I’ve got Nagios 1.3 running as my main server on Solaris 10. I’m trying to test out the use of the NSCA addon to collect passive checks from a remote Linux box… I’ve installed the NSCA daemon on my Nagios main server and also installed the NSCA addon on the linux server without configuring the NSCA daemon, because I only needed to compile the send_nsca agent for my Linux Distribution. Here are my questions:
How do I configue the send_nsca agent to run as a process or service on the linux box, to where it will continually collect information and send it to the main Nagios server?
How do I configure the send_nsca.cfg file accordingly?
Do I need to install the Nagios plugins on the remote host as well in order for the send_nsca utility to work?
The documentation for this setup is very vauge. I’ve read the README file for NSCA and it does not reference how to configure the send_nsca utility on the remote host. Any advice or additional documentation would be appreciated.
nsca does not collect data. See the nagios box on setting up distributed monitoring for more info.
What you will end up with, is ONE local and several remote hosts running nagios. The local nagios server will be running the nsca daemon as a xinetd service. The remote hosts will be running nagios also, and after each service check, they will execute the ocsp_command as defined in the nagios.cfg file. This command will execute the send_nsca script client, which connects to the nsca daemon, gives it the output of the service check, and disconnects. This daemon, then passes the data on to the Central nagios server. This is how distributed nagios servers works and is detailed in the docs. This is also explained in the README with nsca. It tells you how to setup the daemon, client configs, etc.
Thanks for the clarification. I misunderstood how NSCA works. I thought it worked like other monitoring software in which you have to install an agent on remote hosts. From what you are telling me I have to install not only the send_nsca on remote hosts, but also Nagios itself. That sounds like overkill to have to maintain Nagios installs on every server that needs to be monitored. I think I’ll stick with collecting information via snmp and see how that works. I appreciate your help.
You don’t have to install nagios on remote machines, you can gather the data via a cronjob, or use NC_net if it’s a windows box. But having a mini nagios setup is nice on remote systems, since it handles all of the scheduling and retries for you. A cronjob would only make the check you wish, and if it fails, it’s not going to retry.