[Nagios-users] NRPE+xinetd error

[Thiago C. <th...@nt...>]

	Hello, everyone.  I have searched for messages regarding a similar
problem to no avail, so I'm posting here.

	My setup is using check_nrpe (2.0b4, SSL enabled) with xinetd to poll
several servers. It works like a charm, but sometimes Nagios starts to
report this error on some services:

CHECK_NRPE: Error - Could not complete SSL handshake.

	The service status goes to CRITICAL, and (I think) that is expected.=20
Trying to figure out what's going on I saw that xinetd starts to output
error messages like this:

Jul 17 10:49:31 localhost xinetd[8700]: FAIL: nrpe service_limit
from=3Da.b.c.d

	That is normal behavior for xinetd when the service in regard reaches a
certain number of processes.  As expected, netstat -anp shows lots
of nrpe connections:

tcp        0      0 d.e.f.g:5666    a.b.c.d:56103   =20
ESTABLISHED 5825/nrpe
(many of these follows)

	So I figure nrpe is not releasing the connection and dying.  At first I
had the REUSE flag on my xinetd, but after removing I still get that
problem sometimes.  Here is my /etc/xinetd.d/nrpe

        service nrpe
        {
                socket_type     =3D stream
                wait            =3D no
                user            =3D nobody
                server          =3D /usr/local/nagios/bin/nrpe
                server_args     =3D -c /usr/local/nagios/etc/nrpe.cfg -i
                log_on_failure  +=3D USERID
                disable         =3D no
                only_from       =3D a.b.c.d
        }

	Next I'll try to recompile nrpe without SSL support, but I wanted to
see if anyone had this problem before re-deploying on a bunch of
servers.  Any thoughts?

	Thanks in advance!

--=20
  Thiago Figueir=F3          Ger=EAncia de Opera=E7=F5es
th...@nt...        Cel.: (21) 8817-3632=20