Supermicro X9SCM-F PWM fan_control

[twolf]

Does anybody know if it is possible to control the PWM fans on the X9SCM when using ESXi and running unRAID in a VM?

 

[StevenD]

Does anybody know if it is possible to control the PWM fans on the X9SCM when using ESXi and running unRAID in a VM?

 

 

I dont believe so because you cant pass then through to the VM.  Im not really sure how that would even work.  ESXi has its own fan controls, but its really only going to be able to read case and proc temps.

[PeterB]

The original png icon file seems to have disappeared from the url given in the .plg file.

 

It's obviously easy to create another, but it would be good to have the original - can anyone post it somewhere?

[twolf]

I dont believe so because you cant pass then through to the VM.  Im not really sure how that would even work.  ESXi has its own fan controls, but its really only going to be able to read case and proc temps.

 

Yes, that's the dilemma I've found myself in as well. Is there even a way to control the fan speeds on the ESXi side, or are they strictly display-only? If control is available, it might be possible to script a solution somehow (unRAID reading drive temps, and passing reactions over to ESXi).

 

[StevenD]

I dont believe so because you cant pass then through to the VM.  Im not really sure how that would even work.  ESXi has its own fan controls, but its really only going to be able to read case and proc temps.

 

Yes, that's the dilemma I've found myself in as well. Is there even a way to control the fan speeds on the ESXi side, or are they strictly display-only? If control is available, it might be possible to script a solution somehow (unRAID reading drive temps, and passing reactions over to ESXi).

 

 

At work, we load HP agents into ESXi to be able to control the fans and other hardware. i dont believe those exist for most of our unRAID hardware.

[PeterB]

A word of advice for anyone using the fan_speed plugin:  If you have a disk controller card using the mpt2sas driver (and maybe some others?), you must remove the '-d ata' from the smartctl command line, else the drives on the controller card will not respond.  The controller is for SAS and SATA drives - it does not like to be told to behave as an ATA interface!

 

I understand that when Joe first worked on unMENU, a disk controller was found which wouldn't work with the default 'auto' setting in smartctl, hence the switch to force ATA behaviour.

 

I wonder how many people are still using this ATA controller?  I'm sure that there are now many more unRAID installations which bork at this.  Unfortunately, this form of the smartctl command has been copied by others, so the '-d ata' switch gets propogated through any addon which calls smartctl.

[Auggie]

I'm trying out this fan_speed.sh script by aiden located at http://lime-technology.com/forum/index.php?topic=11310.msg108144#msg108144 and it seems to work fine over the past several days, but today I noticed my temps getting into the 50 degree Celcius range, which the script correctly identified and attempted to adjust fan speed, however, the fan speeds are only lounging at 750-ish RPMs and not the normal 1,400-ish RPM's at 100%:

 

Sep 22 10:08:03 UnRAID logger: fan_speed: Highest disk drive temp is: 50C
Sep 22 10:08:03 UnRAID logger: fan_speed: Changing disk drive fan speed from: 116 (45% @ 715 rpm) to: FULL (100% @ 712 rpm)
Sep 22 10:10:14 UnRAID logger: fan_speed: Highest disk drive temp is: 51C
Sep 22 10:10:14 UnRAID logger: fan_speed: Changing disk drive fan speed from: 121 (47% @ 744 rpm) to: FULL (100% @ 743 rpm)
Sep 22 10:12:24 UnRAID logger: fan_speed: Highest disk drive temp is: 51C
Sep 22 10:12:24 UnRAID logger: fan_speed: Changing disk drive fan speed from: 121 (47% @ 738 rpm) to: FULL (100% @ 743 rpm)
Sep 22 10:14:35 UnRAID logger: fan_speed: Highest disk drive temp is: 51C
Sep 22 10:14:35 UnRAID logger: fan_speed: Changing disk drive fan speed from: 121 (47% @ 739 rpm) to: FULL (100% @ 740 rpm)
Sep 22 10:16:46 UnRAID logger: fan_speed: Highest disk drive temp is: 51C
Sep 22 10:16:46 UnRAID logger: fan_speed: Changing disk drive fan speed from: 126 (49% @ 777 rpm) to: FULL (100% @ 774 rpm)
Sep 22 10:18:57 UnRAID logger: fan_speed: Highest disk drive temp is: 51C
Sep 22 10:18:57 UnRAID logger: fan_speed: Changing disk drive fan speed from: 126 (49% @ 770 rpm) to: FULL (100% @ 770 rpm)
Sep 22 10:21:07 UnRAID logger: fan_speed: Highest disk drive temp is: 52C
Sep 22 10:21:07 UnRAID logger: fan_speed: Changing disk drive fan speed from: 126 (49% @ 775 rpm) to: FULL (100% @ 772 rpm)
Sep 22 10:23:18 UnRAID logger: fan_speed: Highest disk drive temp is: 52C
Sep 22 10:23:18 UnRAID logger: fan_speed: Changing disk drive fan speed from: 126 (49% @ 778 rpm) to: FULL (100% @ 776 rpm)

 

I verified the fan RPM's were only in the 750-ish range:

 

root@UnRAID:~# cat /sys/class/hwmon/hwmon1/device/fan4_input
771
root@UnRAID:~# cat /sys/class/hwmon/hwmon1/device/fan5_input
761
root@UnRAID:~# cat /sys/class/hwmon/hwmon1/device/fan6_input
cat: /sys/class/hwmon/hwmon1/device/fan6_input: No such file or directory
root@UnRAID:~# cat /sys/class/hwmon/hwmon1/device/fan3_input
777
root@UnRAID:~#

 

Yesterday, the script was managing the fan speeds just fine:

 

Sep 21 19:11:49 UnRAID logger: fan_speed: Highest disk drive temp is: 38C
Sep 21 19:11:49 UnRAID logger: fan_speed: Changing disk drive fan speed from: FULL (100% @ 1464 rpm) to: 228 (89% @ 1326 rpm)
Sep 21 19:15:58 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 19:15:58 UnRAID logger: fan_speed: Changing disk drive fan speed from: 228 (89% @ 1324 rpm) to: 205 (80% @ 1206 rpm)
Sep 21 19:32:16 UnRAID logger: fan_speed: Highest disk drive temp is: 36C
Sep 21 19:32:16 UnRAID logger: fan_speed: Changing disk drive fan speed from: 205 (80% @ 1204 rpm) to: 182 (71% @ 1076 rpm)
Sep 21 19:34:23 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 19:34:23 UnRAID logger: fan_speed: Changing disk drive fan speed from: 182 (71% @ 1076 rpm) to: 205 (80% @ 1207 rpm)
Sep 21 19:36:29 UnRAID logger: fan_speed: Highest disk drive temp is: 36C
Sep 21 19:36:29 UnRAID logger: fan_speed: Changing disk drive fan speed from: 205 (80% @ 1198 rpm) to: 182 (71% @ 1075 rpm)
Sep 21 19:40:38 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 19:40:38 UnRAID logger: fan_speed: Changing disk drive fan speed from: 182 (71% @ 1084 rpm) to: 205 (80% @ 1209 rpm)
Sep 21 19:42:45 UnRAID logger: fan_speed: Highest disk drive temp is: 36C
Sep 21 19:42:45 UnRAID logger: fan_speed: Changing disk drive fan speed from: 205 (80% @ 1200 rpm) to: 182 (71% @ 1074 rpm)
Sep 21 19:44:52 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 19:44:52 UnRAID logger: fan_speed: Changing disk drive fan speed from: 182 (71% @ 1082 rpm) to: 205 (80% @ 1207 rpm)
Sep 21 19:47:00 UnRAID logger: fan_speed: Highest disk drive temp is: 36C
Sep 21 19:47:00 UnRAID logger: fan_speed: Changing disk drive fan speed from: 205 (80% @ 1201 rpm) to: 182 (71% @ 1075 rpm)
Sep 21 19:57:21 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 19:57:21 UnRAID logger: fan_speed: Changing disk drive fan speed from: 182 (71% @ 1075 rpm) to: 205 (80% @ 1198 rpm)
Sep 21 19:59:28 UnRAID logger: fan_speed: Highest disk drive temp is: 36C
Sep 21 19:59:28 UnRAID logger: fan_speed: Changing disk drive fan speed from: 205 (80% @ 1198 rpm) to: 182 (71% @ 1080 rpm)
Sep 21 20:03:39 UnRAID logger: fan_speed: Highest disk drive temp is: 37C
Sep 21 20:03:39 UnRAID logger: fan_speed: Changing disk drive fan speed from: 182 (71% @ 1082 rpm) to: 205 (80% @ 1201 rpm)

 

I ran a different fan script, unraid-fan-speed.sh by Pauven, http://lime-technology.com/forum/index.php?topic=5548.105, and it correctly set the fan speeds to their normal RPMs at 100%:

 

root@UnRAID:/boot/scripts# unraid-fan-speed.sh
Linear PWM Range is 80 to 255 in 6 increments of 29
Highest temp is: 52
Setting pwm to: 255
root@UnRAID:/boot/scripts# cat /sys/class/hwmon/hwmon1/device/fan4_input
1473
root@UnRAID:/boot/scripts# 

 

This suggests a script issue and not an OS issue, so for the time being, I'm switching back to Pauven's script and cron.

[Auggie]

Wow, happened again; the script suddenly was only setting the RPM's to roughly half the maximum speeds of the capabilities of the fans.

 

I forgot to change the go script so after a reboot this fan_speed.sh script was started up and I decided to let it run.  Then a couple days later when I checked in on the server almost all the drive temps were in the 50 degree Celsius range!  (I got to fix my drive temp warning e-mails).

 

Absolutely unacceptable as this could have resulted in permanent hardware failures.

 

Maybe if I had the time I would scrutinize the script line-by-line, but something is not working with my X9SCM-F board.

 

Attached is the script in case someone wants to gander at what could be causing the RPM issues...

fan_speed.sh.zip

[StevenD]

Wow, happened again; the script suddenly was only setting the RPM's to roughly half the maximum speeds of the capabilities of the fans.

 

I forgot to change the go script so after a reboot this fan_speed.sh script was started up and I decided to let it run.  Then a couple days later when I checked in on the server almost all the drive temps were in the 50 degree Celsius range!  (I got to fix my drive temp warning e-mails).

 

Absolutely unacceptable as this could have resulted in permanent hardware failures.

 

Maybe if I had the time I would scrutinize the script line-by-line, but something is not working with my X9SCM-F board.

 

Attached is the script in case someone wants to gander at what could be causing the RPM issues...

 

 

Don't blame the script for potential hardware failure. You alone are esponsible for monitoring your temps. If you are that concerned, set the fans to high in BIOS, then never worry again.

[Auggie]

Don't blame the script for potential hardware failure. You alone are esponsible for monitoring your temps. If you are that concerned, set the fans to high in BIOS, then never worry again.

 

Sorry, I blame the script.  The other script works just fine, thank you very much.  An answer like yours is surely flame bait as we ALL rely on some technology for monitoring and reporting; including unRAID and SMART and all the other doodads.  Unless you are a hermit who stands by your terminal all day watching and listening and touching your hardware 24/7.

[starcat]

Any update on the fan-speed control script? Anyone tested it with the newer X9SCM-iiF version of the board (actual change is only same network chip for both GigE interfaces, so I suppose it would behave the same). Thanks!

[PeterB]

I use the fan_speed plugin with an X9SCM-iiF, and it works just fine.

[starcat]

Thanks Peter.

Is the X9SCM-iiF (or X9SCM-F) also able to be put into S3 sleep mode?

[Russ Uno]

Not an expert here but Fan "A" on the X9SCM-F is for I/O/Add-on Cards and Fans 1-4 are for CPU/Memory Zone. Don't know if that makes a difference or not. I just installed that board yesterday so I was reading the manual.

 

[PeterB]

I thought that FAN A was for the CPU cooler?

 

I'm not sure about sleep/suspend - I leave my server running 24/7.  Devices in my house, which are always on, require access to the server.  Also, I have a concern about what happens when the server is in suspend mode when a power cut occurs.  Would the server be woken up by the UPS so that it can perform a controlled shutdown?

[starcat]

I never ever had a power cut, so stable is power over here. I even removed all UPS'es around. Just no need for them.

But a good point though!

 

Edit:

The sleep script always sync's memory to the disks prior going into sleep. Theoretically power cut in sleep should not be a problem.

However the most important would be control of fan-speed, the S3 sleep is just a nice to have feature.

[Russ Uno]

I thought that FAN A was for the CPU cooler?

 

That fan info was copied and pasted right from the PDF instructions.

[PeterB]

Edit:

The sleep script always sync's memory to the disks prior going into sleep. Theoretically power cut in sleep should not be a problem.

 

Indeed.  But does unRAID know this?  My expectation is that if the power goes off while the unRAID server is in suspend mode, it will reboot when power returns and start an automatic parity check?

 

Lucky you with the stability of power supply - we are having an average of two outages a day, recently.

 

Until last weekend, my record count of outages in a 24 hour period was 7.  Last weekend, I lost count, but it was at least 10!  One thing is sure - the performance of apcupsd and UPS gets well tested here!

[PeterB]

I thought that FAN A was for the CPU cooler?

 

That fan info was copied and pasted right from the PDF instructions.

 

Ah, okay.  That info appears in the X9SCM manual, but is omitted from the (later) X9SCM-iiF manual.  The iiF manual simply states:

FAN1-4, FANA   - System/CPU Fan Headers

 

Use of FAN A (Fan 1 in software terms) for CPU makes a lot of sense.  The speed of Fans 1-4 cannot be controlled individually - they all respond to the same speed setting.  Fan A is separately controlled and can be left on BIOS control while Fans 1-4 are software controlled.

 

However, looking at the Super Micro FAQs, it seems that, with BIOS control, FAN1-4 are controlled according to CPU temperature and FAN A according to System temperature.  Now, I don't know where 'System' temperature is measured, but it must be somewhere on the mobo.

 

For most of us, we want to control multiple chassis fans, using the plugin, but leave the (one) CPU fan BIOS controlled.  So, it is logical to use FAN 1-4 for chassis, and FanA for CPU.  The only problem with this is that there is a potential discrepancy between CPU and System tempearures.  I guess that it would be possible to control CPU fan, according to CPU Temp, via plugin, but I think that it it probably safer/more reliable to accept that CPU fan is BIOS-controlled according to System temperature.