This is the way we do it, with Parents (not host dependencies).
First we create a virtual object for the VMWare farm. This has a status of UP
if any of the farm servers are up (using check_summary). This virtual 'host'
has several services, using the v0.9 check_vmware, relating to the farm's
alarms, storage volumes, etc. These services have service dependencies on the
VirtualCentre service running on the Virtual Centre host.
The Farm object has ALL of the ESX Servers as Parents.
All the VMs in the farm have the Farm object as a parent. Some of them also
use check_esx3 to alert on Alarms, CPU, and Memory usage within VMWare.
This might seem a bit complex if you've only the one server, but as soon as oyu
have multiple servers in the farm, and use DRS, you have to use a farm object
for parents/dependencies.
It might make more sense for these relationships to be host dependencies rather
than parents i nmost cases, but we have a SAN mirrored environment to a seocnd
ESX farm so that the VMs can be brought up ther ein the event of a complete
farm outage, hence the use of Parents rather than dependencies.
If you have VSphere4 (ESX4.0) with a SNMP-enabled Cisco virtual switch in the
farm, you could probably make the virtual switch the parent device rather than
having to use a farm object.
The VMWare monitoring plugin we're using is v0.9 of check_vmware, from here:
http://www.steveshipway.org/forum/viewtopic.php?f=28&t=1648
check_summary is available from nagiosexchange.org (as is check_esx3 which is
the forerunner of check_vmware)
Steve
________________________________
From: Andrew Davis [ncc...@gmail.com]
Sent: Tuesday, 26 January 2010 9:14 a.m.
To: nagios-users@lists.sourceforge.net
Subject: [Nagios-users] Virtual Machines - define as parent or as host
dependency...
I'm trying to figure out the best way to do this, yet keep things as simple as
possible.
Say I have a server called Saturn running VMWare. I'm monitoring this server
with Nagios.
I also have three VM's on Saturn: Jupiter, Mars, and Pluto
I want to suppress all host and service alerts on Jupiter, Mars, & Pluto if the
host Saturn is down (unreachable). I do NOT want to suppress host or service
alerts from Jupiter, Mars, and Pluto if the VMWare processes (services) are
down on Saturn. Basically, if my VM server is completely unreachable, don't
bother me about its client VM's.
Am I better off doing this with a host dependency? Something like:
define hostdependency {
host_name Host B
dependent_host_name Host C
notification_failure_criteria d,u
}
Or am I better off defining Saturn as the parent of the VM's in the host
config? Something like:
define host {
host_name jupiter
use VMs
alias jupiter
address 172.26.251.60
parents saturn, tpdmzsw1
}
I've successfully used the "parents" directive to define network topology, so I
would think this would work. What might be the risks of defining both?
--
A. Davis
Email: ncc...@gmail.com<mailto:ncc...@gmail.com>
"There is no limit to what a man can accomplish
if he doesn't care who gets the credit." - Ronald Reagan
------------------------------------------------------------------------------
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_______________________________________________
Nagios-users mailing list
Nagios-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting
any issue.
::: Messages without supporting info will risk being sent to /dev/null
