Cheshire Cat Computing

That ones weird; it doesnt help that our VC (and hence API) is slightly old of course. The documentation on the API is seriously sparse, too... it may be that in the latest version, endtime is no longer assumed to be 'now' and needs to be explicitly given?

There is also the slight possibility that your server clocks are out of synch by about 5min and so the previous 'start time' seemed to be in the future. Can you check the clock times on your ESX host, virtual centre and the host running the check_vmware plugin to make sure they are all in synch to the minute or better?

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

This plugin works perfectly ! Congratulations ! I am now checking the whole virtual center, and each hosts individually

Do you know how could I get some perfomance data, for nagios ? I am using pnp4nagios to create graphs, but this plugin does not output performance data. Thanks !

Hi!

I'm trying check_vmware with VC 2.5 (and ESX 3.5) and my services get stuck in "PENDING" state.

In the log it says: "Warning: The check of service [...] looks like it was orphaned (results never came back). I'm scheduling an immediate check of the service...".
And if I try the command (doing copy-paste) of this service in my nagios machine console I get the results.

I would appreciate some help!

Thanks.

The beta version I'm working on now has the date/time fix for the above mentioned problem (ie a workaround for people with badly-synched clocks), plus the Nagios perfdata is output.

Everyone running any monitoring (Nagios, MRTG, etc) should be really, really careful to always have their clocks correctly cynched to a reliable NTP source. Similarly, your ESX servers should always be synching their clocks to NTP. Otherwise, Bad Things (tm) can happen, not just with this plugin but all over.

If your plugin runs OK from the command line, but hangs (gets orphaned) when called from within Nagios, it may well be a permissions problem. Other possibilities:
1. Permissions on the script, config file, etc
2. Permissions on the VMware Perl API files
3. Nagios using embedded perl which cannot access the API correctly - try calling the script without using the embedded perl interpreter

You should also test the script from the command line using the same user that Nagios runs as (in case of permissions problems) and check the PATH settings.

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

Hi Steve,

Looking at the usage of check_vmware I've seen there is no way to pass the "warning" and "critical" thresholds. Am I wrong?

A possible supposition to answer to that is maybe the fact that it uses de VC alarms and we get them in nagios when using the --report=status? For me it's not the same, but is the only possible answer I find..
On the other hand, the check_vmware has a predifined thresholds:

# Default thresholds for Nagios checks
my($WARNSPACE,$CRITSPACE) = (5,3); # in GB
my($WARNCPU,$CRITCPU) = (80,90); # percent max
my($WARNMEM,$CRITMEM) = (80,90); # percent max
my($WARNFAIR,$CRITFAIR) = (90,80); # percent min
my($WARNACTIVE,$CRITACTIVE)=(70,80); # percent max

Is it possible then to change them for each service? Or must they be the same for every kind of service (in case we can only define them in the code)?

Thanks for your help!

And by the way, what happens with the monitoring of Net usage?

In the "usage" of the plugin there is the option, but when using the --generate option the net service is never created. And if you create a service for it it doesn't work (No output returned from plugin).

Thanks very much!

Answers to all the questions:

1. Thresholds. Currently (since there are so many of them) they are hardcoded; soon I'll make them parameterised, which of course makes more sense. Hopefully for this current version which I should be able to release shortly (the NSCA support mentioned later is now working)

2. Net and disk IO monitoring: still to be done.

3. VC alarms come with the 'status' report and are not ideal. The latedt beta version of check_vmware has built-in NSCA support to allow you to pass alerts for individual guests on to the guest object in Nagios via NSCA; this is for the status, mem and CPU checks. The VC alams have a number of problems to do with spike (eg, it goes critical for 5sec then OK again, but nagios has now already received the alarm and alerted).

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

I am new to monitoring ESX 3.0 servers, but not new to nagios.

I am confused how should I have it setup.
1. Where should I have the VMware-Tools-Library ? ( on the Virtual Center, my is windows) where is the link to download?
2. If I call from nagios I will have to do it thru the NRPE agent, correct? (just define a check directly to both .exe?)
3. I would like just info on the ESX server status (CPU,MEM/DISK, and hosts UP/DOWN), the VMs are already being monitored.
4. The perl API SOAP and the others, do I only need it the linux server?

sorry the confusion and thanks for any tip.