LibreElec/Kodi Will Reboot/Freeze

I had a Windows-based Kodi machine that I rebuilt with new hardware and instead used LibreElec after it had gotten to be over a decade old and was having problems. After getting it up and running, at first I had issues where it was taking anywhere from 5 to 20 minutes to recognize and connect to the network SMB media shares on my Debian server (had intended to post about this but didn't, and that's something for another time if I get this resolved), but then I started noticing that if the machine was running for extended periods, everything would freeze up. Usually this would be because I was watching something and had to pause, but never made it back for a long while and forgot it was still running. It's gotten worse. Now it will just straight up freeze without having been running for very long. Doesn't make a difference whether or not a video is playing. It will happen even when on the main screen or in the settings. Sometimes instead of freezing it will do a kind of reboot where it quickly drops to the LibreElec splash screen and then Kodi pops right up again as though it had just been started. This was not something I had ever encountered in the machine's original configuration, nor with another machine I had retired from service that was essentially a repurposed generic desktop PC.

Due to the nature of this, it was hard to get proper logs. There was nothing logged at all during a freeze until the machine was forcibly rebooted. I was finally able to capture some logging when it instead did a reboot/reset. I am on LibreElec 12.0.2 with Kodi 21.2. Here's a link to the log I pulled:

Also, in case this could be a hardware issue, here are some key details about the machine:

- AMD Ryzen 5 7600X

- AMD B450 chipset

- Radeon graphics, OpenGL ES 3.2 Mesa 24.0.9

- 16GB DDR5 RAM

I was poking around in some of the settings when I found some system data that showed CPU temperature. I'm not sure it's being displayed correctly. The numbers look right for C but the unit of measure is F. Around the last failure, it was showing 112F, which would be much lower than I would expect for a CPU if correct, but very concerning if it were in actuality 112C. I guess either way I ought to opening up the machine and poking around inside.

Ok, I've gone the PSU route, since that's the easiest to manage right now. It also makes some sense, given the symptoms of a failing PSU. On top of that, I've been having trouble getting it to turn on lately. Figured it was just the button on the case given that it used to be at floor level and from time to time I'd just bop it with my foot to turn it on. Thought it was a bit messed up and worn out, and that's why it was difficult to turn on. I grabbed a spare PSU I had sitting around and tossed it in. Turned on right away without having to fiddle with the power button. I'll run the machine through its paces, see if that takes care of the issue.

I've had PSUs fail on me in the past, but not like this. In the 20+ years I've been (re)building PCs they've always just flat out failed altogether. This would be the first time I've had one fail in this manner. The kicker of it is that when connecting up my PSU tester, this supply is just fine according to it. I guess it doesn't put enough load on it to trigger a fail condition, it just tells you if each of the connector types is putting out correct voltages.