SuperMicro Long Beep and CPU Temp Tweaks

Lately, my SuperMicro X10SLL-F has lately been producing a loud BEEP in the middle of the night, which is very annoying. The potential cause is a high CPU temperature at 100°C which seems to happen almost every night at 2:00 AM.

My scrub and snapshot tasks are at midnight, and my S.M.A.R.T. tests are at 2AM Sunday only and 3AM twice a month for the long test. This could partly match the pattern but does not account for the daily 2AM spikes in CPU Temp. The CPU does not run hot during normal use, and the heat sink appears to be seated correctly.

I have had issues with this MB alerting for low fan speed before, which I corrected through IPMI using

ipmitool sensor thresh "FANA" lower 50 100 150

Would it be safe to run this and see if it helps? Otherwise, any other solution?

ipmitool sensor thresh "CPU Temp" upper 110 120 130  

Better or more fans for cooling, better heatsink, better airflow case. The highest I have reached in the past month is 41.7C with my xeon. 100C is quite hot! What cpu?

I think the beeping isn’t for your CPU, it’s for the motherboard itself. You need to solve your cooling issues most likely.

1 Like

Oh yes, 100C is way too hot, even if the cpu was running at 100% you should still be able keep it under 75C or in that area.

There are a lot of things that could cause the heat, airflow throufh the case, too small cpu heatsink, improperly installed heatsink, fan spinning too slow.

But that does not address the issue you asked about. Reboot TrueNAS and see if that fixes it. Disable all running VMs, jails, etc.

You need to do some legwork.

Also, SCALE does still have memory leaks so there could be something else goimg on, you might file a bug report.

X10 is older and presume original cooler install. I would look at the cooler area and re-paste the cpu to clean that up. 100c is not healthy and you’d should run fans at full capacity till this is sorted.

Anything environmental change lately? Dust blockage? Construction?

1 Like

Has been already been said, but that’s not enough.
You are reaching really danger temperature!
You should check if thermal paste is fine (if has been ever changed :smiley: ), or if for some reason the cpu fan not working properly

1 Like

Hi all thanks for your replies. Here is some additional info.

  • Processor: Intel Xeon E3-1280V3
  • Cooler: Intel E97379-003 (new with its factory pre-applied paste)
  • CPU fan appears to be spinning normally
  • Not running jails and the only app configured is Plex…
  • Case is a Fractal Node 805 fully loaded with 5 Arctic fans in addition to the two factory fans
  • This build is less than 1 month old so dust is not an issue

In addition, see daily heat graphs from the past 3 days:



CPU really shouldn’t be getting to that temp. It’s right up to tjMax or whatever.

If your cooling was working correctly I’d expect it to cap out at about 70C or so.

Ensure cpu fan is running at max.

I’d reapply the thermal paste I think, if the fan is running at full speed and hasn’t got a bung bearing etc.

I have just configured fans to run at max. Will update this thread if this does not fix the problem. How could the problem be the cooler or thermal paste if they are both new and installed correctly? Is my cooler insufficient for my CPU?

Repaste and reseat that cooler. Despite that cooler being trash, there is no way a CPU with a properly installed cooler should be getting that temps. Something’s not quite right. I would recommend looking into a better cooler like the Thermalright Peerless Assassin 120 SE but check for clearances on that case…

1 Like

Havne’t been able to find what TDP the cooler is rated for, but the TDP of the CPU is 82W.

Generally, I’d expect the 1280 to be i7/i9 class CPU.

This table shows that the i7/i9 CPUs (and some 65W i3s) do not use that cooler… and only the i5/i3s do…

https://www.intel.com/content/www/us/en/support/articles/000058835/processors.html

Could simply be the cooler is not sufficient to dissipate the thermal load, and then the CPU hits junction max, probably 100C, and that means it starts throttling until it cools down to avoid destroying itself.

I personally use Noctua coolers.

1 Like

Max TDP for that cooler is 65W I believe. That may be the issue.

2 Likes

Did you remove the plastic typically covering the paste? I’ve done that and hated myself for it, waste of pre-applied paste, had to use other paste that I thankfully had. I did catch it before the system was powered on but still, stupid mistake.

Yes, and IMHO Is not enough neither for those in all situation.
I think this Is in boxed version of higher tdp CPU

Don’t know for newest CPU.

Still sounds strange to me can be the only reason :thinking: