Hi there! I have a AMD based machine which is experiencing random restarts/freezes and I do not know how I can identify the culprit.

Machine is running debian 10 and open media vault 5. It is an AMD based machine. Until now I have tried with different MoBo, CPU, CPU cooler, PSU but the random restarts are persisting.

CPUs are AMD Ryzen 5 1600 and AMD Ryzen 7 1700 on 2 B450I MoBos one from Asus and one from Gigabyte with Noctua and stock CPU cooler. PSUs are one Silverstone Flex ATX (miserable, but fitting a micro-ITX case) and a BeQuiet Pure Power 11 400W (which should be decent at least).

What I have not changed (so far) is the RAM (memtest with 10 passthrough runs now in the background), the SSD (WD Blue M.2 Nvme which is new, but one never know) and the OS.

My question is if I can dig somewhere in the logs to find the culprit of these restarts.

Thanks!

One problem with crashes is that they may prevent logging. Check "last" to see what it says for the failed session. If there is no stack trace shown on the console, and the system just restarts, then I would suspect power issues. Mine will show two "still running" lines in the event of a sudden power outage, like:
No reason is shown for the 10:11 boot this morning because I accidentally disconnected power after starting the system. The Oct 21 session shows the line for a normal shutdown.

First you will probably have to enable persistent logging, then, after a freeze/restart, issue

It seems that I have the same: the one from 19:33 is the faulty one and system restarted on 28 Oct 20:41... should I suspect then power issues? on the PSU or from the grid?

I have enable it now, and I will have a look after next restart...

Thank you both.

Those power supply brands don't sound top notch quality. Is either under warranty? If not, open up to inspect for leaky or swollen electrolytic capacitors. Visit http://badcaps.net/ to know what to look for. If you find any, you have your answer. Power supplies are notorious sources of random lockups.

Silverstone is an entry level PSU but was the only Flex-ATX that I have found on the market. Be Quiet! is a pretty well established brand here in Europe. Both of them are new and under warranty.

@beje ,the problem with looking at logs because of hard freezes and reboots is when what ever is the cause of it will not show in the logs because all writing stops. I have experienced the same issues you are experiencing. I tried memory tests, replaced the power supply, replaced the video card. This all started after 11 months of my computer build. After another two months of trying to figure out what was causing it the culprit showed itself when my computer would no longer boot up. In my case it turned out to be the motherboard failed which ASUS only had a 12 month warranty. Apparently ASUS quality is not what it once was. I replaced it with a MSI motherboard that ended up dying after two months. I now have a Gigabyte motherboard that has been great so far which is 1 year and 8 months. I would recommend you check in the BIOS for the CPU settings so it doesn't idle because that setting can cause the exact problem you are experiencing with Ryzen CPUs if they idle. Also, make sure you are running Linux kernel 5.3 or higher because earlier kernels can cause those problems with Ryzen CPUs.

If using DDR4 RAM I recommend thorough/overnight testing with memtest86, not memtest86+.

I have done it: 10 cycles and nothing :-)

The last suspect is a (cheap) wattmeter that I have used to measure power consumption. Moved it away and could not reproduce the symptoms and getting 10 days of uptime. Now the respective HW was repurposed... new ideas, new things to do.

Thank you all.

Good news! In that case you might want to mark the thread solved.