VmBix 1.1.1 released

[sgiaicheca]

zabbix_api is already installed

zabbix-api in /usr/lib/python2.6/site-packages

perhaps python version ?

[tatapoum]

What does it say when you try that in python interactive console?

>>> from zabbix_api import ZabbixAPI
>>> zapi = ZabbixAPI(server = "http://yourzabbixserver/api_jsonrpc.php", path = "")
>>> zapi.login("youruser","yourpassword")
>>> print zapi.api_version()
2.2.3
>>>

[sgiaicheca]

i've uploaded my python version to 2.7.5 and uninstall and reinstall zabbix-api

print zapi.api_version() =>>>> 2.2.1

script stop on VMGROUP => cannot resolve VM group

[tatapoum]

You must define a Zabbix group to put the VMs into. It must exist. And a zabbix template to link the VMs to.

[sgiaicheca]

when you say Zabbix group, it's host groups

[tatapoum]

Yes, have a look at the script, you must define :
- a Zabbix hostgroup to put the discovered VMs into
- a Zabbix template to link the discovered VMs to (for example "Template VmBix VM")
I also define a Zabbix proxy, because it sets my environment. If you don't use it, you can just remove the proxy part from the script.

This script wasn't intended to be public. It's just a basic script used in our environment. Use it as a basis for yours.

[sgiaicheca]

i've already understand that your discovery script is not to be implemented as a public script, but it looks useful to discover VMs

I continue to analyze why i've got this error, as i've configured all parameters for my vmware architecture, in log file i've always same errors

Thanks again for your help

[tatapoum]

what's the error ? Could you post the ouput on pastebin or something ?

[sgiaicheca]

Here is my errror

ERROR - vmbix-discovery[16378]: Cannot resolve VM group

I've tried with hosts_group id instead of Name but same error

HostGroup Name is VMs with Linked Template VmBix VM
vmbix-discovery =>
VMGROUP = "VMs"
VMTEMPLATE = "Template VmBix VM"

[roose]

Monitoring multiple ESX hosts

Hi tatapoum,

If I need to monitor multiple ESX hosts, should I put additional info, e.g. username/password/URLs, for the other hosts in the config file? Or should I run multiple vmbix daemons with different config files for each of the ESX hosts I need to monitor?

I can't seem to find the instructions for this. Thanks.

[tatapoum]

Hi, in fact, vmbix connects to the vCenter. So you need to provide credentials to the vCenter server, with the minimum permissions to query the API.

[roose]

Oh, ok. I will try it then. Thanks.

[Tristan]

Hi, in fact, vmbix connects to the vCenter. So you need to provide credentials to the vCenter server, with the minimum permissions to query the API.

Hello,

But what if i have multiple single esx hosts, not conntected to a. Vcenter?

[waclaw]

Hi,
how to read latency information for datastores per ESXi host? I tried this:

zabbix_get -s 127.0.0.1 -p 12060 -k "esx.counter[<ESXi_host_name>,datastore.totalWriteLatency,<Datastore_name>,900]"

but vmbix get only "No returned value"

Ok, I found solution, instead Datasotre_name should by Datastrore_UUID.

[tatapoum]

You found the solution. You need to specify the datastore ID (like naa.xxxxx). I don't know why, but the aggregated instance (here, the aggregation of all the datastores) returns nothing. So this query doesn't work in my setup :

zabbix_get -s 127.0.0.1 -p 12060 -k "esx.counter[<ESXi_host_name>,datastore.totalWriteLatency,,900]"

If you're interested, I have compiled a 1.1.4 version, which adds methods to get the list of instances for a specific counter. Like :

zabbix_get -s 127.0.0.1 -p 12060 -k "esx.counter.instances[<ESXi_host_name>,datastore.totalLatency]"
105 datastore.totalLatency naa.XXXXXXXXXXXXXXXX
105 datastore.totalLatency naa.YYYYYYYYYYYYYYYY
105 datastore.totalLatency naa.ZZZZZZZZZZZZZZZZ

It also includes better error handling.

I can put it on my download page if needed.