Nessus performance on Linux

I'm running nessus 3.0.5 on Debian 4.0 with a 2.6.18 kernel. The 
hardware is a Pentium 4M 2.2 GHz with 1GB of RAM. I'm using nessj on 
another system to connect to this nessus scan engine.

max_checks and max_hosts are both set to 2. I've enabled all plugins 
except DoS and safe_checks are off. I use nmap for port scanning, and 
the results are loaded from a gnmap file.

Occasionally the scanning system becomes very unresponsive, system load 
shoots up to around 10 and the CPU is at 0% idle. Today is especially 
bad with the system spending hours with continuously 60% of CPU time 
spent on "system" with peaks of 80%. If I "kill -STOP" the nessus 
processes, the system goes back to 99% idle. Only 800MB of RAM is used, 
and no swapping occurs.

This causes nessus to take hours to scan a single host with only a few 
open ports. Tcpdump shows that the hosts are still being scanned, but at 
a very slow rate.

What could be the reason that my system is spending so much time on 
kernel processes? Is there any tuning I can do to prevent this from 
happening?

Sincerely,

Richard van den Berg
_______________________________________________
Nessus mailing list
Nessus@list.nessus.org
http://mail.nessus.org/mailman/listinfo/nessus