[Cialug] Slightly OT: Server Room Temperature
major.stubble at gmail.com
Fri Sep 14 13:23:07 CDT 2007
Interesting that this topic should come up. IEEE just had a paper
published on predicting HDD failure that incorporated environmental
factors. I haven't logged in yet to read the article, so I don't
know what the findings are.
My most common failure with high-temperature data centers (if UNI's
machine room can even be considered a 'data center') has been hard
disk drives. So, an article like this may prove useful as we argue
about future expansion.
Hard Disk Drive Reliability Modeling and Failure Prediction
Strom, B. D.; Lee, S.; Tyndall, G. W.; Khurshudov, A.
Magnetics, IEEE Transactions on
Volume 43, Issue 9, Sept. 2007 Page(s):3676 - 3684
A reliability model for the hard disk drive (HDD) is developed,
focusing on head–disk separation as the primary independent variable.
The model is structured to incorporate the theoretical effects of
environmental factors, plus empirical dependence on the product
operating mode. An experimental method based on magnetic spacing loss
theory is used to characterize the head–media separation as a
function of temperature, altitude, humidity, and HDD operating mode.
A statistical model based on these empirical data is developed to
predict HDD reliability for various operating conditions. The
predictions of the model are verified experimentally through
comparison with HDD product reliability test data.
On Sep 14, 2007, at 12:57 PM, Paul Gray wrote:
> Jonathan C. Bailey wrote:
>> I've been considering a shell script to check the environmental
>> monitor in the room and shutdown as needed. Least critical first
>> of course..
> I use weatherduck to monitor temperature, humidity, airflow and
> light levels (eg., when someone comes into the server room). They
> are outstanding monitoring units that sit on the back of a 1U/2U
> case mounted to the serial port. Sitting on the back, I get a
> reading of the hot isle temperature. http://www.itwatchdogs.com/
> I'd also recommend those using server boards that are equipped to
> monitor core temperature to use mbmon. mbmon works exceptionally
> well with my dual-socket, dual-core Opteron boards from Tyan.
> There is a huge difference between ambient temperature and core
> temperature (understatement).
> Cialug mailing list
> Cialug at cialug.org
More information about the Cialug