On various cPanel servers I have witnessed this in the past few days:

I am running Top, and load is drifting along at 1.20, or even as low as something like 0.56. Then within ten seconds the load rockets to over 150, a few second more and it is over 400!!! No exaggeration, it the load really accelerates that fast!

One server barely hung on still running top to give me this before it finally died:

01:19:07 up 4 days, 11:02, 1 user, load average: 937.84, 604.82, 342.23


Okay, so I reboot, check the logs, check the email queue for masses of spam being sent, check the tmp directory for anything suspicious, check the secure, messages, etc. logs looking for some hint as to what may have occured, and nada, zip, nothing. The server just returns to normal after a reboot as if nothing has happened.

And of course I am unable to run netstat or anything else at the time when the load is occuring.

What the heck could be going on here? Anyone have any ideas? Anyway to prevent this?

By the way, we are running WHM 10.8.0 cPanel 10.9.0-R118 and we have very recently updated kernals, with plenty of security implementations otherwise.

I'm sorry and happy to say that none of my cP boxes are doing that.

How many websites are on the server? You need a process list at the time that things are going hay wire. Then analyze the core dump. It's critical that this is done!!!

When the load goes high, check the stats of IO for example
- Disk access
- Memory

if you have defective memory and stuff your load will go high without no reason

Yes. This might very well be the reason. Nice input, X.