The CPU has been posting L1 cache errors in the TrueNAS logs at the same memory address for 2 days, it had trouble posting in the past, but a power cycle usually got it through. This morning, after an unscheduled restart from the server, I power cycled it from S5 to access the BIOS and update it, and it never bame back alive.
The board presents a beep code for no ram when no ram is installed, but doesn’t make it past post. I took that as a sign of life and tried with spare corsair non-ecc memory sticks in all possible correct configurations, nothing makes it post. Reseating the CPU didn’t work either.
I’ll try tomorrow with a 3600X I have laying around, but in your opinion is the motherboard dead or the CPU?
First, L1 cache is inside the CPU, so if there are problems with that, then that would suggest CPU.
Second, having the beep code for no ram, even with trying all slots suggests that the memory controller is having issues. I wouldn’t expect having problems with all of the physical traces to the memory all at once, same with the ram sticks, so that suggests memory controller to me. The memory controller is also inside the CPU package for Ryzen chips.
I tried with a 3600X today, no dice. I don’t know for sure that the bios was updated for it, I don’t remember. However I now also know the 2600X is for sure dead, as it doesn’t boot in my X570 board I also have laying around. I started an RMA request with AMD…
