add service to ping another node from an existing host

[bongmendoza]

Hello,

I'm still in the process of evaluating Nagios Xi and have a question regarding services. All of our remote hosts have VSAT (gateway) unit, then connected to a W2K box, then another NIC connects to another W2K box. So ====>"VSAT GATEWAY"<=====>(1st NIC)-"1st W2K box"-(2nd NIC)<=====>"2nd W2K box". The 2nd W2K box has a specific address 10.0.250.101. Is there a built in service for the remote host to ping this address?

[bongmendoza]

btw, the host is using the NS client

[jsmurphy]

There is the capability for NSClient to work as a proxy for NRPE commands... not specifically ping though as far as I am aware. So assuming NSClient is installed on both servers then you could run a check from Win2k1 to one of the test scripts on Win2k2 and as long as you get a return you can assume the host is alive.

If you want ping specifically you may have to write your own or see if some one else has written a ping for NSclient.

ninja edit: Google found this - https://www.monitoringexchange.org/inve ... -nsclient-

[bongmendoza]

thanks.. this will help...

Do you have a documention or recommendation on how to best configure nagios xi for monitoring... for example, if I already know the services I need to add, is it best to create a service or service group first before adding my hosts?

[jsmurphy]

That's a very open question, as it will depend largely on your network design as to what makes sense for your configuration but I'll do my best.

First of all this: http://nagios.sourceforge.net/docs/nagi ... n/toc.html bookmark it. It has most of the answers and will help you when you get stuck.

Most of the time my approach tends to be a flow down effect where the hosts are at the bottom, that is to say the least configuration (within reason) is done on the host... that way automating the addition of hosts is easy. If I was to diagram the approach it would go something like this:

Host template ===> Host <===> Host Group(s) <=== Service <=== Service template

This leaves out contacts and contact groups but that part really does change very wildly depending on the deployment. Ultimately you want to create as little configuration as possible in the least complex way. so don't have more templates than necessary, if it's going to change often from service to service or host to host (i.e Parents or sometimes host groups or sometimes contacts) then don't put it in the template configure it on the actual host/service. Use the ability to override template definitions as little as possible because it can become a nightmare in large deployments to work out what inherited from who and why and avoid creating complex configuration chains.

Hopefully this helps somewhat, there's a lot of right ways to do it, spend a little bit of time to diagram a design quickly with pen and paper and it will immensely help you catch any short comings before you get to them .

[bongmendoza]

thank you