I kinda started this thread when I mentioned that it appeared I knocked down
a couple of machines w/nmap. I posted a summary of my findings a few days ago,
but in a nutshell, I applied a patch, turned off a few services in inetd,
and changed my nmap options FROM/TO:
FROM: -p "list-of-ports" --initial_rtt_timeout 300 --host_timeout 5000
TO: -p "list of ports" --initial_rtt_timeout 500 --host_timeout 15000 -sT
Note that I'm doing this because I prefer my web interface to return
the results PDQ which is more important to me than total completeness.
My guess would be some half-open connections were left open because of
the agressive timeouts and that confused a few machines. There were a
few people that wrote to me saying nmap had clobbered a few "weak" IP
stacks - one person said they had to buy lots of beer to make up for it! ;-)
I'm in no way "disparaging" nmap (GREAT tool!) ... just letting folks
know my experiences. Since I've done the things above, I've hammered the
crap out of a few dozen Solaris & HP-UX hosts and have had no problems.
alek
Received on Mar 21 2000
Intro
