Temperature Monitoring for ESX servers

[TxTechLaw]

We are looking at AVTech environmental monitors for our server room. From what I can tell, one of these with their PageR software can send shutdown commands to physical Windows servers, as well as Windows virtual machines, based on temp, humidity, and other environmental conditions (like water accumulating on the floor : )

However, I'm not sure how to do the same thing with our ESX hosts, which are Dell R710s. Basically, I want the server to shut itself down if the temperature reaches a certain point. Ideally, it would only do this if the temp were x amount for a given amount of time. Anyone know of a product that will do this? OpenManage? I'm not sure what my options, if any, are with the ESX systems themselves.

Thank you,

Chad

[oreeh]

I've done similar monitoring / shutdown things with HP servers (utilizing the HP ESX Agents - there should be a similar solution available for Dell systems) and external monitoring devices using Nagios / Icinga and some custom scripts / checks.

The monitoring part mostly is straight forward SNMP, the shutdown can be done using SSH or one of the different SDKs.

[couak]

you can try this following python script :

#!/usr/bin/python

import sys

import pywbem

wbemclient = pywbem.WBEMConnection('https://hostname:5989', ('root', 'password'), 'root/cimv2')

for instance in wbemclient.EnumerateInstances('CIM_NumericSensor') :

print "%s %s %s" % (instance, instance, instance)

On one of my R710 server I got this :

System Board 1 System Level 27300 7

Power Supply 2 Voltage 2 22600 5

Power Supply 1 Voltage 1 22800 5

Power Supply 2 Current 2 59 6

Power Supply 1 Current 1 80 6

System Board 1 FAN 5 RPM 360000 19

System Board 1 FAN 4 RPM 360000 19

System Board 1 FAN 3 RPM 360000 19

System Board 1 FAN 2 RPM 360000 19

System Board 1 FAN 1 RPM 360000 19

System Board 1 Ambient Temp 1700 2

According to the CIM SDK reference (http://www.vmware.com/support/developer/cim-sdk/smash/u2/ga/apirefdoc/CIM_NumericSensor.html) the latest line means 17°C

[Texiwill]

Hello,

You may want to approach this using a 'Monitoring Appliance' that gets notified and then can shutdown/migrate the VMs on a host before shutting down the host. In almost all cases of PowerManagement I have found that such a host is invaluable.

Best regards,
Edward L. Haletky VMware Communities User Moderator, VMware vExpert 2009

Now Available: 'VMware vSphere(TM) and Virtual Infrastructure Security'[/url]

Also available 'VMWare ESX Server in the Enterprise'[/url]

Blogging: The Virtualization Practice[/url]|Blue Gears[/url]|TechTarget[/url]|Network World[/url]

Podcast: Virtualization Security Round Table Podcast[/url]|Twitter: Texiwll[/url]