Hi @nickv2002, since you prepared the persistent kernel log, could you upload the appropriate kernel and the full log as described above and provide the URLs, here?
Don’t forget to deactivate this special logging to prevent your root file system to get filled.
If the last log entry was 117C that confirms the shutdown trippoint at 118C would have been triggered.
Had this not happened it would have got a little bit hotter still and then just crashed of its own accord - as you were originally seeing.
Did you only capture the final temperature or was the log maintaining a record of the temperature increase prior to hitting the shutdown trip point ?
To get this hot something must be consuming a lot of CPU time, and at the same time the device must be in a location with poor airflow or nearby heat sources. Only the combination of heavy CPU use and a hot location would cause overheating.
Depending on what addons/skin you use with Kodi it can sometimes start using 100% CPU for no apparent reason, (this happens to me sometimes, and was probably the reason for my system crashing before the thermal limiter was added) this may be enough to trigger the problem you’re seeing.
I’d suggest finding another location to sit it where it is not close to other hot devices and has a bit of free airflow around it.
I spaced my Vero out a bit away from the AppleTV to see if that will help.
Right sorry forgot:
$ sudo journalctl -k -b -2 --no-pager|paste-log
https://paste.osmc.tv/ajunabifan
$ sudo journalctl -b -2 --no-pager|paste-log
https://paste.osmc.tv/arayobaho
Kernel logging now disabled.
From your picture it looks like you could move it to the top of the cabinet, directly under the TV to completely remove heat sources. Why not try that? If it works then you know you have a problem in that cramped space.
Is there any chance you could paste a copy of the temperature log to the OSMC pastebin? I’d like to graph it.
Oh, and the system journal seems to have been lost en-route, though the kernel log made it.
Hmm, not sure what happened to the system journal. I copied and pasted the commands above from my terminal.
Also: I didn’t record temperature over time, just the current value (with the assumption it would stop at some point leaving the last value) so no graphing possible.
I guess I’ll run all the logs again to try to get new data in both these areas.
Thanks. It’ll be a bit easier to import if both measurements are on one line. Every 30 seconds should be enough.
I’ve had a number of freezes recently after a stable period for the last month. Here’s the temp log of time,temp,freq that was recorded before the crash:
2018-06-11T17:04:33-07:00,110000,1512000
2018-06-11T17:05:03-07:00,109000,1200000
2018-06-11T17:05:33-07:00,109000,667000
2018-06-11T17:06:03-07:00,109000,1512000
2018-06-11T17:06:33-07:00,109000,1200000
2018-06-11T17:07:03-07:00,109000,500000
2018-06-11T17:07:33-07:00,110000,667000
2018-06-11T17:08:03-07:00,110000,667000
2018-06-11T17:08:33-07:00,111000,667000
2018-06-11T17:09:03-07:00,112000,667000
2018-06-11T17:09:33-07:00,112000,667000
2018-06-11T17:10:03-07:00,113000,667000
2018-06-11T17:10:33-07:00,113000,667000
2018-06-11T17:11:03-07:00,113000,667000
2018-06-11T17:11:33-07:00,113000,667000
2018-06-11T17:12:03-07:00,113000,667000
2018-06-11T17:12:33-07:00,114000,667000
2018-06-11T17:13:03-07:00,114000,667000
2018-06-11T17:13:33-07:00,115000,667000
2018-06-11T17:14:03-07:00,115000,667000
2018-06-11T17:14:34-07:00,115000,667000
2018-06-11T17:15:04-07:00,115000,667000
2018-06-11T17:15:34-07:00,116000,667000
2018-06-11T17:16:04-07:00,115000,667000
2018-06-11T17:16:34-07:00,115000,667000
2018-06-11T17:17:04-07:00,116000,667000
2018-06-11T17:17:34-07:00,116000,667000
And the kernel logs:
sudo journalctl -k -b 0 --no-pager|paste-log:
https://paste.osmc.tv/ubucewecax
sudo journalctl -b 0 --no-pager|paste-log:
https://paste.osmc.tv/axefogoyar
This crash happened while Kodi was on the home screen. I was not at home or using Vero. Its position has remained unchanged from the photos I pictured above.
So CPU throttling seems to be happening but I still don’t understand why my box is getting so hot?
116c is very hot. 120c is where we induce a critical trip point and turn the device off.
Can you temporarily resituate the Vero 4K and see if it fares better?
Sam
I am also seeing this. - I have my brand new Vero 4K connected to a Sony 930e via a Yamaha RX-A770 AVR.
When I shut the TV and the AVR off and come back later in the day or the next morning the screen is just a pale pink. needs to be hard power cycled to come back.
This is a brand new install with only the HDHomeRun Add-On and a different skin on it. I do have the picture slide show screensaver running, but the machine is kept pretty normal temps without issue.
This is the temp of the device at idle with the slide show running with the TV/AVR off:
temp: 85000
cpu speed: 1512000
I just enabled Kodi debugging logging to see if it catches anything the while its sitting idle right now.
I have seen this issue pop up a few times during the setup as well (with a pale blue/green screen instead of pink). Not sure what the deal is, but hoping it goes away or there is some setting to tweak.
85C at idle is quite hot.
Can you show a photo of the Vero 4K in your environment?
Can you post or message me the modified temp logging script? I’m having similar problems and it would be of use and I’m not that familiar with linux commands to modify it myself from scratch 
@santeri.m.t.niemi I run ./tempLog.sh > temps.csv in a Byobu session (screen, or nohup would probably work too). Here’s the script:
#!/usr/bin/env bash
while [[ true ]]; do
tempValue=$(</sys/class/thermal/thermal_zone0/temp)
freqValue=$(</sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq)
timeValue=`date --iso-8601=seconds`
echo "$timeValue,$tempValue,$freqValue"
sleep 30
done
PS: An update on my Vero: Moving it around to a different location has reduce the frequency that it shuts down due to temperature issues.
1 Like