nrpe server starts too early and fails to bind on IP

[John Plaxco]

When the bind address specified in nrpe.cfg is not available and the binding fails at startup, the startup/init script says that the service was successfully started but really it was not. The server does not stay resident and retry the bind. Therefore, I believe the proper behavior here should be to report a failure.

Steps to reproduce:

apt-get install nagios-nrpe-server
nano /etc/nagios/nrpe.cfg
[change the bind address to one the system doesn't currently have, thus the binding will fail]
service nagios-nrpe-server restart

Expected results:
init script reports [failed] like other services when they fail to start.

Actual results:
init script reports [OK], but service is not running. /var/log/syslog contains the following:

Aug 31 18:24:35 ito nrpe[19683]: Starting up daemon
Aug 31 18:24:35 ito nrpe[19683]: Network server bind failure (99: Cannot assign requested address)

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: nagios-nrpe-server 2.12-4ubuntu1
ProcVersionSignature: Ubuntu 2.6.38-8.42-server 2.6.38.2
Uname: Linux 2.6.38-8-server x86_64
Architecture: amd64
Date: Wed Aug 31 18:29:46 2011
InstallationMedia: Ubuntu-Server 11.04 "Natty Narwhal" - Release amd64 (20110426)
ProcEnviron:
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: nagios-nrpe
UpgradeStatus: No upgrade log present (probably fresh install)

[Simon Déziel (sdeziel)]

I confirmed this on a Lucid container. Setting "server_address=1.2.3.4" and starting/restarting nagios-nrpe-server reports no error but the log says otherwise :

# /etc/init.d/nagios-nrpe-server start
Starting nagios-nrpe: nagios-nrpe.
# echo $?
0
# tail -n2 /var/log/daemon.log
Oct 25 10:54:36 server1030 nrpe[20150]: Starting up daemon
Oct 25 10:54:36 server1030 nrpe[20150]: Network server bind failure (99: Cannot assign requested address)

# lsb_release -rd
Description: Ubuntu 10.04.3 LTS
Release: 10.04

# apt-cache policy nagios-nrpe-server
nagios-nrpe-server:
  Installed: 2.12-4ubuntu1.10.04.1
  Candidate: 2.12-4ubuntu1.10.04.1
  Version table:
 *** 2.12-4ubuntu1.10.04.1 0
        500 http://archive.ubuntu.com/ubuntu/ lucid-updates/main Packages
        100 /var/lib/dpkg/status
     2.12-4ubuntu1 0
        500 http://archive.ubuntu.com/ubuntu/ lucid/main Packages

[Simon Déziel (sdeziel)]

This bug also affects Lucid and Precise and fix/workaround would be greatly appreciated. Maybe converting the initscript to an upstart job would avoid this race ?

[Andrey Konstantinov (andreyinvolute)]

Surprised to see the bug is so old. It only affects people who use server_address directive. We have a Nagios server with ~30 clients connected via OpenVPN and only recently started adding the server_address directive to NRPE configs on clients. All of them will fail to start NRPE on boot since naturally, OpenVPN address isn't up until a while after networking has come up. This happens on both Debian 7 and Ubuntu 12.04. NRPE startup should somehow be moved after networking.

[Bas Couwenberg (sebastic)]

nagios-nrpe (3.0.1-2) includes a systemd service file which should resolve this issue.