On 1/06/2016 03:49, glenvt18 wrote:
From what I can see, everything works as expected. You can safely lessen
the power-down timeout to 3 minutes. This will be enough for EPG scan as you have at most 8 transponders/multiplexes which gives 8 * 1/3 = 2m40s
- see 'EIT scan ...' messages.
I reduced it to 5 mins to give it some margin
As for those small intervals of activity you mentioned, it happens because some 'preparation' work is done before a timer starts. VDR tries to switch to the timer's transponder and gather EPG data. For a VPS timer this begins within VPS margin and an hour before the timer start time. For a non-VPS timer it begins one minute before the timer start time. I've added 30 seconds to the occupied timeout (see my cDevice::SetOccupied() comment) to avoid frequent on/off switching. So, you're likely to see 38-second (or, sometimes, a bit more) periods of activity. If you still worry about it, keep you server idle for 2 hours before a nearest VPS timer starts and paste the log.
OK, I'll keep an eye on it. No adverse effects yet - good work, thanks.
I'm doing some other testing and stuff atm so want to keep the server up - I'll possibly take it down at the weekend, make some power measurements and post them
I've removed 'sudo' from the monitor script: http://pastebin.com/8JDffAXF
Run it as user 'vdr' for some time. It will draw you a nice 'map' of your devices' activity.
If you notice channel switch failures, let me know.
Do your tuners get warm when active?
Yes, the TV tuners always did run warm, and I recall that when VDR is stopped the power is several watts lower (not because of VDR/CPU load itself according to stats, it's when the tuners go active - hence my interest in your patch).
As the server also runs the home automation/lights/heating/security, CCTV, mail server, and file storage/Owncloud it's not an option to shut it down when not recording, and I have t choose when to take it down !
Richard
Thanks again for testing.
Best, Sergey Chernyavskiy.
On Tue, May 31, 2016 at 08:43:01PM +0100, Richard F wrote:
OK Sergey, I've posted a log with your patch applied + log level 3 - about 3 hours worth There was a recording 16:15 to 16:45, other than that idle I can run it for longer if you need it, though it's creating a fair bit of log. I seem to get a lot of erroneous EPGsearch results at level 3, not sure why. http://pastebin.com/8DbYQdTX
Let me know what it tells you Thanks Richard
On 30/05/2016 15:19, glenvt18 wrote:
Could you attach (or, better, upload to pastebin.com) the whole log file. Are you running vdr with --log=3? I need to see 'idle timer' debug messages and some other debug stuff. Could you also apply this change
diff --git a/eitscan.c b/eitscan.c index 3899e00..b1c8c13 100644 --- a/eitscan.c +++ b/eitscan.c @@ -159,7 +159,7 @@ void cEITScanner::Process(void) Skins.Message(mtInfo, tr("Starting EPG scan")); } } - //dsyslog("EIT scan: device %d source %-8s tp %5d", Device->DeviceNumber() + 1, *cSource::ToString(Channel->Source()), Channel->Transponder()); + dsyslog("EIT scan: device %d source %-8s tp %5d", Device->DeviceNumber() + 1, *cSource::ToString(Channel->Source()), Channel->Transponder()); if (lastActivity == 0) // forced scan - set idle timer for each channel switch; // this prevents powering down while scanning a transponder --
Thanks for testing.
Best, Sergey Chernyavskiy.
_______________________________________________ vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr