AIX

Have recently found documentation for the FSP CIM interface.  The CIM allows access to metrics such as inlet/outlet air temperatures, fan speeds, power usage.  The metrics can be retrieved using the wbemcli opensource command compiled on AIX. 

The question is why are some of the servers not returning the full list of metrics?  All these servers are identically configured hardware small Power 9 Model 9009-41A / S914.  Servers are all stand-alone and there is not HMC involved.  An IP connection is available to the HMC port on all FSPs.  All servers were purchased at same time.  All servers and FSPs running AIX updated to the same level and the FSPs have the same firmware version.  So one would expect each FSP would report the same list of metrics.

We recently had 1 server with an issue that was also causing a power supply fault and was missing metric data for the power supplies.  We also noticed missing hardware elements such as a power supply form the CIM data.  The missing hardware was likely the reason for the missing metrics.  This was corrected by replacing a cable in the chassis.

Queries of other FSPs revealed a few servers were returning some but not all of the metrics. 

The inlet air temp has a location code of D1 which would indicate the sensor is in the control panel.

The missing power metric is always at location E3 (systems have 2 power supplies E3 and e4).

Has anyone had any experience with using CIM data and what the missing data might mean? 

This is likely a hardware issue but it seems odd that 30% of these system have a combination of bad control panels, power supplies or cables 

.  

Have recently found documentation for the FSP CIM interface.  The CIM allows access to metrics such as inlet/outlet air temperatures, fan speeds, power usage.  The metrics can be retrieved using the wbemcli opensource command compiled on AIX. 

The question is why are some of the servers not returning the full list of metrics?  All these servers are identically configured hardware small Power 9 Model 9009-41A / S914.  Servers are all stand-alone and there is not HMC involved.  An IP connection is available to the HMC port on all FSPs.  All servers were purchased at same time.  All servers and FSPs running AIX updated to the same level and the FSPs have the same firmware version.  So one would expect each FSP would report the same list of metrics.

We recently had 1 server with an issue that was also causing a power supply fault and was missing metric data for the power supplies.  We also noticed missing hardware elements such as a power supply form the CIM data.  The missing hardware was likely the reason for the missing metrics.  This was corrected by replacing a cable in the chassis.

Queries of other FSPs revealed a few servers were returning some but not all of the metrics. 

The inlet air temp has a location code of D1 which would indicate the sensor is in the control panel.

The missing power metric is always at location E3 (systems have 2 power supplies E3 and e4).

Has anyone had any experience with using CIM data and what the missing data might mean? 

This is likely a hardware issue but it seems odd that 30% of these system have a combination of bad control panels, power supplies or cables 

.  

Have recently found documentation for the FSP CIM interface.  The CIM allows access to metrics such as inlet/outlet air temperatures, fan speeds, power usage.  The metrics can be retrieved using the wbemcli opensource command compiled on AIX. 

The question is why are some of the servers not returning the full list of metrics?  All these servers are identically configured hardware small Power 9 Model 9009-41A / S914.  Servers are all stand-alone and there is not HMC involved.  An IP connection is available to the HMC port on all FSPs.  All servers were purchased at same time.  All servers and FSPs running AIX updated to the same level and the FSPs have the same firmware version.  So one would expect each FSP would report the same list of metrics.

We recently had 1 server with an issue that was also causing a power supply fault and was missing metric data for the power supplies.  We also noticed missing hardware elements such as a power supply form the CIM data.  The missing hardware was likely the reason for the missing metrics.  This was corrected by replacing a cable in the chassis.

Queries of other FSPs revealed a few servers were returning some but not all of the metrics. 

The inlet air temp has a location code of D1 which would indicate the sensor is in the control panel.

The missing power metric is always at location E3 (systems have 2 power supplies E3 and e4).

Has anyone had any experience with using CIM data and what the missing data might mean? 

This is likely a hardware issue but it seems odd that 30% of these system have a combination of bad control panels, power supplies or cables 

.  

Hi Jim,

I've tried to use CIM for checking ambient temperature, but in my case on some boxes CIM is not posible to enabled, on other boxes depends of firmware level I receved less data via CIM (the newer the firmware, the less data I received).

Hoever, yesterday when I read history for VL9xx Levels I found a potentially different method to collect the data I am interested in: Redfish (REST) API.

I'm not familiar wiht redfish, so I put it on my list: 'To read'.

Details:

Power9 System Firmware Fix History - Release levels VL9xx

https://www.ibm.com/support/pages/node/6955591

(Part of details for VL950_119_045 / FW950.70)

• Support for using a Redfish (REST) API to gather power usage for all nodes in watts and the ambient temperature for the system.
  Redfish sample response is as shown below:
  ==>> GET redfish/v1/Systems/<>
  ...
      "Oem": {
          "IBMEnterpriseComputerSystem": {
              ...
              ...
              "PowerInputWatts" : <> ( number in watts),  <<<<============
              "AmbientTemp" : <> (number in Celsius) <<<<============
          }
      },
  ...

(Redfish documents):

https://www.dmtf.org/standards/redfish

Regards,

Michal

Have recently found documentation for the FSP CIM interface.  The CIM allows access to metrics such as inlet/outlet air temperatures, fan speeds, power usage.  The metrics can be retrieved using the wbemcli opensource command compiled on AIX. 

The question is why are some of the servers not returning the full list of metrics?  All these servers are identically configured hardware small Power 9 Model 9009-41A / S914.  Servers are all stand-alone and there is not HMC involved.  An IP connection is available to the HMC port on all FSPs.  All servers were purchased at same time.  All servers and FSPs running AIX updated to the same level and the FSPs have the same firmware version.  So one would expect each FSP would report the same list of metrics.

We recently had 1 server with an issue that was also causing a power supply fault and was missing metric data for the power supplies.  We also noticed missing hardware elements such as a power supply form the CIM data.  The missing hardware was likely the reason for the missing metrics.  This was corrected by replacing a cable in the chassis.

Queries of other FSPs revealed a few servers were returning some but not all of the metrics. 

The inlet air temp has a location code of D1 which would indicate the sensor is in the control panel.

The missing power metric is always at location E3 (systems have 2 power supplies E3 and e4).

Has anyone had any experience with using CIM data and what the missing data might mean? 

This is likely a hardware issue but it seems odd that 30% of these system have a combination of bad control panels, power supplies or cables 

.  

Hi,

If you have faulty Power hardware or even suspect hardware, then I strongly suggest you contact IBM Power hardware support to get that fixed as-soon-as-possible.

I don't know anyone that uses the CIM interface to extract temperature or electrical power use.

This is because most Power customers have HMC's that are network connect to the service processors and that means they are on isolated private networks for security reasons.  The HMC has a different and very complex REST API for that data and performance statistics with a separate bunch of issues. In case, you are interested Google 
If that is of interest find "nextract plus" link here AIXpert Blog from Nigel Griffiths (@mr_nmon) (ibm.com) This is also Youtube videos on the setup.

Sorry, that this does not help much in you case.

Cheers @mr_nmon

Hi Jim,

I've tried to use CIM for checking ambient temperature, but in my case on some boxes CIM is not posible to enabled, on other boxes depends of firmware level I receved less data via CIM (the newer the firmware, the less data I received).

Hoever, yesterday when I read history for VL9xx Levels I found a potentially different method to collect the data I am interested in: Redfish (REST) API.

I'm not familiar wiht redfish, so I put it on my list: 'To read'.

Details:

Power9 System Firmware Fix History - Release levels VL9xx

https://www.ibm.com/support/pages/node/6955591

(Part of details for VL950_119_045 / FW950.70)

• Support for using a Redfish (REST) API to gather power usage for all nodes in watts and the ambient temperature for the system.
  Redfish sample response is as shown below:
  ==>> GET redfish/v1/Systems/<>
  ...
      "Oem": {
          "IBMEnterpriseComputerSystem": {
              ...
              ...
              "PowerInputWatts" : <> ( number in watts),  <<<<============
              "AmbientTemp" : <> (number in Celsius) <<<<============
          }
      },
  ...

(Redfish documents):

https://www.dmtf.org/standards/redfish

Regards,

Michal

Have recently found documentation for the FSP CIM interface.  The CIM allows access to metrics such as inlet/outlet air temperatures, fan speeds, power usage.  The metrics can be retrieved using the wbemcli opensource command compiled on AIX. 

The question is why are some of the servers not returning the full list of metrics?  All these servers are identically configured hardware small Power 9 Model 9009-41A / S914.  Servers are all stand-alone and there is not HMC involved.  An IP connection is available to the HMC port on all FSPs.  All servers were purchased at same time.  All servers and FSPs running AIX updated to the same level and the FSPs have the same firmware version.  So one would expect each FSP would report the same list of metrics.

We recently had 1 server with an issue that was also causing a power supply fault and was missing metric data for the power supplies.  We also noticed missing hardware elements such as a power supply form the CIM data.  The missing hardware was likely the reason for the missing metrics.  This was corrected by replacing a cable in the chassis.

Queries of other FSPs revealed a few servers were returning some but not all of the metrics. 

The inlet air temp has a location code of D1 which would indicate the sensor is in the control panel.

The missing power metric is always at location E3 (systems have 2 power supplies E3 and e4).

Has anyone had any experience with using CIM data and what the missing data might mean? 

This is likely a hardware issue but it seems odd that 30% of these system have a combination of bad control panels, power supplies or cables 

.