Cheshire Cat Computing

This plugin is available from . It also provides output in a MRTG-friendly format (use the -M flag for MRTG, -N for Nagios).

UPDATE: For the latest version, look on . Here you will find examples as well.

Nagios examples:
check_esx -H esx-hostname -C community -l LIST
check_esx -H esx-hostname -C community -v vhostname -l MEM -w 10% -c 5%
check_esx -H esx-hostname -C community -v vhostname -l STATE
MRTG examples:
check_esx -M -H esx-hostname -C community -v vhostname -l CPU
(returns CPU seconds used)
check_esx -M -H esx-hostname -C community -v vhostname -l MEM
(returns Memory avaiable, memory total)

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Plans for updated version:

1) Include extra checks for the host-MIB (number of processes, load average, memory) so that they are all conveniently in one place.
2) Have the CPU check, when run in Nagios mode, take two samples 10 sec apart and extrapolate a percentage from this which can be thresholded.
3) Add checks for virtual network and disk throughput

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Note: This plugin requires the Net::SNMP Perl module to be installed. This is how I access the SNMP. If you get a big error message about 'Cannot locate Net::SNMP' then this is your problem.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

OK, I thought it was obvious, but apparently it isn't.

You need to enable SNMP on your ESX console before this plugin will work.

Login to the console as root and run 'snmpsetup.sh default' to get the daemons running. Then configure /etc/snmp/snmpd.conf and restart them. There is lots about this in Ch7 of the ESX manual. You may also want to change the community from the default of 'public'.
[/b]

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

For some reason, the vhost name you are giving does not match any of the list returned by SNMP.

Remember it is case sensitive. There is an option to check_esx that returns the list of all known vhosts, but I cant remember it off the top of my head (I'm not in the office at the moment). I think you do the host command without giving the vhost name argument, or something similar - Ill check when I'm next in. I suspect it is a simple case or spelling difference.

Another possibility is that you dont have SNMP properly configured on the vmware server and so check_esx is getting no vhost list at all, although I think that would give a different error message.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hmm - check that your VMWare SNMP agents are running.

You should have snmpd running on the VMWare console server, and also the vmware-snmpd helper agent (this is the one that deals with the 6876 enterprise). Try restarting snmpd and vmware-snmpd (in that order).

Make sure you are querying the VMWare console and not the guest OS!

Make sure you can SNMPwalk the agent. IE, check you have the right community string and that you can get a response from the SNMP daemon.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hi Steve,

I came across this page while trying to find out how to use MRTG and vmware. I have a windows XP host running VMWare workstation 5.5 (i.e. not ESX) and inside this I am running a Linux server as a guest OS. I'm looking for a solution whereby I can monitor certain things like CPU use of the host from inside the guest. Your script says it is for ESX, is there a version what will work in my case?

Thanks in advance for any advice.

Regards,
Mark

What we do here is to use a 2-stage process. Each guest is monitored as we would any other standalone server (install agent, set up templates, etc).

Then, we use the check_esx to query SNMP on the vmware console and obtain information like CPU usage by a particular guest, and overall CPU usage on the physical machine.

Finally, due to the dynamic nature of VMware (especially if you use VMotion as we do) I needed to produce a further utility which I have not yet published. That is a perl script which runs every 5 mins, and it queries for all existing guests, plus it has a saved list of past guests. It dynamically recreates some MRTG .cfg files and also fetches CPU, Net and Memory usage information and stores it into multiple .rrd files. MRTG does not actually get involved in the process at all. The .cfg files are then used by routers2 to display some complex userdefined summary graphs over the guests. This has the benefit of holding historical info even for guests no longer in existance or moved to a different server.

I can send you a copy of this last program if you wish? Also, check_esx has been updated and I will upload the new version to the web shortly. The new version is better with CPU stats.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11