DL380 G7 Fan Speed

[rockshox]

We recently purchased 2 DL380 G7's, identical machines. One is running Windows 2008 R2 the second RHEL 5.5 fully patched, BIOS/Firmware updated etc. The machine running Windows 2008 has been flawless, however the machine we have RHEL installed on has had issues since we got it.

The two main issues was big problems with the P410 raid controller and the fan speeds suddenly bumping up in speed to 78% and sticking there no matter if the server has no load or how cool the air conditioned server room is. After going around and around with HP over the raid controller issues, they finally replaced the motherboard which seems to have solved the issues with the P410 controller. We hoped that this would solve the fan issue also, but last night in the middle of the night, no users whatsoever using the box the fan speeds kicked in and changed from 13% to 78%. Checking the Windows 2008 server (which is in the same rack in the same air conditioned server room) shows the fans run at 13-14% constantly, I don't think it's ever gone higher than that even though that server has way more load.

We believe the problem is related to the HP Management drivers as we've found that if we do something to trigger the management agents (pull a network cable, pull a power cord) the fans within seconds instantly spin back down to 13%. Rebooting also spins the fans back down and they'll stay at 13% for severals days, maybe even a week and then something triggers them to speed up. We've had them speed up at all different times of the day, so we haven't been able to see a pattern.

Anyone seen anything like this before or have any suggestions?

[Ali]

HI Rocknox,

You may update the HP System Health Application and Command Line Utilities for Red Hat Enterprise Linux 5 (x86)
http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=us&prodTypeId=15351&prodSeriesId=4091412&prodNameId=4091432&swEnvOID=4006&swLang=8&mode=2&taskId=135&swItem=MTX-33039013d479470688aca7fedf

and
ILO driver to see whether it makes any difference.

thanks,
Aftab

[rockshox]

Hello aftab ali,

Thanks for the reply. We are currently running the most recent versions of the HP Health and ILO drivers.

rpm -qa | grep hp-

hp-health-8.6.0.24-27
hp-OpenIPMI-8.5.0-1.rhel5
hp-snmp-agents-8.6.0.18-18
hp-smh-templates-8.6.0.14-16
hp-ilo-8.5.0-1.rhel5

[Ali]

hi rockshox,

Any other observations?

did you notice any console messages or logs being generated during this time?

log in to ILO to check ILO temperature threshold and IML

I agree it might be due to some management agents as rebooting the server fixes the issue temporarily.

what about stopping the health agents to eliminates the possibility of any h/w failure such as ambient sensor

I have similar server in my environment but it is working fine.. I will check few things and would let you know.

thanks,
Aftab

[rockshox]

Hello aftab ali,

Thanks for the reply. Were continuing to monitor this server but haven't noticed any other observations that will help to resolve the problem. We did notice that at some point yesterday, and we don't know why, the fans slowed themselves down and they are running at 13% now. Completely random, as we hadn't done a thing on this server.

Were continuing to use hplog to monitor the temperatures at random times in case the temperature on one of the sensors spikes, causing the fans to bump up.

[Michael A. McKenney]

Try reseating the fans. Check to see if they are clogged with dust. Make sure you have adequate airflow around the server and out the back.