Plugin locks up

[fuzzweed]

Hi, I'm getting regular lock ups of the plugin. It usually seems to be when scrolling a list (EPG, or search menu both do it).
XBMC is still alive in the background but XNEWA stays stuck on the same screen and won't accept input.


[ATTACH=CONFIG]38963[/ATTACH]

[mvallevand]

I don't see the matching server side logs for the time that you have in xbmc.log, perhaps the box has the wrong time, but the server our your network broke the socket.

Code:

12:27:20 T:2951187264  NOTICE: http://192.168.8.3:8866/control?time=1411558040&key=40&client=XNEWA0x123f587018L
12:27:21 T:2951187264   ERROR: EXCEPTION Thrown (PythonToCppException) : -->Python callback/script returned the following error<--
                                             - NOTE: IGNORING THIS CAN LEAD TO MEMORY LEAKS!
                                            Error Type: <class 'socket.error'>
                                            Error Contents: [Errno 104] Connection reset by peer
                                            Traceback (most recent call last):
                                              File "/storage/.xbmc/addons/script.xbmc.x-newa/resources/src/nextpvr/emulate.py", line 208, in onAction
                                                output.write(jpgfile.read())
                                              File "./Lib/socket.py", line 351, in read
                                              File "./Lib/httplib.py", line 561, in read
                                              File "./Lib/socket.py", line 380, in read
                                            error: [Errno 104] Connection reset by peer
                                            -->End of Python script error report<--

Earlier when running NextPVR failed to send a screen when asked with a 404 which would be odd too but it would cause a delay, so it would be nice to see the server side logs.

Code:

12:26:21 T:2530138944  NOTICE: HTTP Error 404: Not Found

Finally I do see the error at startup that I have asked sub to look at a few times, that typically means a 10 second delay at startup and maybe a low res screen.

Code:

2014-09-25 07:39:07.509    [DEBUG][32]    Got Web Request (192.168.8.6): /control    size=1280x720&client=XNEWA0x123f587018L
2014-09-25 07:39:07.509    [DEBUG][32]    Starting client instance
2014-09-25 07:39:08.512    [DEBUG][32]    Client started
2014-09-25 07:39:08.512    [DEBUG][32]    Unexpected client control error: System.Runtime.Remoting.RemotingException: Failed to connect to an IPC Port: The system cannot find the file specified.


Server stack trace:
   at System.Runtime.Remoting.Channels.Ipc.IpcPort.Connect(String portName, Boolean secure, TokenImpersonationLevel impersonationLevel, Int32 timeout)
   at System.Runtime.Remoting.Channels.Ipc.ConnectionCache.GetConnection(String portName, Boolean secure, TokenImpersonationLevel level, Int32 timeout)
   at System.Runtime.Remoting.Channels.Ipc.IpcClientTransportSink.ProcessMessage(IMessage msg, ITransportHeaders requestHeaders, Stream requestStream, ITransportHeaders& responseHeaders, Stream& responseStream)
   at System.Runtime.Remoting.Channels.BinaryClientFormatterSink.SyncProcessMessage(IMessage msg)

Exception rethrown at [0]:
   at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)
   at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(MessageData& msgData, Int32 type)
   at NUtility.IClientControl.ClientGetScreenImage()
   at Cassini.Request.Control()

Martin

[fuzzweed]

Not sure why the times didn't match up. Fresh logs attached. Have a look around 18:39.
It seems to be particularly if I hold 'down' (even for a fraction of a second) rather than using multiple presses to scroll through a list.

[ATTACH=CONFIG]38965[/ATTACH]

[mvallevand]

it does in web.log like the server became unresponsive and crashed at the time but is probably because unknowingly you are doing a DOS attack on your server sending down keys faster then they can be processed Maybe there is a key repeat setting you can use to slow that down.

Code:

2014-09-25 18:36:36.448    [DEBUG][1]    OnKeyDown() done. Took 68.0039 ms
2014-09-25 18:36:36.945    [DEBUG][1]    OnKeyDown() done. Took 67.0039 ms
2014-09-25 18:36:37.023    [DEBUG][1]    OnKeyDown() done. Took 68.0039 ms
2014-09-25 18:36:37.102    [DEBUG][1]    OnKeyDown() done. Took 69.0039 ms
2014-09-25 18:36:37.178    [DEBUG][1]    OnKeyDown() done. Took 66.0038 ms
2014-09-25 18:36:37.254    [DEBUG][1]    OnKeyDown() done. Took 66.0037 ms
2014-09-25 18:36:37.331    [DEBUG][1]    OnKeyDown() done. Took 67.0038 ms
2014-09-25 18:36:37.409    [DEBUG][1]    OnKeyDown() done. Took 68.0039 ms

Martin

[fuzzweed]

that's funny. I can think of a few companies websites I'd like to bring down just with a one button DOS attack....
Is it possible to code the plugin to handle this? I'd think that holding down a button to scroll was fairly expected behaviour . Would changing a key repeat in the operating system not make all other list scrolling laggy to fix this one??.
Not to mention my own personal torment finding out which Linux file holds the key repeat properties

[fuzzweed]

...tumble weed blowing through the openelec forum.....
http://openelec.tv/forum/105-keyboards/4...peat-delay

[mvallevand]

On my STB if I hold the arrow down, it can continue on for hours after I let go, so hard to say what is expected. From NMT and MVP usage habits I just hit the key many times and get used to using paging and % jumps too.

Martin

[fuzzweed]

On my STB if I hold the arrow down, it can continue on for hours after I let go, so hard to say what is expected.

Martin

Yeah that's annoying too, but at least it's taking in all the commands and executing them (albeit slowly) rather than throwing a complete wobbly.
Anyway, do I take it that's a 'no' it can't be fixed at the server end?

[mvallevand]

Yeah that's annoying too, but at least it's taking in all the commands and executing them (albeit slowly) rather than throwing a complete wobbly.
Anyway, do I take it that's a 'no' it can't be fixed at the server end?

I don't know that would be sub's call, I am only assuming that is why the pipe got broken. I am very reluctant to enforce any minimum time in the client and I do consider this to to be a "Doctor, Doctor it hurts when I do this" problem.

Martin

[fuzzweed]

OK, so I don't think it is key repeats as it locked up around 19:13pm today with very careful single key presses, and it locked up last night while watching TV (somewhere between 9 and 10pm, not sure if this is in the logs) so basically with no key presses. i.e. once the recording has finished and returned to the plugin screen, the plugin wasn't accepting any key inputs.

As I mentioned in another post, the lack of a hardware exit button is compounding the problem as there's no way to exit the frozen plugin via remote, only via a mouse which isn't ideal in an HTPC.

Anyway, I hope the logs shed a bit more light, and you can see if there's anything more that can be done to solve the problem.

[ATTACH=CONFIG]39013[/ATTACH]