How To: Change IPMI Sensor Thresholds using ipmitool

[Ericloewe]

Yes, I can access it through the web interface but there I only have four settings; Standard, Full IO, Full, Optimal. I'm not familiar with any other IPMI utilities. What would you suggest?

See if Supermicro's tools can talk to it. It might be in an inconsistent state

 

[RoboKaren]

Thanks for the quick reply. When I try the next command, ipmitool sensor list all

I get back: Could not open device at /dev/ipmi0 or /dev/ipmi/0 or /dev/ipmidev/0: No such file or directory

You need run this as superuser:

Code:

[moi@mymachine] ~% ipmitool sensor
Could not open device at /dev/ipmi0 or /dev/ipmi/0 or /dev/ipmidev/0: No such file or directory
[moi@mymachine] ~% sudo ipmitool sensor
Password:
System Temp      | 23.000     | degrees C  | ok    | -9.000    | -7.000    | -5.000    | 80.000    | 85.000    | 90.000   
Peripheral Temp  | 27.000     | degrees C  | ok    | -9.000    | -7.000    | -5.000    | 80.000    | 85.000    | 90.000   
CPU Temp         | 0x0        | discrete   | 0x0000| na        | na        | na        | na        | na        | na       
FAN 1            | 600.000    | RPM        | nc    | 300.000   | 450.000   | 600.000   | 18975.000 | 19050.000 | 19125.000
FAN 2            | 1275.000   | RPM        | ok    | 225.000   | 300.000   | 375.000   | 18975.000 | 19050.000 | 19125.000

 

[RemyZ]

Running into the same issue where fans continuously cycling up and down. I have not installed Freenas yet on my build. I am running IPMIUtil on a remote machine running Win7. I changed the lower and upper threshold values to 100/3000 for all 3 case fans only. I left the CPU fan alone. The specs for the fans used in my case are listed here https://www.sanyodenki.com/archive/document/corporatedata/technicalreport/2004/TR17e_c_en.pdf
They make no mention of minimum ratings so I can base my thresholds. After I made changes to the thresholds I made sure to reboot and still no changes


C:\.........\IPMI>ipmiutil sensor -N 192.168.0.180 -U "Username"-P "Password"
-c
ipmiutil ver 2.98
isensor: version 2.98
Connecting to node 192.168.0.180
-- BMC version 1.11, IPMI version 2.0
ID | SDRType | Type |SNum| Name |Status| Reading
0004 | Full | Temperature | 01 | CPU Temp | OK | 39.00 C
0047 | Full | Temperature | 0a | PCH Temp | OK | 27.00 C
008a | Full | Temperature | 0b | System Temp | OK | 24.00 C
00cd | Full | Temperature | 0c | Peripheral Temp | OK | 31.00 C
0110 | Full | Temperature | 10 | VcpuVRM Temp | OK | 35.00 C
0153 | Full | Temperature | b0 | DIMMA1 Temp | Absent | 0.00 na
0196 | Full | Temperature | b1 | DIMMA2 Temp | Absent | 0.00 na
01d9 | Full | Temperature | b4 | DIMMB1 Temp | Absent | 0.00 na
021c | Full | Temperature | b5 | DIMMB2 Temp | Absent | 0.00 na
025f | Full | Fan | 41 | FAN1 | OK | 2500.00 RPM<-------------CPU FAN
02a2 | Full | Fan | 42 | FAN2 | AboveCrit | 6000.00 R<-------------CaseFAN
PM
02e5 | Full | Fan | 43 | FAN3 | AboveCrit | 6100.00 R<-------------CaseFAN
PM
0328 | Full | Fan | 44 | FAN4 | Absent | 0.00 na
036b | Full | Fan | 45 | FANA | AboveCrit | 6200.00 R<-------------CaseFAN
PM
03ae | Full | Voltage | 30 | 12V | OK | 12.13 V
03f1 | Full | Voltage | 31 | 5VCC | OK | 4.97 V
0434 | Full | Voltage | 32 | 3.3VCC | OK | 3.37 V
0477 | Full | Voltage | 33 | VBAT | OK | 2.97 V
04ba | Full | Voltage | 34 | VCPU | OK | 1.15 V
04fd | Full | Voltage | 35 | VDIMMAB | OK | 1.21 V
0540 | Full | Voltage | 38 | 5VSB | OK | 5.05 V
0583 | Full | Voltage | 39 | 3.3VSB | OK | 3.33 V
05c6 | Full | Voltage | 3b | VBMC 1.2V | OK | 1.20 V
0609 | Full | Voltage | 3c | VPCH 1.0V | OK | 0.99 V
064c | Full | Platform Securi | aa | Chassis Intru | OK | 0.00 na
ipmiutil sensor, completed successfully

 

[Ericloewe]

02a2 | Full | Fan | 42 | FAN2 | AboveCrit | 6000.00 R<-------------CaseFAN
PM
02e5 | Full | Fan | 43 | FAN3 | AboveCrit | 6100.00 R<-------------CaseFAN
PM
0328 | Full | Fan | 44 | FAN4 | Absent | 0.00 na
036b | Full | Fan | 45 | FANA | AboveCrit | 6200.00 R<-------------CaseFAN

The upper thresholds seem to be too low (and I hope this is at 100% duty cycle).

They make no mention of minimum ratings so I can base my thresholds. After I made changes to the thresholds I made sure to reboot and still no changes

Sanyo Denki provides excellent datasheets.

That link you provided is not the datasheet (though it does contain much of what is in the datasheet) and it doesn't apply to your case, since it's for a non-PWM fan model.

 

[RemyZ]

I have tried various config changes but it doesn't seem to drop for any reason. The commands I am using are listed below. I tried unplugging and replugging the IPMI ethernet cable and power cycling but still nothing. The only change is instead of cycling it is now at full blast and no commands I input will stop this

C:\.........ipmiutil> ipmiutil sensor -N 192.168.0.180 -U *Username* -P *Password* -n 42 -l 400 -h 3000

Is this the correct syntax ?

 

[Ericloewe]

I have tried different various but it doesn't seem to drop for any reason. The commands I am using are listed below. I tried unplugging and replugging the IPMI ethernet cable and power cycling but still nothing. The only change is instead of cycling it is now at full blast and no commands I input will stop this

C:\.........ipmiutil> ipmiutil sensor -N 192.168.0.180 -U *Username* -P *Password* -n 42 -l 400 -h 3000

Is this the correct syntax ?

Try changing the fan mode in the web interface.

 

[RemyZ]

Yea i tried normal and optimal mode in the web gui and no difference. I am gonna raise the upper threshold and power cycle now and see if that fixes anything

Sent from my SM-G920T using Tapatalk

 

[RemyZ]

I changed the upper threshold values and it is listing as normal now but still loud fans. I changed the limits to 400/7000. After that I unplugged the IPMI, Powercycled and still loud fans.

C:\..........\IPMI>ipmiutil sensor -N 192.168.0.180 -U Username -P Password
-c
ipmiutil ver 2.98
isensor: version 2.98
Connecting to node 192.168.0.180
-- BMC version 1.11, IPMI version 2.0
ID | SDRType | Type |SNum| Name |Status| Reading
0004 | Full | Temperature | 01 | CPU Temp | OK | 24.00 C
0047 | Full | Temperature | 0a | PCH Temp | OK | 26.00 C
008a | Full | Temperature | 0b | System Temp | OK | 23.00 C
00cd | Full | Temperature | 0c | Peripheral Temp | OK | 29.00 C
0110 | Full | Temperature | 10 | VcpuVRM Temp | OK | 32.00 C
0153 | Full | Temperature | b0 | DIMMA1 Temp | Absent | 0.00 na
0196 | Full | Temperature | b1 | DIMMA2 Temp | Absent | 0.00 na
01d9 | Full | Temperature | b4 | DIMMB1 Temp | Absent | 0.00 na
021c | Full | Temperature | b5 | DIMMB2 Temp | Absent | 0.00 na
025f | Full | Fan | 41 | FAN1 | OK | 2700.00 RPM
02a2 | Full | Fan | 42 | FAN2 | OK | 6000.00 RPM<---------No More Above Critical Warning
02e5 | Full | Fan | 43 | FAN3 | OK | 6100.00 RPM<---------No More Above Critical Warning
0328 | Full | Fan | 44 | FAN4 | Absent | 0.00 na
036b | Full | Fan | 45 | FANA | OK | 6200.00 RPM<---------No More Above Critical Warning
03ae | Full | Voltage | 30 | 12V | OK | 12.19 V
03f1 | Full | Voltage | 31 | 5VCC | OK | 5.00 V
0434 | Full | Voltage | 32 | 3.3VCC | OK | 3.37 V
0477 | Full | Voltage | 33 | VBAT | OK | 2.97 V
04ba | Full | Voltage | 34 | VCPU | OK | 0.79 V
04fd | Full | Voltage | 35 | VDIMMAB | OK | 1.21 V
0540 | Full | Voltage | 38 | 5VSB | OK | 5.05 V
0583 | Full | Voltage | 39 | 3.3VSB | OK | 3.33 V
05c6 | Full | Voltage | 3b | VBMC 1.2V | OK | 1.21 V
0609 | Full | Voltage | 3c | VPCH 1.0V | OK | 0.99 V
064c | Full | Platform Securi | aa | Chassis Intru | OK | 0.00 na
ipmiutil sensor, completed successfully

 

[RemyZ]

Pouring over the ipmi util documentation.
http://ipmiutil.sourceforge.net/docs/UserGuide

 

[RemyZ]

I checked the logs and received all kinds of threshold errors. I powercycled and no new threshold events being listed.
I then set my thresholds again but this time using the "-p" persist setting and powercycled but no change at the moment. Changed the Fan Mode on the webgui but still nothing. I have listed some ipmiutil outputs. I am at a stopping point since I don't know what to try next other than moving one case fan from FANA to Fan 4 but dont think that would help. Any help would be appreciated Thanks

C:\.........IPMI>ipmiutil health -N 192.168.0.180 -U **** -P ****-c
ipmiutil ver 2.98
ihealth ver 2.98
Connecting to node 192.168.0.180
BMC manufacturer | 002a7c (SuperMicro), product | 0884
BMC version | 1.11, IPMI v2.0
IPMI driver type | 6 (lan)
Power State | 00 (S0: working)
Selftest status | 0055 (OK)
Chassis Status | 21 00 40 00 (on, see below)
chassis_power | on
pwr_restore_policy | last_state
chassis_intrusion | inactive
front_panel_lockout | inactive
drive_fault | false
cooling_fan_fault | false
Power On Hours | 1 hours (0 days)
BMC LAN Channels | 1
Chan 1 AuthTypes | MD2 MD5 Straight_Passwd
ipmiutil health, completed successfully


C:\...............IPMI>ipmiutil sel -N 192.168.0.180 -U **** -P **** -c
ipmiutil ver 2.98
isel: version 2.98
Connecting to node 192.168.0.180
-- BMC version 1.11, IPMI version 2.0
SEL Size = 512 records (Used=0, Free=512)
RecId | Date/Time | SEV | Src | Evt_Type | Sensor | Evt_detail
Firmware Log (SEL) is empty
ipmiutil sel, completed successfully


C:\..............ipmiutil sensor -N 192.168.0.180 -U **** -P ****
-c
ipmiutil ver 2.98
isensor: version 2.98
Connecting to node 192.168.0.180
-- BMC version 1.11, IPMI version 2.0
ID | SDRType | Type |SNum| Name |Status| Reading
0004 | Full | Temperature | 01 | CPU Temp | OK | 41.00 C
0047 | Full | Temperature | 0a | PCH Temp | OK | 28.00 C
008a | Full | Temperature | 0b | System Temp | OK | 23.00 C
00cd | Full | Temperature | 0c | Peripheral Temp | OK | 30.00 C
0110 | Full | Temperature | 10 | VcpuVRM Temp | OK | 35.00 C
0153 | Full | Temperature | b0 | DIMMA1 Temp | Absent | 0.00 na
0196 | Full | Temperature | b1 | DIMMA2 Temp | Absent | 0.00 na
01d9 | Full | Temperature | b4 | DIMMB1 Temp | Absent | 0.00 na
021c | Full | Temperature | b5 | DIMMB2 Temp | Absent | 0.00 na
025f | Full | Fan | 41 | FAN1 | OK | 2700.00 RPM<---------CPU Fan
02a2 | Full | Fan | 42 | FAN2 | OK | 6000.00 RPM<---------Case Fans
02e5 | Full | Fan | 43 | FAN3 | OK | 6100.00 RPM<---------Case Fans
0328 | Full | Fan | 44 | FAN4 | Absent | 0.00 na
036b | Full | Fan | 45 | FANA | OK | 6200.00 RPM<---------Case Fans
03ae | Full | Voltage | 30 | 12V | OK | 12.13 V
03f1 | Full | Voltage | 31 | 5VCC | OK | 5.00 V
0434 | Full | Voltage | 32 | 3.3VCC | OK | 3.38 V
0477 | Full | Voltage | 33 | VBAT | OK | 2.97 V
04ba | Full | Voltage | 34 | VCPU | OK | 1.16 V
04fd | Full | Voltage | 35 | VDIMMAB | OK | 1.22 V
0540 | Full | Voltage | 38 | 5VSB | OK | 5.05 V
0583 | Full | Voltage | 39 | 3.3VSB | OK | 3.33 V
05c6 | Full | Voltage | 3b | VBMC 1.2V | OK | 1.21 V
0609 | Full | Voltage | 3c | VPCH 1.0V | OK | 0.99 V
064c | Full | Platform Securi | aa | Chassis Intru | OK | 0.00 na
ipmiutil sensor, completed successfully

 

[RemyZ]

Sorry for all my rants. All is well now and appreciate any and all help. Power cable needed to be unplugged for a few minutes for it to register the changes in ipmi. Thanks again

 

[Glorious1]

Sorry for all my rants. All is well now and appreciate any and all help. Power cable needed to be unplugged for a few minutes for it to register the changes in ipmi. Thanks again

Yes, I reported the same thing earlier in the thread. Had to completely remove power from the motherboard, not just restart.

 

[megagator]

The only change is instead of cycling it is now at full blast and no commands I input will stop this

I'm struggling with this exact problem now. I have an X8DTN+ whose fans will not respond to any sensor thresholds I set and insist on running full force. Currently, ipmitool reports that all the fans have no upper threshold set ("n/a"), but do have lower thresholds, 200 300 400. I tried the -p flag that ipmiutil offers but that didn't have any effect either. I have been pulling the power cord out each cycle to get ipmi to restart, but no matter what ipmitool always reports the same sensor thresholds.

# ipmitool sensor | grep Fan

Fan1 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan2 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan3 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan4 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan5 | 4900.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan6 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan7 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan8 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

 

[Ericloewe]

I'm struggling with this exact problem now. I have an X8DTN+ whose fans will not respond to any sensor thresholds I set and insist on running full force. Currently, ipmitool reports that all the fans have no upper threshold set ("n/a"), but do have lower thresholds, 200 300 400. I tried the -p flag that ipmiutil offers but that didn't have any effect either. I have been pulling the power cord out each cycle to get ipmi to restart, but no matter what ipmitool always reports the same sensor thresholds.

# ipmitool sensor | grep Fan

Fan1 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan2 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan3 | 6400.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan4 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan5 | 4900.000 | RPM | ok | 200.000 | 300.000 | 400.000 | na | na | na

Fan6 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan7 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

Fan8 | na | RPM | na | 200.000 | 300.000 | 400.000 | na | na | na

The BMC on your board might not have an upper threshold for fan speeds (because it's of dubious value), so that might be a red herring.

Apparently, mixing PWM and non-PWM fans has been known to cause weird issues, so you might want to look into that, if it applies to you.

 

[belisarius]

I've been fighting with the same problem with my X8DTN+ as well. It's in an SC847 case (with 7 jet-engine fans up front for airflow), and half of the PWM headers on the motherboard are on the far side of the case, so I used a breakout cable to use just one header to try and control all the fans. I can't get them to drop below 2900 RPM (based off the reporting of FAN2, which is what the breakout cable is hooked up to), even though I've set the fan profile in the BIOS to be the "ES" mode for quietest operation. And like megagator, all my upper bounds are set to N/A.

I've been using ipmitool from Linux. I'm playing around with it now, and I'm ordering PWM extension cables to try and hook up each of the front fans individually, and I'll see how that works.

 

[Ericloewe]

The fans might simply not go lower than 2.9kRPM. It sounds about right for 10k RPM fans. Or the BMC simply doesn't want to make them go lower (and the thresholds won't help you there).

 

[belisarius]

I was worried that might be the case as well. But I haven't been able to find information on them regarding what the minimum speed is, and I've seen a number of posts around talking about lower RPM fans in other setups (this is the first I've come across that specifically references both the X8DTN+ and IPMI settings).

But if the fans simply won't go any lower, that leaves me the options of either replacing the fans or do some sound proofing.

I've actually been able to noticeably reduce the noise by removing a fan or two, without noticeable adverse temperature effects, so I might do that and see if I can deal with the noise.

 

[Ericloewe]

I was worried that might be the case as well. But I haven't been able to find information on them regarding what the minimum speed is, and I've seen a number of posts around talking about lower RPM fans in other setups (this is the first I've come across that specifically references both the X8DTN+ and IPMI settings).

But if the fans simply won't go any lower, that leaves me the options of either replacing the fans or do some sound proofing.

I've actually been able to noticeably reduce the noise by removing a fan or two, without noticeable adverse temperature effects, so I might do that and see if I can deal with the noise.

It's better to replace the fans than to remove some.

Assuming you have 10k RPM fans, going with 7k should work nicely, assuming you've got enough margin in the thermal design for the lower maximum airflow.

Supermicro should have some that are ready to go, but you can also get the bare fan (no connector on the cable, even) from Sanyo Denki (I think Mouser carries them).

 

[Kevin Horton]

I, like many others, had my fans stick at full speed after making one too many changes to the thresholds using ipmitool. I didn't want to do a full system power cycle, as I was 24 hours into badblocks testing (I was only attempting to adjust the fans because I noted that the hard drive temperatures were too high).

A lucky hit on Google led me to try:

Code:

ipmitool bmc reset warm

That command reset the BMC, without disturbing the rest of the X10SL7-F motherboard, so the bad blocks testing contined unabated. If that hadn't worked, my next attempt would have been

Code:

ipmitool bmc reset cold

 

[Kevin Horton]

I discovered that it is possible to control the IPMI fan speed using raw commands in ipmitool.

On my X10SL7-F board, the following command will check the fan speed:

Code:

ipmitool raw 0x30 0x45 0x00

The result should be 00, 01 or 02.
00 = Normal,
01 = Full and
02 = Optimal.

The fan speed may be changed by:

Code:

ipmitool raw 0x30 0x45 0x01 0x00 #set to Normal
ipmitool raw 0x30 0x45 0x01 0x01 #set to Full
ipmitool raw 0x30 0x45 0x01 0x02 #set to Optimal

I'm pondering writing a script that checks hard drive temperatures, and increase the fan speeds if any of them were 40 deg C or higher, and decrease the fan speeds if all were 38 deg or lower. I'd use cron to run the script every X minutes. I'll add a thread to the Scripts section if I get this working.