Nagios Support Forum

As the subject states, I'm going to possibly be adding 4,000 hosts which will have about 7 services each by default. Some of course will have more. So, at my math, that is 28000 checks being added. I'll be using the bulk cloning tool and adding half at once and then the other half(prod+dev). Oh, and these will be all WMI checks on windows servers.

In my main area I have two workers using mod_gearman, the primary nagios server and another worker. Should I just go ahead and turn up another worker now or will those two workers be enough to add that many checks? Any other items I need to think about?

Thanks!

p.s. It feels so good to be this close to winning the war of what does the Windows monitoring in my company

Couple thing to think about. Before adding all of them with the bulk cloning wizard I would set (or increase) max_input_vars in /etc/php.ini
Also, you will likely need a higher than default max_execution_time then As for the workers, if you have strong hardware, and 5 minute check intervals, I would say that the 2 mod_gearman workers should do it, but might not be a bad idea to have a 3rd on the ready.

If you haven't done so, I would strongly suggest setting up a RAM disk in a large environment like this, it can really reduce the disk I/O which is the one thing that can bring your XI server to it's knees.
http://library.nagios.com/library/produ ... n-nagiosxi

Is it odd that I don't have the max_input_vars in my php.ini?

And what about memory limit?

Also, my two dat files are 14MB right now, how much will they be growing when I add the many more hosts and services? I want to make sure I size the ram disk properly. My current numbers are 552 hosts and 4100 services. Can I simply do math to figure out how large the files will be growing or is it more complicated than that?

You could math it out to predict what usage would be under optimal circumstances. But, I would give it some extra beyond that just in case the added checks slow the server down, causing anything else using the ramdisk (rrdcached/perfdata) to use more spooling space.

abrist wrote:You could math it out to predict what usage would be under optimal circumstances. But, I would give it some extra beyond that just in case the added checks slow the server down, causing anything else using the ramdisk (rrdcached/perfdata) to use more spooling space.

Works for me....just waiting on the answer to my php configuration question now.

Memory limit should be in the php config by default, and would be slightly strange if it is not. The max_input_vars is not standard to have in the config by default and generally needs to be created.

sreinhardt wrote:Memory limit should be in the php config by default, and would be slightly strange if it is not. The max_input_vars is not standard to have in the config by default and generally needs to be created.

Oh, memory limit is in there, I was asking should I increase that from the 128M that it is currently set at.

I would suggest yes. If your server has 8gb of ram or more, you should be alright moving it to 256mb as that would still allow for 40+ LARGE (256mb) sessions.

abrist wrote:I would suggest yes. If your server has 8gb of ram or more, you should be alright moving it to 256mb as that would still allow for 40+ LARGE (256mb) sessions.

Ok, 16GB, so made it 256M for PHP. Also made my ram disk 500M and think it is all working. Does just doing an "apply configuration" after everything is done, make it all take effect?