Cheshire Cat Computing

I have just created a new utility for monitoring VMWare servers.

This is a script that is run every 5 mins, and queries a VMWare server via SNMP. It extracts all details about Guests, HBAs, network interfaces etc, combines this with some cached information, and creates/updates a directory full of RRD databases.

It also creates a MRTG .cfg file intended for use by routers2 to display this information. The .cfg file is never used by MRTG, since this script takes care of data extraction and updating the .rrd files. This file changes as new Guest OSs are created or moved (eg with VMotion), and it uses the routers2 facilities to generate many userdefined summary graphs. It should also work with other frontends like 14all, but not as neatly

So far, it's looking good here. Anyone interested in trying it out? You would need VMWare of course, plus routers2, and a bit of Perl and SNMP knowledge to set it up.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hi!

I have tried to get your vmware_monitor2.pl to work but i fail.
I run a RedHat box with routers2.cgi and it works fine, now I really want to monitor our ESX boxes with routers2.cgi.

When I run the script I get the "snmp lookup of vmids" the nothing, no cfg file is created.
./vmware_monitor2.pl -H xxx1vm01 -C xxxxx -d /var/www/html/mrtg/rrdfiles
(snmp lookup of vmids)

Is there anything I have to do on the ESX, I can run cfgmaker and that works.

Please help.

/Skalman

You need to make sure you have SNMP enabled on your ESX server.

1) Enable SNMP
2) Enable the vmware-snmpd agent as well

Now the vmIDs lookup should work, and you'll start getting .cfg files created and no error messages. However, for added detail, you should also

3) Install the vmware-stats SNMP extension script (including the modifications to snmpd.conf) on the ESX server. This is available from http://www.nagiosexchange.org, just search for the check_esx plugin.

Once you have done this, you will also get readytime stats and more detailed memory stats collected.

If you still don't get anything coming out, the error message should help. Try to snmpwalk your ESX server. Do you get anything at all? If not, SNMP is not set up correctly or you have the wrong community. Try to SNMPwalk the enterprises OID tree, and see if the ESX tree is there. If not, you don't have the vmware-snmpd running.

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hi Steve!

Thanx for your reply, A snmpwalk works.
I started vmware-snmpd and now i also get "table recieved" bur nothing else. I get no error message. Any idea?

/M

Hi.

[root@gbg1log1 vmware]# ./vmware_monitor2.pl -H xxxxvm02 -C xxxxx -d /var/www/html/mrtg/rrdfiles -c /tmp -m vmware.cfg
(snmp lookup of vmids)
Table retrieved

No files are created, please help!
I realy want to get this to work.

Thanx!
/M

I'm away in Oz this week, but I can take a look when I get back.

One thing I notice is that you're running in debug mode (the -d flag) but passing a parameter. IIRC that should be -D to specify directory? I may be wrong, I don't haveanything handy to check on. Run with the -h flag to get the help text and check the syntax to be sure you're using the right options - it may be that you're creating the files in a different place to where you're expecting.

Steve

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

... vmware_monitor2.pl

Hi again, I tried the other script vmware_monitor.pl and it works I dont get disk usage and network graphs, is that because I have´nt installaed the Nagios stuff on the vm-box yet? or what do you get with the nagois stuff?

Any idea why the vmware_monitor2.pl does not work.

/M

vmware_monitor.pl uses just the basic SNMP and gives a more restricted output.
vmware_monitor2.pl uses the SNMP plugin and can graph things like memory split, and ready times.

So, you need the vmware-stats SNMP plugin installed on the ESX server for it to work correctly (though it should fail back)

_________________
Steve Shipway
UNIX Systems, ITSS, University of Auckland, NZ
Woe unto them that rise up early in the morning... -- Isaiah 5:11

Hi Steve!

As I'm not a perl expert I can't really understand your install instructions.
Se my comments in the text.

1. Copy the snmp.conf to /etc/snmp/snmpd.conf on your ESX server, and change
it as required (at least change the system contact, location, and
SNMP community name)
(No problem here)
2. Copy the vmware-stats script to the /etc/snmp directory, permission 755
On the ESX server right?
3. Copy the check_esx2 to the plugins directory. You may need to change the
perl location and the statefile location in the script.
Plugins directory? Where is this usually on a ESX server?
4. Add the definitions in checkcommands.cfg to your checkcommands.cfg file
This file I don't understand where to put. Is this just for Nagios?
5. You can now add services as in the example template!
Nagios again?
6. vmware_monitor2.pl is a script to generate .rrd files and a dummy MRTG .cfg
file that graph lots of things and can be viewed from a MRTG frontend such
as routers2. This script should be called with appropriate params every 5min
if you want to do the graphing. It uses the same agent as the Nagios stuff.
Dont worry if you dont want to use it, it isnt compulsory!
When I run the vmware_monitor2.pl as you se here I get no cfg file, the vmware_monitor.pl works fine.
[root@xxx1log1 vmware]# ./vmware_monitor2.pl -d -H xxx1vm02 -C xxxxx -D /var/www/html/mrtg/rrdfiles -c /tmp -m /etc/mrtg/xxxx/vmware/xxx1vm02test.cfg
(snmp lookup of vmids)
[root@xxx1log1 vmware]#


Please help!
Best
/M