Problem
The nrpe daemon (for Nagios) does not start at all or does not start at system boot.
Environment
UCS 4.2 or newer with systemd.
Solution
Check file permissions for /etc/nagios
and /var/run/nagios
and set them accordingly.
systemctl stop nagios-nrpe-server.service
chown -R root:root /etc/nagios
chown root:nogroup /etc/nagios/nrpe.univention.d
find /etc/nagios -type d -exec chmod 0755 {} \;
find /etc/nagios -type f -exec chmod 0655 {} \;
rm -rf /run/nagios
systemd-tmpfiles --create
echo "systemd-tmpfiles --create">>/etc/rc.local
systemctl stop nagios-nrpe-server.service
Note Instead of using rc.local you should make sure your user called “nagios” is a local user (see below).
Root Cause
For the files and directories under /etc the permissions have changed due to an unknown reason.
For /var/run the root cause is a domain user named “nagios”. At system boot the systemd-tmpfiles could not get the user-id for a user named “nagios” as at this early stage a connection the the LDAP server was not yet possible. So tmp-file creation failed. Once booted the system can easily get the user’s id.