[mythtv-users] strange real-time commercial flagging behavior

[Peter Lee]

On 6/2/05, Chris Pinkham <cpinkham at bc2va.org> wrote:
> > Basically, on some recordings (about 1 in 10, I would say), the
> > real-time flagging for a program appears to "finish" well before the
> > program has actually finished recording.  At least this is what the
> > log says.  For example, here are some log entries from today, to
> > illustrate the phenomenon:
> 
> Can you try turning off the ~realtime flagging and verify that this
> you aren't still getting corrupt recordings.  If the recording is
> failing on it's own for some reason then it makes sense that the
> flagger would finish early.  I haven't noticed this issue but I may
> have missed it, so I'd like to verify you aren't getting corrupt
> recordings which is causing flagging problems.

Hi, Chris.

Well, I ran my system this weekend with the ~realtime flagging turned
off, and every recording (9 in total, I think) was made perfectly. 
So, while not totally conclusive, it does appear that turning off the
~realtime feature fixes my problem.

However, I wonder if I might be having some kind of performance
problem?  Maybe the extra CPU load or disk activity of ~realtime
flagging causes recordings to fail for some reason on my system? 
Seems unlikely, but maybe another experiment to run is to do some big
nuvexport jobs for a couple of days, to see if this has any impact on
the recordings.

Anyway, the bottom line is that I no longer get backend crashes or
corrupt recordings after turning off the realtime flagging.

Peter