Quantum networking

[Joel Hammer]

We all know that electrons behave in the manner that the experimenter
expects, thus electrons are waves or particles depending on how you
measure their properties.

I had a similar experience today in networking.  This is worth passing
on because I haven't seen this explicitedly stated elsewhere.

To make a long story short, I have a double homed server, one NIC facing the
hostile internet, the other relaxing in  the friendly waters of my home
network.

My firewall is set to cover up most ports in the NIC facing the internet,
while leaving everything open on the other NIC.

I had to make this my mail server, too.  I had a very tough time getting
netscape, running of this server, to download mail from this server. I
got no error messages with netscape, just no mail downloading.

Obviously, since mutt was working fine, I assumed it must be netscape. Wrong.

The box was and is called jhammer6. Previously, this name resolved
to a 192.168.1.xx ip. Now this name resolved to a 68.33.8.xx ip.  So,
since netscape was configured to look for jhammer6, it was sending its
request virtually via the external NIC, whereupon the firewall DENIED
it. (Thus no error message. I believe DENY just lets the packet drop
on the floor.) Changing the incoming mail server to "localhost" solved
my problem. This was only a problem for software running on the mail
server itself. Other computers in the network could contact jhammer6
just fine without interference from the firewall, since their packets
were arriving on the trusted internal NIC.

So, there's my quantum analogy. On the server, if you ask for jhammer6,
you are treated differently than if you ask to speak to localhost. The
packet is trusted or not trusted, depending on who it is asking to
speak to.

There must be an acronym in linux, like RTFM, which expresses the idea
that you just learned a lot of stuff you really didn't want to know but
had to learn anyway so stop complaining and just suck it up.
(JLALOSIRDWTKBHTLASSCAJSIU)

Joel