I ran a firewall test called LeakTest ver1.1 and it says that my firewall failed or have none.. but this is not true i use stronger firewall script found on TrinityOS web page is there a flaw within this script or should I seek a much stronger firewall script to run Also this runs on a dedicated
redhat 7.2 DSL- router/firewall Ive just about got everything turned off as far as open ports
the p/c is never logged in, I also have set a 16 bit alpha
numeric password for root some advise would be nice I just got this running last week so I'm pretty new to it

does it give any details on what failed, and did you run it from outside of the firewall?

It just said my firewall failed or have none at all I ran the test from my Win 2000 p/c which is behind the
Redhat DSL router /firewall are you familiar with Trinitys web page and the script I utilize ? stronger version

Firewall Penetrated!

LeakTest WAS ABLE to connect to
the main GRC.COM Web Server!

here's exactly what came back after the test

Firewall Penetrated!

LeakTest WAS ABLE to connect to
the main GRC.COM Web Server!

OOOOP's sorry dont normally stutter

Alright now I'm really confused I ran the leak test program from http://grc.com/ and it said my firewall failed or have none
but then I run the Shields Up utility and it says I'm running in complete stealth mode even the ports all reported stealth mode

You could check out this documentation, which has links to some sites to test the security of your firewall.

http://www.linuxsecurity.com/docs/colsfaq.html

Leaktest is nothing more than a simple app that tries to establish a connection to grc's ftp server, receive 13 chars as proof and disconnect.
What it tries to make you aware of is, in the case of a workstation, not server, sw/ppl don't allow inbound connections tru the firewall that aren't part of an established connection, or a connection they didn't initiate to deny access to local services, but don't apply some kind of filter to outbound traffic. Ofcourse this could be usefull in case of Ad/malware trying to "phone home", but *not* if you want to browse, ftp etc, etc in which case you *are* initiating the connection.

In the GRC example ZoneAlarm is able to create a signature for each app that's allowed (outbound) traffic, but this will only work on the box itself, for the apps on that box. On Linux we don't have this kind of checking mechanism, and if we had I'd think it would bog the system down.

Filtering outbound traffic to certain addresses (like oscar.yahoo.com) or ports (like 1214 for KaZaa type sharing) will get you a long way, but some apps allow for using proxies or being able to connect to another port (like 8000, 8080), and adware riding http.

[x] don't press submit twice :-]

try this

ipchains -P input DENY
ipchains -P output DENY
ipchains -P forward DENY

then run the test

I believe that GRC.com is pretty much for winblows boxes. I went ther without a firewall and my machine just refused every attempt to connect with it.
Which I think is what a linux machine should do. I will set up my firewall and report back the differences.

the firewall made it complete stealth mode. Then tried my winblows machine without the firewalls (I run 3 on it) and it took it like a cheap hooker. Kinda makes me feel safer with my linux box.

1 members found this post helpful.