How to diagnose lockup/crash? (LibreELEC 9.2.0 (RPi4.arm), 4GB, Hauppauge 950q USB tuner, TVheadend)

1st Official Post

    Very new to RPi, LibreELEC, TVheadend and Hauppauge 950q USB tuner....but pretty familiar with run of the mill Linux/Unix.

    The USB tuner seems very standard and works fine on Windows 10. On the above set up it works .... for a while? I can get to TVheadend via a browser, can log in to LibreELEC via SSH putty. TVheadend on RPi serves up video to various machines on the LAN via VLC. But....

    After a while (sometimes hours) the video stream stops and the RPi seems to "lock up"

    * I am running the RPi headless (no kb, mouse, HDMI....just hardwire to LAN, and the TV tuner)

    * No response from the browser to TVheadend software after lock up

    * No ability to log in to LibreELEC via SSH Putty after lock up

    * Have to power cycle the RPi to get things back to working.

    I see much discussion around firmware drivers for tuners. Pretty sure there are two possible here in the LibreElec standard distribution:

    dvb-fe-xc5000-1.6.114.fw <== this one seem to get loaded at present

    dvb-fe-xc5000c-4.1.30.7.fw <==not sure if my Hauppauge 950q USB tuner can/should be running this one?

    For "starter" diagnostics I booted the RPi without the tuner. Then installed it. Then ran dmesg and captured the tail of that below.

    Is there a troubleshooting procedure I can follow to pin down why the lockup? I'm not even dead certain it's the TV Tuner locking up the whole RPi, but it seems like the area of greatest probability?

    Rick

    tail of dmesg, just after plugging in usb tuner

    • Official Post
    Quote from rjlabs

    I'm not even dead certain it's the TV Tuner locking up the whole RPi, but it seems like the area of greatest probability?

    Thats something I had seen in the past years maybe 1 or 2 times. So this is very unlikely, not impossible.

    Just to name the biggest problems for this kind of issue:

    - SD card

    - power supply

    - heat/temperature

    If it is kind of reproduce able I would change those parts (if possible) and have a look.

    Also you could write some small script and add it to crontab -e

    * * * * * /storage/crash.sh >/dev/null 2>&1

    /storage/crash.sh

    Bash
    #!/bin/bash

file="/storage/crash_$(date +%m-%d-%Y-%H_%M_%S).txt"

dmesg | tail -n40 > "$file"
echo -e "\n\n\n== RAM ================" >> "$file"
free -m >> "$file"
echo -e "\n\n\n== CPU ================" >> "$file"
mpstat >> "$file"
    • CF card Samsung, very current. Tested prior to install. Just tested again, unused space only. Tested fine.
    • power supply brand new 3 amp. Since RPi runs fine much of the time it seems unlikely. RPi red light remains lit with no interruption during OS lock up.
    • heat/temperature unlikely because RPi is in an aluminum total heat-sink case with two fans running. Seems to run about body temperature. TV Tuner however does run warmer. TV tuner was purchased used and is the only part of the system that isn't very new. (Purchased used in hopes the driver for that would be well established.)

    Created cron above and ran. Copied logs to: Index of /RPi Pretty sure 22 is the last cron job to run just ahead of the crash.

    • Official Post
    Quote from rjlabs

    Pretty sure 22 is the last cron job to run just ahead of the crash

    those logs look pretty normal :(

    do you know if it happens in a similar timeframe ?

    reboot and maybe cron some script

    uptime >> /storage/uptime.txt


    Also can you try different SD card and powersupply just to rule that out, we saw already very strange things.

    Do you have some other devices around ? Maybe try it there if the same happen, it could be some kernel problem from the RPi, can't rule that out.