Nagios Support Forum

I have a BPI setup to which I added a hostgroup. Within that HG I marked several hosts as essential members (EM). I also added a service, also as an EM.

Before my eyes, the HG spontaneously disappeared from the BPI. I had both the HG and the service listed. The screen blinked as if updating, and the HG was gone. Before and after shots, below. Placing in-line put them out of order, but . . .

This happened twice. The first time, I put one of the EM hosts in the HG into a critical state (changed its IP to a bogus number). The expected result was that the HG would go critical, and my whole BPI would go critical, or at least warn. Instead, the HG disappeared from the BPI, poof!.

Thinking I, found a bug, I re-added the HG, marked the EM hosts, wrote config. Without me doing anything further, the HG just disappeared again as I was looking at it. WTHeck?

I just caused it to happen a third time. Re-added the disappeared HG, wrote config. Marked the EM's of the HG, wrote config. Watched for several minutes. All was good, nothing vanished.

Then reconfigured one of the EM's to critical (bogus IP). Went back to BPI, expanded the HG, expecting to see the EM therein changed to critical. It was still green, but then -- Blip -- entire HG disappeared from the BPI. I suppose at the moment the BPI updated after the host went critical.

This sounds really odd to me. I tried replicating it, going in with some preconceived notions that turned out to be false. I was concerned that changing the IP address took the object out of the host group it was being referenced and that might be why it was being removed from BPI, but upon trying to replicate this, I found that changing the IP address of a host in a host group just caused my BPI to show that host within the host group under the new IP, but as "Unreachable" in a yellow, non critical status.

Do you happen to have any error logs that correlate with this? What do you see in the "Config Management" tab within BPI when these changes are made?

Finding when I apply config, HG's disappear from all BPI's that I've created. Looking in config management, it reflects those removals.

I can re-add HG to a BPI, and write config. But as soon as I apply config (which of course is done any time a CCM change is made). HG is removed again. So this makes it impossible to include a HG in a BPI.

bpi.log is empty - nothing recorded at all. No record of this deletion of HG's from BPI in nagios.log

Also only this in /usr/local/nagiosxi/var/cmdsubsys.log. Doesn't seem to be an error:

Fri, 25 Oct 2024 10:36:17 -0400 CMDLINE=php /usr/local/nagiosxi/html/includes/components/nagiosbpi/api_tool.php --cmd=syncall

Found how to solve the problem. In the Settings tab, there is are two undocumented options in the Sync on Apply Config section. When I un-checked them, and applied config, added HG's were not removed.

So the documentation states: "The information on the settings screen is self explanatory and does not need duplicating here." Really, Nagios? Neither of those check boxes indicates that HG's added to a BPI (and the work done to tag EM's, set thresholds and priorities) will be summarily and without warning deleted at the next apply config. Why would you not document everything in BPI?