Installing NRPE

[toleolu]

I installed CentOS 6.5 and the NRPE agent on an old laptop for testing. I ran the NRPE monitoring wizard, but except for ping, all the services error with connection refused.

I got the right IP entered during the agent install, however, when I looked in nrpe.cfg the only allowed host was 127.0.0.1. I added a second line allowed hosts=xxx.xxx.xxx.xxx, but I'm thinking I should have just added a space and the ip on the same line. I get a little nervous editing these config files so I thought I would stop and ask and the pros before I go any further.

Mahalo

[Box293]

When NRPE runs as the xinetd service, the allowed hosts is defined in the following file:

Code: Select all

/etc/xinetd.d/nrpe

It's the line "only_from".


As a test, execute the following command

Code: Select all

/usr/local/nagios/libexec/check_nrpe -H <centos_laptop>

If NRPE is listening and your Nagios Host is allowed, then it should report something like:

Code: Select all

NRPE v2.14

[toleolu]

Yep, get the response back as you indicated. I went ahead and removed that entry from nrpe.cfg.

Still get "Connection refused or timed out".

Do I need to install anything else on the CentOS laptop other than just the agent?

[toleolu]

Interesting development.

Saw a post about the same issue, suggested turning the firewall off, so I did, now I get "Connection refused by host" whereas before, with the firewall on, I was getting "Connection refused or timed out".

Not sure if that means anything, but thought I would mention it.

[toleolu]

This is strange, I saw a post here that gives you the command to get the nrpe pid. netstat -at | grep nrpe

When I run that from the laptop, I don't get a pid, I get:

tcp 0 0 *:nrpe *:* LISTEN

Is anything in there the pid?

[Box293]

NRPE is clearly listening, we just need to work out why the checks aren't

Can you attach your nrpe.cfg file and also show your test command from the CLI.

[toleolu]

Thanks, here's the command:

[root@hhscnagios ~]# /usr/local/nagios/libexec/check_nrpe -H 10.5.190.51 -t 600 -p 15677 -c check_load
Connection refused by host

Just FYI, Ping does work.

I double checked with our network team and we don't block any ports on the internal network, but while troubleshooting, I turned the firewall off on both the Nagios XI server and the laptop, but that didn't fix the problem.

The config wizard prompts for SSL, says it's set by default, I tried it both ways but neither worked either. I suspect if that were the problem though I would get a different error message.

Nrpe.cfg is attached.

Mahalo

[abrist]

You are specifying port 15677 in your check_nrpe command, but the nrpe.cfg you sent uses port 5666.
1) Are you using xinetd?
1b) If so, did you configure it to use port 15677?
2) Lets run a port scan on the remote nrpe host from the XI server:

Code: Select all

nmap <remote host ip> -p 5666,15677 

[toleolu]

I didn't do anything with the ports, that port 15677 is in check_nrpe command. I ran nmap like you suggested and it shows port 5666 open and 15677 closed so I changed the port number in the command to 5666. Now it connects, but it looks like I have a plugin issue. I ran the standard nrpe monitoring wizard and just selected load, users, and processes, just the defaults that come up in the wizard and I get:

Load says: Status Unknown check_load: Could not parse arguments.
Usage:check_load [-r] -w WLOAD1,WLOAD5,WLOAD15 -c CLOAD1,CLOAD5,CLOAD15

Users says: Status Unknown (No output returned from plugin)

Processes says: Status Critical NRPE:Command 'check_total_procs' not defined.

To back up here, all I originally set out to do what test installing the client for your Linux servers. So with that in mind, is this best way to install a Linux client, or is there a better way to set up monitoring on a Linux server?

Mahalo;

[abrist]

The best method is to install our linux agent on the linux host (you can download the agent tarball/install from the second page of the linux wizard).
http://assets.nagios.com/downloads/nagi ... _Agent.pdf