Support forum for Nagios Core, Nagios Plugins, NCPA, NRPE, NSCA, NDOUtils and more. Engage with the community of users including those using the open source solutions.
It looks like when the plugins were configured the configuration process could not figure out how to parse your ping executable's output and is therefore griping that it couln't understand it. What platform are you running your nagios instance on, and what does your ping executable produce, both for success and for failure?
I have OpenSuse 11.4 . Nagios pluging were configured with option --with-ping-command=ping (without this plugs did't want configure). Don't understand second question. Normal ping work ok
You mention that if you didn't supply the "--with-ping-command" option at configuration time the configure would not complete. What error was it giving you?
On a Debian Linux system the configuration reported an auto-detected ping-command of "/bin/ping -n -U -w %d -c %d %s"; you might want to try that (the %d and %s parameters get replaced with numbers and the target address). Different systems can have different command-line syntaxes for "ping"; when in doubt, the manuals for your system will take precedence. In any event, what you want for a basic "ping" check is to send a single echo request with a distance-appropriate timeout period.
Well, there's been some progress. What you may be seeing now is a different way of calling the "check_ping" command for a host check than a service check. What nagios command are you using for the host check, and what is that command's translation into the system call? Compare that to what the service check's translation is and make modifications to the host check command definition appropriately. The other thing you might want to contemplate is to get "fping" and tell the configuration process that you have that facility and prefer it to the basic system-supplied "ping".
The "quick and dirty" would likely be to make the host check simply use the service "ping" check and be done with, but that will likely have pitfalls to it. Recall that in the modern world a "ping" is usually only useful to point out that the target is reachable, not "up and healthy".
This is rather interesting as I've never seen the basic "./configure" process miss the ping syntax so badly. Once you get this sorted you may want to publish your results so others might learn from them.