How do Nagios clients on Windows communicate?

2006-06-01 00:00:00

After reading through my small write-up on Nagios clients on UNIX you may also be interested in the same story for Windows systems.

Since Nagios was originally written with UNIX systems in mind, it'll be a little bit trickier to get the same amount of information from a Windows box. Luckily there are a few tools available that will help you along the way.

For a quick introduction the Nagios clients, read the write-up linked above. Or pick it from the menu on the left.

---

A quick comparison

	NSClient	NRPEnt	NSClient++	SNMP	SNMP traps	NC_net **
Connection initiation	Srv -> Clnt	Srv -> Clnt	Srv -> Clnt	Srv -> Clnt	Clnt -> Srv	Clnt -> Srv Srv -> Clnt
Security	Password	Password Encryption	Password Encryption * ACL	Access List Password	Access List Password	Encryption ACL
Configuration	On client	On client	On client	On client	On client and On server	On client
Difficulty	Moderate	Moderate	Moderate	Hard	Hard	Moderate
Resource usage ***	unknown	unknown	9MB RAM	unknown	unknown	30MB RAM
Available	Here	Here	Here	Here	Here	Here

*: Thanks to Jeronimo Zucco for pointing out that encryption in NSClient++ only works when used with the NRPE DLL.

**: Thanks to Anthony Montibello for pointing out recent changes to NC_Net, which is now at version 3.

***: Thanks to Kyle Hasegawa for providing me with resource usage infor on the various clients.

---

NSClient

NSClient was originally written to work with Nagios when it was still called NetSaint: a long, long time ago. NSClient only provides you with access to a very small number of system metrics, including those that are usually available through the Windows Performance Tool.

Personally I have no love for this tool since it is quite fidgetty to use. In order to use NSClient on your systems, you will need to do the following.

• Install NSClient on your Windows box.
• Define check commands based on the check_nt plugin.
• Use these check commands for your service checks.

You can now set up your services.cfg in such a way that each remote service is checked like so:

define service{

   host_name remote-host

   service_description D_ROOT

   check_command check_nt_disk!C!85!95

}

Your check command definition would look something like this:

define command {

   command_name check_nt_disk

   command_line /usr/local/nagios/libexec/check_nt -H $HOSTADDRESS$ -p 1248 -v USEDDISKSPACE -l $ARG1$ -w $ARG2$ -c $ARG3

}

NRPEnt

NRPEnt is basically a drop-in replacement for NRPE on Windows. It really does work the same way: on the Nagios server you run check_nrpe and on the Windows side you have plugins to run locally. These plugins can be binaries, Perl scripts, VBScript, .BAT files, whatever.

To set things up, you'll need the same things as with the normal NRPE.

• Download and install NRPEnt on each client,
• Install all check scripts on the system, and
• Configure the NRPE daemon to have a check command for each check you would like to perform.

You can now set up your services.cfg in such a way that each remote service is checked like so:

define service{

   host_name remote-host

   service_description D_ROOT

   check_command check_nrpe!check_root

}

And in nrpent.cfg on the client you would need to include:

command[check_root]=C:\windows\system32\cscript.exe //NoLogo //T:10 c:\nrpe_nt\check_disk.wsf /drive:"c:/" /w:300 /c:100

NSClient++

Due to the limited use provided by NSClient, someone decided to create NSClient++. This piece of software is a lot more useful because it actually combines the functionality of the original NSClient and that of NRPEnt into one Windows daemon.

NSClient++ includes the same security measures as NRPEnt and NSClient, but adds an ACL functionality on top of that.

On the configuration side things are basically the same as with NSClient and NRPEnt. You can use both methods to talk to a client running NSClient++.

SNMP

Unfortunately I haven't yet worked with SNMP on Windows systems, so I can't tell you much about this. I'm sure though that things won't be much different from the UNIX side. So please check the Nagios UNIX clients story for the full details.

To make proper use of monitoring through SNMP you'll need to:

• Install an SNMP daemon / agent on your system,
• put all the check scripts you want to run on the system as well,
• register a private Enterprise ID with IANA to hold your custom objects,
• configure the SNMP daemon in such a way that the results of the check scripts are placed in custom objects, and
• configure basic security on the SNMP daemon.

Ufortunately the check_snmp script that comes with Nagios isn't flexible enough to let you monitor custom SNMP objects in a nice way. This is why I wrote the retrieve_custom_nagios script, which is available from the menu. Your service definition would look like this:

define service{

   host_name remote-host

   service_description D_ROOT

   check_command retrieve_custom_snmp!.1.3.6.1.4.1.6886.4.1.4

}

As I said, I haven't configured a Windows SNMP daemon before, so I really can't tell you what the config would look like. Just look for options similar to "EXEC", which allows you to run a certain command on demand.

Just as is the case with UNIX systems you will need to dig around the MIB files provided to you by Microsoft and you hardware vendors to find the OIDs for interesting metrics. It's not an easy job, but with some luck you'll find a website where someone's already done the hard work for you :)

SNMP traps

SNMP doesn't involve polling alone. SNMP enabled devices can also be configured to automatically send status updates do a so-call trap host. The downside to receiving SNMP traps with Nagios is that it takes quite a lot of work to get them into Nagios :D

To make proper use of monitoring through SNMP traps you'll need to:

• install an SNMP daemon / agent on your system,
• define all the SNMP traps you would like to send on the client,
• install an SNMP trap daemon on your server,
• configure the SNMP trap daemon to tell it what to do with the incoming traps,
• install something that makes a translation between SNMP traps and Nagios service definitions.

There are -many- ways to get the SNMP traps translated for Nagios' purposes, 'cause there's many roads that lead to Rome. Unfortunately none of them are very easy to use.

• SNMPtt, an open source tool.
• EventDB, a database driven piece of software by Netways GMBH.
• My crappy-ass solution, that just cross-references a list of OIDs to a list of Nagios actions.

NC_net

NC_net is another replacement for the original NSClient daemon. It performs the same basic checks, plus a few additional ones, but it is not exentable with your own scripts (like NRPEnt is).

So why run NC_net instead of NSClient++? Because it is capable of sending passive check results to your Nagios server using a send_nsca-alike method. So if you're going all the way in passifying all your service checks, then NC_net is the way to go.

I haven't worked with NC_net yet, so I can't tell you anything about how it works. Too bad :(

UPDATE 31/10/2006:
I was informed by Marlo Bell of the Nagios mailing list that NC_net version 3.x does indeed allow running your own scripts and calling them through the NRPEnt interface! That's great to know, as it does in fact make NC_net the most versatile solution for running Nagios on your Windows.

Also, Anthony Montibello (lead NC_Net dev) tells me that NC_Net 3 requires dotNET 2.0.

kilala.nl tags: tutorial, sysadmin, nagios, windows,

View or add comments (curr. 7)