Nagios Support Forum

G'day Guys,

Scenario: 1000+ hosts in Nagios XI.

Just wondering what functions are disabled during the reload configuration process following the addition of a new host? E.g. are notifications still processed? Passive checks received?

Is there a document on this?

Cheers,
John Hobbins
Australia

In the time it takes to run: New checks aren't launched or processed while this is down. Previously submitted events/notifications will still be processed by XI's event queue. Passive checks can only be received if they're coming in with NRDP.

G'day mate,

Thanks for that.

Is it technically possible to create new hosts by adding new records directly into NDOUtils mySQL database?

Got a scenario where new servers are added to the network and we need checks to be performed while hosts are added.

Hosts may be added many times per hour and we can't afford for checks not to be processed while Nagios restarts.

If so, do you have any doco on this?

Cheers, John

Is it technically possible to create new hosts by adding new records directly into NDOUtils mySQL database?

No, ndoutils isn't authoritative on this, it gets all of that information from the Core engine.

Nagios will wait up to 10 seconds for issued checks to come back with results and process them before shutting down. If you need that time to be longer you can extend the wait time in the /etc/init.d/nagios script, there's a timed loop that waits for submitted checks, and then just kills them off if there are any left after 10 seconds. *Most checks do not take that long, so generally you're not actually losing any check results.

We have multiple users, 1 I know of in particular with highly elastic environments, where dozens, if not hundreds of hosts will be added and removed daily from their environments. The resource to look at for this is this document below.
http://assets.nagios.com/downloads/nagi ... gement.pdf

Nagios retains the check schedule across a process restart, so restarting with a new configuration effectively pauses the monitoring process more than it drops it. Does that make sense?

G'day Mike,

I'm guessing that is your name

Thank you for the information you have provided.

A little background to our system.

Nagios XI (shared environment) will receive passive checks from Nagios Core systems located at each customer site.

We have a mechanism to ensure unique hostnames are enforced on the Master (Nagios XI).

Assuming we get customers for our service we could have a scenario where:

1. Many customers create hosts at the same time meaning reloads would occur sequentially.

2. More customers could then add hosts and more reloads would occur.

3. Master system basically is reloading more times than it is processing passive checks.

Do you know of any environments doing something similar?

We could modify our system that schedules host creation and reloads for 0200 in the morning but that would be undesirable.

I hope I haven't confused you!

Regards, John

Yeah, that makes sense actually. If you're using NRDP for the passive checks, they will actually be dropped into the result queue whether or not the monitoring engine is running or not. That might be a good way to ensure that it's always getting the results no matter what. Just make sure there's something keeping track of the monitoring engine status, because I have seen a user with thousands of passive checks with NRDP coming in, and the monitoring engine was down for multiple days and the hard drive filled up :/

Understand that even though the Apply Configuration process might take a while to fully apply with over 1000 hosts, the downtime of the monitoring engine itself is still literally a few seconds, so the check results are almost always being processed unless you have multiple configuration changes coming in per minute, in which case there's a custom script that I can send you which will make sure the new config loads are synchronous and don't overlap.

Legend mate.

Thank you for the advice.

If you can send that script that would be great.

Thumbs up to the team at Nagios. Best support I've encountered in 20 years of being in the IT game.

Thumbs up to the team at Nagios. Best support I've encountered in 20 years of being in the IT game.

Thanks so much from all of us! That's great to hear!



Here's that script I was talking about. This sends the Apply Configuration command to the XI command queue, so even if there are near simultaneous submissions, they'll still happen in a synchronous order.