All sorts of crashes and exceptions after fresh flash of 3.2.023

[panch]

Hi,

My vu+duo2 has been working flawlessly before updating to the 3.2.023 version, but after doing a fresh install with no settings restoration, The box suddenly began to crash when starting up. The exception was something about the setupTranscoder. After shutting i down and rebooting, it failed to start the background update check, which resulted in an exception at startup. I could disable the checks in the settings, but when I then go into the plugin software downloads, the box crashes again. Something about the directory: '/var/lib/opkg' not existing.
I could mend this, but it would crash again...

Has this something to do with the GST 1.6.1 update? Tried the build 021, but this one apparently also has the 1.6.1 version, according to the info box.

I have attached 4 error logs, hopefully some you clever guys can help me out.

[abu baniaz]

Unikley to be GST related



Traceback (most recent call last):
File "/usr/lib/enigma2/python/Plugins/SystemPlugins/ViX/SoftcamManager.py", line 548, in softcam_check
IOError: [Errno 2] No such file or directory: '/tmp/cam.check.log'
< 414.939354> [ePyObject] (CallObject(<bound method SoftcamAutoPoller.softcam_check of <Plugins.SystemPlugins.ViX.SoftcamManager.SoftcamAu toPoller instance at 0x619256c0>>,()) failed)
]]>
</enigma2crashlog>
</crashlogs>
</openvix>




Traceback (most recent call last):
File "/usr/lib/enigma2/python/mytest.py", line 660, in <module>
runScreenTest()
File "/usr/lib/enigma2/python/mytest.py", line 466, in runScreenTest
plugins.readPluginList(resolveFilename(SCOPE_PLUGI NS))
File "/usr/lib/enigma2/python/Components/PluginComponent.py", line 104, in readPluginList
File "/usr/lib/enigma2/python/Components/PluginComponent.py", line 30, in addPlugin
File "/usr/lib/enigma2/python/Plugins/SystemPlugins/Hotplug/plugin.py", line 67, in autostart
File "/usr/lib/python2.7/site-packages/twisted/internet/posixbase.py", line 408, in listenUNIX
File "/usr/lib/python2.7/site-packages/twisted/internet/unix.py", line 293, in startListening
CannotListenError: Couldn't listen on any:/tmp/hotplug.socket: [Errno 2] No such file or directory.



Traceback (most recent call last):
File "/usr/lib/enigma2/python/Components/ActionMap.py", line 68, in action
File "/usr/lib/enigma2/python/Components/ActionMap.py", line 48, in action
File "/usr/lib/enigma2/python/Screens/Menu.py", line 55, in okbuttonClick
File "/usr/lib/enigma2/python/Tools/BoundFunction.py", line 9, in __call__
File "/usr/lib/enigma2/python/Screens/Menu.py", line 70, in runScreen
File "/usr/lib/enigma2/python/Screens/Menu.py", line 76, in openDialog
File "/usr/lib/enigma2/python/mytest.py", line 303, in openWithCallback
dlg = self.open(screen, *arguments, **kwargs)
File "/usr/lib/enigma2/python/mytest.py", line 316, in open
self.execBegin()
File "/usr/lib/enigma2/python/mytest.py", line 232, in execBegin
c.execBegin()
File "/usr/lib/enigma2/python/Screens/Screen.py", line 82, in execBegin
File "/usr/lib/enigma2/python/Screens/SoftwareUpdate.py", line 204, in checkNetworkState
File "/usr/lib/enigma2/python/Screens/SoftwareUpdate.py", line 218, in startCheck
File "/usr/lib/enigma2/python/Components/Ipkg.py", line 85, in startCmd
OSError: [Errno 2] No such file or directory: '/var/lib/opkg'
< 73.140540> [ePyObject] (CallObject(<bound method NumberActionMap.action of <Components.ActionMap.NumberActionMap instance at 0x59c90af8>>,('OkCancelActions', 'ok')) failed)
]]>
</enigma2crashlog>
</crashlogs>
</openvix>



Traceback (most recent call last):
File "/usr/lib/enigma2/python/Components/ActionMap.py", line 68, in action
File "/usr/lib/enigma2/python/Components/ActionMap.py", line 48, in action
File "/usr/lib/enigma2/python/Screens/Menu.py", line 55, in okbuttonClick
File "/usr/lib/enigma2/python/Tools/BoundFunction.py", line 9, in __call__
File "/usr/lib/enigma2/python/Screens/Menu.py", line 70, in runScreen
File "/usr/lib/enigma2/python/Screens/Menu.py", line 76, in openDialog
File "/usr/lib/enigma2/python/mytest.py", line 303, in openWithCallback
dlg = self.open(screen, *arguments, **kwargs)
File "/usr/lib/enigma2/python/mytest.py", line 316, in open
self.execBegin()
File "/usr/lib/enigma2/python/mytest.py", line 232, in execBegin
c.execBegin()
File "/usr/lib/enigma2/python/Screens/Screen.py", line 82, in execBegin
File "/usr/lib/enigma2/python/Screens/SoftwareUpdate.py", line 204, in checkNetworkState
File "/usr/lib/enigma2/python/Screens/SoftwareUpdate.py", line 218, in startCheck
File "/usr/lib/enigma2/python/Components/Ipkg.py", line 85, in startCmd
OSError: [Errno 2] No such file or directory: '/var/lib/opkg'
< 378.598858> [ePyObject] (CallObject(<bound method NumberActionMap.action of <Components.ActionMap.NumberActionMap instance at 0x59931468>>,('OkCancelActions', 'ok')) failed)
]]>
</enigma2crashlog>
</crashlogs>

[panch]

But if it isn't GST related, what could then be the problem, when I'm doing everything as a fresh install? Skin selection/hard drive issue?
---EDIT---
I have just tried installing the 3.2.018 version (but restoring from a backup of 023), the info box still states it is GStreamer 1.6.1 - or is that not the same as the GST version? I thought the 1.6.1 was introduced in the 023 build?
45910 (dont know why it is upside down)

[birdman]

My vu+duo2 has been working flawlessly before updating to the 3.2.023 version, but after doing a fresh install with no settings restoration, The box suddenly began to crash when starting up.By "fresh install" do you mean "flashed a new version"?

Tried the build 021, but this one apparently also has the 1.6.1 version, according to the info box.It had 1.6.0 according to the info box on my system. 1.6.1 came in at 023. To get back to 021 you'd need to reflash 021 (or restore a full saved-image of 021, not just a backup from it).

The logs in post #1 are odd. The first two shows files in /tmp missing (or /tmp itself ??) while the second two show /var/lib/opkg missing (which seems to be because you pressed some keys before it got to the bit where /tmp would cause a problem).

You could login to the system to find out what state the file-system is actually in, but beyond that I'd suggest a fresh flash (from USB) of 3.2.024

[abu baniaz]

024 is not available as a download. 025 should be transferred to server after feeds are updated on Friday

[Huevos]

The logs in post #1 are odd. The first two shows files in /tmp missing (or /tmp itself ??) while the second two show /var/lib/opkg missing (which seems to be because you pressed some keys before it got to the bit where /tmp would cause a problem).Logs show an IOError. Probably out of RAM for some reason.

[panch]

Logs show an IOError. Probably out of RAM for some reason.

Couldn't IO error be because of the missing file/directory as it shows?

IOError: [Errno 2] No such file or directory: '/tmp/cam.check.log'

By fresh install, I meant installing the image using USB, and setting everything up as if it was a new box. But I'm not sure if there are ways to properly clean out the system prior to the image install? Because it seems I cannot get back to GST1.6.0 or lower, even by installing an older image.

The error about the opkg dir missing, I could go in and create the directory, and manually call the opkg upgrade. That would solve the problem issued by one of the logs.

A filesystem check (using the GUI) of the harddrive did not reveal any problems (or nothing was posted on the screen afterwards).

[DaMacFunkin]

On a fresh install certain things such as EPG default to internal memory, could this be your problem?
I know this doesn't help but I fresh flashed Duo2 to 023 and set up from scratch then online updated to 024 with no problems.

[panch]

On a fresh install certain things such as EPG default to internal memory, could this be your problem?

Do you mean internal flash? Since memory gets erased when powered off. I absolutely have no idea, I have never changed the default way of using the EPG data.

[DaMacFunkin]

I choose to have my EPG data save to HDD but after a fresh install it changes to flash, could it be loading your EPG into the flash and filling it?

[panch]

I don't know about that, is there a way to check the remaining space on the flash?

[DaMacFunkin]

Just go into EPG settings and change the location to somewhere else then reboot, it is one of the first things I do I do on a clean install.

[panch]

Just go into EPG settings and change the location to somewhere else then reboot, it is one of the first things I do I do on a clean install.

I will try that with the 025 build when it is released - but it doesn't answer the question of how I get to know how much space is left on the internal flash? I'm not that well into linux, so if it can be done via putty, I would need some hints on the commands to call.

[abu baniaz]

df -h


A guide to linuux/unix commands should you need it: http://www.world-of-satellite.com/showthread.php?40948-How-to-use-Putty-for-Telnet-commands

[panch]

Thanks, I'll try that. What could be filling up the flash, other than EPG data?
But to be realistic - the box has 1GB flash storage - so the EPG data should take up >900MB? I have a hard time believing this.

I would very much like to have the question answered to, how I downgrade the GST from 1.6.1 to 1.6.0 or lower, since this is apparently not bein done when flashing an older image.

[DaMacFunkin]

It depends what epg source you use, if you have a few rytec sources it can be massive.

[Sicilian]

since this is apparently not bein done when flashing an older image.

What on earth are you going on about? Flashing PRE 3.2.023 DOES give GST 1.6.0.

You issue is not caused by GST.

Disconnect your network cable and leave disconnected.

Re-flash and setup fresh todays 025 release. Do not restore anything. Leave ethernet or wifi off.

Test at free to air for hours without ANY network connection.

Also confirm, do you have any port forwarding setup to the receiver?

[panch]

What on earth are you going on about? Flashing PRE 3.2.023 DOES give GST 1.6.0.

I was just concluding based on the infobox, if you see my (although rotated) image a few posts up, it says that it is build 3.2.018, and the GStreamer is 1.6.1 - if GStreamer and GST is not the same, then I'm sorry for jumping to conclusions, but how can I then see what the GST version is?


Disconnect your network cable and leave disconnected.
Re-flash and setup fresh todays 025 release. Do not restore anything. Leave ethernet or wifi off.
Test at free to air for hours without ANY network connection.
Thank you, I will try that.


Also confirm, do you have any port forwarding setup to the receiver
I am forwarding 8000-8002, as well as port 80 to from the router right now. That hasn't been changed since before 3.2.023

Lastly, I want to add that I tried to re-enable WOL since this has not been working for quite some time. I tested this and it worked. Could that cause any issues?

[Huevos]

These STBs are not hardened for exposure to the public internet. Close the ports on your router and the problem will be gone.

[birdman]

A guide to telnet commands should you need it: http://www.world-of-satellite.com/showthread.php?40948-How-to-use-Putty-for-Telnet-commandsThose are Unix/Linux commands. Telnet is a network protocol, and telnet commands are something completely different.
I'd access the box using SSH (even when using Putty).

[birdman]

Logs show an IOError. Probably out of RAM for some reason.Or (guessing a bit here) it could be the result of something going wrong with the flash and bad blocks being used. So try a USB flash from fresh to see how that fares.

[panch]

Ok, so did a full reflash of the newest 3.2.025 build - disabled all network and let it run for 4-5 hours without problems. Also I disabled port forwarding on my router to prevent possible incoming connections.
After that I enabled networking, and installed the sundtek control panel, as well as the Magic-FHD skin.
I selected to save EPG data on the harddrive, and checked that there was only used 10% space of the internal flash.

Suddenly I got an exception that made the box halt completely - it couldn't even restart after the 10s timeout. I shut down the box and restarted it, and when it came up another exception occurred.
The two exceptions have been attached:
1: Suddenly this occurred: 45925
2: After rebooting: 45926

I don't use softcam so if that's causing any issues, it should be disabled - but I guess it is when performing a clean flash?
The 2nd exception is the background check, again with the directory it cannot create, and also somehow this is the issue with the first exception.

Can any of you make anything out of this?

--EDIT---
I can confirm, that the /var/lib/opkg directory is missing - which means that the plugins list cannot be downloaded. If I manually create the directory, I can mend this problem. But why is it removed in the first place??

--EDIT 2---
I can also confirm, that the /tmp/hotplug.socket file is actually there. But the size of it is 0. I have performed a "init 4 && opkg update && opkg upgrade && init 6" call from putty, and after restarting the exception has now not occurred after restarting. But I'm pretty convinced that it's a matter of time before some exception occurs again.

[abu baniaz]

I don't use softcam so if that's causing any issues, it should be disabled - but I guess it is when performing a clean flash?


That is checking that location and not finding it. This causes the crash. Anything that looks at /tmp location will cause a crash. As you say that the location/path should be there.

[panch]

So what can I do from here? This exception happened suddenly whiile watching TV...

[panch]

Just while I was writing the last post, another exception occurred. They look a bit the same as the previous ones, file/directory errors. Again, I had to turn it off at the back, and after rebooting, the second exception occurs.

1: Exception that suddenly happenned: 45930
2: Exception after rebooting: 45931

[abu baniaz]

Do you have teh following location? /var/volatile/tmp

Do you have a symlink for tmp?

[panch]

Do you have teh following location? /var/volatile/tmp

Do you have a symlink for tmp?

As you can see here, that should be in order:

login as: root
root@vuduo2:~# ls
logs
root@vuduo2:~# cd ..
root@vuduo2:/home# cd ..
root@vuduo2:/# ls
autofs boot etc home media opt run share tmp var
bin dev hdd lib mnt proc sbin sys usr
root@vuduo2:/# cd tmp
root@vuduo2:/var/volatile/tmp# ls
cam.check.log hotplug.socket lcd4linux nxserver_ipc
root@vuduo2:/var/volatile/tmp#

[abu baniaz]

Purely speculating here.... Someone else had a similar issue with his Duo2. Have a look here:
http://www.world-of-satellite.com/showthread.php?46301-Box-starting-to-crash-after-about-24-Hours-use

[birdman]

Again, I had to turn it off at the back, and after rebooting, the second exception occurs.

1: Exception that suddenly happenned: 45930
2: Exception after rebooting: 45931Since you have appear to have access to the running system uploading the dmesg and messages file from /var/log after an exception might be useful.

[birdman]

1: Exception that suddenly happenned: 45930
2: Exception after rebooting: 45931That second one isn't after a reboot - it's after enigma2 (the GUI) has restarted (the system hasn't rebooted).

Both have the symptoms of /tmp having gone away. Do you have any process running that might remove it in some way?
Mind you, the /var/lib/opkg wouldn't be affected by that.


I can also confirm, that the /tmp/hotplug.socket file is actually there. But the size of it is 0.You'd expect it to have a size of 0. It's a (Unix) socket, so all that is there is an inode entry in the file-system (so processes can all call it by the same name).

[abu baniaz]

What is the folowing?
nxserver_ipc

[Huevos]

Panch, you are not listening. Close the ports on your router. That is what is causing this problem.

[birdman]

What is the folowing?
nxserver_ipcProbably part of FreeNX - perhaps it comes along with the "sundtek control panel" that he mentions. #22 indicates that all is well before this is installed.

[twol]

Dont think he checks opkg is before installing sundtek.
Perhaps he could reflash :) check that the opkg directory is there, if so copy it to the pc, install sundtek ... Check again.
If opkg is deleted, he could then ftp back the opkg directory ..... Would get him up running perhaps ... If it is sundtek.

[Sicilian]

Power off your router and modem for 10 minutes. Re-flash again with Ethernet disconnected. Make sure all ports forward to receiver are closed.

If you get the issue again try using a network switch instead of connecting direct to router.

[panch]

Thanks for all your replies, I will take them one at a time tomorrow when I have time to look at it again. Definitely some good pointers!


Sent from my iPhone using Tapatalk

[birdman]

If opkg is deleted, he could then ftp back the opkg directory....If installing X deletes the opkg directory then the solution is to not install X.

[twol]

If installing X deletes the opkg directory then the solution is to not install X.

Absolutely correct, but if he loses the tuner then maybe it will all work if he puts back the missing directory...... Short term until it is fixed:)
A short term resolution can sometime be better than .... Nothing:)

[panch]

Ok, so yesterday, I got to reflash the box before leaving it for 12 hours. I have left it almost untouched from the original image, only installing channels and not adding plugins or skins. Everything is working as it should.

Also, I cut all ports from the box, beacuse I actually realised that everyting began to go wrong after setting the DMZ on the router to the IP of the VuDuo2. To sum up - these were the things that could make the box behave weirdly:

Install and use the Magic FHD skin
Installing the sundtek driver (although this has worked flawlessly the last year)
Enabling WOL and trying to perform a resume from WOL
Install Samba (and NFS)
Opening ports 80 and 8001-8002 to the receiver
Opening all other ports to the receiver (for debug)



So now I will take every step one at a time in the specified order. Since it has been mentioned numerous times that the ports should stay closed, I will try opening those if everything else works. The #28 links to another member having some issues with the samba server, which is very similar to my issue (tmp folder problems), and the issues for me also started after I installed the samba server. It will be very interesting to see what provokes the crashes. I will take some days, though, because I want to be absolutely positive that each of the points in the list is not affecting performance. At least the problems occurred quite fast, so I will expect that I can locate the error within a few days.


Panch, you are not listening. Close the ports on your router. That is what is causing this problem.
What is the purpose of a networked receiver, if you cannot schedule recordings on the go, or stream recordings while outside the house? Do you really mean that ports 80 and 8001-8002 should stay closed?
One of the other error sources could be, that I recently setup outside access to the channels using Serviio DLNA server and using DLNA to stream live tv from the box. This could correlate with the SetupTranscoder crashes I have seen. I will investigate this also, when I eventually open the ports that are used by the transcoder.

By the way, the receiver is located behind a 5-port switch which is connected to the switch in the cable modem.

[Sicilian]

What is the purpose of a networked receiver, if you cannot schedule recordings on the go, or stream recordings while outside the house? Do you really mean that ports 80 and 8001-8002 should stay closed?
One of the other error sources could be, that I recently setup outside access to the channels using Serviio DLNA server and using DLNA to stream live tv from the box. This could correlate with the SetupTranscoder crashes I have seen. I will investigate this also, when I eventually open the ports that are used by the transcoder.

By the way, the receiver is located behind a 5-port switch which is connected to the switch in the cable modem.

Opening the ports imo will cause the crash.

Try using a VPN.

[panch]

Because of unauthorized access to the receiver?


Sent from my iPhone using Tapatalk

[Sicilian]

Because of unauthorized access to the receiver?
Sent from my iPhone using Tapatalk

Yes that too.

I replicated your issue few weeks back, closing the ports stopped the crash.

Please try with a VPN and report back if possible.

[panch]

Yes that too.

I replicated your issue few weeks back, closing the ports stopped the crash.

Please try with a VPN and report back if possible.

I have been running with the ports 80 and 8001-8002 opened since I first bought the receiver, so I highly doubt that these are the culprits. However, i very recently (actually when the problems started to show) set the receiver to DMZ in the router because I had some trouble with the DLNA playback on a TV located outside the home network. This was a debug setting, but I had forgotten all about opening everything, so thanks a bunch for clearly stating it.
Hopefully this is the reason, but is there no plans for debugging why this occurs when opening the ports?

[Sicilian]

I have been running with the ports 80 and 8001-8002 opened since I first bought the receiver, so I highly doubt that these are the culprits.

I had telnet and 80 open when I managed to replicate, no crash since.

By port forward any ports you are opening your receiver to being hacked. :trustme:

DMZ is not the culprit, but would not help anyway.



Hopefully this is the reason, but is there no plans for debugging why this occurs when opening the ports? Been trying to track it down for a while. It's present in many 3rd party team images.

Try a VPN if you can.

[panch]

Just an update on the current situation: The list of steps that I wanted to take one at a time (listed in #39) have all been performed, except for the last two that deals with the port forwarding issues.
The receiver still runs robustly, so I have now opened the port 80 (mapping externally from port 50000) to have the web interface exposed to the internet. The next thing will be to have ports 8001-8002 exposed in order to have the streaming running so that I can use the Vu+ PlayerHD work from my iDevice.

I'm not sure if I want to set up DMZ for the receiver, since this was unintentional and I really only want to expose a minimum of ports to the outside world. Then we will see how that performs.

[ArowonA]

I had telnet and 80 open when I managed to replicate, no crash since.
Did you have a secure password? I would assume it's caused by automated scanners trying to bot routers, but it would be interesting to see why they are deleting the opkg folder..maybe to stop people seeing what was installed, or to prevent an automated fix..

[abu baniaz]

http://www.world-of-satellite.com/showthread.php?46802-Guide-Setting-up-SSH-with-public-private-keys

Maybe this guide will be useful

[panch]

http://www.world-of-satellite.com/showthread.php?46802-Guide-Setting-up-SSH-with-public-private-keys

Maybe this guide will be useful

Although a but cumbersome, but it that's the only way to get external access, then be it.
What if I want to use Serviio DLNA server on an external PC that streams live TV channels from my receiver? Would I then also create a SSH tunnel using putty before starting the server on the PC? And what about the beforementioned iOS Vu+ Du2 PlayerHD app that streams TV and performs scheduling?

[judge]

The receiver still runs robustly, so I have now opened the port 80 (mapping externally from port 50000) to have the web interface exposed to the internet. The next thing will be to have ports 8001-8002 exposed in order to have the streaming running so that I can use the Vu+ PlayerHD work from my iDevice.

You're just waiting to have your tuners hacked so. Changing ports might add up to a second on a scan, no means of security.

[kartingzz]

I tried different images, didn't install any plugins but still my Zgemma 2s kept crashing with this log:

Traceback (most recent call last):
File "/usr/lib/enigma2/python/mytest.py", line 660, in <module>
runScreenTest()
File "/usr/lib/enigma2/python/mytest.py", line 466, in runScreenTest
plugins.readPluginList(resolveFilename(SCOPE_PLUGI NS))
File "/usr/lib/enigma2/python/Components/PluginComponent.py", line 104, in readPluginList
File "/usr/lib/enigma2/python/Components/PluginComponent.py", line 30, in addPlugin
File "/usr/lib/enigma2/python/Plugins/SystemPlugins/Hotplug/plugin.py", line 67, in autostart
File "/usr/lib/python2.7/site-packages/twisted/internet/posixbase.py", line 408, in listenUNIX
File "/usr/lib/python2.7/site-packages/twisted/internet/unix.py", line 293, in startListening
CannotListenError: Couldn't listen on any:/tmp/hotplug.socket: [Errno 2] No such file or directory.

But when I replaced USB Wifi Dongle to Ethernet cable- I don't have any problems. Still didn't have a chance to run for longer than 3 hours but looks like box works well

[abu baniaz]

There is no support for Zgemma receivers on this forum. You should not hijack threads either.

[jtrus]

panch
Think of your home network like your house.
Suppose you have allowed your neighbour to use your STB while you are away.
"Don't worry I have left the front door unlocked so you can get in at any time" Of course you wouldn't. That's like having an unsecured network.
"I will leave the kitchen window closed but unlocked so you can climb in" That's like leaving an open port on your network.
Please get yourself a router capable of supporting VPN. Then when you are away, you can link to your home network securely as if your were in the house. All communication is encrypted.