Re: [vdr] Device power saving feature - power measurements

2016-06-30 Thread Richard F
On 28/06/2016 03:14, glenvt18 wrote:
> Hi Richard,
>
> As there are no debug messages in your log I tried to reconstruct some
> of them (power-related) - see my comments. Note that an extra second is
> added to all timeouts due to the 'greater than' comparison. It was done
> on purpose to add some save margin. Power-saving looks to work OK. But.
> I don't see any 'timer ... start/stop' or 'record ...' messages in your
> log. So, I have some questions:
Yes, I turned off debug as it's been running awhile.
> 1) Did you fail to record 'The Big C & Me' at 01:05 and 'Greenberg' at
> 01:35?
No, that was just a (standard) notification from VDR that the running
status of a program had changed
> 2) Did you cancel 'Greenberg'?
No, VDR was idle before I started up Kodi/VNSI at around 01:00.

Actually looking back through the log some more, it seems that Kodi had
one of it's regular hiccup startup problems here.
Basically it starts, then crashes, then starts, often 2 or 3 times. This
is a known issue - the Kodi devs say they've changed the startup
mechanism in newer versions, but these aren't stable enough for regular
use for various other reasons.
> 3) Was your streaming interrupted?
> 4) What is your vnsi streaming priority?
Not sure I have a VNSI priority, at least there isn't a setup variable
for it.
> 5) What is your recording priority?
Generally 50 or 99. But in this case it wasn't recording, just streaming.
> 6) Are the channels you recorded and the channels you streamed on the same
> transponder? I'm asking this because I see device 1 being idle all that
> time.
As above, there was only 1
> 7) What happens with your VDR when a vnsi streaming is interrupted by a
> higher priority recording on the same device but on another transponder?
Not tested.
> 8) What timers (their dates) did you have from Jun 24 00:00 to Jun 24
> 04:00? 
None
> 9) What is your Vps Margin setting?
600

On reflection I think this is really a VNSI problem, possibly a
side-effect of locking up the VNSI process.
I have had a similar situation in the past (before adding your patch)
where VNSI server plugin locks up after a client hiccup start, then any
client trying to connect keeps trying to attach and times out. This time
the client didn't time out (as I recall), but it was late, so I left it
and went to bed...!

I'll monitor for any other effects.

thanks
Richard
>> Jun 24 01:12:59 ha-server vdr: [8690] loading 
>> /etc/vdr/plugins/vnsiserver/allowed_hosts.conf
>> Jun 24 01:12:59 ha-server vdr: [8690] VNSI: Client with ID 75 connected: 
>> 10.0.0.90:52606
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: Welcome client 'XBMC Media 
>> Center' with protocol version '8'
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: LiveStreamer::Close - close
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: close video input ...
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: Successfully found following 
>> device: 0x91cef0 (1) for receiving
>> Jun 24 01:12:59 ha-server vdr: [15789] dvb tuner: power-up - opening 
>> frontend 0/0
> streaming started on device 0, wake up device 0
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: activate live receiver: 1, pmt 
>> change: 1
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: Successfully switched to 
>> channel 1 - BBC ONE Lon
>> Jun 24 01:12:59 ha-server vdr: [15789] VNSI: Started streaming of channel 
>> BBC ONE Lon (timeout 5 seconds)
>> Jun 24 01:12:59 ha-server vdr: [15793] VNSI: Created stream for pid=101 and 
>> type=7
>> Jun 24 01:12:59 ha-server vdr: [15793] VNSI: Created stream for pid=102 and 
>> type=2
>> Jun 24 01:12:59 ha-server vdr: [15793] VNSI: Created stream for pid=106 and 
>> type=2
>> Jun 24 01:12:59 ha-server vdr: [15793] VNSI: Created stream for pid=105 and 
>> type=9
>> Jun 24 01:13:00 ha-server vdr: [8686] channel 3 (BBC TWO) event Fri 
>> 24.06.2016 01:05-02:05 (VPS: 24.06. 01:05) 'The Big C & Me' status 4
>> Jun 24 01:13:05 ha-server vdr: [8679] dvb tuner: power-down - closing 
>> frontend 1/0
>> Jun 24 01:13:10 ha-server vdr: [8679] dvb tuner: power-up - opening frontend 
>> 1/0
> tune to a transponder on device 1 to refresh EPG before a (VPS) timer starts, 
> call cDevice::SetOccupied(), set device 1 idle timer to 38 sec
>> Jun 24 01:13:46 ha-server vdr: [15789] VNSI-Error: cxSocket::read(fd=18): 
>> read() error at 0/4
>> Jun 24 01:13:46 ha-server vdr: [15793] VNSI: exit streamer thread
>> Jun 24 01:13:46 ha-server vdr: [15789] VNSI: LiveStreamer::Close - close
>> Jun 24 01:13:46 ha-server vdr: [15789] VNSI: close video input ...
>> Jun 24 01:13:46 ha-server vdr: [15789] VNSI: activate live receiver: 0, pmt 
>> change: 0
>> Jun 24 01:13:46 ha-server vdr: [15789] VNSI: close video input ...
>> Jun 24 01:13:46 ha-server vdr: [8692] VNSI: Client with ID 75 seems to be 
>> disconnected, removing from client list
>> Jun 24 01:13:49 ha-server vdr: [8679] dvb tuner: power-down - closing 
>> frontend 1/0
> device 1 has been idle for 38 sec - power it down
>> Jun 24 01:13:51 ha-server vdr: [8690] l

Re: [vdr] Device power saving feature - power measurements

2016-06-30 Thread glenvt18
> No, that was just a (standard) notification from VDR that the running
> status of a program had changed
Sorry. I was a bit sleepy that night:) It was a false alarm.
> Not sure I have a VNSI priority, at least there isn't a setup variable
> for it.
It's under Settings->Add-ons->My add-ons->PVR clients->
VDR VNSI Client->Configure->General->Priority. I have it set to 0.
> > 8) What timers (their dates) did you have from Jun 24 00:00 to Jun 24
> > 04:00? 
> None
> > 9) What is your Vps Margin setting?
> 600
In this case your VDR might have been trying to update EPG for the
timers scheduled later, probably, next night. Anyway, it's all about
VDR's internal logic for updating timers' EPG events and has nothing to
do with power saving. Every time a VDR decides to update EPG for a
channel with a timer scheduled the device's frontend must be opened
prior to zapping and kept open for a while (38 sec as you see in the
log). If you have power-related issues you're likely to see ioctl()
errors in the log and emergency exits after a recording starts (VDR will
be restarted in this case).
> I'll monitor for any other effects.
Let me know.

-- 
glenvt18

___
vdr mailing list
vdr@linuxtv.org
https://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr