Forum OpenACS Development: Re: sec_sweep_sessions can hang forever if database breaks

Yes I was about to point that out that "excessive time" entry as its exactly what happened to us with two scheduled events sharing the main thread.

Its actually easy to setup a schedule in its own thread you just use the -thread switch thus:

ad_schedule_proc -thread t -debug t 600 backoffice::export::send_promos
or
ad_schedule_proc -thread t -debug t -schedule_proc ns_schedule_daily {04 30} backoffice::export::send_stock_status

The downside I believe is that it sets up and tears down its own interp - but thats not usually too much of a problem.

- Steve

Steve, yes, the -thread option to ns_schedule_proc (which ad_schedule_proc just passes along) is what I referred to above. If I remember correctly, that starts up a new thread for every individual run of the job, and then kills it. If the main scheduler thread is locked up, will the thread for the new job ever get spawned? I'd need to test it, but I suspect the answer is no, it will not.