242 replies to this topic

[paneologist]

I will have look. Please in the future exclude calls to futex (-e !futex) because they tend to fill up the log, without adding any information.

 
ok now. running like that now "strace -f -tt -e "!futex" -p 1522 > /media/hdd/strace/strace6.txt 2>&1"
 

Can you change the setting "show zap errors" to "disable" (just to check)?

I changed "Hide zap errors" to true.

[paneologist]

box got stuck again in the evening.

 stb1debuglog17.txt   386.84KB   3 downloads

 stb1strace7.txt.bz2   1.06MB   1 downloads

[paneologist]

just for info. I didn't have a single event since then.

The only thing I changed is installing openvpn including lots of dependencies. Among them coreutils.

I hope I am not crowing too soon.

 

but perhaps possible that it has to do with coreutils? any call to those in the python code? where busybox or coreutils programs differ?
maybe I'll dive into those straces.

Going to be away for ~10 days without physical access to the stb.

Edited by paneologist, 30 August 2023 - 09:37.

[WanWizard]

Thanks for all your efforts sofar!

[paneologist]

something weired happened.

 

I have the topic's problem now on my Xtrend et10000 (openPLi release 8.3) too - happened several times when zapping.

this box has no active tuners, using the Vu+ as fallback remote tuner.

2023-09-01T08:45:32+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'InfobarChannelSelection' -> Action = 'keyUp'.
2023-09-01T08:45:32+02:00 172.17.0.98 enigma2: [Skin] Processing screen 'ChannelSelection_summary' from list 'ChannelSelection_summary, SimpleSummary', position=(0, 0), size=(128 x 32) for module 'SimpleSummary'.
2023-09-01T08:45:32+02:00 172.17.0.98 enigma2: [Screen] Showing screen '['ChannelSelection_summary', 'SimpleSummary']'.
2023-09-01T08:45:32+02:00 172.17.0.98 enigma2: [Screen] Showing screen 'ChannelSelection'.
2023-09-01T08:45:36+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'ChannelSelectBaseActions' -> Action = 'showFavourites'.
2023-09-01T08:45:38+02:00 172.17.0.98 enigma2: [eDVBServicePlay] pts of eit change: 42f04210, fixup_pts: 521f70, first_pts: 429e22a0
2023-09-01T08:45:38+02:00 172.17.0.98 enigma2: [eDVBServicePlay] pts of eit change: 42f04210, fixup_pts: 521f70, first_pts: 429e22a0
2023-09-01T08:45:39+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'OkCancelActions' -> Action = 'ok'.
2023-09-01T08:45:46+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'ChannelSelectBaseActions' -> Action = 'keyRight'.
2023-09-01T08:45:51+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'ChannelSelectEPGActions' -> Unknown action 'showEPGList'! (Typo in keymap?)
2023-09-01T08:45:51+02:00 172.17.0.98 enigma2: [Screen] Showing screen 'UnhandledKey'.
2023-09-01T08:45:52+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'OkCancelActions' -> Action = 'cancel'.
2023-09-01T08:45:52+02:00 172.17.0.98 enigma2: [Screen] Showing screen 'InfoBarSummary'.
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [ActionMap] Keymap 'InfobarActions' -> Action = 'showRadio'.
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [eDVBServicePlay] timeshift
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [Navigation] playing:  1:0:A:28A5:40F:1:C00000:0:0:0:
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] read got interrupted by signal, stop: 0
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: message repeated 477 times: [ [eFilePushThreadRecorder] read got interrupted by signal, stop: 0]
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] stopping thread: 100
2023-09-01T08:45:53+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] stopping thread: 99
[...]
2023-09-01T08:45:56+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] stopping thread: 70
2023-09-01T08:45:56+02:00 172.17.0.98 enigma2: [gRC] main thread is non-idle! display spinner!
2023-09-01T08:45:56+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] stopping thread: 69
[...]
2023-09-01T08:46:03+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] stopping thread: 1
2023-09-01T08:46:03+02:00 172.17.0.98 enigma2: [eFilePushThreadRecorder] thread could not be stopped!

It seems very strange and unique to me?

what unusual things do have both boxes have in common?

-IPv4 only networks

-NFS mounts and exports

-syslog

-mostly same packages on the boxes. some more on the Xtrend, didn't clean it up, when it became my 2nd box.

 

 

[WanWizard]

If this box doesn't have any tuners, what is using the FilePushThreadRecorder?

[paneologist]

maybe timeshift?

[WanWizard]

Timeshifting a stream? Brr....

[paneologist]

Timeshifting a stream? Brr....

as written, I didn't clean the box up, when it became the secondary.
on the next thought: why not timeshifting something from a remote fallback tuner?

 

updated the box to the 9.0-rc.

had another issue when streaming from the Vu+ and on start (import bouquets and epg) it sometimes fails. The Vu+ server gives sometimes HTTP 401 "not authorized". sometimes it works when I restart the Vu+ box.

they are in different ip subnets, enabled option "Enable access from VPNs" and https (with real ssl cert/key).

but it states "Without auth only local access is allowed!" - maybe this is a problem?

didn't see an option when setting up the remote fallback tuner to set up authorization.

would be nice if you could set up access control like in a normal web server, to specify which address ranges are allowed to access.

[WanWizard]

The problem with permanent timeshift is the fact that its implementation is crap, it is not done as a normal recording, which could very well be the root cause of the problem.
 
If you have authentication on the server box, you need to configure that on the client, there is an "additional OpenWebIf options" or so on the configuration screen. If you enable that, it folds out to reveal additional input fields, like username, password and TCP port.
 

would be nice if you could set up access control like in a normal web server, to specify which address ranges are allowed to access.

Yes, that would be nice, but unfortunately there is no web server running on the box, it is just a python twisted thread with a socket listener. Which is also the reason we don't want the box on the internet, it isn't secure at all.

It will also be difficult to implement in a seperate process, as need needs all sorts of access to internal Enigma data structures. So it would require a complete rewrite.

[Erik Slagter]

Yes indeed. I would suggest that anyone having recording issues, for the moment disable timeshifting (either automatic or manual). If the error is gone then, we know where to search.

 

Also, the huge amounts of interrupted reads seem to be a sign that the stopping, later on, will not succeed. Can you confirm that? That during a normal, succesful recording, you do not get the huge amount of interrupted reads? That would also be an interesting pointer.

[paneologist]

The problem with permanent timeshift is the fact that its implementation is crap, it is not done as a normal recording, which could very well be the root cause of the problem.
 
If you have authentication on the server box, you need to configure that on the client, there is an "additional OpenWebIf options" or so on the configuration screen. If you enable that, it folds out to reveal additional input fields, like username, password and TCP port.
 

would be nice if you could set up access control like in a normal web server, to specify which address ranges are allowed to access.

Yes, that would be nice, but unfortunately there is no web server running on the box, it is just a python twisted thread with a socket listener. Which is also the reason we don't want the box on the internet, it isn't secure at all.

It will also be difficult to implement in a seperate process, as need needs all sorts of access to internal Enigma data structures. So it would require a complete rewrite.

thanks, found the option for the authentification.

 

hm. just some thoughts.. rewrite... If I had to implement something like openpli I would split the processes as much as possible. A main thread with an api socket, any programs as sockets, and any recording a seperate thread, etc... just thoughts.

 

Yes indeed. I would suggest that anyone having recording issues, for the moment disable timeshifting (either automatic or manual). If the error is gone then, we know where to search.

 

Also, the huge amounts of interrupted reads seem to be a sign that the stopping, later on, will not succeed. Can you confirm that? That during a normal, succesful recording, you do not get the huge amount of interrupted reads? That would also be an interesting pointer.

I disabled the timeshift on the secondary box. But the Vu+ had the issue twice again. Box in standby, so there shouldn't run timeshift.

 

May it have to do with slow i/o maybe / timeing issue?
never had the issue on the xtrend box. now it is connected via a vpn to the Vu+, as I am not at home, but somewhere else, and the connection speed doesn't seem be very good - download 35Mbit/s and 10Mbit/s upload. And the box running openvpn creates a noticable load. Since this setup I had this issue often when timeshift was being stopped.
 

on the Vu+, hm. don't know if minidlna (which uses inotify afaik - (according to lsof dbus daemon, udevd, avahi daemon and streamproxy also use inotify)) and nfs exports affect i/o speed.

 

what is the sign of those interrupted reads?

Edited by paneologist, 4 September 2023 - 17:13.

[Abu Baniaz]

Although I can't reproduce the lockup at the moment, when I could do, there was no timeshift involved.

[WanWizard]

hm. just some thoughts.. rewrite... If I had to implement something like openpli I would split the processes as much as possible. A main thread with an api socket, any programs as sockets, and any recording a seperate thread, etc... just thoughts.

 

We've been having these plans for about 10 years.

 

But no people, no appitite, no commitment for such a large project.

[Huevos]

 

hm. just some thoughts.. rewrite... If I had to implement something like openpli I would split the processes as much as possible. A main thread with an api socket, any programs as sockets, and any recording a seperate thread, etc... just thoughts.

 

We've been having these plans for about 10 years.

 

But no people, no appitite, no commitment for such a large project.

 

Maybe 10 years ago would have been a good time to do it. These days most developers have lost interest in the project.

Edited by Huevos, 5 September 2023 - 20:30.

[WanWizard]

It has become a self-fulfilling profecy.

[Huevos]

It has become a self-fulfilling profecy.

Enigma just got interesting to the general public because of card sharing, but that is not so fruitful these days, and anyway everyone has fat internet connections now. Satellite is in it's sunset phase so not surprising developers have moved on.

[littlesat]

...

Edited by littlesat, 6 September 2023 - 14:11.

[WanWizard]

But the patch from dimitrij made it int…. So it looks like it should resolve it.

 

Wrong thread?

[littlesat]

What is happening here... indeed wrong thread....