Nagios 4.4.6 on CentOS 8.2: undefined symbol last_command_check

[ezoltan (Zoltan Erszenyi)]

Hello,

I am running a brand-new Nagios 4.4.6 server installed from RPM, on CentOS 8.2. Installed livestatus using check-mk-raw-1.6.0p14-el8-38.x86_64.rpm as per https://checkmk.com/cms_install_packages_redhat.html. omd version returns the expected output.

Added the broker_module line to nagios.cfg.

Nagios fails to load, error in nagios.log: undefined symbol last_command_check

Forum posts indicate that I must compile livestatus with the --with-nagios4 switch to force using the correct headers. I am not aware that I have this option when installing from RPMs.

Please help me make it work.

Thank you,
Zoltan

[r.sander (Robert Sander)]

Nagios 4.4.6 is not part of check-mk-raw-1.6.0p14-el8-38.x86_64.rpm, have you installed that for yourself?

I do not think that you can use it with the broker module from that package as the module is for Nagios 3.5

Have you read https://checkmk.com/cms_intro.html and are you familiar with the OMD concept of monitoring sites?

Have you run omd create mysite and omd start mysite?

[ezoltan (Zoltan Erszenyi)]

Hi r.sander, wow, that was quick! Didn’t expect a reply so soon

Let me start by saying that I am not a Linux person in that I have some experience but my background lies in Microsoft enterprise and cloud solutions.

I started by installing a brand-new CentOS 8 box using the latest OS ISO image. Then I installed the latest stable Nagios Core version, which is 4.4.6, released on 28th of April 2020, using the Nagios Core install guide published on the Nagios site.

I am trying to install Thruk, following the instructions at Thruk - Installation, where, when it comes to Livestatus, it says :

Refer to h_t_t_p_s_://_checkmk.com/cms_livestatus.html on how to install livestatus onto your monitoring box.

I’ve done exactly that and I arrived to the site, which is not exactly an installation manual. I had no option but to search for the install instructions which I found at https://checkmk.com/cms_install_packages_redhat.html, for my specific environment. I followed the instructions, to the letter, specific for CentOS 8. The instructions make no mention whatsoever of version constraints. or specific requirements. In fact the word Nagios is not mentioned once on the page.

If there are specific product version requirements then I believe they should be in the documentation. If following the documentation to the letter ends in failure then there is either a problem with the product, the documentation, or both. I believe in this instance it is a bit of both as the RPM installer fails to look for Nagios dependencies (if it does any detection at all), and the documentation also fails to mention any version-related constraints. Given the popularity of Nagios, I believe it would deserve at least a mention, even more so that the RPM is compiled for a pre-historic version of it.

I am a bit disappointed that Nagios 4 has been released on the 20th of September 2013, and 7 years on, Livestatus still expects to find Nagios 3.5, which, by any measure in IT parlance, is a dinosaur.

The bottom line is that I need it installed in my environment on a current version of CentOS running a current version of Nagios, using RPMs, and I can’t.

If I must use the source code to compile it specifically for Nagios 4 then that’s OK too, I just need some consistent guidance that actually works. However I’d prefer an RPM if that’s possible at all.

NOTE - due to restrictions on the number of links I can include as a new member, I had to alter some URLs.

Thank you.

[tavanez (Pedro Tavanez)]

Hi @ezoltan,

Are you still struggling ? Basically you’ve come across a fork of Nagios. Check Mk is more user friendly and have a nicer UI.
If you are still trying to keep up with this, from Nagios/Thruk install :
broker_module=/opt/local/livestatus/livestatus.o /tmp/live.sock
Did you check to see if you have the Livestatus in the specified path ? (Most probably isn’t). Since I suspect you’ve also installed Check MK, when you say you followed “https://checkmk.com/cms_install_packages_redhat.html”.
Grep in your system for Livestatus and change the Nagios.cfg path to the right one and check if it runs or not.
As per my opinion, you should give a try to Check MK and see how easy you can get your monitoring up and running further customization is also achievable.