I spent enough time to contact you with my problem.
I am setting up my Truenas, but unfortunately my server runs on an AMD processor. As I understand it, for the correct display of the temperature in the system, you need to add sysctl with the command dev.amdtemp.0.sensor_offset. But it gives an error when adding “Sysctl ‘dev.amdtemp.0.sensor_offset’ does not exist in kernel.”
Now the question is, how to add amdtemp to in kernel?
I am a user in this and do not understand how to do this and what commands to execute.
I understand that I need to run commands to install amdtemp through the shell, but I can’t figure out how.
If possible, write what needs to be run through the shell.
I’m using an amd ryzen 3700x and my temp gets reported correctly.
And to answer your question: truenas is an appliance os and it’s not supported or advices to add anything manually to the install.
My advice - raise a Bug ticket, upload the dump-file and see what iX say about this. Best case they can advise how to make it work - worst case they will hopefully build it into the next minor version.
You can’t install a FreeBSD device driver in Linux. Whatever guide or steps you have found are likely for TrueNAS Core (FreeBSD based) and not TrueNas Scale (Linux based).
Let’s say there is a way to set the temperature offset in Scale, what value would you set it to? How do you know the temperature is incorrect and what the correct temperature offset should be?
But that should only be a problem on core, for scale even my oldest 1600x never had problems with reporting the correct temps - even when scale was still in its first late beta phase.
I completely agree, that’s why I created a topic here, and I hope for the community. Which can provide a solution.
I thought maybe there is a possibility to shift the temperature data.
My server at 100% CPU load shows a temperature close to 100C (although the real temperature does not exceed 50C) and in the TrueNAS SCALE system the temperature protection is triggered and turns off the server.