[olug] System monitoring

J.R. Wessels jwplex at gmail.com
Sat Sep 9 21:42:21 UTC 2006

I work with Bill on our monitoring system.  We have an older Nagios system
I'm moving over to Groundwork Opensource which is basically Nagios with a
Mysql backend and web frontend to do configuration/generate reports.

For the monitoring, I have it grabbing info using the Nagios plugins for
HVAC, Windows, Linux, and Netware.  For monitoring drive space on the
Netware systems, we have CIFS shares enabled so I just tell Nagios to check
those shares like it would a Windows box.  Same for the Linux boxes.  On CPU
usage, we don't do that, but there are Nagios pluggins that would let the
monitoring box ssh into the target and check memory and cpu usage.
Novacoast gave a demo on the Groundwork full version (there is a free open
source version, then a paid version with extras) and I think I saw a Windows
plugin to check via RPC call for CPU, memory and service status.  There may
be a free version of that plugin someplace.

Otherwise, most of the Nagios checks we have are just over the network to an
IP address.  The simplest ones are to check if a certain TCP port is up.
The more complex ones like our HVAC checks ask for the humidity and
temperature from the sensor's builtin web service, or pulling a webpage from
a website that requires authentication over ssl.

When you setup a Nagios system, there is a libexec directory in the install
directory.  In that directory are all the plugins that were created, and you
can add your own there.  These are just regular executables that you can run
from the commandline and pass flags to get a feel for how you could use them
in Nagios.  The handiest flag is -h to give how the plugin works.

I recently just setup a VMware virtual machine with Groundwork Opensource
for a friend in Seattle.  Everything was setup for him in advance and all he
as to do is change the root password and set the machine's new ip address
(it is a suse linux vm, so type yast at the command line as root).  Beyond
that all he has to do is in the web interface and doesn't have to know much
about the underlying Linux os.  I have it up on my website for download at:

Root password is sample66 (you'll want to change that of course) and the web
interface login is admin/sample66

You can ignore the zentrack.zip download - that is a web based opensource
helpdesk ticketing system I setup for my friend. You are welcome to it, but
it isn't a monitoring system.  Root password is sample66, web interface
login is Administrator/sample66

If you want to do more advanced virtual machine monitoring (such as with
VMware's Virtual Infrastructure 3 or VMware Server) you could do it 3 ways.
Have the Nagios box SSH into the target machine and have it run a script
that gives either the output OK, CRITICAL or WARNING with perhaps a little
more text after those.  You'd have to write the script to do the checks with
the various vmware commands in the background (like with vmware-cmd or
esxtop) since Nagios doesn't know how to talk those.

Another way would be to use passive checks, were the target machine instead
talks to the Nagios checking machine and gives updates on how things are
doing.  Bad thing about passive checks is if the box submitting the checks
is down, Nagios never receives an update and still thinks things are running
fine.  So a combo of active and passive checks would be recommended.

The final way would be to use Vmware's API to check on the status of virtual
machines. This would be the most powerful and could get into the nitty
gritty of things like how is the vmware cluster doing.

On 9/9/06, Rob Townley <rob.townley at gmail.com> wrote:
> On 9/6/06, Craig Wolf <CJWolf at mpsomaha.org> wrote:
> > So I am working on a project at the office when my boss says that in
> conjunction with our new SAN deployment, he wants to have some system
> monitoring.  He wants to be able to goto a central computer/screen and be
> able to look at the overall health of the network/servers.
> >
> > Other than Big Brother (which Adam has repeatedly tongue lashed me for
> NOT running), what other software do people use??  Not just Linux, has to be
> able to pull info from Windows, Linux, and even traditional NetWare
> (although over the next 6-8 months that will start to diminish as we move to
> OES Linux).
> >
> > So, what is everyone using???  Thanx to all in advance and ask me for
> more info if needed!
> >
> > Craig Wolf
> > Linux Web Server Support
> > Desktop/Network Specialist
> > 402-894-6283
> >
> >
> > _______________________________________________
> > OLUG mailing list
> > OLUG at olug.org
> > http://lists.olug.org/mailman/listinfo/olug
> >
> i don't have much experience with this sort of software, so i did not
> want to comment.  But does BB or Nagios cover Virtual Machines?  i
> didn't think it would matter as long as the guest OS is on the
> supported list of OSs until i read the latest SourceForge newsletter,
> which mentions OpenQRM.org because it covers Xen, VMWare, and QEMU as
> well as Linux, FreeBSD, Solaris and Windows and has plugins for Nagios
> and Webmin.  Does NOT cover Netware.
> The other open source projects would be the glpi-project.org as a
> plug-in to glpi-project.org, but again i have tested them.
> _______________________________________________
> OLUG mailing list
> OLUG at olug.org
> http://lists.olug.org/mailman/listinfo/olug

More information about the OLUG mailing list