DoN. Nichols wrote:

If there is an existing default route, you will either need to
remove it first. For the above line, that would be done by:

route delete default 10.0.0.50

followed by adding the default route to your ppp interface.

route add default (IP-address of the other end of your PPP interface)

Syntax of the route command varies between Linux and Solaris quite regularly,
to the point that I get bitten in the rear by it pretty frequently at work.
Linux often requires things like an extraneous 'gw' keyword when specifying
default routes, which is annoying to say the least. Plus it can be
distribution specific, which is even worse.

Anyway -- on both systems, there is an /etc/hostname.(interface-name)
file for each interface. In your case, you would need two. One for the
eth0 interface, and one for the ppp0 one. *Assuming that Linux does it
like Solaris and OpenBSD.

This is one area where Linux is pretty reliably different from Solaris. I
don't think I've ever seen that format used on any Linux distributions...
a lot of them now seem to inherit from the Red Hattishness that is putting
stuff under /etc/sysconfig, although I seem to recall that SuSE used to use
a gigantic rc file in /etc someplace. In the end, it can all be changed since
it's based on what the various start scripts launched by init during the
boot process were written to do, which is one of the reasons why Linux
distributions have such variability between them.

I think the one common thing I've seen with everything is how PPP utilities
function, though. It seems like they all simply launch a shell script for
bringing the link up and down, and that script is generally responsible for
switching the default route, plumbing (to use a Solaris term) interfaces,
and possibly even loading kernel modules. I'll have to dig through my
workstation, I might have kppp docs on it already. It'd be a nice break
from dealing with firewalls and VPN tunnels.