When refactoring a check_memory I wrote a few years ago I bumped into the feared

NRPE: Unable to read output

error message on our nagios instance.

When looking for a solution I went through most possible debug steps I could think of and which are nicely described by nagios support but didn't found any solution.

I almost grabbed to some anti depressants when I thought of the thing I always forget about.

SELINUX

When crawling through the audit log it became clear I forgot to configure the proper selinux context type for the new script.

type=PATH msg=audit(1518702310 …

It seems like a very simple job, and in fact it is. But I had an issue when I tried to change this in my Cent OS 6.3 setup.

After some digging on the internet I found out selinux was the blocking factor.

The configuration of the new port has to be done in the lighttpd conf file.

/etc/lighttpd/lighttpd.conf

server.port = 2080

When I changed the config file and restarted the /etc/init.d/lighttpd service I got following error:

(network.c.379) can't bind to port:  2080 Permission denied

I checked that I added the …