Posts by mgouin

    For what it's worth, from tvheadend web UI (http://ip:9981) in the debug log, I was able to witness a playback stopped automatically. This is what is displayed. It seems that the connection is lost and then restored. But why...


    Full tvheadend log here: http://sprunge.us/CqEdeV

    Hello,

    I'm about to try to downgrade to a previous LibreELEC version in order to solve my issue, but before doing so, I'd love to have some experienced user input about my issue.

    I started running LibreELEC 9.2.0 on a Raspberry Pi 3 Model B since a while. I use it mainly as a PVR with tvheadend running locally on the PI. My setup is as follows:

    • 2 USB ATSC tuner connected to the Pi via a powered USB hub
    • 1 external Hard Drive connected to the Pi via a powered USB hub
    • tvheadend running locally on the PI

    I know that in the early versions of 9.2.X, I was able to successfully sustain 2 simultaneous recordings while watching an already recorded program.

    Now at LE 9.2.6, with the exact same physical setup, in about 10-15 minutes, the playback stops when there are 2 active recordings.

    I can fairly easily reproduce the problem. I don't have a lot of addons running, beside tvheadend, the epg grabber and tvheadend client.

    If useful, you can have have a look at my full debug log (from boot) here: http://sprunge.us/jqKa5p

    In this log, I started playing an already recorded movies at:

    2021-03-02 21:39:33.994 T:1937283520 NOTICE: VideoPlayer::OpenFile: pvr://recordings/tv/active/_-fond-de-train_2016-05-29, TV%20(Radio-Canada), 20160530_020000, 1026801670.pvr

    Then, the playback stopped around this line:

    2021-03-02 21:52:42.965 T:1494913920 ERROR: AddOnLog: Tvheadend HTSP Client: pvr.hts - Command fileRead failed: No response received 

    I'm no expert, but it seems as if tvheadend stops responding?

    Where should I start from there?

    Thanks a lot in advance for your help!

    Mathieu

    Hey no problem! Thanks for sharing your passion :)

    Unfortunately, I did not go this deep in the development. I simply took the pre-compiled image for raspberry pi 3 from LibreELEC official download.

    I ended up looking at the source code because I was trying to analyze the crash dumps generated by kodi, especially the stack trace.

    Thanks!

    Hello,

    Thanks for the reply. I'm curious, do you experience similar crashes?

    On my side, I did not get much progress beside a few more tests (refer to the github xbmc issue I raised).

    Regarding your suggestion, I don't understand what you want me to try. The post refers to the command line argument length, vs what I thought my issue was is an internal C++ string length.

    Could you please elaborate on your idea? I'm sincerely willing to try stuff as I'm mostly wandering in the dark with this issue...

    Thanks again for your help!

    I created the following script to get every second the CPU, Temperature & Free RAM that I execute as a nohup.

    Then, I analyzed the data with a quick python script:

    Code
    import pandas as pd
    import matplotlib.pyplot as plt
    
    # Read to dataframe
    df = pd.read_csv('all.log', index_col='DATE', parse_dates=['DATE'])
    
    # Plot relevant information data
    df[['%usr','%nice','%sys','%iowait','%soft','TEMP']].plot()
    plt.show()

    Here is a screen shot of the plot, zoomed in where the crash occured:

    Imgur: The magic of the Internet

    Notice how the IOWait graph line jumps around the time my last crash dump was generated (kodi_crashlog_20200229052148.log)

    Is there any way to debug iowait?

    If it can help in debugging, I'm using the my LibreElec Pi as a DVR recorder with tvheadend running locally on the Pi. However, when the crash occurs, no recordings were in progress at this time.

    thanks

    Hello,

    I again got a very similar crash. Let me know if you need more logs, but the crash dump was at the same place (pasted below).

    I do embedded programming as a living, so I tried to see if I could find anything.

    I found this:

    It seems as if the xbmc/GUIInfoLabel.cpp at master · xbmc/xbmc · GitHub : Get()) method is causing a too big string.

    From: libstdc++-v3 Source: basic_string.tcc Source File

    Code
    00462     _S_create(size_type __capacity, size_type __old_capacity,
    00463           const _Alloc& __alloc)
    00464     {
    00465       typedef basic_string<_CharT, _Traits, _Alloc> __string_type;
    00466       // _GLIBCXX_RESOLVE_LIB_DEFECTS
    00467       // 83.  String::npos vs. string::max_size()
    00468       if (__capacity > _S_max_size)
    00469     __throw_length_error(__N("basic_string::_S_create"));

    Please anyone... Where do I continue from here?

    Thanks again!

    Hello,

    With debug log enabled I was able to catch another nighly crash. It ran for 4 days, so bear with the long debug log...

    dmesg: http://ix.io/2c68

    kodi.log: http://ix.io/2c69

    kodi.old.log: Kodi old log for post https://forum.libreelec.tv/thread/21384-kodi-crash-randomly-overnight-and-reboot-rpi3 · GitHub (I think I reached the paste limit)

    kodi_crashlog_20200218083716.log: kodi_crashlog_20200218083716.log · GitHub

    Let me know if this is sufficient.

    Thanks a lot!

    Quote

    It's always good to enable debugging so the Kodi log is more detailed. Do so, restart Kodi, and reproduce the error(s) for a new log file.

    It's funny, I enabled debug log yesterday, restarted and my system ran overnight without a crash. It did not restart by itself.

    However, as a related question: is there a way to keep debug log enabled, but without the annoying top left overlay displayed constantly on the screen. I think I saw that in a custom xml, but I cannot find it back.

    As I cannot yet pinpoint to something specific I did to cause the numerous crash every night, I'd like to keep running in debug log to hopefully catch the issue again.

    Thank you in advance ! :)

    Hello,

    I have LE 9.2.0 running on a RPi3.

    Recently, I realized that my system Kodi uptime was always a few hours, instead of days. I then noticed that every night the system reboots Kodi restarts by itself during the night. Normally, at the end of the evening, I just close the TV and let the RPi3 running.

    When trying to look at the kodi logs (in /storage/.kodi/temp), I found the following interesting files:

    Code
    -rw-r--r--    1 root     root       62.5K Jan 19 22:28 kodi_crashlog_20200119222837.log
    -rw-r--r--    1 root     root       79.7K Jan 21 03:34 kodi_crashlog_20200121033352.log
    -rw-r--r--    1 root     root      379.7K Jan 23 01:45 kodi_crashlog_20200123014549.log
    -rw-r--r--    1 root     root       70.3K Jan 24 05:15 kodi_crashlog_20200124051550.log
    -rw-r--r--    1 root     root       98.1K Jan 29 01:55 kodi_crashlog_20200129015542.log
    -rw-r--r--    1 root     root       69.7K Jan 30 02:58 kodi_crashlog_20200130025802.log
    -rw-r--r--    1 root     root       66.3K Feb  1 00:43 kodi_crashlog_20200201004302.log
    -rw-r--r--    1 root     root      277.6K Feb 10 12:38 kodi_crashlog_20200210123826.log
    -rw-r--r--    1 root     root      119.9K Feb 13 03:29 kodi_crashlog_20200213032858.log
    -rw-r--r--    1 root     root       56.3K Feb 14 00:49 kodi_crashlog_20200214004853.log

    If it can be useful, here are the pastebin of the last 2 crashes:

    Let me know if more crash dumps would help?

    I briefly looked at the crash dumps, and the cause all seems to be inside the thread 1 of Kodi. Except the last crash, the previous ones all seem the same.

    How do I start from here?

    Thanks in advance!

    Hello,

    I'm using a Raspberry Pi 3, with LE LibreELEC-RPi2.arm-9.2.0.img. I have setup TVheadend.

    When I reboot the PI, both Kodi & tvheadend web ui don't list my recordings anymore. However, they are physically still stored on a 2TB USB external hard drive (/var/media/RED_2TB).

    I think this relates to the tread: TVheadend problems after new update - lose list of recordings - Raspberry Pi - OSMC Forums

    Is there a way to delay TVH when starting up?

    FYI, my dmesg log is here: http://ix.io/27YL

    Thanks!