Re: Nessus performance on Linux

Ron Gula wrote:
> I'm surprised you are seeing such a high system load, and not CPU cycles
> shown for nessusd.

The two nessusd processes are both competing for CPU time. They 
continiously add up to 85-95% of CPU time. Which is fine. What I do not 
understand is why my system is indicating to spend 60-80% CPU time on 
system (=kernel). This is definitely caused by the nessusd processes.

> If you are scanning just a few
> hosts and it's taking hours, this is also something that isn't normal.

This could be because I'm scanning across a WAN with 350 rtt times. 
However, if the CPU usage was lower at least I could scan more hosts in 
parallel. See my reply to Doug Nordwall.

> I'd look at local environmental issues like a firewall or IPS running on
> your system.

Unfortunately there are no such things on my system. That would be an 
easy fix.

> the built-in port scanner(s) for Nessus. I'd also kick your max checks
> to something much higher like 20 and see if your scan times are different.

I can't try that right now, but when I did that in the past, the system 
load increased so high that the system became totally unresponsive and 
nessus started showing many false positives.

Thanks for the suggestions.

Sincerely,

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