View Single Post
Originally Posted by curt.clifton View Post
I wonder if there's a seconds/milliseconds error. 180000ms in the future, would by 3 minutes, a much more reasonable interval.
Hmm… 180000 isn't a units problem, it does refer to seconds: that's 50 hours, which is 48 hours (the configured "due soon" interval) plus two hours (the maximum amount of time the scheduler is supposed to sleep between refreshing the schedule), so that's the limit date in its database query.

But wait a minute! That's it, isn't it? It looks like that log shows the scheduler incorrectly using that limit date as its wakeup time, so it goes to sleep for 50 hours rather than two.

I'll go check the code…