OK, you may remember I wrote a message a couple days ago about a not-so-good RAQ4.

I suspended a suspect account and it was better.
I unlocked it tonight, but with CGI disabled.

Anyyyway, 15 mns ago, the server became HORRIBLY slow.
After 10 mns or so, it's back to normal.
The horrible part is 'top' never show a particular application as using all the CPU.

Well, look at this 'top' header:
CPU states: 3.6% user, 5.7% system, 0.0% nice, 90% idle
Mem: 257636 K av, 255252 K used, 2384 K free, 219048 K shrd, 5264 K buff
Swap: 131532 K av, 60136 K used, 71369 K free, 43268 K cached

Now, for vmstat:
procs memory swap io systemcpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 22 1 86288 2936 5240 23812 3 1 210 86 110 145 17 6 77

Please note that:
-every time % user goes us, % system goes up the same
-with these values the system actually seems faster than when it's about to die, of course

What's causing all this? Could this be CPU overload?
Or CPU overload caused by running out of memory?

Thanks,

The RAQ has sent me an email, because the CPU was heavily loaded (68%).

I never get to see this overload when I run 'top', but I can see it with 'uptime'.
Of course since 'top' doesn't show it, I can't spot any faulty app!