1. When a task goes into TIMEOUT, it needs to be carefully reviewed. TIMEOUT means that something that you needed isn't there... your case seems to contemplate a missing run, but if it's so, you need to manage the "absence" in your code (i.e. it must be eventually "planned"). To summarize, a TIMEOUT task is something that you DIDN'T plan to go on timeout....if instead you are cool with that, you should manage it into the task itself, e.g. raising an error if n seconds are passed (so the task gets into the FAILED status)
2. if you modify directly the scheduler_task table, pretty much ALL fields are important, because every one of them specifies how, when, etc a task needs to run 3. if you need it monthly, why the need for "number of seconds" to be variable ? In any case you'd better off queuing something that gets fired every once in a while and schedules the task only when it's needed On Friday, April 25, 2014 7:00:18 PM UTC+2, David Rager wrote: > > I'm just a tad confused about the scheduler - just 3 items I haven't been > able to figure out and handle: > > 1. When a scheduler_task times out I see the status of TIMEOUT. It > appears the task will not run again for the next scheduled run. Is this > the expected behavior or am I misinterpreting? Because I work with > external resources that are not always reliable I expect timeouts but would > like the task to run on the next scheduled try if possible. > > 2. When entering or modifying scheduler_task items is it important to > update the Next Run Time to a future data in order for the task to run? I > have been doing this because it seems logical. For instance - after a > TIMEOUT when I update the task to QUEUED do I also need to modify the Next > Run Time to get it to go again? I believe this is the behavior but wanted > to clarify. > > 3. Is there a method to run a task monthly (where number of seconds is > variable)? > > I'm contemplating writing a schedule manager of a sort which would: > 1. capture the status of all the tasks (just a snap of the table) > 2. flip the TIMEOUTs to QUEUED for me every morning so they will continue > to run > 3. update the Next Run Time on timed out jobs and monthly jobs on the flip > of the month > > I would use a scheduler_task to do this (with the hope that task would not > timeout or stop executing for whatever reason) - but perhaps I'm > misinterpreting how this is working from playing with it this week. > > Appreciate any clarification - this is my first attempt to harden the > scheduler for production use. > > TY! > -- 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 web2py+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
