ESXi 4.0 Hardware and Software Monitoring VMWare

[dima_dm]

OK, i have read almost all of this post, now i have some questions...

I have ESXi 5.1u1(free edition), will the scripts posted here work?

Yes

I also don't have HP environment, can the agents from hp be install on our Intel environment to work this scripts?

You must use agents from your hardware vendor. On other hardware vendor is used other WBEM class. You can rewrite script /etc/zabbix/zabbix_wbem_vmware.pl

In ESXi 5.1 you have SNMP capabilities, but it will only show you your hardware, can you somehow get used recourse true CIM?

I don't use SNMP. I use
WBEM for hardware monitoring
VMWare API for ESXi monitoring

[steveboyson]

Dima,
are you sure to multiply 8192 to get from KBps to bps?
It's around line 340:

	   if ($units eq "KBps")
	    {
	     if (lc($desc) eq "data_transmit_rate" ) {$desc="Network_Data_Transmit_Rate";}
	     if (lc($desc) eq "data_receive_rate" ) {$desc="Network_Data_Receive_Rate";}
            $API{$name.".".$desc.".".$type}=$Value[0] * [COLOR="red"]8192[/COLOR]; # in bps
	    }else{
		 $API{$name.".".$desc.".".$type}=$Value[0]; # Number 
		 }

Or is it from "kilobytes per second" to "bits per second"? Then you're correct.

But according to https://www.zabbix.com/documentation...fig/items/item the unit "Bps" stands for "bytes per second). So I would expect a multiplier of 1024, not 8192.

[steveboyson]

Okay, according to http://pubs.vmware.com/vsphere-51/in...C269C8529.html
the unit sent from ESX is "Megabits per second" for network rate".

Funny thing, all unit labels returned by the Perf API which are measuring a rate are "KBps" or "kiloBytesPerSecond".

But then again the multiplier seems not to be correct?

EDIT: Things get even more weird.
https://www.vmware.com/support/devel..._counters.html states that the rate unit is "kiloBytesPerSecond"

I'm completely confused now.

[dima_dm]

Dima,
are you sure to multiply 8192 to get from KBps to bps?

Yes. This counter in KBps.
counter * 8 ( in bit) * 1000 ( in K) = in bps
Correct to multiply 8000.

[steveboyson]

Yes. This counter in KBps.
counter * 8 ( in bit) * 1000 ( in K) = in bps
Correct to multiply 8000.

Thanks, so 8000 is correct and 8192 is not?

[dima_dm]

Thanks, so 8000 is correct and 8192 is not?

Yes. 8000 is correct.

[steveboyson]

Dima, thank you and sorry for asking again.

We are using vSphere 5.1 (soon to be upgraded to 5.5) and the docs say that the unit is "kiloBytesPerSecond":

http://pubs.vmware.com/vsphere-51/index.jsp?topic=%2Fcom.vmware.wssdk.apiref.doc%2Fnetwork_counters.html

I don't think this unit value changed since 4.1 and will ever change.
Also, a deep inspection in the transmitted counters show also "kiloBytesPerSecond".

I assume, with "kilo" they mean 2^10 = 1024 and not 10^3 = 1000

So I think a multiplier of 1024 is correct to achieve "bytes per second".

What do you think about that?

Thanks for your help!

[kblackwell]

No problems, just a template question

I've gotten this to work. But I have a question about the template.

In the directions, you indicate that the template macro {$HOST} is the IP of the ESXi server.

If that is the case, I have to create a copy this template for each ESXi host, and I won't be able to use the template on multiple hosts.

Is that correct?

Kevin

[pc99096]

you need to define the macro on the server (host) level, not on the template level.
create one host for each esx, assign the same template and define macro separately for each host.

[kblackwell]

pc99096,

Smart Man.

Thanks

[kblackwell]

Are others having small issues?

I'm wondering if anyone else is having similar issues.

I've set both the Server and Agent Timeout values to 30, but periodically if I list the /tmp directory, the vmware file will be of size 0. Not sure if it's a timeout thing or not.

Then the item becomes not supported.
This is a bit of a problem.

[pc99096]

you can always try to run the script manually:
./zabbix_perl_api_vmware.pl VMWARE_IP PASSWORD USERNAME

if you want to measure the time it takes to execute, try
time ./zabbix_perl_api_vmware.pl VMWARE_IP PASSWORD USERNAME

[pc99096]

just found out that after upgrading from esxi 5.1 to 5.5 the perl script stopped working, even after vmware perl sdk upgrade to 5.5

[steveboyson]

Hmmm, it works here without any probs in a 5.5+ latest patchset environment.

Although we are running the scripts on VMware's vMA ("vSphere Management Assistant") commonly used as a Perl framework appliance for monitoring and backup.

[pc99096]

# ./zabbix_perl_api_vmware.pl 10.10.10.10 password root
SOAP request error - possibly a protocol issue: <?xml version="1.0" encoding="UTF-8"?>
<soapenv:Envelope xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<soapenv:Body>
etc
etc



instead of returning "0" and creating /tmp/vmware_api_10.10.10.10.txt, it returns
a huge XML file and no txt file created. i suspect it might be something with https certificates after the upgrade? or probably some perl incompatible modules?


EDIT:
this one seemed to help - downgrade libwww-perl

# cpan GAAS/libwww-perl-5.837.tar.gz