TrueNAS Scale - Random Reboots Under Full CPU Load

TrueNAS General
,
1

Hey everyone,

I’m running TrueNAS Scale on a system that was upgraded a few months ago, and since around mid-October, I’ve been dealing with random, sporadic reboots, which have become more frequent (every few hours, pretty randomly).

Under normal usage, everything seems fine. CPU temps, usage, and logs show nothing suspicious. However, I’ve noticed something consistent: if I put the CPU under 100% load, the system will reboot after about 60 seconds. CPU temps never exceed about 69°C, and power draw at the wall is around 160W right before it goes down.

At first, I suspected the second CPU power cable might be at fault, but it’s properly connected. Interestingly, if I disconnect that cable, the system reboots even faster (within 2–3 seconds of full load).

There is nothing in the system logs that indicates what’s happening. I’m out of ideas at this point and would really appreciate any suggestions on how to debug or pinpoint the cause of these reboots.

Hardware details:

  • Motherboard: MSI X570-A PRO

  • CPU: Ryzen 7 3700X

  • RAM: 32G Corsair Vengeance LPX DDR4

  • PSU: Corsair CX750M

  • Storage: 4x WD Red 6TB

  • Cooling: stock AMD cooler

Thanks in advance for any help!

2

If i may say the first thing i would check in your situation… try swap PSU (with cable!).

Second thing Is a memtest, after have unplugged everything not essential

1 Like
3

Thanks for the suggestion!

I had some time to do some testing…

Memtest passed. It ran for almost 5 hours, it did 4 passes of all the tests with all CPU cores and gave no problems.

Trying each stick of RAM independently and detaching each of the HDDs didn’t solve the issue.

I’m now ordering a replacement PSU… even if it was able to run fine during the memtest with a sustained load of ~140W and a peak of 175W, which is more than it draws when it shuts down :person_shrugging:

4

Memtest Is a good indicator to understand if (very short) RAM and CPU are working well together, but in reality not stress so much the CPU. I would try a CPU oriented stress test also, if the culprit Is hardware side should have the same instability. Sound strange to me now also point to the PSU if It could handle the test without issue, i know i have suggested to check It but maybe was too early to order one :smile: Hope in case you can send It back without pay anything :weary:

5

Hmm, is there any automatic overclocking done in the bios which suddenly got turned on ?
Asking because of this.
Simple CPU stress test:

for i in $(seq $(getconf _NPROCESSORS_ONLN)); do yes > /dev/null & done

stop with killall yes

1 Like
6

Lo and behold… it was the PSU :’)

2 Likes