Re: [gentoo-user] nagios - check_nrpe missing

[Alan McKinnon <alan.mckinnon@gmail.com>]

On Sat, 24 Nov 2012 23:31:30 -0700
Joseph <syscon780@gmail.com> wrote:

> On 11/24/12 10:18, Alan McKinnon wrote:
> >On Fri, 23 Nov 2012 10:44:59 -0700
> >Joseph <syscon780@gmail.com> wrote:
> >
> >> On 11/23/12 08:40, Alan McKinnon wrote:
> >> >On Thu, 22 Nov 2012 23:04:17 -0700
> >> >Joseph <syscon780@gmail.com> wrote:
> >> >
> >> >> I just installed "nagios" but I can not seem to find: check_nrpe
> >> >> and there is no ebuild: "nagios-nrpe"
> >> >
> >> >net-analyzer/nrpe ?
> >> >
> >> >--
> >> >Alan McKinnon
> >> >alan.mckinnon@gmail.com
> >>
> >> So I think this has to bee installed on the monitoring server and
> >> the client, isn't it?
> >>
> >> How to do you use those plug-ins?  I have installed
> >> "nagios-check_logfiles" plugin and I don't how to use it, is it
> >> suppose to show up on web-interface?
> >
> >I honestly don't know, I don't remember ever installing nrpe on
> >Gentoo. Everywhere I've used it, it's been on some other distro.
> >
> >-- 
> >Alan McKinnon
> >alan.mckinnon@gmail.com
> 
> I gave up on this "nagios" too hard to set it up and/or find any
> decent instructions how to set it up correctly :-)
> 

Nagios itself is quite easy to understand and grasp, but it can be
fiddly to implement it - mostly because you end up with all kinds of
little scripts doing useful work and they can all be very different in
how they expect to be used.

Reading install HOWTO docs for Nagios is a useless activity. You will
almost always end up finding a doc that describes a different version
to what you use, and on a different system. I prefer to just understand
what Nagios does and how it's set up on my machines, then read the
plugin's code, that tells me how to install and use it.

First thing to understand is what Nagios is, and it's not a monitoring
tool! That just happens to be the problem people use it to solve.
Nagios is a state tracking and notification engine. It compares the
state of something now to the state it was in 5 minutes ago and if
things changed it sends a notification.

All the monitoring stuff it can do is actually plugins and little
scripts sitting in various places. To check the state of a ping test,
Nagios runs a ping test script. That's if Nagios does the test itself,
you can also have plugins feed information back into Nagios.

Then there's the tests you want to do that Nagios can't see directly,
like load on a host. You have to log into the host to see that (which
is dangerous). So there's a daemon running locally on the host called
nrpe, and Nagios queries this daemon asking it for results. The results
are the state that Nagios is tracking.

Like I said, script authors like shoving their scripts in any old damn
place and this is a PITA to sort out when it goes wrong. Rather just
decide for yourself where things need to go and put them there yourself.

There are so many things Nagios could do much better than it does, but
the maintainer is highly resistant to adding any features that he
doesn't use himself. So there are many forks around, why don't you try
one of those? Some of them are rather well coded. There's many, and
searching Google for "Nagios forks: will turn up useful projects.




-- 
Alan McKinnon
alan.mckinnon@gmail.com