2011-10-19, 01:39
tsp42 Wrote:I'm trying to reproduce the EPG/Channel crash but not luck so fare. But I did uncover a deadlock in the PVR code.
I think I might have only started getting the crashes when I started increasing the number of days in of EPG data to save to 10 or 14 days. (i.e. with the default of 2 days, I didn't see them, at least not as much). I did notice my CPU usage went up considerably while doing the reset EPG and the "save to db" option enabled when I had a larger number of days. Perhaps the problem is at the sqlite level?
Ideally I'd want all 14 days because SchedulesDirect provides the Mythbackend up to 14 days in advance, so I'd want that exposed to the PVR