CHECK_NRPE: Error receiving data from daemon

[cshaffer]

I am setting up a RHEL4 server as a client. check_nrpe runs OK from the server and the client.
From the client: check_nrpe -H localhost returns NRPE V2.12
From the server: check_nrpe -H 10.206.201.25 returns NRPE v2.12
From the client, the check_disk command works OK
check_disk -w 20% -c 10% -p / returns DISK OK etc.
But when I try running the check_disk from the client as a command of check_nrpe an error occurs.
check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /' -n returns CHECK_NRPE: Error receiving data from daemon.
When I check the system log /var/log/messages I see Error: Could not complete SSL handshake. 1
I thought that the -n option in the check_nrpe command told it not to use SSL?

The command definition in /usr/local/nagios/etc/nrpe/common.cfg is command[check_disk]=/usr/local/nagios/libexec/check_disk $ARG1$
The service nrpe definition in /etc/xinetd.d/nrpe contains the line only_from = 127.0.0.1 10.206.201.19 (localhost and nagios server)
In the file /usr/local/nagios/etc/nrpe.cfg
allowed_from=127.0.0.1,10.206.201.19 (localhost and server)
command[check_disk]=/usr/local/nagios/libexec/check_disk -w $ARG1$ -c $ARG2$ -p $ARG3$

Any help would be appreciated.

[tmcdonald]

Do you have dont_blame_nrpe set to 1 in the NRPE config?

EDIT: Link to previous thread: http://support.nagios.com/forum/viewtop ... =6&t=27051

[cshaffer]

Yes. That was the default and I left it alone. By the way, I saw two other users with a similar question on the general discussion list.

[lmiltchev]

Instead of this:

Code: Select all

./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /' -n

you should be running this:

Code: Select all

./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /'

If for some reason, you complied nrpe with no ssl, you could try this instead:

Code: Select all

./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p / -n'

[cshaffer]

When I tried ./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /'
I got a different error message. CHECK_NRPE: Received 0 bytes from daemon. Check the remote server logs for error messages.
I hadn't seen this error message before.
But the system log still had the Error: Could not complete SSL handshake. 1, the same as the other error messages. This seems to be the root problem.
When I tried using the -n option with the check_disk command, ./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p / -n'
I got the Received 0 bytes error again. Could not complete SSL handshake. 1

I think SSL is compiled into nagios. But it is certainly possible that something went wrong when I installed it. Can I reinstall it over the old installation?

[lmiltchev]

Have you restarted xinetd after making changes in the nrpe file? Try:

Code: Select all

service xinetd restart
./check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /'

[cshaffer]

Thanks. I have been restarting xinetd, but to make sure, I did it again and tried the command again. This time the error was Received 0 bytes from daemon. Check the remote server logs for more error messages.

Checked the /var/log/messages file and it was the Error: Could not complete SSL handshake. 1 again.

Since the check_nrpe and the check_disk commands run OK when run standalone, the problem must be with the check_disk command being processed through the check_nrpe command. Is it possible that it is a bug? I am totally guessing here.

[lmiltchev]

Is NRPE running as a standalone daemon? I thought, you were running it under xinetd... What is the output of the following two commands (run on the remote box)?

Code: Select all

ps axuw | grep nrpe
netstat -at | grep nrpe

[cshaffer]

Here is the output from the commands

-bash-3.00# ps axuw | grep nrpe
root 21555 0.0 0.0 5224 644 pts/5 S+ 14:24 0:00 grep nrpe

-bash-3.00# netstat -at | grep nrpe
tcp 0 0 *:nrpe *:* LISTEN

When I set this up on the client, I used an archive called. linux-nrpe-agent-el4.tar.gz. This created a directory called linux-nrpe-agent-el4. In that directory there is a script called fullinstall. It told me that the install had successfully completed. I notice that there is some code in fullinstall that checks to make sure that the redhat version is 5 or above, but it is commented out. Is RedHat EL4 supported as a Nagios client host? The el4 in the tarball's name leads me to think it is. Maybe I should run the install again? Or is there a better way to do the install?

[abrist]

-bash-3.00# netstat -at | grep nrpe
tcp 0 0 *:nrpe *:* LISTEN

Looks like nrpe is running.

The el4 in the tarball's name leads me to think it is.

Yeah, this is that last of the el4 packages.

Lets test nrpe from the nagios XI server:

Code: Select all

/usr/local/nagios/libexec/check_nrpe -H <ip of remote host>

You should see version output.
Now lets alter the nrpe.cfg on the remote host, from:

Code: Select all

command[check_disk]=/usr/local/nagios/libexec/check_disk -w $ARG1$ -c $ARG2$ -p $ARG3$

To:

Code: Select all

command[check_disk]=/usr/local/nagios/libexec/check_disk $ARG1$

Restart nrpe on the remote system.
Looking at your first post, try running the check withouut ssl as the version can be pulled without it:

Code: Select all

check_nrpe -H localhost -c check_disk -a '-w 20% -c 10% -p /'