Re: [vdr] broken recordings

2016-08-08 Thread cedric.dew...@telfort.nl 

>
>You might also try to replace your raspberry Pi - maybe there is
>another problem somewhere (if you already did this, sorry, I
>must have over-read it).
>
Hi,

If you want to replace hardware, may I suggest using a board based on the 
allwinner A20, like the bananapi or the A20-OLinuXino-LIME? I had one of those, 
and it's a powerful beast. It has dedicated network and SATA, so the USB ports 
are free to talk to the tuner. One thing though, it does have accelerated video 
output, (the sintel trailer can be played while the CPU only deals with the 
audio), but I don't know if VDR already is capable of taking advantage of it.

Cheers,
Cedric

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

Re: [vdr] broken recordings

2016-08-08 Thread Patrick Boettcher 
On Mon, 8 Aug 2016 22:51:13 +0200
Patrick Boettcher  wrote:

> Hi,
> 
> On Mon, 8 Aug 2016 20:30:33 +0200
> Matthias Bodenbinder  wrote:
> 
> > Hello Christoph,
> > 
> > based on your feedback I made another test. The USB HD performance
> > seems to be ok (see my other reply). But anyways I made a test with
> > recording directly to the Flash SD card. And that works pretty well.
> > 15 min without issue. So it looks like it is indeed an issue with
> > USB on the Raspberry PI 2. Any idea how to solve that?  
> 
> It _could_ be the write-cache-flush which saturates the bus and then
> dramatically decreases I/O of the overall system.
> 
> Try
> 
> hdparm -W 0 /dev/
> 

You could also try iotop, which should I/O activity of all processes,
maybe there is something going on.

sudo apt install iotop
sudo iotop

Thanks,
-- 
Patrick.

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

Re: [vdr] broken recordings

2016-08-08 Thread Patrick Boettcher 
Hi,

On Mon, 8 Aug 2016 20:30:33 +0200
Matthias Bodenbinder  wrote:

> Hello Christoph,
> 
> based on your feedback I made another test. The USB HD performance
> seems to be ok (see my other reply). But anyways I made a test with
> recording directly to the Flash SD card. And that works pretty well.
> 15 min without issue. So it looks like it is indeed an issue with USB
> on the Raspberry PI 2. Any idea how to solve that?

It _could_ be the write-cache-flush which saturates the bus and then
dramatically decreases I/O of the overall system.

Try

hdparm -W 0 /dev/

which might work. I had mixed results with this command on a PC back
then when I wanted to store a continuous stream which came in at 33
MiB/s and lead to data drop due to write-caches on an SSD...

My parents use your setup with an HDD connected to USB and receiving
all DVB over network (which is USB on the RPI2) and have no problems.

You might also try to replace your raspberry Pi - maybe there is
another problem somewhere (if you already did this, sorry, I
must have over-read it).

regards,

-- 
Patrick.

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

Re: [vdr] broken recordings

2016-08-08 Thread VDR User 
It really does sound like you're saturating the usb bus. Are you sure
you're comparing against real world performance and not
technical/theoretical?

On Mon, Aug 8, 2016 at 11:30 AM, Matthias Bodenbinder
 wrote:
> Hello Christoph,
>
> based on your feedback I made another test. The USB HD performance seems to 
> be ok (see my other reply). But anyways I made a test with recording directly 
> to the Flash SD card. And that works pretty well. 15 min without issue.
>
> So it looks like it is indeed an issue with USB on the Raspberry PI 2. Any 
> idea how to solve that?
>
> Matthias
>
>
> Am 08.08.2016 um 11:10 schrieb C.Scheeder:
>> Hi,
>> it meuns your system is not able to process all the data received in time.
>> where do you write your recordings?
>> I guess its a usb-harddisk connected to your raspi, correct?
>> It's probably just to slow to write all the data vdr throws at it.
>> Or your usb-bus is saturated with all the data flowing in from your receiver 
>> and back out to the harddisk.
>> Christoph
>>
>> Am 07.08.2016 um 16:34 schrieb Matthias Bodenbinder:
>>> One more hint to the problem.
>>>
>>> Another test with recording Arte HD gives the following output in user.log:
>>>
>>> Aug  7 16:31:14 raspberry vdr: [28148] timer 1 (5 1503-1655 'Galaxis 
>>> Milchstraße') set to event Son 07.08.2016 15:05-16:45 (VPS: 07.08. 15:05) 
>>> 'Galaxis Milchstraße'
>>> Aug  7 16:31:19 raspberry vdr: [6438] i/o throttle activated, count = 1 
>>> (tid=6438)
>>> Aug  7 16:31:22 raspberry vdr: [6438] buffer usage: 70% (tid=6448)
>>> Aug  7 16:31:24 raspberry vdr: [6438] buffer usage: 80% (tid=6448)
>>> Aug  7 16:31:26 raspberry vdr: [6438] buffer usage: 90% (tid=6448)
>>> Aug  7 16:31:29 raspberry vdr: [6438] buffer usage: 100% (tid=6448)
>>> Aug  7 16:31:29 raspberry vdr: [6438] ERROR: 1 ring buffer overflow (1 
>>> bytes dropped)
>>> Aug  7 16:31:35 raspberry vdr: [6438] ERROR: 26189 ring buffer overflows 
>>> (4923532 bytes dropped)
>>> Aug  7 16:31:41 raspberry vdr: [6438] ERROR: 32283 ring buffer overflows 
>>> (6069204 bytes dropped)
>>> Aug  7 16:31:47 raspberry vdr: [6438] ERROR: 36593 ring buffer overflows 
>>> (6879484 bytes dropped)
>>>
>>>
>>> What does that mean?
>>>
>>> Matthias
>>>
>>> ___
>>> vdr mailing list
>>> vdr@linuxtv.org
>>> https://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>>>
>>
>>
>> ___
>> vdr mailing list
>> vdr@linuxtv.org
>> https://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>
>
> ___
> vdr mailing list
> vdr@linuxtv.org
> https://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr

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

Re: [vdr] broken recordings

2016-08-08 Thread C.Scheeder 

Hi,
it meuns your system is not able to process all the data received in time.
where do you write your recordings?
I guess its a usb-harddisk connected to your raspi, correct?
It's probably just to slow to write all the data vdr throws at it.
Or your usb-bus is saturated with all the data flowing in from your receiver and back out to the 
harddisk.

Christoph

Am 07.08.2016 um 16:34 schrieb Matthias Bodenbinder:

One more hint to the problem.

Another test with recording Arte HD gives the following output in user.log:

Aug  7 16:31:14 raspberry vdr: [28148] timer 1 (5 1503-1655 'Galaxis 
Milchstraße') set to event Son 07.08.2016 15:05-16:45 (VPS: 07.08. 15:05) 
'Galaxis Milchstraße'
Aug  7 16:31:19 raspberry vdr: [6438] i/o throttle activated, count = 1 
(tid=6438)
Aug  7 16:31:22 raspberry vdr: [6438] buffer usage: 70% (tid=6448)
Aug  7 16:31:24 raspberry vdr: [6438] buffer usage: 80% (tid=6448)
Aug  7 16:31:26 raspberry vdr: [6438] buffer usage: 90% (tid=6448)
Aug  7 16:31:29 raspberry vdr: [6438] buffer usage: 100% (tid=6448)
Aug  7 16:31:29 raspberry vdr: [6438] ERROR: 1 ring buffer overflow (1 bytes 
dropped)
Aug  7 16:31:35 raspberry vdr: [6438] ERROR: 26189 ring buffer overflows 
(4923532 bytes dropped)
Aug  7 16:31:41 raspberry vdr: [6438] ERROR: 32283 ring buffer overflows 
(6069204 bytes dropped)
Aug  7 16:31:47 raspberry vdr: [6438] ERROR: 36593 ring buffer overflows 
(6879484 bytes dropped)


What does that mean?

Matthias

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




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