40 replies to this topic

[Martin_B]

maybe it was intended to do so.
in my case it seems that tuner (or lnb) is turnded off when not in use.
not only in standby, just when not in use.

So if i record on chanel A and switch to chanel B, when different tuners are used for A and B
i get this "dropouts" when box turns on or off used/not used tuner (lnb).

option dvb lnb_power_off_in_standby=0
always keep both tuners on, even when not in use (even in standby unfortunately this will consume some additional power in standby)
but for my configuration it's worth
no more dropouts!

BUT:
There are a lot more causes for dropouts.

Your description i guess is is rooted to dvb-power-option

If you take recording from Chanel A and have dropouts after 10-120 seconds, without switching, etc. it would be rooted to ext4 bug.

Regards
Martin

[teleradio]

maybe it was intended to do so.
in my case it seems that tuner (or lnb) is turnded off when not in use.
not only in standby, just when not in use.

So if i record on chanel A and switch to chanel B, when different tuners are used for A and B
i get this "dropouts" when box turns on or off used/not used tuner (lnb).

option dvb lnb_power_off_in_standby=0
always keep both tuners on, even when not in use (even in standby unfortunately this will consume some additional power in standby)
but for my configuration it's worth
no more dropouts!

BUT:
There are a lot more causes for dropouts.

Your description i guess is is rooted to dvb-power-option

If you take recording from Chanel A and have dropouts after 10-120 seconds, without switching, etc. it would be rooted to ext4 bug.

Regards
Martin

Ok, I will try it. Wgich is the command line to connect to my xt-9200 via telnet?

[Martin_B]

First it is important to edit file
/etc/modules-load.d/_et9x00.conf (THIS IS FOR PLI3)
/etc/modutils/_et9x00 (THIS IS FOR PLI2x)

Use Filezilla (or other FTP program) to connect to box and open/edit the file as described.

Save the file to box.


I think, it is not absolutely necessary to connect via telnet and run command "update-modules" but
if you made a mistake when editing, "update-modules" will show an error and you can revoke your changes.

To connect via telnet (from Windows 7) press keys <windows> + r
and type cmd
press ok
then type
telnet 192.168.137.10 (YOU MUST ENTER THE IP ADDRESS OF YOUR BOX)
login as
root
and type
update-modules

IF YOU GET AN ERROR MESSAGE HERE , REVOKE YOUR CHANGES TO
/etc/modules-load.d/_et9x00.conf

IN THAT CASE:
DONT REBOOT BEFORE YOU REVOKED YOUR CHANGES.


Regards Martin

[pieterg]

but again, if the situation happens on a single transponder as well, you can forget about all this, because the lnb is not switched off as long as at least one of the services is kept active.

[Martin_B]

@pieterg
yes problably you are right; it was just a suggestion, could have been tested very quickly.
Seems to be difficult at the moment:)

Regards Martin

[MiLo]

I wonder if the HDD had fragmentation issues?

No, the HDD had nothing to do with it. Use a remote disk or USB stick for recording if you still don't believe that linux is actually capable of reliably writing to a harddisk.

A HDD dropout/overrun will result in AT LEAST 1MB data loss. That's usually several seconds of video. They're also easy to detect: Just run enigma2 in foreground (init 4; enigma2) and it will tell you when it occurs. And when a recording ends, it willl give you a summary of buffer usage and total overruns (if any).

That formatting the HDD seemed to help is probably related to a different pattern of power consumption or slightly different timings in case of a software bug.

Edited by MiLo, 21 September 2012 - 10:59.

[Martin_B]

@teleradio

please tell me your results. I'm very interested.
If you don't cope with telnet, i will help you.
Maybe by Forum Personal Messenger, not to bore other users:))

Regards Martin

[teleradio]

@teleradio

please tell me your results. I'm very interested.
If you don't cope with telnet, i will help you.
Maybe by Forum Personal Messenger, not to bore other users:))

Regards Martin

Hello Martin!
First of all thank you for your interest and help!
I have done the changes in "_et9x00.conf" as you suggested by using Filezilla. After a reboot, there are no changes in record behavior. The dropouts, especially when you change from one scrambeld broadcast to another, still remain. Maybe the cause could be the condiguration of cccam.

greeting
teleradio

[Erik Slagter]

I thought you told us earlier on that the problem was also present on FTA-channels?

[teleradio]

I thought you told us earlier on that the problem was also present on FTA-channels?

You are right! It happens also on FTA-channels, but not not in the measure such as it do between scrambeld channels or between a scrambeld and a FTA-channel! After I have reintitialized the HDD the dropouts are less but they aren't disappeared. For example: if you record a scrambeld channel like ORF-HD and you switch to a FTA - channel or a scrambeld channel and then you switch back or to other channels, your record will present some minor dropouts. When you are recording a scrambeld channel like ORF-HD and you switch to another scrambeld channel like SF1-HD (suisse TV), your record will present a freeze which correspond exactly to the moment you have made the switch between the two channels.

greeting
teleradio

[Erik Slagter]

Maybe the harddisk draws too much current when it spins up? Can you try it with the harddisk running permantly?

Did you also already try it with PiP (if you have that available)?

Edited by Erik Slagter, 22 September 2012 - 14:27.

[teleradio]

Maybe the harddisk draws too much current when it spins up? Can you try it with the harddisk running permantly?

Did you also already try it with PiP (if you have that available)?

Hello Erik!
That was the solution! Nomore dropouts as described after running HDD permanently!

What I don't understand is why, because once you have started the first record the HDD should be already running if you start a second record. Or not?

greeting
teleradio

[Erik Slagter]

You can configure the harddisk to stop spinning after some timeout. If the timeout is very short, the harddisk may stop spinning even while recording. What also may be the case that at the start of the recording both the harddisk AND the tuners draw more power than usual, I've heard similar experiences.

You wouldn't have a dm8000 with one or two dm800 DVB-S2 tuners in it, by chance?

[teleradio]

You can configure the harddisk to stop spinning after some timeout. If the timeout is very short, the harddisk may stop spinning even while recording. What also may be the case that at the start of the recording both the harddisk AND the tuners draw more power than usual, I've heard similar experiences.

You wouldn't have a dm8000 with one or two dm800 DVB-S2 tuners in it, by chance?

Hello Erik!

Thank you for interest and your help: No I unfortunatly do not have such receivers!

I would have a further question related to the HDD timeout setup. Do you think a timeout setting of 1 minute would also avoid or better solve the dropout problems?

greeting teleradio

[Erik Slagter]

I think that if you have problems, you'd better keep the harddisk spinning all of the time.

[ims]

may be it is same problem ? http://git.opendream...2a605e0d3533795

"some optimizations to prevent record drop outs on other hdd activity during recording"

[Martin_B]

A HDD dropout/overrun will result in AT LEAST 1MB data loss. That's usually several seconds of video. They're also easy to detect: Just run enigma2 in foreground (init 4; enigma2) and it will tell you when it occurs. And when a recording ends, it willl give you a summary of buffer usage and total overruns (if any).

Hi MiLo

can I log this output to a file?
I'd like to find out under which conditions these "buffer underruns" start to occure.
(issue: Buffer underruns while recording FTA channels)

Regards

[MiLo]

may be it is same problem ? http://git.opendream...2a605e0d3533795
"some optimizations to prevent record drop outs on other hdd activity during recording"

Probably it means that they copied some of our changes again.

[Flaxtor]

Hello!

I have the same problem and the problem is not S2-tuners. I have T tuners and i also get this problem when i record on one channel and change channels in the meantime. When i then play up the record i get dropouts and i can refer it to channelchange. I flashed my dm8k a couple of days ago and i used to use Gemini2 v5.10 with no record problems at all. I have not played around with the sleep settings on the HDD. I use 20121020 image.

Have you got any solutions on this? Everything else on the image works great, i now use a T2 usb tuner, the EPG data is very good, I really like it!

Keep up the good work!

[Flaxtor]

No new info on this?