server shutdown due to oveheating

[Cherie Galloway]

How can I get infomation on the internal temperature of the system when it has shut off because of overheating? System is a V2250. Our coolers broke down and we had the system crash when the room got up to 107%. Now we what to see what the internal temp was.

[S.K. Chan]

What wil happen is system will warn if the temperature reaches 95F. A second warning at 104F will shut the OS. So the temperature of the server has to be above 104F.
I do not think you can monitor the temperature of the system. However take a look at /etc/envd.conf which defines what are the action the OS will take in the event of OVERTEMP signal being detected. The "envd" daemon started in /sbin/init.d/envd monitor this event.

[Tom Geudens]

Hi,
You can use EMS to give warnings. If you're in time you can then do a "gracefull" shutdown in order to prevent the crash.
However, it is impossible to get the temperature from the server itself, I recently asked this myself, see http://forums.itrc.hp.com/cm/QuestionAnswer/0,,0x83d335067c18d6118ff40090279cd0f9,00.html

Another solution (we use it, but it looks a bit stupid :-) is to install an SC10-rack. A rack like that DOES have a measureable temperature (you can get it from STM).

Hope this helps,
Tom Geudens

[Patrick Wessel]

Ron,
The V-Class doesn't log the temperature information during the system operation. If you like to know the current temerature and the threshold for over-temp warning and over-temp shutdown you can use following command on the teststation:
ds1620 -n 0

You should find an entry in the /spp/data/eventlog file with the pce_util output from the over temperature event. This entry includes the temperature in the moment when the system issued the over-temp warning and performed the over-temp power-off