CPU load triggered all linux machines

[reincarne]

Hi,
I have a weird problem with the CPU load check. Once it is triggered on the Nagios machine, almsot all Linux machines are triggered and alerting on high CPU load.
I'm sure there is no real problem, but something related to a sort of bug or bad configurations.


Thanks in advance.

[BanditBBS]

What command do you have setup for checking the CPU_Load on the remote machines? Are you using the check_cpu or whatever? If so, that's checking the local machine only, even though you think it is checking the remote machines. You need to be using nrpe or some other agent to check the remote machines.

[sreinhardt]

Sounds like bandit is right on as usual. Are you using nrpe or another command to check the remote systems? Do all the systems have the same load, which also happens to be the same as the nagios machine?

[reincarne]

Sounds like bandit is right on as usual. Are you using nrpe or another command to check the remote systems? Do all the systems have the same load, which also happens to be the same as the nagios machine?

I will check it.
But from what I see now, I'm checking the load on Nagios machine using the "check_local_load", which make sense as it checks the local cpu load.
As about the rest of the servers, I'm using check_nrpe --> check_load.

[slansing]

Your other service checks for load should not trigger based on another one unless this is set up, its likely coincidence.

[reincarne]

Your other service checks for load should not trigger based on another one unless this is set up, its likely coincidence.

The only reasonable thing I came with which triggers the CPU load on all the Linux machine is something that might be related to a network or sort of service which triggers it.
Anyways, I will check that.

[slansing]

That is what I would suspect, were they going into a state that had nothing to do with CPU usage, say, throwing an out of bounds error?