Monitor Vmware ESX

david35 · Post by **david35** » Wed Dec 03, 2008 8:57 am

Now every thing seems ok except one thing.

Sometimes Cacti is instabled. I have empty graphs.
More often with CPU graph.
But not for all VMs in the same time.
It's a bit weird.
Does anybody has the same pb ?

david35 · Post by **david35** » Wed Dec 03, 2008 10:08 am

Ok, I've found my pb.
My VMs are on a cluster of ESX.
My graph are for only one ESX but my VM have been switch to another.

So if I have 2 ESX, to graph CPU of one VM I have to generate 2 graphs.

YoMarK · Post by **YoMarK** » Wed Dec 03, 2008 1:00 pm

Then don't monitor your VM's with ESX(only use this template to monitor ESX and Service Console), there is no reason to do this.
Just treat them like normal physical machines, and then you could use all the other templates on this forum.

gsaray101 · Post by **gsaray101** » Mon Dec 08, 2008 2:45 pm

I have esx servers 3.5. When I try to walk the enterprise mibs, this is what I get:

D:\cacti\scripts>snmpwalk -v2c -c <password> <IP Address> .1.3.6.1.4.1.6876
SNMPv2-SMI::enterprises.6876.1.1.0 = STRING: "VMware ESX Server"
SNMPv2-SMI::enterprises.6876.1.2.0 = STRING: "3.5.0"
SNMPv2-SMI::enterprises.6876.1.3.0 = OID: SNMPv2-SMI::enterprises.6876.60.1.3.5.0
SNMPv2-SMI::enterprises.6876.1.4.0 = STRING: "113339"
SNMPv2-SMI::enterprises.6876.3.1.1.0 = INTEGER: 4
SNMPv2-SMI::enterprises.6876.3.2.1.0 = INTEGER: 8383808
SNMPv2-SMI::enterprises.6876.3.2.2.0 = INTEGER: 278528
SNMPv2-SMI::enterprises.6876.3.2.3.0 = INTEGER: 8105280
SNMPv2-SMI::enterprises.6876.4.1.1.1.0 = STRING: "yes"

I dont get any cpu, memory or interface statistics. Has anybody experience this problem?

Thanks,

phalek · Post by **phalek** » Mon Dec 08, 2008 3:04 pm

did you try everything i posted here:
http://forums.cacti.net/post-150198.htm ... ht=#150198
?

gsaray101 · Post by **gsaray101** » Mon Dec 08, 2008 3:31 pm

yes, I did everything you recommended, still not getting any data back from the enterprise mib.

phalek · Post by **phalek** » Tue Dec 09, 2008 4:45 am

hm, I had the same problem, mainly because the upgrade to 3.5.0 killed the localhost entry in the hosts file. Other than the stuff i provided I did nothing to fix the problem I had ...

well, except the VM stats are missing as lone as they are not running/poweredOn ...

dmarshall99 · Post by **dmarshall99** » Thu Dec 18, 2008 11:15 pm

I found the snmp info very lacking. I've scripted esxtop. I've written up what I did here:

http://bable.cybermarshall.com/2008/12/ ... ith-cacti/

gsaray101 · Post by **gsaray101** » Wed Feb 04, 2009 5:03 pm

this is not working for me, after I've made the recommenation. cpu, dis, mem and network is not coming back with any entry. However, when I manually walk the oids listed in the scripts, it is coming back with some data. Anybody faced with this?

gsaray101 · Post by **gsaray101** » Thu Feb 05, 2009 12:45 pm

what about disk space for vm in esx? Does anybody has a script to graph vm's disk space?

Wannes · Post by **Wannes** » Fri May 22, 2009 4:19 am

I'm currently monitoring the network traffic of the VM's within my ESX, but sometimes I tend to get "bogus" peaks of 93Gb network traffic. Which ultimately ruin my graphs.
Is there any way to prevent this? And is this an ESX problem or a Cacti problem?

netocool · Post by **netocool** » Wed Jun 03, 2009 8:48 am

this template seen do not support for cacti-0.87d

dose any body update it?

thanks.

Throbulator · Post by **Throbulator** » Fri Jun 12, 2009 3:01 pm

Bumping thread... Does the ESX VMWare Server Template work with Cacti 0.8.7d?

(edited for clarification)

taliz · Post by **taliz** » Sat Jun 20, 2009 10:26 am

scpetand wrote:Hi

I have created a howto to graph Vmware ESX IO by using a nagios plugin developed by op5.
Take a look at http://www.it-slav.net/blogs/?p=262

Regards
Peter Andersson
peter@it-slav.net

check_esx3.pl is a nice script and I've started using it to create graphs of cpu, memory etc. Although it consumes enormous amounts of cpu when running, any idea why?
Doesn't matter which command you run, everything takes 3-7 seconds on my 2x quad 2,5ghz machine, and it consume 100% cpu. This means that I cannot poll/graph as many hosts and vm's as I'd like unfortunately..

EDIT: I'm not a programmer myself, but I ran DProf and dprofpp on it and noticed that it makes more than 550 000 calls to LibXML. Doesn't seem right..

EDIT2: Ran the sample scripts by vmware, and they behave just as bad, so I guess vmware toolkit is the culpŕit. I've posted in the vmware forums as well to see if I can get an answer to this.

EDIT3: http://communities.vmware.com/thread/213515?tstart=0

kenny29 · Post by **kenny29** » Fri Jul 03, 2009 10:43 am

Hi,

I'm currently monitoring the network traffic of the VM's within my ESX and I have exactly the same problem that Wannes. I'm using Cacti 0.8.6j.

I was testing the network bandwith with a ftp transfert lock to 512 kbytes/s and on the graph there was 512 Kbytes/s.

But during the transfert I don't know why on the graph there is a peak to 13 Gbytes/s.... and if I zoom on the graph I can see that after this peak the traffic returned to normal state (512 Kb/s)

So I will ask the same questions "Is there any way to prevent this? And is this an ESX problem or a Cacti problem?

Thanks

Cacti

Monitor Vmware ESX

esx servers

esx servers

How I graph ESX with Cacti

esx 3.5

disk space monitoring for vm servers within esx

Bump

Re: Create Cacti or op5 Statistics graphs with op5 ESX3 plug

Bog peaks network traffic

Who is online