Monitor Vmware ESX

nintehol · Post by **nintehol** » Fri Aug 21, 2009 4:25 am

Hello,

i've tried to import the VM Ware Template but i get the message:

Import Results

Cacti has imported the following items:

[fail] m [new]

I use Cacti Version 0.8.7d on Suse Linux 10

Mabye any ideas?

Tiago Cruz · Post by **Tiago Cruz** » Wed Aug 26, 2009 4:48 pm

Where can I find esxiograph.sh ?
I cant found it…

Dr_Boo · Post by **Dr_Boo** » Tue Sep 08, 2009 4:22 am

Hi Everybody, I have upgraded my Debian cacti server to lenny and the cacti Installation from 0.86j to 0.87d. All works fine, except the monitoring of the virtual esx machines. I can monitor the ESX host Maschine, but no virtual server (cpu mem etc.). Result: 0 rows 0 items ?!? Its very important for our business.

sjw7 · Post by **sjw7** » Mon Sep 21, 2009 10:05 am

Has anyone had any luck graphing VMWare vSphere servers? I added a device and it retrieves the system information and in the status column for the graphs it can see there is data returned.

When i run a verbose query though for one of them i get the following.

Code: Select all

+ Running data query [33].
+ Found type = '6 '[script query].
+ Found data query XML file at '/var/www/cacti/resource/script_server/vmware_cpu.xml'
+ XML file parsed ok.
+ Executing script for list of indexes '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: index'
+ Executing script query '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: query index'
+ Found item [vmIndex='No more variables left in this MIB View (It is past the end of the MIB tree)'] index: No more variables left in this MIB View (It is past the end of the MIB tree)
+ Executing script query '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: query displayname'
+ Found item [vmDisplayName='No more variables left in this MIB View (It is past the end of the MIB tree)'] index: No more variables left in this MIB View (It is past the end of the MIB tree)
+ Executing script query '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: query vmid'
+ Found item [cpuvmID='No more variables left in this MIB View (It is past the end of the MIB tree)'] index: No more variables left in this MIB View (It is past the end of the MIB tree)
+ Executing script query '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: query gueststate'
+ Found item [vmGuestState='No more variables left in this MIB View (It is past the end of the MIB tree)'] index: No more variables left in this MIB View (It is past the end of the MIB tree)
+ Executing script query '/usr/bin/php -q /var/www/cacti/scripts/ss_vmware_cpu.php 172.24.200.33 229 2:161:500:1:10:public:::::: query vmstate'
+ Found item [vmState='No more variables left in this MIB View (It is past the end of the MIB tree)'] index: No more variables left in this MIB View (It is past the end of the MIB tree)
+ Found data query XML file at '/var/www/cacti/resource/script_server/vmware_cpu.xml'
+ Found data query XML file at '/var/www/cacti/resource/script_server/vmware_cpu.xml'
+ Found data query XML file at '/var/www/cacti/resource/script_server/vmware_cpu.xml'

I know there have been big changes in SNMP in vSphere but before i start digging i wondered if somone had already crossed this bridge.

Many thanks in advance.

DickyS · Post by **DickyS** » Fri Oct 09, 2009 4:23 am

Hi,

I finally got round to pointing Cacti (0.8.7e) at an ESX (3.5 U3) host and had a go with ymartin59's scripts posted Nov 17 2006. Unfortunately these don't work with the modified SNMP capabilities that arrived with the 0.8.7 code train, resulting in "Success [0 Items, 0 Rows]" for the ESX host's "Associated Data Queries".

Modifiying the XML and PHP to cope with the new requirements of cacti_snmp_walk and cacti_snmp_get was easy and working files for 0.8.7 are attached. Note that the graph templates are untouched and are as they were three years ago - I've yet to decide if the information they provide is reliable, bearing in mind the previous discussion in this thread.

I've tested the modified files with SNMPv3 (supported from 3.5 U2 I believe) authPriv and all is working OK. I did need to extend Cacti's SNMP Timeout for the ESX host (to 2000ms) to make sure the discovery scripts had enough time to execute properly.

Hope this helps someone,

Cheers,
DickyS

ottow · Post by **ottow** » Mon Oct 12, 2009 7:01 am

Has anyone gotten this working under vSphere? Or maybe an alternative exists?

/Otto

vpl · Post by **vpl** » Mon Oct 12, 2009 12:28 pm

Thanks DickyS! I used your template and got everything working without having to modify anything.

damccumb · Post by **damccumb** » Tue Oct 27, 2009 12:24 pm

I am still having trouble with some of the graphing. I am able to graph all of the ESX host queries. I don't seem to be able to graph any of the individual VM's running on the ESX host. I have attached my verbose query, which seems ok to me, and the graph debug that throws the error.

it seems to be that the data query is not making the rrd file. I an not very familiar with cacti and would love any help with troubleshooting this graphing problem thanks..

-David

ottow · Post by **ottow** » Mon Nov 02, 2009 6:06 am

Using this documentation from Dell, http://en.community.dell.com/groups/del ... 27351.aspx I am able to snmpwalk my vSphere server from my Cacti system. However, the output from .1.3.6.1.4.1.6876 is very limited, it seems vSpere is not publishing all the data that the older ESX versions did.

snmpwalk output:
SNMPv2-SMI::enterprises.6876.1.1.0 = STRING: "VMware ESX"
SNMPv2-SMI::enterprises.6876.1.2.0 = STRING: "4.0.0"
SNMPv2-SMI::enterprises.6876.1.4.0 = STRING: "164009"
SNMPv2-SMI::enterprises.6876.3.1.1.0 = Gauge32: 1
SNMPv2-SMI::enterprises.6876.3.2.1.0 = Gauge32: 33552720
SNMPv2-SMI::enterprises.6876.3.2.2.0 = Gauge32: 425984
SNMPv2-SMI::enterprises.6876.3.2.3.0 = Gauge32: 33126736
SNMPv2-SMI::enterprises.6876.3.5.1.0 = INTEGER: 2
SNMPv2-SMI::enterprises.6876.3.5.2.1.2.1 = STRING: "vmhba0"
SNMPv2-SMI::enterprises.6876.3.5.2.1.2.2 = STRING: "vmhba32"
SNMPv2-SMI::enterprises.6876.3.5.2.1.3.1 = INTEGER: 11
SNMPv2-SMI::enterprises.6876.3.5.2.1.3.2 = INTEGER: -1
SNMPv2-SMI::enterprises.6876.3.5.2.1.4.1 = INTEGER: 1
SNMPv2-SMI::enterprises.6876.3.5.2.1.4.2 = INTEGER: 1
SNMPv2-SMI::enterprises.6876.3.5.2.1.5.1 = STRING: "Smart Array E200i"
SNMPv2-SMI::enterprises.6876.3.5.2.1.5.2 = STRING: "iSCSI Software Adapter"
SNMPv2-SMI::enterprises.6876.3.5.2.1.6.1 = STRING: "cciss"
SNMPv2-SMI::enterprises.6876.3.5.2.1.6.2 = STRING: "iscsi_vmk"
SNMPv2-SMI::enterprises.6876.3.5.2.1.7.1 = STRING: "0b:08.0"
SNMPv2-SMI::enterprises.6876.3.5.2.1.7.2 = STRING: "UNKNOWN - NULL PCI DEV IN VMKCTL"

And since the OIDs the script queries does not exist, I am getting the dreaded "No SNMP data returned" in Cacti for all VMware queries.

To me it seems as if VMware has removed the abilty to monitor system performance data via SNMP in vSphere...
But I may be completly wrong of course.

/Otto

DHornby · Post by **DHornby** » Wed Nov 04, 2009 4:48 am

Any updates on this?
I'm running Vsphere 4 and would like to know if I can monitor via snmp.

tchiew · Post by **tchiew** » Fri Nov 27, 2009 2:58 pm

Hi. I'm using the scripts on esx 3.5 u4 and cacti doing a pretty good jod on grahing the data. I have question about the memory usage. Now is graphing service console (not doing much) and I'm looking for graphing and actual esx memory usage. I'm kinda new on this Cacti monitoring. I'm thinking of changing the esx_mem_cos but I cannot find the esx snmp variable on memory. Can anyone help me?

massi · Post by **massi** » Wed Dec 09, 2009 10:20 am

All the data for the host is correctly graphed, however the VM php scripts seems to not work [CentOS5 - Zend 2.1.0 - PHP 5.1.6 - 0.8.7e - ESX 3.5]:

Code: Select all

php /var/www/html/cacti/scripts/ss_vmware_cpu.php ss_vmware_esx_cpu 1.0.0.5 5 2:161:3000:1:10:public:::::: get usage 14 
PHP Notice:  Undefined offset:  1 in /var/www/html/cacti/scripts/ss_vmware_cpu.php on line 26
PHP Notice:  Undefined offset:  2 in /var/www/html/cacti/scripts/ss_vmware_cpu.php on line 27
PHP Notice:  Undefined offset:  3 in /var/www/html/cacti/scripts/ss_vmware_cpu.php on line 28
PHP Notice:  Undefined offset:  4 in /var/www/html/cacti/scripts/ss_vmware_cpu.php on line 29
PHP Notice:  Undefined offset:  5 in /var/www/html/cacti/scripts/ss_vmware_cpu.php on line 47

The array seems not well defined.

computer_guru · Post by **computer_guru** » Thu Dec 10, 2009 1:04 pm

Has anybody found a way for cacti to monitor used space on an ESX VMFS volume?

http://forums.cacti.net/viewtopic.php?p=178536

giancarlob · Post by **giancarlob** » Sat Feb 13, 2010 6:44 pm

I have expanded the script ESXIOGRAPH.PL with cpu, mem and net data plot for VM and VS. It works with ESXi 4U1 and Cacti/Ubuntu
You can found it and new templates here: http://www.to.cnr.it/index.php?option=c ... &Itemid=84

ottow · Post by **ottow** » Mon Feb 15, 2010 9:34 am

giancarlob, I have tried to implement your suggested solution, only to run up against an issue I can't get past; the check_esx3.pl gives me an error message when executing, like this:
ESX3 UNKNOWN - Missing perl module VMware::VIRuntime.

I have installed the VMware Infrastructure (VI) Perl Toolkit (VMware-vSphere-SDK-for-Perl-4.0.0-161974.x86_64.tar.gz), but it seems the script is unable to detect the Perl module...
I am seing this under Fedora Core 9.

Anyone have any suggestions?

/Otto

Cacti

Monitor Vmware ESX

Problem with template importing

Cant monitor virtual machines after upgrade to 0.87d

ymartin59's original scripts modified for Cacti 0.8.7[x]

vSphere

Monitor used space on VMFS

Added cpu, mem and net to esxiograph.pl

Who is online