ZenPacks : VMware Virtual Machines : Comments

RE: VMware Virtual Machines

community@zenoss.org — Fri, 23 Oct 2009 19:02:15 GMT

A couple of helpful instructions here.. there is one pre-install task, and two post-install tasks.

Pre-Install:

1. Create /Devices/Server/ESX Device Class

Post Install:

2. Assign the VirtualMachines collector plug-in to the ESX Device Class.

3. Open the /Devices/Server/ESX zproperty and add zPythonClass = ZenPacks.vmware.VirtualMachines.VMDevice

RE: VMware Virtual Machines

community@zenoss.org — Mon, 25 Jan 2010 18:03:03 GMT

I have also run into a problem, when adding a device in the class /Device/Server/ESX I get this from ZenDisc:

ERROR

zen.ZenDisc

Error: [Failure instance: Traceback (failure with no frames): Products.ZenHub.PBDaemon.RemoteException: Remote exception: exceptions.AttributeError: 'VMDevice' object has no attribute 'primaryAq'Traceback (most recent call last): File "/usr/local/zenoss/zenoss/Products/ZenHub/PBDaemon.py", line 75, in inner return callable(*args, **kw) File "/usr/local/zenoss/zenoss/Products/ZenHub/services/ModelerService.py", line 79, in remote_getDeviceConfig device = device.primaryAq() AttributeError: 'VMDevice' object has no attribute 'primaryAq' ]

Any ideas?

RE: VMware Virtual Machines

community@zenoss.org — Thu, 04 Feb 2010 17:02:53 GMT

Has anyone run into issues where the list of Virtual Machines tab does not populate?

The ZenPack is not able to pull SNMP values for memory statistics.

Any assistance related to these issues would be appreciated.

- Ken

RE: VMware Virtual Machines

community@zenoss.org — Mon, 08 Feb 2010 23:40:56 GMT

The ESX server SNMP sub-agent must be configured. See attached link for details.

Per daveborg98->

"I have found a document that describes how to setup the SNMP in ESX 4, which greatly differs from past versions. Please see http://www.vmware.com/pdf/vsp_4_snmp_config.pdf."

- Ken

RE: VMware Virtual Machines

community@zenoss.org — Fri, 12 Feb 2010 19:06:19 GMT

snmpwalk -v2c -On -c public esx01 .1.3.6.1.4.1.6876 | grep .1280
.1.3.6.1.4.1.6876.2.1.1.2.1280 = STRING: "hostname"
.1.3.6.1.4.1.6876.2.1.1.3.1280 = STRING: "/path/to/host.vmx"
.1.3.6.1.4.1.6876.2.1.1.4.1280 = STRING: "rhel5_64Guest"
.1.3.6.1.4.1.6876.2.1.1.5.1280 = INTEGER: 384 ### MemConfigured
.1.3.6.1.4.1.6876.2.1.1.6.1280 = STRING: "poweredOn"
.1.3.6.1.4.1.6876.2.1.1.7.1280 = INTEGER: 1280 ### VMID
.1.3.6.1.4.1.6876.2.1.1.8.1280 = STRING: "running"
.1.3.6.1.4.1.6876.2.1.1.9.1280 = INTEGER: 1
.1.3.6.1.4.1.6876.2.2.1.3.1280.1000 = STRING: "scsi0"
.1.3.6.1.4.1.6876.2.2.1.4.1280.1000 = STRING: "lsilogic"
.1.3.6.1.4.1.6876.2.3.1.3.1280.2000 = STRING: "scsi0:0"
.1.3.6.1.4.1.6876.2.4.1.3.1280.4000 = STRING: "ethernet0"
.1.3.6.1.4.1.6876.2.4.1.4.1280.4000 = STRING: "10.x.x.x Subnet"
.1.3.6.1.4.1.6876.2.4.1.5.1280.4000 = STRING: "monitor_dev"
.1.3.6.1.4.1.6876.2.4.1.6.1280.4000 = STRING: "true"
.1.3.6.1.4.1.6876.2.4.1.7.1280.4000 = Hex-STRING: 00 50 FF FF FF FF
.1.3.6.1.4.1.6876.2.5.1.3.1280.8000 = STRING: "/dev/fd0"
.1.3.6.1.4.1.6876.2.5.1.4.1280.8000 = STRING: "false"
.1.3.6.1.4.1.6876.2.6.1.3.1280.3002 = STRING: "/usr/lib/vmware/isoimages/linux.iso"
.1.3.6.1.4.1.6876.2.6.1.4.1280.3002 = STRING: "false"

That's an SNMP walk of VirtualMachine ID 1280. We have the latest vSphere 4.0 U1, its configured to proxy the VMWare SNMP agent as jenkinskj posted above.

It seems to me that the Memory Configured counter has moved to ...6876.2.1.1.5.$VMID. The Memory Consumed counter and CPU Utilized counters do not appear to be available for Virtual Machines with ESX 4/vSphere. I am able to pull the Virtual Machine inventory, and run reports with the templates in this ZenPack. However, the CPU/Memory graph templates for VirtualMachines seem to be moot.

If someone knows a workaround, or additional configuration needed on vSphere, please post it to this thread so others can find it.

Update (3/4/2010):

The counters for VM CPU/Memory/Disk/Network can be polled by the vSphere SDK for Perl on vSphere/ESX 4.x.

One of my co-workers came across a thorough example using this type of monitoring with Unnoc (demo here). We are going to temporarily use this, until there is a working Zenpack, or we figure out how to build a working example for Zenoss ourselves. I am pretty new to Zenoss, and just learning the ropes.

RE: VMware Virtual Machines

community@zenoss.org — Fri, 12 Feb 2010 19:10:20 GMT

I am running ESX 4/vSphere. Thanks for the find.

As Eric indicated, help to get this working is greatly appreciated.

- Ken

RE: VMware Virtual Machines

community@zenoss.org — Thu, 11 Mar 2010 20:37:11 GMT

On the ESX host, do i need to install net-snmp for it to work properly? And if so, where did you get the RPM in order to install it?

Thansk.

RE: VMware Virtual Machines

community@zenoss.org — Thu, 11 Mar 2010 20:47:51 GMT

Stephan, your ESX should have net-snmp installed by default. I do not recall being able to select individual packages to install, or not install, when doing an ESX install. You can check with rpm -qa net-snmp, and it should return something like net-snmp-<version>.<distro release>_<patchlevel>

As mentioned previously in this thread, the VM performance template no longer works with ESX 4.

RE: VMware Virtual Machines

community@zenoss.org — Wed, 02 Jun 2010 12:46:05 GMT

Hi there,

1. Create /Devices/Server/ESX Device Class: DONE
2. Zenpack Installation: DONE
3. Assign the VirtualMachines collector plug-in to the ESX Device Class.: DONE

4. Open the /Devices/Server/ESX zproperty and add zPythonClass = ZenPacks.vmware.VirtualMachines.VMDevice : DONE

But I can't see the VMWare tab on my device page

(running Zenoss 2.5.2)

RE: VMware Virtual Machines

community@zenoss.org — Wed, 02 Jun 2010 14:44:20 GMT

Did you restart Zenoss?

ESX server SNMP sub-agent must be configured as indicated in this discussion.

- Ken

RE: VMware Virtual Machines

community@zenoss.org — Thu, 24 Jun 2010 00:02:26 GMT

Would anyone have the correct syntax to test the CPU Utilization Template?

${here/ZenPackManager/packs/ZenPacks.vmware.VirtualMachines/path}/libexec/check_esx2 -N -H ${dev/id} -C ${here/zSnmpCommunity} -v ${here/vmDisplayName} -l CPU -w 90 -c 95 |sed -e 's/.*CPU\susage\sis\s/|cpuUtil=/;s#%##;s/on.*(/cpuSec=/;s#s##;s/average.*$//'

When I run a test on a Virtual Machine I get this error, also:

Argument "SERVER01" isn't numeric in numeric lt (<) at /usr/local/zenoss/zenoss/ZenPacks/ZenPacks.vmware.VirtualMachines-1.2.3-py2.4.egg/ZenPacks/vmware/VirtualMachines/libexec/check_esx2 line 712.

Argument "SERVER01" isn't numeric in numeric le (<=) at /usr/share/perl5/Net/SNMP/Message.pm line 1028.

Unable to retrieve CPU tatistics for SERVER01: Received noSuchName(2) error-status at error-index 1

I want to run it from the command line because I believe it is failing due to a syntax error.

This is my error:

prodState	500
stateChange	2010/06/23 11:06:04.000
facility	unknown
eventClassKey
agent	zenperfsnmp
dedupid	esx1.company\|SERVER01\|/Perf/Snmp\|\|1\|Error reading value for \"SERVER01\" on esx1.company (oid .1.3.6.1.4.1.6876.3.2.4.1.3.640 is bad)
manager	zenoss.company
Location	/Location
ownerid
firstTime	2010/06/23 16:23:06.000
eventClass	/Perf/Snmp
message	Error reading value for "SERVER01" on esx1.company (oid .1.3.6.1.4.1.6876.3.2.4.1.3.640 is bad)
DevicePriority	3
suppid
monitor	localhost
priority	-1
DeviceClass	/Server/ESX
eventState	0
evid	c430caf8-46a2-419b-9f1c-50a70e949ebd
eventClassMapping
component	SERVER01
clearid
DeviceGroups	\|/ESX/ESX1 Virtual Machines
eventGroup	SnmpTest
device	esx1.company
severity	1
count	3
ntevid	0
summary	Error reading value for "SERVER01" on esx1.company (oid .1.3.6.1.4.1.6876.3.2.4.1.3.640 is bad)
eventKey
lastTime	2010/06/23 17:06:04.000
ipAddress	10.1.1.153
Systems	\|/VMWare/ESX

RE: VMware Virtual Machines

community@zenoss.org — Mon, 06 Sep 2010 07:06:51 GMT

Tried this on Zenoss 3 and ESX 4. Thanks to all the previous comments that provided useful tips. Here are my own comments:

On using in Zenoss 3

There is a duplicate of "Virtual Machines" on the left hand pane. This is different from "VirtualMachines" under Components. It's annoying as hell and I haven't figured out how to get rid. Incidentally, the same thing happened when I tested ZenossVirtualHostMonitor + VMwareESXMonitor (which I think is the basis of this Zenpack here).
Under Components | VirtualMachines, each VM guest is on a permanent Down status. How does it check that the guest is up, I can't find it in any of the py scripts?

On monitoring vSphere 4

I deleted the original performance templates for CPU and memory. As per Eric Wolfe above, the OIDs used by this Zenpack are no longer valid.
I created Command-based templates using OP5's check_esx3 plugin for CPU, memory and network. Sample command is: check_esx3 -H ${dev/id} -u user -p pass -N ${here/vmDisplayName} -l cpu -s usage -w 90 -c 95
check_esx3 needs the vSphere SDK for Perl

RE: VMware Virtual Machines

community@zenoss.org — Mon, 18 Oct 2010 09:16:41 GMT

for Python 2.6.2

http://dev.zenoss.org/zenpacks/

ZenPacks.vmware.VirtualMachines-1.2.3-py2.6.egg