Apply configuration for a single host

[venkitesh]

hi team,

if i make any changes in a host, i have to do apply configuration which is again re-initializing all the hosts .

Is there any way, instead of apply configuration we can reinitialize only the hosts we made changes than re initializing all the hosts and services.

[Box293]

There is no alternative.

Is there something causing you issues? What version of XI are you using?

[martial86]

Sorry for digging out this old thread, but I need some clarification.

"Apply Configuration" seems to re-initialize all hosts / configuration files and restarts Nagios, if successful. Does this have a potential impact on the monitors (active & passive) when the are sending data in the moment Nagios is restarted? Or does Nagios queue the results and apply them once Nagios is back.

I am concerned, as we want to use Nagios as a self-service tool so customers can create their own monitors, but if we have 200 customers logged in at the same time trying to apply their configurations, we might end up with a Nagios instance that is constantly being restarted.

Thanks in advance!

[lmiltchev]

What is the version of Nagios XI that you are using? After Nagios XI 2014R1.1 the apply configuration process only writes out the "changed" configuration files (not all).

- Changed the Apply Configuration process to only write out changed configuration files -SW, AB

https://assets.nagios.com/downloads/nag ... S-2014.TXT

[martial86]

Thanks for your feedback! We are running the latest version of XI 2.7.

Ok, so only changed configurations are processed - but the Nagios service is still restarted when hitting "Apply Configuration" or not?
I mean it is unlikely that 200 customers change their configs/monitors at the same time, but the more users we have - the higher the likelihood of Nagios being restarted very often.

Any information on the question what happens with data (NRDS / NRPE) that is sent in the moment Nagios is restarted by the Apply Configuration?

[jdalrymple]

For active checks - in process worker threads will be killed off and then respawned after restart. This is because Nagios (by default) remembers after a restart when the last check result came in and knows that it's time to get a new one.

For passive checks - it kind of depends on the passive method in play, but as a genarlization, all of the passive types will get queued somewhere and not get lost.

A couple things to point out:

1) You can use the max_check_attempts and retry_interval directives to avoid false positives for when you have many clients fiddling with your nagios process.

2) If you are offering mufti-tenency in your XI environment, I'd be wary of allowing people from one organization the ability to fiddle with Nagios configs attached to the same XI instance that another organization relies on. Improper configuration can break monitoring for everyone. The proper way to sandbox it would be for each organization to have their own XI instance.