Minisforum MS-01 BIOS

[updatelee]

can someone check if this update fixes the amt nic eating the dhcp server responses if hosted on that nic ?
i think the response gets rerouted to the amt and than droped by the amt(mebx)

This is an issue with all Intel I226-LM I've ever used. I226-V do not have this issue (but dont support AMT), so for now if Im going to need DHCP I use the i226-V models. I cant see how a bios update would fix this as Im posititive its an issue in the card itself.

 

[reneil1337]

It's been a week on v1.27 without any crashes. I didn't downgrade the memory speed to 4400 (as that didn't help in 1.24 either) but reduced the maximum CPU frequency as suggested in the readme. So far so good, will report back after 1 month.

Update: July 22nd still running without a single issue after 14 days

 

[Tarwater]

It's been a week on v1.27 without any crashes. I didn't downgrade the memory speed to 4400 (as that didn't help in 1.24 either) but reduced the maximum CPU frequency as suggested in the readme. So far so good, will report back after 1 month.

Update: July 22nd still running without a single issue after 14 days

Hi, which CPU do you have? I have 12600H and tempted to install 1.27, I am on 1.26 but put off because of high idle in 1.27

 

[reneil1337]

I have the i9-13900H and modified couple things you can ready my write up here Unraid fueled MS-01 Homelab Server - HackMD

 

[seraphimserapis]

I'm on 1.27 (latest microcode, too) with my 13900h (96GB RAM). I changed the Turbo Ratio Limits as instructed in the README and set the RAM frequency to 4400, too. I still see massive temperature spikes (usually on a single core) that hit 90+ degrees celsius.

I already redid the thermal paste (using Noctua NT H2) but that is still not really helping either. While I don't get regular crashes during idle (e.g., running a Ubuntu VM with some docker containers and 2-3 LXCs), I get crashes due to temperature reaching 100 degrees Celsius as soon as I run things in parallel like updating both the Proxmox host and the VM in parallel.

Any ideas on what else to try would be welcome. Worst case, I need to go through the RMA process.

Here is my temp at idle:

sensors
coretemp-isa-0000
Adapter: ISA adapter
Package id 0: +48.0°C (high = +100.0°C, crit = +100.0°C)
Core 0: +46.0°C (high = +100.0°C, crit = +100.0°C)
Core 4: +45.0°C (high = +100.0°C, crit = +100.0°C)
Core 8: +45.0°C (high = +100.0°C, crit = +100.0°C)
Core 12: +45.0°C (high = +100.0°C, crit = +100.0°C)
Core 16: +44.0°C (high = +100.0°C, crit = +100.0°C)
Core 20: +45.0°C (high = +100.0°C, crit = +100.0°C)
Core 24: +47.0°C (high = +100.0°C, crit = +100.0°C)
Core 25: +47.0°C (high = +100.0°C, crit = +100.0°C)
Core 26: +47.0°C (high = +100.0°C, crit = +100.0°C)
Core 27: +47.0°C (high = +100.0°C, crit = +100.0°C)
Core 28: +48.0°C (high = +100.0°C, crit = +100.0°C)
Core 29: +48.0°C (high = +100.0°C, crit = +100.0°C)
Core 30: +48.0°C (high = +100.0°C, crit = +100.0°C)
Core 31: +48.0°C (high = +100.0°C, crit = +100.0°C)

nvme-pci-5900
Adapter: PCI adapter
Composite: +32.9°C (low = -0.1°C, high = +85.8°C)
(crit = +87.8°C)
Sensor 1: +32.9°C (low = -273.1°C, high = +65261.8°C)

acpitz-acpi-0
Adapter: ACPI interface
temp1: +27.8°C

nvme-pci-0100
Adapter: PCI adapter
Composite: +43.9°C (low = -0.1°C, high = +99.8°C)
(crit = +109.8°C)

nvme-pci-5800
Adapter: PCI adapter
Composite: +31.9°C (low = -0.1°C, high = +85.8°C)
(crit = +87.8°C)
Sensor 1: +31.9°C (low = -273.1°C, high = +65261.8°C)

 

[reneil1337]

reporting in: 1 month online with bios v1.27 and still no crash

 

[tomado09]

First; thanks for all the comments here! It's really amazing to read all the small findings, to get some ideas on where and what to look for, when troubleshooting issues with the MS-01 (which is an amazing piece of machine/server - when it's stable..). Here is some of my recent findings related to kernel/process crashes, and findings on temperature - hopefully it can help!

BIOS 1.27 AND suggested adjustments in BIOS possibly solved random crashes
I can confirm that BIOS 1.27 (MS-01-AHWSA-V1.27_4_28_V2.zip), in addition to follow the steps mentioned in "0-READ-ME.txt" (downgrade memory speed to 4400, disable overclocking lock and reduced max CPU frequency/turbo ratio limit option), seem to have fixed my seemingly random kernel and process crashes on Proxmox 8.4.1.

Based on my current research, most crashes/debug traces pointing towards memory corruption (which according to memtest86+ after 18 hours is healthy.. but good thing I ran it, cause that was where I noticed the high temperatures). Since performing the adjustments suggested in "0-READ-ME.txt", I've not experienced any issues. It's now been 8 days without any issues (I follow the journal logs/kernel closely using loki), whereas I previously experienced processes crashing "randomly" earlier, like pvestatd crashing.

PS: I also noticed that process crashes/issues did not start before I loaded the server to about 20% average CPU (based on Proxmox metrics from pvestatd, which ironically was one of the process resulting in regular "general protection fault").

I also have two (2) MS-01 still on 1.26 without any issues (yet), but they are loaded less (between 1-8 %), and CPU is running way cooler, e.g. mean values celsius 55-63C (mean) on less loaded nodes vs 76-80C (mean) and spiking to 90-99C on the MS-01 with 20-25% load (still do/identical to on 1.26 BIOS).

My hardware configuration
For reference, I have four (4) MS-01 with the Intel i9-13900H configuration, upgraded RAM to 96GB on each, with added 2x NVMe (using all three M.2 slots, on different speeds), and using both the vPro NIC (226-LM) and both 10Gb NICs (X710). No added expansion card.

High temperatures (CPU)
As mentioned, I've ran memtest86+ for about 18 hours (on BIOS 1.26), and while the results was a "PASS" every time (see screenshot below), I also noticed the high temps, sometimes over 90C, which got me a bit worried (CPU is rated to 100C max, AFAIK). For now, the MS-01 "servers" are physically located in the basement, with about 18-23C room temperature (summer in Norway..), only with stock options for fan/thermal paste, etc. (not replaced thermal paste.. yet).

View attachment 44373

I've installed lm-sensors and correlated the package/core IDs with output from node_exporter to closer monitor the temperatures (using Prometheus) for each P/E-core, and noticed only "minor to none difference" for temperature on both P/E-cores, both on 1.26 and after 1.27 at approx. same CPU-load - but the crashes stopped!

For some more reference, here is last 24h, sorted on mean temp on each core, first on a node with ~20% CPU load avg. (now on BIOS 1.27 with adjustments suggested in, not experiencing any crashes anymore).

View attachment 44374

And another node with ~1-3% CPU load avg (on BIOS 1.26, but not experienced any crashes yet).

View attachment 44375

And just to compare temperatures on 1.26 vs 1.27, I've highlighted "temp6" (from node_exporter, which correlates to P-core 4 in lm sensor), I don't see much difference (time ~16-18 is with BIOS 1.26, then it crashed, and ~20:45+ is on BIOS 1.27). IMO, no changes in temperature, but the crashes stopped, which was my main issue on the node.

View attachment 44376

And from the logs on the node the experienced crashes, here is a quick overview of count on "critical" and "coredump" from journal. Also note that issues seemed to escalate after upgrading to kernel 6.8.12-11-pve vs previously 6.8.12-10-pve, but then later I experienced issues on all 6.8-kernels (tried multiple versions), even tried rolling back to 6.5.13. But then, after BIOS 1.27 + adjustments was performed 2025-06-26, I've not seen any crashes (yet).

View attachment 44378

And for reference, from 0-READ-ME.txt, BIOS 1.27 for Minisforum MS-01, extracted from MS-01-AHWSA-V1.27_4_28_V2.zip:

This BIOS Improved stability.
If you have had problems with crashes, restarts, and hangs before, you can flash this BIOS.
If your problem still happens.
1. You may need downgrade memory speed to 4400.
    There have a menu in Onboard Devices Setting.

    This BIOS will take more idle power consumption.
    Linux: 14.5W-> 17.5W
    Windows: 14.5W -> 22W
    (Data tested with 32G 5200Mhz RAM x 2 and 1xPCIE4.0 SSD)

（Recommend)
2. Disable Overclocking Lock:  Advaned -> CPU configuation ->Overclocking Lock:[ Disable]
Reduce the maximum CPU frequency:
Advaned -> CPU configuation ->Turbo Ratio Limit Options.
P-Core Trubo Ratio Limit Ratio0: 51
P-Core Turbo Ratio Limit Ratio1: 51
E-Core Trubo Ratio Limit Ratio0: 40
E-Core Turbo Ratio Limit Ratio1: 40
E-Core Turbo Ratio Limit Ratio2: 40
E-Core Turbo Ratio Limit Ratio3: 40
Save and reboot enter into BIOS. Advaned -> CPU configuation ->Turbo Ratio Limit Options. Check is it max frequency has been reduced.

Reboot into System. Try is it more stable.



If all of this cannot solve problem, you may need do a request RMA.

I'd be genuinely curious if you can clock the CPU back up to normal with a repaste...especially with the stability issues occurring for you with higher CPU loads and with the abhorrent MS-01 stock paste jobs that I've seen reported.

I'm getting my MS-01 tomorrow - finally delivered after about a month of waiting. Should have gotten it on Amazon, lol.