On Thu Nov 17 2005 at 07:27, [EMAIL PROTECTED] wrote:

> Does it mean that the process is actually killed by 
> some other plugin and not this one ?
> Under what circumstances can this hole be a false positive ?

I got a strange behaviour once. An agent opened two TCP ports. One of
them was a web server, the other one something that Nessus did not
know.
miscflood attacked the unknown service and crashed the agent. So I got
an alert on the web port. I tried to reproduce this (unpublished)
attack against this web server and could not; I wondered what happened
until I noticed that there was _also_ an alert on the unknown service,
from check_ports.nasl. 
"This port was detected as being open by a port scanner but is now closed.
 This service might have been crashed by a port scanner or by a plugin"

For whatever reason, miscflood did not see that it crashed the
service. I don't remember the details, maybe there was a delay...

In short, also the DoS attack are run one at a time, the messages may
be "mixed" under some weird circumstances. The only way to avoid this
would be to slow down those generic attacks, and they are already not
quick :-/
Anyway, if I were you, I'd investigate further: you obviously have
something fragile here. It might even be an exploitable buffer...
_______________________________________________
Nessus mailing list
[email protected]
http://mail.nessus.org/mailman/listinfo/nessus

Reply via email to