Nagios Support Forum

I’ve been looking at Nagios on and off for a while now without ever implementing anything. However I would now like to get something going that would allow us to monitor our network better. The current scenario is that in our office basement server/comms room we have 4 x Cisco 2950G Series switches which we use as core switches (in a fail over configuration) for our server/firewall connections. These switches then connect to 2 fibre media convertors. These convertors connect to the rest of the building network via a fibre cable which runs up the length of the building terminating on each of the 3 floors of our office. Thus every floor operates independently of any other. On each floor there is a fibre transceiver which connects to a Cisco 2950 Series switch (not G series).

My questions are:
1. Is there any reason not to implement Nagios as a VMware guest over a physical server?
2. What is the best setup to monitor our VMware ESXi hosts and associated VM guests (no ESXi clustering involved, just stand alone)?
3. What is the best setup to monitor the failover status of the core switches as well as bandwidth over the uplinks?
4. What is the best setup to monitor the activity/bandwidth over the fibre transceivers which are connected to each floor’s switch uplink port?
5. I’ve also had a look at Fully Automated Nagios (FAN). Do you think this would be the best way to get started or building something from scratch?
6. If using FAN is not a good idea, which O/S would be the best to run Nagios on (we use Solaris and RHEL mostly)?
7. If using FAN is not a good idea, which add-ons etc. would be useful for visuals (GUI, graphs etc.) and configuration?

I'm really sorry for all the noobie like questions (there will probably be more at some point). Many thanks for your time.

I don't usually point people on the Core forum to our commercial product, but for this situation Nagios XI might be your best option for getting up and running much more quickly. We have a pre-installed VM with a 60 day trial, and there are wizards for network switches and routers, as well as a VmWare wizard. If you'd prefer to stick with a free product we'll be happy to point you in the right direction as well, but I'm thinking you could pretty much pay for the license several times over with the man hours it would take to learn and set these things up with Nagios Core.
http://nagios.com/products/nagiosxi

Many thanks for the reply. I was hoping to use Nagios Core rather than Nagios XI - but I do appreciate what you are saying. Even if I could get the infrastructure monitoring going (switches, firewalls and router) going relatively quickly, I could build the rest around that over a period of time.

Is there any chance my initial questions could be answered and I can take it from there? Obviously if I run into any kind of major show stopper then I'll turn my attention to the XI version.

Sure thing.

1. Is there any reason not to implement Nagios as a VMware guest over a physical server?

The main limitation that shows up with VM's is disk I/O, but this will only really matter if you're running a larger install, like over 10,000+ hosts+services. VM's are a nice solution because of the ability to use snapshots and make easy backups.

2. What is the best setup to monitor our VMware ESXi hosts and associated VM guests (no ESXi clustering involved, just stand alone)?

Look on exchange.nagios.org, I think the plugin we use for XI is check_esx3.pl. Be warned though, it's a bit CPU intensive compared to a simple ping check.

3. What is the best setup to monitor the failover status of the core switches as well as bandwidth over the uplinks?
4. What is the best setup to monitor the activity/bandwidth over the fibre transceivers which are connected to each floor’s switch uplink port?

We use mrtg / SNMP under the hood for routers and switches. I would look at mrtg with Nagios docs for details on how to set this up. Check Nagios Exchange as well, monitoring switches and routers is extremely popular with Nagios.

5. I’ve also had a look at Fully Automated Nagios (FAN). Do you think this would be the best way to get started or building something from scratch?

I haven't tried this out myself, but if it works, I'd say roll with it! ; )

6. If using FAN is not a good idea, which O/S would be the best to run Nagios on (we use Solaris and RHEL mostly)?

I think you're going to have the easiest time with community support on either a RHEL or CentOS system.

7. If using FAN is not a good idea, which add-ons etc. would be useful for visuals (GUI, graphs etc.) and configuration?

This really varies by environment. I recommend a graphing tool (PNP or Nagiosgraph), a Config Manager (either nagiosql or nconf), and since I'm a bit biased because I wrote it, Nagios BPI can do some cool things with Business Process monitoring ; )

Front Ends:
http://www.nagios.org/download/frontends/

Many thanks for your thoughts.