TrueNAS Core going offline everyday

[jbarry14]

I built a truenas core server several months ago and it has been working flawlessly up until a few days ago. It is running TrueNAS-13.0-STABLE. It now either crashes or goes offline everyday. The power light on the rack goes out, but the disk lights are still on. I have to force shut down the machine and restart it to come online. When it comes back up, it seems like everything is fine and works great. I don't have any error messages that I am aware of. It just says "the system had an unscheduled system reboot. The operating system successfully came back online"

What could be causing this issue?

 

[HoneyBadger]

I'm unsure if EPYC processors necessitate the same adjustment of power states (disable C6 states, dynamic power control methods like Cool N Quiet, eRP control, etc in BIOS) as Ryzen, but given that they're both AMD processors it would seem a prudent place to start.

Do you see anything unexpected in the IPMI log of the system? ipmitool sel elist should show this from the shell if you don't have it set up, and then ipmitool sel get <recordid> to pull more details if anything is present.

 

[jbarry14]

I'm unsure if EPYC processors necessitate the same adjustment of power states (disable C6 states, dynamic power control methods like Cool N Quiet, eRP control, etc in BIOS) as Ryzen, but given that they're both AMD processors it would seem a prudent place to start.

Do you see anything unexpected in the IPMI log of the system? ipmitool sel elist should show this from the shell if you don't have it set up, and then ipmitool sel get <recordid> to pull more details if anything is present.

I was reading other posts about that bios change for AMD processors. I will make that change. I will also update the OS to 13 U2.

I typed that command into the shell. This is what I got, Just a lot of messages about the fans, but nothing else. - 42f6 | 10/14/2022 | 14:15:18 | Fan FAN1 | Lower Critical going low | Deasserted | Reading 1960 < Threshold 420 RPM