false critical alerts on all hosts

[rnjie]

am not sure why this is an issue but i get this when i check nagios status
service nagios status
dximonp1.transplace.com nagios[705]: WARNING: RLIMIT_NPROC is 39248, total max estimated processes is 67018! You should increase your limits (ulimit -u, or limits.conf)

the limit is currently set to 100,000

* hard nofile 100000
* soft nofile 100000

root hard nofile 100000
root soft nofile 100000

nagios hard nofile 100000
nagios soft nofile 100000

any idea what this mean and dhow i resolve it

[scottwilkerson]

This isn't an actual issue, it is a calculation done on possible max processes. You should not get anywhere close to this limit.

Our developers are aware of this miscalculation, but it will have no impact on the actual running of nagios

[rnjie]

thank you, but i am still getting high load averages on the server even though i have enough resources and i do not think its a licensing issue because i have a max of 1000 hosts and unlimited services, but each time i add another hosts, these processes become baclup and the server becomes slower

[scottwilkerson]

What exactly are you thinking is high load? Here's my previous responses:

The profile you sent before showed this in the top.txt

Code: Select all

load average: 3.18, 4.21, 4.83

On a system with 16cpu cores this isn't a high load average, there is no waiting taking place at all
If you were regularly sustaining a load above 15, that would be concerning to me

One thing worth pointing out, you said this is a VM, if you have over-provisioned your VM's all with a large amount of CPUs this can make the hypervisor spend a considerable amount of CPU resources just determining which processor to run each operation on.

[rnjie]

it goes up all the way up to 60 sometimes, its concerning

[scottwilkerson]

it goes up all the way up to 60 sometimes, its concerning

when is it at 60 can you run top to see what process is using the CPU?

[scottwilkerson]

and is 60 a 1 min average or a sustained average load?

[rnjie]

average load, and its always mysqld and httpd thats highest

[scottwilkerson]

Can you send us a profile.zip the next time is does this so we can take a look

thanks

[rnjie]

profile before.zip

i have attached two profiles, one before i added network switches and one after i added network switches, you say resource is not an issue but i do not understand why the load sky rockets when i add a new host, if you can help troubleshoot this i will be grateful because i know my system has enough resources. i only added 3 switches and the load went from 0.9 to about 124 immediately after