Power meter readings incorrect in iLO2 v1.70 ?

[Ian Y]

I've just updated all the firmware in a DL380 G5 and have noticed a substantial difference in the reported power usage levels.

Before updating the iLO reported it was using around 300W, after updating the server is reporting it is using around 80W !?

Nothing else was changed with the server, only the firmware. I have used a power reading clamp device which indicated the server was drawing around 1.2A (which at 240V UK voltage give a wattage of 288W which is approx what the previous version of iLO was reporting), we will have some more accurate inline power meters arriving shortly so I will be able to verify the iLO reading is bogus.

We have an identical DL380 G5 with the same System ROM (11/01/2008), older disk controller firmware (1.2) and iLO2 v1.61 which is reporting around 300W (see end of post for details).

Please can someone at HP look into this.

System info:
HP DL380 G5
System ROM: P56 11/01/2008
License Type: iLO 2 Advanced
iLO 2 Firmware Version: 1.70 12/02/2008

24-hour History (DL380 G5 iLO2 v1.70):
Average Power: 84 Watts
Maximum Power: 366 Watts
Minimum Power: 84 Watts

24-hour History (DL380 G5 iLO2 v1.61):
Average Power: 339 Watts
Maximum Power: 392 Watts
Minimum Power: 339 Watts

[Shiraj Fernando]

Hi,

If you look at the fixes made with the iLO2 Firmware 1.70 it mentions that "Improved the gathering of historical power consumption data"

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=15351&prodSeriesId=1121516&swItem=MTX-7b457ac480b9439c944576ef2a&prodNameId=3288134&swEnvOID=1005&swLang=8&taskId=135&mode=5

Also with System ROM v.11/01/2008 it mentions about the following enhancement.

"Added support for Dynamic Power Capping. Dynamic Power Capping allows reclaiming power and cooling capacity by safely limiting server power consumption using a fast-acting, hardware-based capping algorithm. Dynamic Power Capping requires updating the iLO 2 Firmware in addition to the System ROM."

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=15351&prodSeriesId=1121516&swItem=MTX-2a94ee936b09442fab036d42b1&prodNameId=3288134&swEnvOID=1005&swLang=8&taskId=135&mode=5

Regards,
Shiraj.

[Ian Y]

Although it does indeed improve the look of historical power data and does give facilities for power capping, I'm not using the power capping and the look of the data should not change the data!

This also does not explain the independent power readings which seem to validate the approx 300W consumption being the more accurate value.

Therefore it is still unknown, via another channel I've heard from an HP support person who is also surprised at this, however they are unable to offer any reason for it or a solution.

[juan quesada]

hi,

bad news, http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&objectID=c01130986&jumpid=reg_R1002_USEN

basically there is no fix yet, my recommendation is to log a call with HP and have the case elevated for further resolution.

[Ian Y]

This is not a sporadic drop to 0 and back to normal again, this is consistently reading low.

After logging a call with HP support (and after they suggested the usual turn it off and on again), I still have the problem. So they then suggested a downgrade of iLO to 1.61.

I have now downgraded to 1.61 and still am having the same problem.

[KarloChacon]

I think it is the same issue here


http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1313420


basically what our friend did was reseat ilo

I had the same issue and we had to remove power cable for 3 minutes and ilo power meter came bac to normal