Cheshire Cat Computing

Hi,

I'm running vcenter 5.0 with one esxi 4.1.0 and on esxi 5.0. However I'm having trouble getting values from the VMs.

The only info I got with check_vmware.pl was the number of running vm. (My final goal is to poll this from cacti, but it doesn't really change the usage there).

I'm running debian 6.0.3, vsphere sdk Perl 5.0.0-422456.x86_64, and check_vmware 1.13.

Here is what I get from my commands : (I censored machine name an vm name, but it's returned correctly, note that I have space and dash in the vm name)





Any idea ? I'm running to " young " versions ?

Edit : FYI I'm using a readonly user, maybe it's rights aren't high enough ? I don't think so, as I'm able to see my usage in vcenter.

The 'net' reports (and the ones for disk io) are not working yet; there are problems correctly identifying the relevant items.

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

Just wanted to say thanks for the plugin!
So far it's working great!

Guys... What to do...
I have ESXi 5 hosts.
Also I have the same problem with cpus and NICs

CRIT: CPU usage at 297586.56% (need more CPU allocation?)
CRIT: Network usage at 80537144Kbps (need more resource shares or physical interfaces?)|net=80537144;7680;10240;0;

What should I do? How to resolve this bug.
Maybe I should downgrade to SDK 4? Or not. I guess not...
Because I can monitor correct values from ESXi 4.1

So what to do?

Hello everybody,

I had tested your script and for me it's works fine except that as you said it is resource intensive. Have you any idea why it consume so many resources ?
I'm not a developper so maybe it's a stupid question.

Indeed, I tested your script with FAN (Fully Automated Nagios: Nagios + Centreon) and all dependencies necessary (Perl Kit, ...). The VM was configured with 1 vCPU (2,8GHz) and 1GB Memory, no CPU or memory contention is the ESXi (vSphere 5).
I scheduled 4 check (CPU, Memory, Network and Disk), so 4 executions of the script, every 2 minutes and each execution consume between 75% and 90% of CPU.

Is in the future version it will be optimize to consume less CPU resource ?

Actually, Is it possible to optimize it ?


Sebastien.

Just curious how I can use your check_vmware.pl script with Nagios to examine the status of multiple vCenter servers. Our environment is multi-tennant with multiple ESX clusters each managed by an independent vCenter server. I'm new to Nagios so any direction would be helpful.

Bob Rains

the script is working fine on my environment except --report=net. It gives me exactly the same result (0) everytime!


[root@nagiosrv VMWare]# ./check_vmware.pl --config=configfile --report net --debug
Starting.
Connecting
Connected
Server Time : 2012-04-30T11:23:14.980296Z
Report type requested is [net]
Base is Datacenters
Running network report
Retrieving PerfMgr data
Selected interval is: 300
Retrieving list of hosts...
Making new propertyspec
Making new filterspec from
Retrieving vim_service
Retrieving properties from VimService
Checking faults on SoapResponse
Processing entities:
host-18
host-9
host-15260
host-8843
host-8848
host-15265
Creating query for MORef host-18
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Creating query for MORef host-9
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Creating query for MORef host-15260
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Creating query for MORef host-8843
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Creating query for MORef host-8848
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Creating query for MORef host-15265
Start time: 2012-04-30T11:18:00Z
End time : 2012-04-30T11:23:00Z
Retrieving data...
Perfstats retrieved...
Results for value host-18
Perfdata object: 2012-04-30T11:25:00Z
Results for value host-9
Perfdata object: 2012-04-30T11:25:00Z
Results for value host-15260
Perfdata object: 2012-04-30T11:25:00Z
Results for value host-8843
Perfdata object: 2012-04-30T11:25:00Z
Results for value host-8848
Perfdata object: 2012-04-30T11:25:00Z
Results for value host-15265
Perfdata object: 2012-04-30T11:25:00Z
Disconnecting...
Exiting with status (0)
Network usage at 0Kbps|net=0;7680;10240;0;


Do you have any idea? I am using SDK for Perl 4.1 and check_vmware_1.13. thanks~

The network usage part is not yet working - it is still under development.

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

I've tried multiple ideas, but have come up with nothing.

From my testing, datastores can only be checked for integer values, and only in GB.

Am I missing using a percentage, for example, I've been asked to check for 75% and 85%.

Well, it appears I can only use integers. For large datastores, this works fine.

But one of the datastores I've been asked to monitor is only 2GB, so it appears the only thing I can check is 1GB, which isn't at all what the other department wants or needs.

Checking for 0.75GB would be acceptable, but alas, it's looking for an integer.

Please help!


Dave

This is a limitation of the current code; at the moment, the parameter can only be integer, and for disk space is in GB. I had not anticipated a datastore of a smaller size than 100GB...
Unfortunately this is not such a quick fix because Im using the VMWare tools API option parsing functions. I'll need to look into it... in the meantime, you always have the option to exclude this datastore from the checks, should you wish.

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