Supermicro fan control for Linux (home) servers.

This is a systemd service running on Linux that can control fans with the help of IPMI on Supermicro X10-X13/H10-H13 (and some X9) motherboards.

• a Supermicro motherboard with ASPEED AST2400/2500/2600 chip
• Python 3.10-3.14
• a Linux distribution with:
  • systemd and bash
  • coretemp kernel module for Intel CPUs or k10temp kernel module for AMD CPUs
  • drivetemp kernel module (kernel version 5.6+ required) for SATA HDDs/SSDs
• ipmitool
• optional: smartmontools for SAS/SCSI disks and standby guard feature
• optional: nvidia-smi for GPU fan controller

1. Set up the IPMI threshold values for your fans (see chapter 6. for more details)
2. Optional: enable advanced power management features for your CPU and SATA hard disks for lower power consumption, heat generation and fan noise.
3. Load kernel modules (coretemp/k10temp and drivetemp)
4. Install smfc service (see chapter 9. for more details) or run smfc in docker (see more details in Docker.md)
5. Edit the configuration file /etc/smfc/smfc.conf and command line options in /etc/default/smfc (see chapter 10. for more details).
6. Start smfc service (see chapter 11. for more details)
7. Check results in system log
8. Leave feedback in discussion #55

Feel free to visit Discussions and raise your questions or share your experience on this project.

This service was designed for Supermicro motherboards with IPMI functionality. It implements fan controllers that control fan speed dynamically in one or more IPMI zones. The service operates the fans in IPMI FULL mode, where the fan rotation level can be adjusted with IPMI raw commands (read more details here).

Key features:

• Five independent fan controllers (CPU, HD, NVME, GPU, CONST) that can be enabled/disabled and combined freely
• User-defined control function mapping temperature intervals to fan level intervals with configurable discrete steps
• Support for multiple IPMI zones with automatic shared zone arbitration (highest fan level wins)
• Temperature calculation methods: minimum, average, or maximum across multiple devices
• Temperature smoothing with configurable moving average window to reduce fan speed oscillation
• Sensitivity threshold to avoid unnecessary fan speed changes on small temperature fluctuations
• Standby guard feature for SATA hard disk arrays organized in RAID
• Support for SATA, SAS/SCSI, and NVMe disks with automatic HWMON/smartctl fallback
• Nvidia GPU temperature monitoring via nvidia-smi
• Platform abstraction for different Supermicro motherboard generations (X9-X13/H10-H13)
• Runs as a systemd service or in Docker
• Safe shutdown: all fans are set back to 100% speed at service termination

IPMI zone is a logical term, representing a cooling zone, where there are predefined fans having the same rotation speed. Please note that the fan assignment to an IPMI zone is predefined on the motherboard, it cannot be changed (Supermicro does not provide individual fan configuration features in IPMI, while other vendors do it). On a typical Supermicro motherboard, there are two IPMI zones:

• CPU or System zone (IPMI zone 0) with fan names: FAN1, FAN2, etc.
• Peripheral or HD zone (IPMI zone 1) with fan names: FANA, FANB, etc.

On Supermicro server boards, there could be more IPMI zones with different fan names (see issue #31).

smfc v3.8.0 and earlier versions implemented a feature (called Swapped Zones) to swap IPMI zone 0 and 1. From smfc v4.0.0 the IPMI zones can be assigned freely to fan controllers providing more freedom and convenience for the user (see ipmi_zone= parameter for more details).

In smfc, the following fan controllers are implemented:

These fan controllers can be enabled and disabled independently. They can be used in a free combination with one or more IPMI zones. Multiple fan controllers can share the same IPMI zone -- smfc will automatically apply the highest fan level requested by any controller in that zone (see chapter 1.3 for details). CONST fan controller is an exception here, it does not require a temperature source, it can provide a constant fan level for one or more IPMI zones. In smfc configuration file each fan controller has an individual section.

In smfc, a temperature-driven fan controller implements the following control logic:

1. it reads the temperature
2. it calculates a new fan level based on the user-defined control function and the temperature value
3. the service applies the fan level for the IPMI zone(s) with IPMI commands (i.e. ipmitool)

If the temperature source has multiple instances (e.g. multiple CPUs, HDDs, NVMEs or GPUs) then the user can configure a calculation method (i.e. minimum, average, maximum) for the calculation of the final temperature value (see temp_calc= parameter).

Please note that smfc will set all fans back to 100% speed at service termination to avoid overheating!

When multiple fan controllers are assigned to the same IPMI zone, smfc detects this at startup and automatically switches to a two-phase arbitration loop for those controllers. Controllers on non-shared zones are not affected -- they apply their fan levels directly.

At startup, smfc logs the detected shared zones at INFO level:

Shared IPMI zone 1: ['HD', 'NVME', 'CONST']

For shared zones, the control loop uses a two-phase approach in each iteration:

1. Compute phase: each fan controller on a shared zone reads its temperature source and calculates its desired fan level, but defers the IPMI call.
2. Apply phase: the service collects all desired levels, groups them by IPMI zone, and applies the maximum level per zone. Only one IPMI command is sent per zone, and only when the level has actually changed.

Controllers on non-shared zones skip the apply phase entirely -- they execute their own IPMI calls directly during the compute phase, just like they would if no sharing existed.

This means the hottest component always wins on a shared zone. For example, if HD fan controller wants 45% on zone 1 and NVME fan controller wants 70% on the same zone, smfc will set zone 1 to 70%. When the NVME cools down below the HD temperature, the HD controller's level will take over.

The CONST fan controller also participates in the arbitration -- its constant level acts as a guaranteed minimum for its zone(s). For example, configuring [CONST] level=40 on zone 1 ensures that zone never drops below 40%, even if all temperature-driven controllers would request a lower value.

When a shared zone's level changes, the log output at INFO level shows the winning controller and lists all other controllers with their requested levels:

Shared IPMI zone [1]: new level = 70% (winner: NVME=70%/52.0C, losers: HD=45%/38.5C, CONST=40%)

For non-shared zones, only the applied level is logged:

IPMI zone [0]: new level = 60% (CPU=45.0C)

Fan controllers use user-defined control functions that map a temperature interval to a fan rotation level interval.

The following five parameters will define such a function:

 min_temp=
 max_temp=
 min_level=
 max_level=
 steps=

In this way, a fan controller can map any new temperature value to a fan level (from Celsius degrees to % value).
Changing the fan rotational speed is a very slow process (it could take several seconds depending on the fan type and the requested amount of change), so we try to minimize these kinds of actions. Instead of setting fan rotational speed continuously, we define discrete fan levels based on steps= parameter.

The fan controllers implement the following strategies to avoid/minimize the unnecessary change of fan rotation speed:

1. When the fan rotational speed is changed, it always applies a delay time (defined in [IPMI] fan_level_delay= configuration parameter) to let the fan implement the physical change.
2. There is a sensitivity threshold parameter (sensitivity=) in the fan controller configuration. While the temperature change is below this value, the fan controller does not react.
3. The configuration parameter polling= defines the frequency of the temperature reading. The bigger polling time, the lower frequency of the fan speed change.
4. The smoothing= parameter enables a moving average window for temperature readings. When set to a value greater than 1, the fan controller averages the last N temperature readings before making fan level decisions. This reduces fan speed oscillation caused by brief temperature spikes. For example, smoothing=4 averages the last 4 readings. Set to 1 (default) to disable smoothing.

For the HD fan controller, an additional optional feature was implemented, called Standby guard, with the following assumptions:

• SATA hard disks are organized into a RAID array
• the RAID array will go to standby mode recurrently

This feature monitors the power state of SATA hard disks (with the help of smartctl) and will put the whole array into standby mode if a few members have already stepped into that state. With this feature, the situation can be avoided where the array is partially in standby mode while other members are still active. SCSI disks are not compatible with this feature.

The following table summarizes how the temperature is read for different disk types:

The smfc service was originally designed for SATA hard drives, but from smfc v3.0.0, it is also compatible with NVME and SAS/SCSI disks. smfc v5.0.0 introduced a new NVME fan controller and separated the use of SATA and SAS/SCSI HDDs/SSDs and NVME SSD disks.

Some additional notes:

• For SATA disks the drivetemp kernel module should be loaded. This is the fastest way to read disk temperature, and the kernel module can report the temperature while hard disks are in sleep mode!
• For SAS/SCSI disks the smartctl command will be used to read disk temperature.
• If drivetemp module is not loaded or an HDD is not compatible with drivetemp module then smfc will use smartctl automatically.
• Different disks types can be mixed in hd_names= configuration parameter but the Standby guard feature will not be supported in this case.
• For NVME SSDs, no kernel driver needs to be loaded; the kernel can handle this disk type automatically
• NVME SSDs can be used in [NVME] fan controller and [HD] fan controller does not accept them anymore.

Feel free to create a short feedback in discussion #55 on your compatibility experience.

Originally, this software was designed to work with Supermicro X10-X12/H10-H12 motherboards with IPMI functionality, where the BMC chip is ASPEED AST2400 or AST2500. Most motherboards in this set are compatible with the Supermicro IPMI raw commands used here and are supported out of the box.

Newer X13/H13 motherboards with AST2600 chips are also compatible with smfc (see more details in issue #33 about an X13SAE-F motherboard). The only difference is in the implementation of thresholds: the AST2600 chip implements only the Lower Critical threshold, so setting up thresholds is different.

Some motherboards require platform-specific IPMI raw commands for fan control. smfc implements a platform abstraction (since smfc v5.1.0) that handles these differences. The platform is auto-detected from BMC product name, or can be overridden with the platform_name= configuration parameter (see chapter 10.2). Currently supported platforms:

With this abstraction layer, new Supermicro motherboards can also be added to smfc with a good understanding of their IPMI raw commands and fan control logic.

For the newer X14/H14 motherboards, compatibility is still being investigated. There are some issues (#98) and discussions (#92, #106) about this to get better understanding.

Some X9 motherboards are supported (since smfc v5.2.0) via the genericx9 platform, provided they support the specific IPMI raw commands used for fan control. There is no auto-detection for X9 boards, so you must set platform_name=genericx9 in the configuration file.

The earlier X8 motherboards are NOT compatible with this software. They do not implement IPMI FULL mode, and they cannot control fan levels with IPMI raw commands.

On Supermicro X10-X11 motherboards IPMI uses six sensor thresholds to specify the safe and unsafe fan rotational speed intervals (these are RPM values rounded to the nearest hundreds, defined for each fan separately):

Lower Non-Recoverable  
Lower Critical  
Lower Non-Critical
Upper Non-Critical  
Upper Critical  
Upper Non-Recoverable

but newer Supermicro X13 motherboards (with AST2600 BMC chip) have only one sensor threshold:

Lower Critical  

Originally, this chapter was created for Supermicro X10-X11 motherboards, but can be easily adapted to X13 motherboards as well (see more details in #33).

Like many other utilities (created by NAS and home server community), smfc also uses IPMI FULL mode for fan control, where all fans in the zone:

1. initially configured to full speed (100%)
2. then their speed can be safely configured in [Lower Critical, Upper Critical] interval
3. if any fan speed oversteps either Lower Critical or Upper Critical threshold then IPMI will generate an assertion event and will set all fan speeds back to 100% in the zone

Please also consider the fact that fans are mechanical devices, their rotational speed is not stable (it could be fluctuating). To avoid IPMI's assertion mechanism described here please follow the next steps:

1. Per fan: check the minimum and maximum rotational speeds of your fan on its vendor website
2. Per fan: configure proper IPMI sensor thresholds adjusted to the fan speed interval
3. Per zone: define safe min_level/max_level values for smfc respecting the variance of all fans in the IPMI zone (it could take several iterations and adjustments)

Here is a real-life example for a Noctua NF-F12 PWM fan:

Upper Non-Recoverable = 1800 rpm
Upper Critical = 1700 rpm
Upper Non-Critical = 1600 rpm
Lower Non-Critical = 200 rpm
Lower Critical = 100 rpm
Lower Non-Recoverable = 0 rpm
Max RPM = 1500 rpm
Min RPM = 300 rpm
max_level = 100 (i.e. 1500 rpm)
min_level = 35 (i.e. 500 rpm)

Notes:

• Use the following ipmitool command to display the current IPMI sensor thresholds for fans:

  root@home:~# ipmitool sensor|grep FAN
  FAN1             | 500.000    | RPM        | ok    | 0.000     | 100.000   | 200.000   | 1600.000  | 1700.000  | 1800.000  
  FAN2             | 500.000    | RPM        | ok    | 0.000     | 100.000   | 200.000   | 1600.000  | 1700.000  | 1800.000  
  FAN3             | na         |            | na    | na        | na        | na        | na        | na        | na        
  FAN4             | 400.000    | RPM        | ok    | 0.000     | 100.000   | 200.000   | 1600.000  | 1700.000  | 1800.000  
  FANA             | 500.000    | RPM        | ok    | 0.000     | 100.000   | 200.000   | 1600.000  | 1700.000  | 1800.000  
  FANB             | 500.000    | RPM        | ok    | 0.000     | 100.000   | 200.000   | 1600.000  | 1700.000  | 1800.000  
  

• Use the following ipmitool command to list assertion events:

  root@home:~# ipmitool sel list
     1 | 10/19/2023 | 05:15:35 PM CEST | Fan #0x46 | Lower Critical going low  | Asserted
     2 | 10/19/2023 | 05:15:35 PM CEST | Fan #0x46 | Lower Non-recoverable going low  | Asserted
     3 | 10/19/2023 | 05:15:38 PM CEST | Fan #0x46 | Lower Non-recoverable going low  | Deasserted
     4 | 10/19/2023 | 05:15:38 PM CEST | Fan #0x46 | Lower Critical going low  | Deasserted
     5 | 10/19/2023 | 05:20:59 PM CEST | Fan #0x46 | Lower Critical going low  | Asserted
  

• Use the following ipmitool commands to specify all six sensor thresholds for FAN1:

  root@home:~# ipmitool sensor thresh FAN1 lower 0 100 200
  root@home:~# ipmitool sensor thresh FAN1 upper 1600 1700 1800
  

• You can also edit and run ipmi/set_ipmi_threshold.sh to configure all IPMI sensor thresholds

• If you install new BMC firmware on your Supermicro motherboard, you have to configure IPMI thresholds again

• If you do not see fans when executing ipmitool sensors, you may want to reset the BMC to factory default using the Web UI or using ipmitool mc reset cold

• Noctua specifies the variance of minimum and maximum fan rotational speeds (e.g. see the specification of Noctua NF-F12 PWM). For example:

  • Rotational speed (+/- 10%) 1500 RPM: 1350-1650 RPM interval
  • Min. rotational speed @ 20% PWM (+/-20%) 300 RPM: 240-360 RPM interval

  Please note that LNA/ULNA cables or Y-cables can modify the rotational speed calculations here and the required IPMI sensor thresholds too.

You can read more about:

• IPMI fan control: STH Forums and TrueNAS Forums
• Change IPMI sensors thresholds: TrueNAS Forums

If low noise and low heat generation are important attributes of your Linux box, then you may consider the following chapters.

Most modern CPUs have multiple energy-saving features. You can check your BIOS and enable them to minimize the heat generation.

Intel(R) CPUs:

• Intel(R) Speed Shift Technology
• Intel(R) SpeedStep
• C-states
• Boot performance mode

AMD(R) CPUs:

• PowerNow!
• Cool`n`quiet
• Turbo Core

With this setup the CPU will change its base frequency and power consumption dynamically based on the load.

In case of SATA hard disks, you may enable:

• advanced power management
• spin down timer

With the help of command hdparm you can enable advanced power management and specify a spin down timer (read more here):

hdparm -B 127 /dev/sda
hdparm -S 240 /dev/sda

In file /etc/hdparm.conf you can specify all parameters persistently:

quiet

/dev/sda {
    apm = 127
    spindown_time = 240
}
/dev/sdb {
    apm = 127
    spindown_time = 240
}
...

Important notes:

1. If you plan to spin down your hard disks or RAID array (i.e. put them to standby mode) you have to set up the configuration parameter [HD] polling= to at least twice as big as the spindown_time specified here.
2. In file /etc/hdparm.conf you must define HD names in /dev/disk/by-id/... form to avoid inconsistency.

One or more of the following Linux kernel modules need to be loaded for smfc:

• coretemp: temperature report for Intel(R) CPUs
• k10temp: temperature report for AMD(R) CPUs
• drivetemp: temperature report for SATA hard disks (available from kernel 5.6+ version)

Use /etc/modules file for persistent loading of these modules.

Notes:

• Reading drivetemp module is the fastest way to get the temperature of the hard disks, and it can read temperature of the SATA hard disks even if they are in standby mode.

For the installation and uninstallation, you need root privileges. There are several ways to install and uninstall smfc, this chapter will show them.

There is an installation script (bin/install.sh) which can install smfc in two different ways:

• remotely from the GitHub repository (no cloning required)
• locally from a git repository (GitHub repository needs to be cloned)

The installation script requires curl, pip, gzip, and mandb commands, and it has the following command line parameters:

user@host:~$ ./install.sh --help
usage: install.sh [-h|--help] [-k|--keep-config] [-l|--local] [-v|--verbose]
           -h, --help         help text
           -k, --keep-config  keep original configuration file
           -l, --local        installation from a local git repository
           -v, --verbose      verbose output

The default location of the installed files:

Notes for the script:

• Different Linux distros install python package to different folders (it was tested on Ubuntu 24.04, Debian 13, Suse Leap 15, Proxmox 9, and Arch Linux)
• Installation process will stop for any error
• The default installation method is the remote installation
• Using --local parameter will do installation locally, from the current folder (i.e. GitHub repository needs to be cloned)
• The default action is the following:
  • the existing previous configuration file will be renamed
  • a new configuration file will be installed
  • hd_names= configuration parameter will be pre-filled with the list of existing hard disks for user's convenience (please check/edit this parameter!)
• Using --keep-config parameter, the original configuration file will be preserved
• Using --verbose parameter, the phases of the installation will be displayed

For remote installation the script can be executed (as root user) this way:

curl --silent https://raw.githubusercontent.com/petersulyok/smfc/refs/heads/main/bin/install.sh|bash /dev/stdin --verbose

or if you want to preserve your existing configuration file:

curl --silent https://raw.githubusercontent.com/petersulyok/smfc/refs/heads/main/bin/install.sh|bash /dev/stdin --verbose --keep-config

For local installation follow these steps (here uv command is also required):

git clone https://github.com/petersulyok/smfc.git
cd smfc
uv sync
uv build
./bin/install.sh --local --verbose --keep-config

Here we clone the GitHub repository and make the smfc package locally before installation.

On the other hand, there is an uninstallation script (bin/uninstall.sh) which can uninstall smfc. This script has the following command line parameters:

user@host:~$ ./uninstall.sh --help
usage: uninstall.sh [-h|--help] [-k|--keep-config] [-v|--verbose]
           -h, --help         help text
           -k, --keep-config  keep original configuration file
           -v, --verbose      verbose output

This script can be executed locally and remotely as described for installation above. Here is an example for remote execution:

curl --silent https://raw.githubusercontent.com/petersulyok/smfc/refs/heads/main/bin/uninstall.sh|bash /dev/stdin --verbose

The script will remove the installed smfc files and the python package.

Pre-built DEB and RPM packages are available as build artifacts from the GitHub packages workflow on each release. You can also build them locally from the repository.

For Debian/Ubuntu systems:

sudo dpkg -i smfc_*.deb

For Fedora/RHEL systems:

sudo dnf install smfc-*.rpm

The packages install the same files as the manual installation (service unit, configuration, man page, and sample configs). Configuration files under /etc/ are preserved on upgrade. See PACKAGES.md for build instructions and compatible distributions.

smfc is also available as a docker image, see more details in Docker.md. In this case, your job is only to provide your configuration file on the host computer, smfc will be executed automatically when the container is starting.

After successful installation, create/edit your new configuration file. If you just upgraded to a new smfc version, you can preserve the existing one.

You have to think over and answer the following questions:

1. What are the most important heat sources in your machine? Typically, these could be CPU(s), hard disks, or GPUs.
2. Which fan controller would you like to use and configure in smfc?
3. What is the expected temperature interval (minimum/maximum C degree) for the selected temperature source(s)? Use some test tools to measure it (e.g. s-tui, fio, iozone) if you don't have their track records.
4. Which IPMI zone(s) will be connected to these fan controllers/temperature sources)? Check how many IPMI zones you have, how the fans are connected on your motherboard, and how they are cooling the selected temperature source(s). Multiple controllers can share the same zone -- the highest requested level will be applied automatically.
5. What is the stable level interval for fans in the selected IPMI zone(s)? Probably this part requires the most patience! You have assumptions here that need to be verified. If you experience IPMI assertions and your fans are spinning up then you have to refine the level interval or threshold configuration and try again. You will have several cycles here, this is normal.

The configuration file contains sections. The first one for IPMI configuration, the rest for fan controllers. Edit /etc/smfc/smfc.conf and specify your configuration parameters here:

#
#   smfc.conf (C) 2020-2026, Peter Sulyok
#   smfc 5.x service configuration parameters
#
#   Please read the documentation here: https://github.com/petersulyok/smfc
#

# Ipmi specific parameters.
[Ipmi]
# Path for ipmitool (str, default=/usr/bin/ipmitool)
command=/usr/bin/ipmitool 
# Delay time after changing IPMI fan mode (int, seconds, default=10)
fan_mode_delay=10
# Delay time after changing IPMI fan level (int, seconds, default=2)
fan_level_delay=2
# IPMI parameters for remote access (string, default='')
#remote_parameters=-U USERNAME -P PASSWORD -H HOST
# Supermicro platform (string, default='auto')
# Potential values:
#  auto       - automatic discovery based on BMC information
#  generic    - Generic Supermicro X10-X13/H10-H13 platform
#  genericx9  - Generic Supermicro X9 platform
#  X10QBi     - Supermicro X10QBi platform
platform_name=auto


# CPU fan controller: works based on CPU(s) temperature.
[CPU]
# Fan controller enabled (bool, default=0)
enabled=1
# IPMI zone(s) (comma- or space-separated list of int, default=0))
ipmi_zone=0
# Calculation method for CPU temperatures (int, [0-minimum, 1-average, 2-maximum], default=1)
temp_calc=1
# Discrete steps in mapping of temperatures to fan level (int, default=6)
steps=6
# Threshold in temperature change before the fan controller reacts (float, C, default=3.0)
sensitivity=3.0
# Polling time interval for reading temperature (int, sec, default=2)
polling=2
# Minimum CPU temperature (float, C, default=30.0)
min_temp=30.0
# Maximum CPU temperature (float, C, default=60.0)
max_temp=60.0
# Minimum CPU fan level (int, %, default=35)
min_level=35
# Maximum CPU fan level (int, %, default=100)
max_level=100
# Moving average window size for temperature smoothing (int, default=1, 1=disabled)
smoothing=1


# HD fan controller: works based on SATA or SAS HDDs/SSDs temperature.
[HD]
# Fan controller enabled (bool, default=0)
enabled=1
# IPMI zone(s) (comma- or space-separated list of int, default=1))
ipmi_zone=1
# Calculation of HD temperatures (int, [0-minimum, 1-average, 2-maximum], default=1)
temp_calc=1
# Discrete steps in mapping of temperatures to fan level (int, default=4)
steps=4
# Threshold in temperature change before the fan controller reacts (float, C, default=2.0)
sensitivity=2.0
# Polling interval for reading temperature (int, sec, default=10)
polling=10
# Minimum HD temperature (float, C, default=32.0)
min_temp=32.0
# Maximum HD temperature (float, C, default=46.0)
max_temp=46.0
# Minimum HD fan level (int, %, default=35)
min_level=35
# Maximum HD fan level (int, %, default=100)
max_level=100
# Moving average window size for temperature smoothing (int, default=1, 1=disabled)
smoothing=1
# Names of the HDs (str multi-line list, default=)
# MUST BE specified in '/dev/disk/by-id/...' form, for example:
# hd_names=/dev/disk/by-id/ata-WDC_WD100EFAX-68LHPN0_8CH7T91E
#	/dev/disk/by-id/ata-WDC_WD100EFAX-68LHPN0_8CH7T813
#	/dev/disk/by-id/ata-WDC_WD100EFAX-68LHPN0_8CHUFFRE
hd_names=
# Path for 'smartctl' command (str, default=/usr/sbin/smartctl).
smartctl_path=/usr/sbin/smartctl
# Standby guard feature for RAID arrays (bool, default=0)
standby_guard_enabled=0
# Number of HDs already in STANDBY state before the full RAID array will be forced to it (int, default=1)
standby_hd_limit=1


# NVME fan controller: works based on NVMe SSD(s) temperature.
[NVME]
# Fan controller enabled (bool, default=0)
enabled=0
# IPMI zone(s) (comma- or space-separated list of int, default=1))
ipmi_zone=1
# Calculation of NVMe temperatures (int, [0-minimum, 1-average, 2-maximum], default=1)
temp_calc=1
# Discrete steps in mapping of temperatures to fan level (int, default=4)
steps=4
# Threshold in temperature change before the fan controller reacts (float, C, default=2.0)
sensitivity=2.0
# Polling interval for reading temperature (int, sec, default=10)
polling=10
# Minimum NVMe temperature (float, C, default=35.0)
min_temp=35.0
# Maximum NVMe temperature (float, C, default=70.0)
max_temp=70.0
# Minimum NVMe fan level (int, %, default=35)
min_level=35
# Maximum NVMe fan level (int, %, default=100)
max_level=100
# Moving average window size for temperature smoothing (int, default=1, 1=disabled)
smoothing=1
# Names of the NVMe devices (str multi-line list, default=)
# MUST BE specified in '/dev/disk/by-id/...' form, for example:
# nvme_names=/dev/disk/by-id/nvme-ADATA_LEGEND_650_2OFF29AO8DKR
#	/dev/disk/by-id/nvme-CT4000P3PSSD7_2446E89408FA
nvme_names=


# GPU fan controller: works based on Nvidia GPU(s) temperature.
[GPU]
# Fan controller enabled (bool, default=0)
enabled=0
# IPMI zone(s) (comma- or space-separated list of int, default=1))
ipmi_zone=1
# Calculation of GPU temperatures (int, [0-minimum, 1-average, 2-maximum], default=1)
temp_calc=1
# Discrete steps in mapping of temperatures to fan level (int, default=5)
steps=5
# Threshold in temperature change before the fan controller reacts (float, C, default=2.0)
sensitivity=2.0
# Polling interval for reading temperature (int, sec, default=10)
polling=2
# Minimum GPU temperature (float, C, default=40.0)
min_temp=40.0
# Maximum GPU temperature (float, C, default=70.0)
max_temp=70.0
# Minimum GPU fan level (int, %, default=35)
min_level=35
# Maximum GPU fan level (int, %, default=100)
max_level=100
# Moving average window size for temperature smoothing (int, default=1, 1=disabled)
smoothing=1
# GPU device IDs (comma- or space-separated list of int, default=0)
# These are indices in nvidia-smi temperature report.
gpu_device_ids=0
# Path for 'nvidia-smi' command (str, default=/usr/bin/nvidia-smi).
nvidia_smi_path=/usr/bin/nvidia-smi


# CONST fan controller: sets constant fan level (without any heat source) for IPMI zones(s).
[CONST]
# Fan controller enabled (bool, default=0)
enabled=0
# IPMI zone(s) (comma- or space-separated list of int, default=1))
ipmi_zone=1
# Polling interval for checking/resetting level if needed (int, sec, default=30)
polling=30
# Constant fan level (int, %, default=50)
level=50

Important notes:

1. [Ipmi] remote_parameters=-U USERNAME -P PASSWORD -H HOST parameter can be used for remote access for the IPMI interface. It could be useful for a VM setup where the hard disks are configured with PCI passthrough (e.g. a TrueNAS running in a VM on Proxmox), but IPMI needs to be accessed "remotely". Please note that the HOST is the BMC network address (not the VM host address).
2. [HD] hd_names= is a compulsory parameter for HD fan controller, and it must be specified in /dev/disk/by-id/... form. Please note that the /dev/sda form is not persistent and could change after a reboot!
3. [NVME] nvme_names= is a compulsory parameter for NVME fan controller, and it must be specified in /dev/disk/by-id/... form. Please note that the /dev/nvme0n1 form is not persistent and could change after a reboot!
4. [CPU] / [HD] / [NVME] min_level= / max_level= should be configured in alignment with threshold configuration (see more details in this chapter). Be patient, several refinement cycles could happen.
5. Several sample configuration files are provided in ./config/samples folder.
6. Save/backup your configuration file when you've got the final version. Avoid overwriting if you upgrade to a new version of smfc.

After manual installation, smfc can be started and stopped as a standard systemd service. Remember to reload systemd configuration after a new installation or if you changed the service definition file:

systemctl daemon-reload
systemctl start smfc.service
systemctl stop smfc.service
systemctl restart smfc.service
systemctl status smfc.service
● smfc.service - Supermicro Fan Control
     Loaded: loaded (/etc/systemd/system/smfc.service; enabled; preset: enabled)
     Active: active (running) since Tue 2026-03-03 21:32:22 CET; 7min ago
 Invocation: b613d841e1cf43f4ace80d472623ed4c
   Main PID: 82521 (smfc)
      Tasks: 1 (limit: 154231)
     Memory: 12M (peak: 19.5M)
        CPU: 1.271s
     CGroup: /system.slice/smfc.service
             └─82521 /usr/bin/python3 /usr/local/bin/smfc -c /etc/smfc/smfc.conf -l 3

Mar 03 21:37:27 nas smfc.service[82521]: Shared IPMI zone [0]: new level = 67% (winner: CPU=67%/51.0C, losers: NVME=35%/37.9C)
Mar 03 21:37:30 nas smfc.service[82521]: Shared IPMI zone [0]: new level = 35% (winner: CPU=35%/31.0C, losers: NVME=35%/37.9C)
Mar 03 21:38:46 nas smfc.service[82521]: Shared IPMI zone [0]: new level = 67% (winner: CPU=67%/49.0C, losers: NVME=35%/37.9C)
Mar 03 21:38:49 nas smfc.service[82521]: Shared IPMI zone [0]: new level = 35% (winner: CPU=35%/31.0C, losers: NVME=35%/37.9C)

The smfc program has the following parameters:

root@nas$ smfc --help
usage: smfc [-h] [-c CONFIG_FILE] [-v] [-l {0,1,2,3,4}] [-o {0,1,2}] [-nd] [-s] [-ne]

options:
  -h, --help      show this help message and exit
  -c CONFIG_FILE  configuration file (default is /etc/smfc/smfc.conf)
  -v, --version   show program's version number and exit
  -l {0,1,2,3,4}  set log level: 0-NONE, 1-ERROR(default), 2-CONFIG, 3-INFO, 4-DEBUG
  -o {0,1,2}      set log output: 0-stdout, 1-stderr, 2-syslog(default)
  -nd             no dependency checking at start
  -s              use sudo command
  -ne             no fan speed recovery at exit

smfc command-line options can be specified in /etc/default/smfc file if you run smfc as a systemd service.

If you are testing your configuration, you can start smfc directly in a terminal (logging to the standard output on debug log level):

smfc -o 0 -l 3

In case of Docker installation, smfc will be executed automatically when the container is started. Its command-line parameters can be specified in the docker-compose file.

All messages will be logged to the specific output and the specific level. With the help of command journalctl you can check logs easily. For example:

1. listing service logs of the last two hours:

    journalctl -u smfc --since "2 hours ago"
   

2. listing service logs from the last boot:

    journalctl -b -u smfc
   

Most probably, there was an assertion (i.e., the rotational speed of a fan went above or below an IPMI threshold) and IPMI switched back that zone to full rotational speed. You can check the current fan rotational speeds:

ipmitool sdr

and you can also check IPMI event log and list assertion events:

root@home:~# ipmitool sel list
   1 | 10/19/2023 | 05:15:35 PM CEST | Fan #0x46 | Lower Critical going low  | Asserted
   2 | 10/19/2023 | 05:15:35 PM CEST | Fan #0x46 | Lower Non-recoverable going low  | Asserted
   3 | 10/19/2023 | 05:15:38 PM CEST | Fan #0x46 | Lower Non-recoverable going low  | Deasserted
   4 | 10/19/2023 | 05:15:38 PM CEST | Fan #0x46 | Lower Critical going low  | Deasserted
   5 | 10/19/2023 | 05:20:59 PM CEST | Fan #0x46 | Lower Critical going low  | Asserted

If the problematic fan (causing the alert) is identified, then you must adjust its threshold. This process could take several adjustment cycles. Be patient :) You may read this chapter for more details.

The configuration is the following:

• Supermicro X11SCH-F motherboard

• Intel (R) Xeon (R) E-2276G processor

• 128 GB ECC DDR4-2666MHz RAM

• Fractal Design Node 804 case, with separate chambers for the motherboard and the hard disks:

  

• Proxmox 9

• 8 x WD Red 12TB (WD120EFAX) hard disks in ZFS RAID

• 4 x Noctua NF-F12 PWM fans (FAN1, FAN2, FAN3, FAN4) in IPMI CPU zone

• 2 x Noctua NF-F12 PWM on an Y-adapter + Noctua NF-A14 PWM fans (FANA, FANB) in IPMI HD zone

Further readings:

• BMC IPMI User's Guide 1.1b (X10/X11/H11)
• X9 SMM IPMI User's Guide
• Supermicro Management Software download
• IPMI utilities overiew
• IPMICFG User's Guide (1.18)

• [STH forums] Reference Material: Supermicro X9/X10/X11 Fan Speed Control
  • [STH forums] Addition to X9 motherboards
• [TrueNAS forums] How To: Change IPMI Sensor Thresholds using ipmitool
• [TrueNAS forums] Script to control fan speed in response to hard drive temperatures
• [Pcfe's blog] Set fan thresholds on my Supermicro H11DSi-NT

• IPMI v2.0 specification by Intel
• ipmitool — CLI for managing IPMI-enabled devices
• smartmontools — S.M.A.R.T. monitoring tools for hard disks (smartctl)
• nvidia-smi — NVIDIA System Management Interface for GPU monitoring
• hwmon subsystem — hardware monitoring framework used for temperature readings
• coretemp — Intel CPU temperature monitoring
• k10temp — AMD CPU temperature monitoring
• drivetemp — SATA disk temperature monitoring, and its GitHub repository

• [GitHub] Kevin Horton's nas_fan_control
• [GitHub] sretalla's fork nas_fan control
• [GitHub] supermicro-fancontrol by jvdillion
• [GitHub] PureCypher's supermicro-fans
• [GitHub] luckylinux's supermicro-fan-control
• [GitHub] mrstux's hybrid_fan_control

Written with StackEdit.