's all it's
needed.....
So, if that's your only function, you can either use that line in a
controller, or just use the appadmin inserting a *single* row on the
scheduler_task table. Repeating functions will be repeated with just ONE
row in the scheduler task (where the repeats column != 1) and they'll
honour period. status, next_run_time, last_run_time, times_run,
times_failed will be updated accordingly to executions on that *single* row.
NB: If the function returns something (and doesn't raise an exception),
every execution will create a row in the *scheduler_run* table.
I think that you have myscheduler.queue_task() in your models, and
basically myfunction gets queued every time over and over (because models
are executed at every request). That's why you have multiple records in the
scheduler_task table and observing "strange" behaviours.
Please let me know if this explanation made sense and if not, feel free to
ask for further details.
On Monday, January 21, 2013 2:31:48 PM UTC+1, Stormcrow wrote:
>
> I have created a *scheduler.py *in /models/ which looks something like
> this:
>
> def my_function():
> ### function body ###
>
> scheduler = Scheduler(db)
> scheduler.queue_task(
> function='my_function',
>
> task_name = 'my_function',
>
> repeats = 0, # unlimited runs
>
> period = 300 # every 5 minutes
>
> )
>
> The problem I'm seeing is that multiple tasks are being created in the
> database for this one function at totally incorrect *next_run_time* with
> respect to *start_time* and *period*. The multiple tasks created in the
> db are also not at the correct time interval apart. In addition, changing
> the repeat value to something other than zero has not effect, tasks
> continue to be created and executed unrelentingly according to what I see
> in the database.
>
> I am very keen to use the Scheduler, it seems like an elegant solution to
> what I'm trying to do but with all these tasks sprouting and executing at
> (what appears to be) random it is basically broken. Apologies if I
> misunderstood the usage of the scheduler, I was proceeding as close as I
> could to what I deciphered from the book.
>
> Thanks in advance for the help!
>
--
