Network problems

[Timothy Murphy]

I've been trying to move from one home network server to another
(because of hard disk problems with the first).

I've found from this that I don't really understand how networking works,
and I'm writing now to ask for some help with this.

Basically, I have this setup:
  ADSL modem/router <-> server <-> Linksys WiFi router.
I'm running dhcp and shorewall on the server.
The ADSL modem has address 192.168.1.254 .
The computers, printers, etc on the LAN have addresses 192.168.2.* ,
eg the laptop I'm using has address 192.168.2.7 ,
and has default gateway 192.168.2.2 , which is the server.
The default gateway on the server is 192.168.1.254 .
(I am running CentOS on the servers, and Fedora-16 on everything else.)

Now this is my question:
Suppose I want to access the internet, say www.google.com .
Presumably my packets go first to the Linksys router,
then to server, and then on to the ADSL modem/router.

I have the lines
  #INTERFACE SOURCE ADDRESS PROTO PORT(S) IPSEC MARK
  eth0       eth1
in /etc/shorewall/masq on my server, with
  #ZONE INTERFACE BROADCAST OPTIONS
  net   eth0      detect    dhcp,tcpflags,routefilter,nosmurfs,logmartians
  loc   eth1     192.168.2.255
  vpn   tun0     192.168.6.255
in /etc/shorewall/interfaces .

Am I right in thinking that the masq entry causes packets
arriving at the server along the eth1 (192.168.2.*) LAN
to be re-directed along the eth0 (192.168.1.*) interface,
and thence to the ADSL modem?

What has been happening in practice is that when I change server
(with the new server at 192.168.2.5), alter all the relevant addresses
restart shorewall and dhcpd on the new server,
and check "route -n" on all the computers involved
I'm not able to access the internet from my laptop.
In fact I cannot access anything on the eth0 (192.168.1.*) network.

At this point I have a cup of tea, then re-start everything,
re-boot my laptop, etc, and after some time it all starts working.

What I'd really like is to trace packets as they go through the system,
and see how they change.
Or alternatively, read some document which will explain to me
exactly how all the parts of the system fit together.

I really would be most grateful for any advice or suggestions on this.

-- 
Timothy Murphy  
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin