2018-10-16

NOTICE: Major problem with nrpe-3.2.1-6 in EPEL-7

During the summer, I worked on updating nrpe to a newer version and made changes to the systemd startup to match the provided one. Part of this was adding PIDfile so that systemd could send signals and monitor the correct nrpe daemon as there had been bugs where systemctl was unable to restart the daemon.

I tested nrpe-3.2.1-6 on my systems and had no problems, and then put it in epel-testing for a couple of months waiting for some testing. This is where I made a mistake and forgot about it and also I did not thoroughly test nrpe updates from very old versions of nrpe. My tests of updates had been with more recent versions which had a line in the start up for


pid_file = /var/run/nrpe/nrpe.pid

which made sure that my tests worked fine. The daemon started up and it ran without problems, created the file in the correct place etc etc. However if you had a configuration management system with an older template for the file, or had touched your /etc/nagios/nrpe.cfg previously you have problems. yum update will fail to restart the nrpe and other errors will occur.

One fix would be to update the config file to the newer version in the 3.2.x series, but that is not going to work for a lot of people.

I have worked with Andrea Veri to work out a functional change which will allow for systemctl to work properly without needing the pid_file. This is by removing the PIDfile and making the startup a simple versus forking daemon. I have built nrpe-3.2.1-8 and it should show up in epel-testing in the next day or so.

Please, please test this and see if it works. If it really works (aka after 24 hours of an update it is still running, please add karma to it in

https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2018-7f7330f37a 

Thank you.