[Nagiosplug-help] check_nrpe problem

[Robert Lowe]

Hi,

I'm having trouble starting nrpe via xinetd.  It starts but
quickly fails; trying to run check_nrpe reveals an error
about which there is a FAQ (SSL handshake).  However, I
think I'm seeing something different than what the FAQ
describes.  I can run nrpe on the remote host in daemon
mode, and all is fine.  If xinetd launches it, it fails.
I run xinetd in debug mode and get:

Service defaults
         Instances = 60
         CPS = 30 25
         Logging to syslog. Facility = authpriv, level = info
         Log_on_success flags = HOST PID
         Log_on_failure flags = HOST

Service configuration: nrpe
         id = nrpe
         flags = REUSE IPv4
         socket_type = stream
         Protocol (name,number) = (tcp,6)
         Groups = 0
         Server = /usr/sbin/nrpe
         Server argv = nrpe -i -c /etc/nrpe.cfg
         Only from:  143.44.4.13(NUMERIC)
         Logging to syslog. Facility = authpriv, level = info
         Log_on_success flags = HOST PID
         Log_on_failure flags = HOST USERID

Service configuration: sgi_fam
         id = sgi_fam
         flags = IPv4
         type = RPC UNLISTED
         socket_type = stream
         Protocol (name,number) = (tcp,6)
         Groups = 0
         Bind = localhost
         Server = /usr/bin/fam
         Server argv = fam
         RPC data
                 program number = 391002
                 rpc_version = 2
         Logging to syslog. Facility = authpriv, level = info
         Log_on_success flags = HOST PID
         Log_on_failure flags = HOST

Service configuration: ftp
         id = ftp
         flags = IPv4
         socket_type = stream
         Protocol (name,number) = (tcp,6)
         Nice = 10
         Groups = 0
         Server = /usr/sbin/in.ftpd
         Server argv = in.ftpd -l -a
         Logging to syslog. Facility = authpriv, level = info
         Log_on_success flags = HOST DURATION PID
         Log_on_failure flags = HOST

06/1/30 at 09:45:00: DEBUG: {cnf_start_services} Started service: nrpe
06/1/30 at 09:45:00: ERROR: {activate_rpc} pmap_set failed. service=sgi_fam 
program=391002 version=2
06/1/30 at 09:45:01: DEBUG: {activate_rpc} Registered 0 versions of sgi_fam
06/1/30 at 09:45:01: DEBUG: {cnf_start_services} Started service: ftp
06/1/30 at 09:45:01: DEBUG: {cnf_start_services} mask_max = 7, 
services_started = 2
06/1/30 at 09:45:01: NOTICE: {main} Started working: 2 available services
06/1/30 at 09:45:01: DEBUG: {main_loop} active_services = 2
06/1/30 at 09:45:39: DEBUG: {main_loop} select returned 1
06/1/30 at 09:45:39: DEBUG: {server_start} Starting service nrpe
06/1/30 at 09:45:39: DEBUG: {main_loop} active_services = 2
06/1/30 at 09:45:39: DEBUG: {exec_server} duping 9
06/1/30 at 09:45:39: DEBUG: {main_loop} active_services = 2
06/1/30 at 09:45:39: DEBUG: {main_loop} select returned 1
06/1/30 at 09:45:39: DEBUG: {check_pipe} Got signal 17 (Child exited)
06/1/30 at 09:45:39: DEBUG: {child_exit} waitpid returned = 31672
06/1/30 at 09:45:39: DEBUG: {server_end} nrpe server 31672 exited
06/1/30 at 09:45:39: DEBUG: {child_exit} waitpid returned = -1
06/1/30 at 09:45:39: DEBUG: {main_loop} active_services = 2

...with no indication of a problem in syslog.

My nrpe xinet configuration looks like:

# default: on
# description: NRPE
service nrpe
{
         disable         = no
         flags           = REUSE
         socket_type     = stream
         wait            = no
         user            = root
         server          = /usr/sbin/nrpe
         server_args     = -i -c /etc/nrpe.cfg
         log_on_failure  += USERID
}

I'm running it as 'root' for testing purposes only.

The FAQ covers five cases:

1. Different versions.
    If this were the case, I could not run nrpe in daemon mode.

2. SSL is disabled.
    Again, I'm able to successfully run nrpe in daemon mode, and I
    have not turned off SSL on either end via CLI switches.

3. Incorrect file permissions.
    Running as root.

4. Pseudo-random device files are not readable.
    Running as root.

5. Unallowed address.
    Not a factor... no tcp_wrappers that I can tell, and host.allow/.deny
    look fine to me.  If it were rejected because of this, I'd see some
    evidence of it in syslog.

What else can it be??  The two lines from the xinetd startup concern me,
but I've never seen them before, and don't seem to be able to find out
any information about them:

06/1/30 at 09:45:00: ERROR: {activate_rpc} pmap_set failed. service=sgi_fam 
program=391002 version=2
06/1/30 at 09:45:01: DEBUG: {activate_rpc} Registered 0 versions of sgi_fam

Or am I just overlooking something?  Any clues greatly appreciated!

-Robert