Operating System - HP-UX
1753620 Members
6432 Online
108797 Solutions
New Discussion юеВ

Temperature monitoring in SuperDome

 
SOLVED
Go to solution
leoh_1
Regular Advisor

Temperature monitoring in SuperDome

How to monitor the temperature in SuperDome. I heard of a HP Mib called SnmpHpunix . Is it the solution? Is there any command line for it?
7 REPLIES 7
Jaime Bolanos Rojas.
Honored Contributor

Re: Temperature monitoring in SuperDome

Leoh,

In a supperdome you can monitor the temp throught the EMS hardware monitor utilities or through the management processor ( MP ) and you can do both of them remotely.

Regards,

Jaime.
Work hard when the need comes out.
leoh_1
Regular Advisor

Re: Temperature monitoring in SuperDome

Thank you jaime.
Before start reading about EMS or MP manuals in HP-UX, would you tell me if there is a simple comandline to retrieve this information (temperature)?

Ideal world

#ems -temp
35C

:))
Bill Hassell
Honored Contributor

Re: Temperature monitoring in SuperDome

I will answer this before Clay does: Computers are LOUSY thermometers!! Even computers like the (very expensive) SuperDome. You can determine if a temperature warning has occurred (apx. 85-90 deg F) or an emergency has occurred (95 to 105 deg F) at which point the SuperDome shuts itself off. Wonderful. The computer protected itself but the disk arrays, tape drives, network appliances and other servers, PCs, printers, etc will meltdown. Oh, they won't always melt, their electronics is permanently damaged so they seldom work right if at all.

DO NOT relegate temperature monitoring to equipment that was never designed for this job. There are a several companies that provide not only accurate and multipoint temperature monitoring but also check on water on the floor (ALWAYS a problem in computer rooms with poor air conditioning) and humidity, and can connect into central alarm systems (email is far too unreliable for a problem that can destroy millions of dollars of equpment).

These alarm systems are monitored by the same people that watch for fire and burglary, etc. The advantage is that a live body will keep calling until someone is located to take care of the problem. You can be sure heads will roll if there are 10 emails warning about airconditioning failure in the sysadmins mailbox on Sunday evening that were not read until Monday morning. Even pager notification is far too unreliable -- no verification of receipt or acknowledgement that action is taking place.

And even with a solid environmental monitoring system, you still need a thermostatic circuit breaker that removes all power once the room reaches 100 deg F. Even with 24x7 it may be too late to start powering things down, especially if the airconditioning cannot be fixed in 10 minutes.

By the way, a typical data center that loses all airconditioning will rise in temperature to 100 degrees in less than 10 minutes, 140 degrees in less than 20. At 140 deg F, you call the insurance company to cover your loss.


Bill Hassell, sysadmin
leoh_1
Regular Advisor

Re: Temperature monitoring in SuperDome

We have selected machines in strategic places in data center to alarm us that the temperature is getting high. SuperDome is in one of this strategic places. If we get the temperature of different places in datacenter, and calculate the delta T, then we can predict an airconditioning problem and we can act in 3 minutes or less.
Jaime Bolanos Rojas.
Honored Contributor
Solution

Re: Temperature monitoring in SuperDome

Leoh,

According to this thread you can run a command like this:

#/opt/resmon/bin/resdata -R 110100482 -r /system/events/core_hw/core_hw -n 110100492 -a

You can also check the log file: /var/resmon/log/event.log if there is any h/w issue.

On that thread they had a problem where the server overheated and went down,

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=950169

Regards,

Jaime.
Work hard when the need comes out.
leoh_1
Regular Advisor

Re: Temperature monitoring in SuperDome

Thank you jaime, you guided me to the right place.

URL with tutorial on monitoring system from HP hardware

http://docs.hp.com/en/diag.html#EMS%20Hardware%20Monitors

URL with events from core hardware, including Temperature issue:
http://docs.hp.com/en/diag/ems/dm_core_hw.htm#33

URL with FAQ showing were is the notification file
http://docs.hp.com/en/diag/ems/ems_faq.htm#defaultnotification

Althought I did not discover how to retrive the current Temperature of the cabinet...
Jaime Bolanos Rojas.
Honored Contributor

Re: Temperature monitoring in SuperDome

Leoh,

About checking current temp for a box, is not possible, those machines are going to tell you if they are too hot, but never how hot, even thought a couple of people have been interested on that feature already.

Regards,

Jaime.
Work hard when the need comes out.