Distributed SNMP monitoring.

[Mike McClure]

Hi Terry,

> Hello.
>
> Hoping I can get some thoughts on this from those experienced with
> Nagios.
>
> I'm looking to deploy a network monitoring solution primarily to monitor
> host resources such as disk space, processor usage, and so forth via
> SNMP, and also to receive SNMP traps and notify accordingly.  The hosts
> are Open/FreeBSD, Solaris, Linux, and Windows.

Actually, Nagios has agents (NRPE, NCSA, NSClient) for all those platforms that are
arguably easier to use than most of the SNMP-based systems I've seen.  I recommend
that you tunnel the various protocols through stunnel (SSL TCP tunnels), especially
NRPE and NSClient.  Those two do not encrypt their authentication data.

If you must be able to do something based on SNMP traps, then you might have quite a
bit of work ahead of you.  They aren't integrated into Nagios (as is discussed in
the docs), and probably never will be for various reasons.  Nagios is based on
polling.

> I'm in need of a
> solution that supports tiering -- I need an external server to monitor
> exernal devices and an internal server to monitor internal devices, with
> the external server pumping its data to the internal server, making the
> internal server the central/master server.

That's certainly possible with Nagios, however, is this absolutely required? 
There's no way for the internal system to connect to the external devices?

The way I'd implement that, assuming you can't do it all from one box, is have the
secondary system run the Nagios daemon (but not the CGI programs), and pass the data
back with NSCA, MySQL, or scp it and parse the contents into the external command
pipe.  You'd have to keep the configurations on each system in sync, but not check
external devices from the internal device, and vice-versa.  It could be a messy
situation.

>
> I first looked an OpenNMS, but it doesn't have tier support yet.  Then I
> ran across Nagios, which does seem to have tier support, but also seems
> to be geared more towards up/down monitoring than SNMP monitoring.

It depends on what you mean by "SNMP monitoring".  Nagios (and it's specialized
agents) can collect resource statistics, just like with SNMP.  Disk, memory and
processor utilization are all available.  Traps really aren't necessary, since
collection is polled regularly.

However, Nagios is really built to be a polling/alerting system.  There's very
little collection or visualization of statistical information.  For a solution to
that, I'd recommend Cacti, Cricket or MRTG, in that order.  Integration with Nagios
is possible.

>
> So, I'm wondering what those who have used Nagios think of its
> appropriateness (or lack thereof) for what I'm trying to accomplish.
> From the documentation I've read so far, it appears to me that tiered
> host resource monitoring is possible with the NRPE daemon running on
> each monitored host.  Is this accurate?  If so, does this daemon work on
> Open/FreeBSD?  These OS's represent the majority of our hosts.

Yes, NRPE works just fine on xBSD.

>
> Any help/advice on this will be greatly appreciated.
>
> Thanks,
> Terry

-- 
Mike McClure, CCIE # 5125, CISSP # 30232
PNE Services, Inc. -  http://www.pneservices.com
mmcclure at pneservices.com
mobile: 913-636-5590



-------------------------------------------------------
This SF.net email is sponsored by: Get the new Palm Tungsten T 
handheld. Power & Color in a compact size! 
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en