Hey guys new here but very excited to be here. I set up my new unraid box and am having a ton of issues the past two days.

Problems:

System crashes with unclean reboots roughly every 45–90 minutes (sometimes faster). Crashes happen at idle with no load, no parity running, and even with all Docker containers stopped. No apparent trigger.

The most annoying symptom, syslog mirror to flash captures nothing most of the time. The system dies so instantly that it has no time to write any log entries. I even have enabled mirror to flash on my syslog, and don't really get anything meaningful most of the time. Occasionally I got a lead that I would track down (listed below) but so far nothing has panned out.

Things I think I have ruled out:

• RAM — No errors, memtest ran several hours and several passes.

• Flash drive — Healthy 120GB Samsung, not being filled up with logs (despite persistent logging enabled).

• Incoming power — UPS shows stable 120V in/out, only one logged event (intentional unplug). Other devices on same UPS function fine and are not experiencing any issues. Also have whole rig wired through a Killowatt device and the readouts at the wall all look normal.

• Loose connectors — All reseated (24-pin ATX, 8-pin CPU, SATA power/data, RAM).

• Software (mostly?) - Crashes even if docker daemon is completely disabled, system totally idle, temps all well below 40c, drives not reading/writing anything. Several leads software related have turned out to be nothing (rabbit holes below).

Rabbit holes I have gone down:

Autofan: My first instinct was autofan. I had to do some extra tinkering to get it to work with my setup to begin with. Initially nct6683 was what was being used and was only giving a few temp monitor options. I installed the nct6687 plugin from the community apps and defaulted autofan to use that. That resolved my issue and allowed me to get accurate temp readouts and gave unraid the ability to push PWM commands to the motherboard. I also created a user scripts to auto run autofan on array start. When the crashes started, it seemed like a logical spot to start as it seemed to be the most invasive thing I had done to the system up to that point (outside of setting up standard docker containers). Completely removed the whole plugin, got rid of the script that was autorunning on boot, removed the nct6687 plugin, and reverted back to the nct6683 driver (also tried blacklisting this at one point).

Tailscale Daily Cron: I noticed in one log that the crash happened a bit after this was running so I added a line in my go file to remove it (not sure if this was smart so I added it back as it didn't seem to have an effect). I thought it was something to do with updating tailscale automatically and was curious if that was causing some lockups (again, just the most recent thing in the logs at the time).

rm -f /etc/cron.daily/tailscale-daily 2>/dev/null

Docker Container Network: With all the crashing, the docker containers were coming up and down a lot. At one point in the logs the most recent thing was Homarr attempting to poll some containers for some integrations. Those containers had not come back up after Gluetun rebooted because they depend on Gluetun for their networking. Homarr was getting a bunch of errors back and I thought maybe that was causing the crashing. Started with turning off containers one or two at a time. Eventually turned off all docker containers and even killed the entire docker daemon process.

Where I am at right now:

I kind of am worried it is something hardware. It just feels that way to me, but I am very new to unraid and this could be par for the course. I really don't know how I can confirm this theory though...outside of buying a different motherboard and PSU. They are both within return windows, I just need to know how best to isolate this further. Also, is mirror to flash the best we can get as far as logging goes? Not sure if there is something I am missing with logging to get a better picture of what exactly is going on.

Any help is greatly appreciated. Glad to be apart of the community.

System Build:

Unraid Version: 7.2.4
Motherboard: MSI PRO B650M-A WiFi
CPU: AMD Ryzen 5 8500G
RAM: 32GB DDR5
Array: 3× 28TB Seagate ST28000NT000 HDD + 1TB Samsung 990 EVO Plus NVMe cache

Unassigned: Samsung 840 EVO 250GB (backup drive)

PSU: Corsair RM1000x

UPS: CyberPower CP1500PFCLCD (connected via USB, running apcupsd)

Boot Drive: 120GB Samsung Flash Drive

Edited April 8Apr 8 by Hobbit

Attach Diagnostics ZIP to your NEXT post in this thread.

On 4/8/2026 at 1:50 AM, trurl said:

My friend, I could give you a hug right now. This was the issue. Some combination of my RAM being set at the full speed of 6000Mhz and the global C states being enabled seems to have been causing the crashing. I was not able to find any option in my BIOS for the Power Supply Idle Control settings, but turning off global C states and setting RAM back down to 4800 Mhz seems to have resolved my issues.

Thank you again, I spent hours and hours looking into this. I need to be better about using the forums, I just wasn't aware just how good they were.

Overclocking a file server doesn't really make much sense anyway.

1 minute ago, trurl said:

Overclocking a file server doesn't really make much sense anyway.

I agree, I just always have built personal rigs where I overclock everything. Kind of thought it made sense to max it out for things like running docker containers and what not. I am just so happy it is stable so I can properly play (until I break it again) :D