242 replies to this topic

[WanWizard]

I'm a puzzled as you are .

 

Can you tell us:

- your EXACT setup (hardware, tuners, how configured)?

- what you EXACTLY do to reproduce it?

 

If I can move my recordings to my Solo 4K (only one FBC, so hopefully enough), I can free up my Duo 4K and try to build an identical setup here, and retrace your steps.

 

Same question for @Abu.

[paneologist]

I'm a puzzled as you are .

 

Can you tell us:

- your EXACT setup (hardware, tuners, how configured)?

- what you EXACTLY do to reproduce it?

 

If I can move my recordings to my Solo 4K (only one FBC, so hopefully enough), I can free up my Duo 4K and try to build an identical setup here, and retrace your steps.

 

Same question for @Abu.

easiest way to provide config files?
i try to summarize:

hardware:
-Vu+ Duo 4K SE
   with 2 twin-FBC tuners DVB-S2 (one cable to each twin tuner connected the jess switch)

   4TB HDD Seagate ST4000LM024-2AN17V

-CAM 803 CI-Module (but problem exists also without)

-USB DVB-T stick (not yet working, but also, problem exists without)

-Jultec JPS1712-8M (Jess) with 4 Quattro LNBs for SATs 13.0E, 19.2E, 23.5E, 28.2E (64 tuners/sats set up accordingly)

-openwebif with ssl certificate (als without https, the prblem existed)

-2 nfs mounts and 1 nfs export (mounted by other stb) - can also try if problem exists without nfs mounts

-plugins configured: epgrefresh and autotimer, smartd/smartmontools, ntp,  minidlna... hm

-enigma2-plugin-settings-hans-13e-19e-23e-28e

-full list of packages installed:  stb1currentlyinstalled.txt   30.94KB   1 downloads

 

ask, if I missed something

 

to reproduce:

zap around (active timeshift)

or

create timers (by autotimer or manually in epg) (recording to harddisk)

or

run epgrefresh (with fake recording option - not had an issue yet using PIP)

-> by an estimated chance of less than 1% or 80% for epgrefresh the problem occurs at end of a recording or during epgrefresh

hm. maybe this could be a thing. minidlna is also set up to share the hdd contents. uses inotify for changed inodes? hm

on one occasion I captured lsof output. just noticed, that an inode/file is opened twice by enigma2 (a radio recording):  stb1lsofenigma2.txt   15.96KB   0 downloads

Edited by paneologist, 11 August 2023 - 14:59.

[WanWizard]

Differences with my current setup:

- I don't have an SE

- My local disk is an NVMe in a USB enclosure

- Different Jultec model switch to FBC1, Inverto Unicable switch to FBC2

- a lot less plugins, no minidnla, no vtuners ( either USB or SAT>IP )

- a self scanned / build channellist

- no active timeshift

 

I'll try to find time to crawl into the attic and swap my Duo with the Solo, at a moment there aren't any recordings for a few hours...

[Abu Baniaz]

HDD is an internal mechanical device.

 

Reception hardware:

Diseqc switch with 28.2, 19.2, 13 hooked up to Tuner A

28.2 hooked up to Tuner B.

 

There is a high isolation switch in line as the unicable output comes from another signal wire. I was meant to do a report for a bug that affects ATV/ATV. VIX closed it as allegedly fixed and users are told to use unicable. 

 

Tuner cards:

Slot 1. 45308X FBC unit (A-H)

Slot 2. 45308 FBC unit ((I-P)

 

For purpose of test

Tuner A = Simple, single 28.2

Tuner B = Simple Single 28.2

 

 

Image flashed is last nights developer build. I was previously on June image and used "flash image". No plugins installed, I used the overscan wizard to resize screen , in expert mode. I scanned 28.2 using Enigma2 and did not use ABM or AB 28.2. I created a 28.2 bouquet from the service list. I changed show recording as red. No other menus or options used.

 

Using the 28.2 bouquet, I recorded following for 15 minutes

4music, 11265 H 27500

4seven, 11127 V 22000

5 USA 11307 H 27500

5 Action, 11307 H 27500

92 News 11538 V 22000

Aljazeera HD 11507 V 23000

Arise News 11112 H 22000

BBC Alba 10803 H 22000

BBC Alba  10803 H 22000

BBC Alba HD 11023 H 23000

BBC Alba HD 11023 H 23000

BBC Four 10803 H 22000

BBC News 10803 H 22000

BBC News HD 11023 H

BBC One 10803 H 22000

BBC One 10803 H 22000

BBC One 10803 H 22000

BBC One 10803 H 22000

BBC One Scotland HD 11023 H 23000

BBC One Wal HD 11023 H 23000

BBC One WM  11023 H 23000

BBC Parliament 10803 H 23000

One RB  1 10803 H 23000

BBC Scotland 10803 H 

BBC Three 10803 H 22000 (Started getting blank screens on subsequent channels but had audio. even going back to 4 seven)

BBC Three 10803 H 22000

BBC Two 10803 H 22000

BBC Two 10803 H 22000

Brit Asia TV 11112 H 22000

Channel 4 HD 11127 V 22000

 

At 19:34 when the recordings start to end, it locked up. I dozed off and my wife woke me up at 20:45, it was still locked up.

 

 

 

 

Attached Files

•  192.168.1.179-20230811-190313.log.txt   2.14MB   3 downloads

[Abu Baniaz]

This is a lot fewer channels being recorded. It did not hang as the three recordings stopped, but it still had the 

[eFilePushThreadRecorder] read got interrupted by signal, stop: 0

 

Attached Files

•  192.168.1.179-20230811-205831.log.txt   699KB   1 downloads

[Abu Baniaz]

For this test, I stopped all the EPG readers. I created a bouquet "to record" with several FTA services.  Restarted so options are saved. Then init 4 and manually deleted epg.dat. I started receiver again and started recordings.

 

So no EPG readers active, no EPG data, and I still got the e-filepush errors.

[eFilePushThreadRecorder] read got interrupted by signal, stop: 0

 

An aside, why does it ignore the time to record when there is no event info. Makes it impossible to record 5 minutes without any EPG to exclude EPG acquisition and corrupt EPG to invoke the lockup when recordings stop.

 

 

Attached Files

•  192.168.1.179-20230811-212848.log.txt   262.58KB   2 downloads

Edited by Abu Baniaz, 11 August 2023 - 21:42.

[Stan]

[...] why does it ignore the time to record when there is no event info [...]

It seems to do this even when there is an event info. [Link]

[Abu Baniaz]

Latest log attached. Picture of EPG settings attached. 

 

 

Attached Files

•  1_0_1_5857_82E_2_11A0000_0_0_0_20230812183716.jpg   76.46KB   0 downloads
•  192.168.1.179-20230812-182516.log.txt   751.02KB   2 downloads

[WanWizard]

I see nothing wrong here. The recording is stopped, the threadrecoder waits until SIGUSR1 is received, and closes the recording.

[eFilePushThreadRecorder] read got interrupted by signal, stop: 0
[eFilePushThreadRecorder] stopping thread: 100
[eFilePush] SIGUSR1 received
[eFilePushThreadRecorder] read got interrupted by signal, stop: 1
[eDVBRecordFileThread] waiting for aio to complete
[eDVBRecordFileThread] buffer usage histogram (40 buffers of 188 kB)
[eDVBRecordFileThread]   1:    855
[eDVBRecordFileThread]   2:    854
[eFilePushThreadRecorder] THREAD STOP

Only possible comment you can make is that it takes a bit long before SIGUSR1 is received.

 

But that seems to be standard on VU+, everything takes long, even pressing the on/off button as a two-second lag before the box switches on or off.

 

The Solo 4K has the same issue.

[paneologist]

I tried again to install the nightly build. I removed usb dvb drivers beforehand. It worked well. But I didn't get CI+ to work (copied /etc/ciplus/ directory from old install with the certificates). But only got MMI messages "The set-up is now being finalized, this can take a few minutes..." and decoding won't work. Also tried to do the authorization progress again. same result.

so reverted back  to release version again
and ciplus doesn't work there either anymore :-(

that's a tragedy I didn't expect

[paneologist]

CAM also doesn't work in inserted in my TV anymore. same message. :-(
installing nightly build again as this issue seems unrelated.

[paneologist]

this seems more like topic.
after timers ending I get greenscreens:

Traceback (most recent call last):
  File "/usr/lib/enigma2/python/timer.py", line 228, in calcNextActivation
  File "/usr/lib/enigma2/python/timer.py", line 301, in processActivation
  File "/usr/lib/enigma2/python/RecordTimer.py", line 919, in doActivate
  File "/usr/lib/enigma2/python/RecordTimer.py", line 1438, in cleanupDaily
  File "/usr/lib/enigma2/python/RecordTimer.py", line 1069, in saveTimer
UnicodeEncodeError: 'ascii' codec can't encode character '\xfc' in position 40: ordinal not in range(128)
[ePyObject] (CallObject(<bound method Timer.calcNextActivation of <RecordTimer.RecordTimer object at 0xb0e77bf8>>,()) failed)

 

 enigma2_crash_1691928180.log   21.31KB   1 downloads

 enigma2_crash_1691927940.log   58.55KB   1 downloads

[Dimitrij]

 

print u'\xfc'
ü

U+00FC	ü	\xc3\xbc	ü	ü 	LATIN SMALL LETTER U WITH DIAERESIS

 

Edited by Dimitrij, 13 August 2023 - 13:37.

[paneologist]

And again.
Where's use of ASCII enforced where previously unicode was allowed?

doesn't make sense to me.

[WanWizard]

It doesn't, it looks more like a py2 -> py3 conversion issue.

[littlesat]

Does dimitrij’s big revert patch put back vu streamproxy back for broadcom transcode? Sorry this merge request is such huge that I lost what exactly might be the difference.

Edited by littlesat, 13 August 2023 - 14:18.

[paneologist]

and which "ü" causes it?
have 33 lines with this letter in timers.xml

[WanWizard]

Does any of you have FCC active by any chance? If so, does the problem go away if you switch it off?

[paneologist]

Does any of you have FCC active by any chance? If so, does the problem go away if you switch it off?

installed yes. configured no - active don't know?

[WanWizard]

It is installed by default on any VU+ box with FBC capability. If not configured then it is not active. So we can rule that out too.