NRPE Works From Command Line; "Incorrect command line arguments" from Nagios

Question

I'm trying to setup NRPE with a basic example as a first step towards deploying it. For simplicity, I'm running both NRPE and Nagios locally. NRPE and Nagios both seem to be working, but using check_nrpe from Nagios fails mysteriously. Edit: I'm using version 4.1.1 of Nagios and 2.15 of NRPE.

I think that NRPE is working because I can run check_nrpe manually:

/s/l/n/nagios_instance ❯❯❯ libexec/check_nrpe -H 127.0.0.1 -p 5667 -c check_total_procs      ⏎ 
PROCS CRITICAL: 536 processes | procs=536;150;200;0;

I think that Nagios is working (and supplying correct arguments) because I have written this simple check in python, file named check_hello_world.py that lives in the libexec directory:

#!/usr/bin/env python

import sys
print str(sys.argv)
raise SystemExit, 2

The relevant lines of my nagios config look like this:

define host {
    use                             linux-server
    host_name                       yourhost
    alias                           My first Apache server
    address                         127.0.0.1
}

define command {
    command_name check_procs_nrpe
    command_line $USER1$/check_hello_world.py –H "$HOSTADDRESS$" -p 5667 -c "check_total_procs"
    #command_line $USER1$/check_nrpe –H "$HOSTADDRESS$" -p 5667 -c "check_total_procs"
}


define service {
    use                             generic-service
    host_name                       yourhost
    service_description             Check using the hello world plugin
    check_command                   check_procs_nrpe
    }

As you can see, I define the check_procs_nrpe command with one of two commands, either the simple python script, or the nrpe command to check procs that worked from the command line. When I use the python script, within a couple of minutes I get an email with text:

Additional Info:

[/spare/local/nagios/nagios_instance/libexec/check_hello_world.py, –H, 127.0.0.1, -p, 5667, -c, check_total_procs]

This seems to show that Nagios is working correctly, triggering alerts, sending emails, and correct resolving arguments. When I change the command to use check_nrpe and rerun Nagios, I get an email with text:

Additional Info:

Incorrect command line arguments supplied

I really don't know where to go next. I don't have root on this box and thus cannot easily reconfigure rsyslog to try to get logging information. Happy to provide additional info upon request. Any ideas?

Answer 1

In conclusion, the issue turned out to be that the dash in the '-H' was not actually a dash. It was an en-dash. You can actually see it quite clearly in the text above if you look for it. Unfortunately, Source Code Pro (the font I use in my terminal, and therefore in vim) does not really differentiate dashes and en dashes. I reopened an issue regarding this, in the meanwhile I'm looking for a new font (and adding dash differentiation to my list of criteria).