Two unrelated questions: NRPE port and a plugin issue

[snapon_admin]

Ok, so I have 2 unrelated questions, but didn't want to create 2 separate posts for them. If I should do that instead, just let me know.

First question: we have an app here that uses port 5666, so we need to change all of our nrpe.cfg files to a different port on our unix hosts. Since it appears that our windows hosts are all using port 12489, I figured I would just change it to that. So I know how to change it in nrpe.cfg on the hosts, but where would I change this on the nagios server? I assume it would need to be changed somewhere, but I have no idea where.

Second question: I tried running the Solaris monitoring wizard to add a service check for a specific process and I got this: NRPE: Command 'check_init_service' not defined. So I checked the nrpe.cfg and sure enough there's no command definition for check_init_service, which is what the wizard puts in for a command. I also checked the plugins folder and there's a lot of plugins in there, but I don't know if any of them are the right one. So the question is, which plugin does the Solaris Monitoring wizard look for when looking for stats on a specific process?

[lmiltchev]

First question: we have an app here that uses port 5666, so we need to change all of our nrpe.cfg files to a different port on our unix hosts. Since it appears that our windows hosts are all using port 12489, I figured I would just change it to that. So I know how to change it in nrpe.cfg on the hosts, but where would I change this on the nagios server? I assume it would need to be changed somewhere, but I have no idea where.

If you are using xinet, you can change the port in the "/etc/xinetd.d/nrpe" file from:

Code: Select all

port            = 5666

to

Code: Select all

port            = <your new port>

Restart xinetd:

Code: Select all

service xinetd restart

Note:
If you are using NRPE as a standalone daemon, you will have to modify the "server_port= " line in the nrpe.cfg file and restart the service.

Your windows hosts are probably NOT using 12489 port with check_nrpe. It is used with check_nt. I would NOT recommend using 12489. Try using a different port.

Second question: I tried running the Solaris monitoring wizard to add a service check for a specific process and I got this: NRPE: Command 'check_init_service' not defined. So I checked the nrpe.cfg and sure enough there's no command definition for check_init_service, which is what the wizard puts in for a command. I also checked the plugins folder and there's a lot of plugins in there, but I don't know if any of them are the right one. So the question is, which plugin does the Solaris Monitoring wizard look for when looking for stats on a specific process?

How did you install the linux agent on the client? Did you use our installer?

http://assets.nagios.com/downloads/nagi ... _Agent.pdf

If you did, you should have a file named "common.cfg". The "check_init_service" is defined in it:

Code: Select all

command[check_init_service]=/opt/nagios/libexec/check_init_service $ARG1$

[snapon_admin]

Ok, I will look at that for the port change, thank you. I'm assuming the "/etc/xinetd.d/nrpe" change is for the Nagios server and the "modify the "server_port= " line in the nrpe.cfg file and restart the service" change is done on the monitored host correct?

Our windows hosts aren't using nrpe, they're using nsclient++ so that's why I thought it was using port 12489. As for the unix hosts using nrpe, I'll pick a different port somewhere up there in the 12,000s for those.

I was brought in to this project after nagios was installed and after nrpe had been setup, but I do not believe it was installed using that process. I think there was some sort of incompatibility with our solaris boxes that wouldn't allow us to use that install process. All of our Unix servers have the nagios agent installed in a /vendor/nagios directory, so maybe that's why it was done that way? I'm not exactly sure, but the consultant we had install it was recently diagnosed with thyroid cancer, so he's helpful but sparingly (for good reason). I'm going through our nrpe.cfg file now trying to modify it so that it will accept inputs from the wizard and I have everything working except for this one check. It appears that he had just hard coded some commands in for Mem, CPU, Root, and Swap checks and then left it at that, so running the wizard didn't work on anything until I went through and modified the nrpe.cfg file. I'm kind of learning on the fly here...heh.

[lmiltchev]

If you don't have the common.cfg, most probably you haven't used our installer. That's OK - you can define this command in the nrpe.cfg:

Code: Select all

command[check_init_service]=/opt/nagios/libexec/check_init_service $ARG1$

Test the command on the Nagios XI server from the CLI:

Code: Select all

/usr/local/nagios/libexec/check_nrpe -H <client ip> -c check_init_service -a <your service>

[snapon_admin]

Yeah, we don't have the common.cfg file. And the problem with adding that to our nrpe.cfg file, is that our plugins aren't in that directory (/opt/nagios/libexec/) and the directory our plugins are in doesn't have that script.

Where our plugins are stored on our Unix servers is here: /opt/csw/libexec/nagios-plugins/

Code: Select all

kenapps11g$ ls /opt/csw/libexec/nagios-plugins/
check_apt                   check_http                  check_mrtgtraf              check_pgsql                 check_time
check_breeze                check_icmp                  check_mysql                 check_ping                  check_udp
check_by_ssh                check_ifoperstatus          check_mysql_query           check_pop                   check_ups
check_clamd                 check_ifstatus              check_nagios                check_procs                 check_users
check_cluster               check_imap                  check_nntp                  check_real                  check_wave
check_dhcp                  check_ircd                  check_nntps                 check_rpc                   check_zone_cpu.sh
check_dig                   check_jabber                check_nrpe                  check_sensors               check_zone_mem.sh
check_disk                  check_ldap                  check_nrpe_1k               check_simap                 check_zpool_fs
check_disk_smb              check_ldaps                 check_nrpe_8k               check_smtp                  negate
check_dns                   check_load                  check_nt                    check_snmp                  pst3
check_dummy                 check_log                   check_ntp                   check_solaris_cpu_usage.sh  urlize
check_file_age              check_mailq                 check_ntp_peer              check_spop                  utils.pm
check_flexlm                check_mem.pl                check_ntp_time              check_ssh                   utils.sh
check_fping                 check_mem.pl.megabyte       check_nwstat                check_ssmtp
check_ftp                   check_mem.pl.orig           check_oracle                check_swap
check_hpjd                  check_mrtg                  check_overcr                check_tcp

[snapon_admin]

Is this the check_init_services script that Nagios is looking for when it creates this service check?

https://github.com/bryanwb/chef-nagios/ ... it_service

[lmiltchev]

It doesn't seem you have this plugin. You can create a new file in your plugins directory:

Code: Select all

cd /opt/csw/libexec/nagios-plugins
vi check_init_service

Copy and paste the following text:

Code: Select all

#!/bin/sh

PROGNAME=`basename $0`

print_usage() {
        echo "Usage: $PROGNAME"
}

print_help() {
        echo ""
        print_usage
        echo ""
        echo "This plugin checks the status of services normally started by the init process."
        echo ""
        support
        exit 0
}


case "$1" in
        --help)
                print_help
                exit 0
                ;;
        -h)
                print_help
                exit 0
                ;;
        *)

                if [ $# -eq 1 ]; then
                        /sbin/service $1 status
                        ret=$?
                        case "$ret" in
                             0)
                                exit $ret
                                ;;
                             *)
                                exit 2
                                ;;
                        esac
                else
						 echo "ERROR: No service name specified on command line"
                        exit 3
                fi
                ;;
esac

Set the permissions:

Code: Select all

chown root:nagios check_init_service
chmod 755 check_init_service
chmod u+s check_init_service

Then you can set up your command in the nrpe.cfg:

Code: Select all

command[check_init_service]=/opt/csw/libexec/nagios-plugins/check_init_service $ARG1$

[snapon_admin]

K, we did that and retried running the script. We got an error when trying to read /sbin/services. Did some looking with one of our Unix admins and it looks like /sbin/services is a linux directory. Our boxes are solaris, and /sbin/services doesn't exist. The Unix equivalent seems to be /etc/services, so we changed the script to reflect that. Running the script (from the client, not the nagios server) again we get this:

Code: Select all

kenapps11g$ ./check_init_service SSH
/dev/fd/3: /etc/services: cannot execute

Looking at the script again, it appears the script is trying to execute something in /sbin/services. I'm not sure what it's trying to execute or if whatever it's trying to execute exists in /etc/services in Solaris. This script appears to be for linux, not unix. Is there anyway to modify this to make it work with unix, or is there another script we should use?

[sreinhardt]

Well, if you just changed the:

Code: Select all

/sbin/service $1 status

to

Code: Select all

/etc/service $1 status

it would be executing /etc/service SSH status, which I think you mentioned is actually a directory? Do you maybe want to use

Code: Select all

/etc/init.d/$1 status

instead? That would be standard linux services without the service command, so it still might be different for solaris, sorry I don't have one around to test on.

[snapon_admin]

/etc/service is not a directory, I miss-read the email that the Unix admin sent me. He says it's not a directory, but it's not an executable file either. This file basically just gives names to ports. If I run a netstat command instead of showing me port 22, it shows me ssh for example. Are there any plugins that can check the output of a 'ps' command and use that to see if a process is running? The only ones I've found that use 'ps' need to read PID numbers or PID files, but that won't work for us because if the service is restarted it gets a different PID number (from what I understand), and some of our processes don't create PID files.