Nagios plugin runs manually, returns error when run automatically

Question

To monitor Mailman mailing lists, I've added the check_mailman_qfiles plugin to an Icinga host.

With only the hostname as an argument, the plugin works fine when run manually:

user@icinga:~$ /usr/lib/nagios/plugins/check_mailman_qfiles mail.example.com 

all normal

However, when run periodically as part of Icinga, the plugin complains that it can't stat "archive", the first file in the qfiles directory:

Notification Type: PROBLEM
Service: Mailman
Host: example.com mail server
Address: 10.1.2.3
State: UNKNOWN
Date/Time: Fri Jun 26 16:09:16 PDT 2015

Additional Info:
**ePN /usr/lib/nagios/plugins/check_mailman_qfiles: Cant stat /usr/local/mailman/qfiles/archive: No such file or directory

Initially I thought this was because I ran the manual command from my shell, while the nagios user has /bin/false as its shell. But I tried using bash for nagios (temporarily), and the command runs fine as the nagios user.

What else would cause the plugin to fail when run automatically?

I've pasted below the mailman command, service, and hostgroup definitions as well as a link to the plugin.

Also, this is for Icinga 1.7.1-7 running on Debian wheezy. Thanks!

Command definition, from /etc/icinga/commands.cfg:

define command{
    command_name    check_mailman_qfiles
    command_line    $USER1$/check_mailman_qfiles $ARG1$
    }

Service definition, from /etc/icinga/objects/services_icinga.cfg:

define service{
    service_description     Mailman
    use                     generic-service
    hostgroup_name          mailman-servers
    is_volatile             0
    check_period            24x7
    max_check_attempts      3
    normal_check_interval   3
    retry_check_interval    1
    contact_groups          admins
    notification_interval   120
    notification_period     24x7
    notification_options    w,u,c,r
    check_command           check_mailman_qfiles
    }

Hostgroup definition, from /etc/icinga/objects/hostgroups_icinga.cfg:

define hostgroup{
    hostgroup_name  mailman-servers ; The name of the hostgroup
    alias           Mailman servers ; Long name of the group
    members         mail.example.com
}

Plugin is at https://exchange.nagios.org/directory/Plugins/Email-and-Groupware/Mailman/check_mailman_qfiles/details

score 2 · Accepted Answer · answered Jul 06 '15 at 21:27

2

Your problem is the embedded perl interpreter (ePN = embedded perl nagios). That plugin may not be (apparently isn't) ePN-safe.

Disable it globally (because it has terrible memory leaks), or disable it just for that script per the documentation.

Also, your check command should be check_mailman_qfiles!mail.example.com to match your manual test.

answered Jul 06 '15 at 21:27

Keith

4,627
14
25

Thanks! Disabling ePN did the trick. One caveat is that this plugin checks _only_ for stale files in Mailman's qfiles directory. It does not verify that Mailman is running. – user8162 Jul 09 '15 at 14:07

Nagios plugin runs manually, returns error when run automatically

1 Answers1