My ppp vs. Eth0

My ppp vs. Eth0

Post by Wayne Steiner » Thu, 02 Jul 1998 04:00:00



I have RedHat 5.0 installed (for about a week) with a modem and NE*000
ethernet card . I notice that if I have eth0 active, then I can't access ppp
(I'm not sure about the other way around because I am still messing with
Samba). Do I need to set something to make them on different 'ports' or
similar thing?

Thanks in advance.

Wayne

 
 
 

My ppp vs. Eth0

Post by Stev » Thu, 02 Jul 1998 04:00:00




>I have RedHat 5.0 installed (for about a week) with a modem and NE*000
>ethernet card . I notice that if I have eth0 active, then I can't access ppp
>(I'm not sure about the other way around because I am still messing with
>Samba). Do I need to set something to make them on different 'ports' or
>similar thing?

>Thanks in advance.

>Wayne

I ran into the same problem was I was setting up ethernet on my Linux
LAN.  The problem is most likely because you're not setting up your
routing table properly.  On my (Slackware) system, the example route
commands given in rc.inet1 didn't fit the bill, so I had to experiment
a bit.

The *only* commands I use to initialize the ethernet card are:

-----------------------------
/sbin/ifconfig eth0 10.0.0.1 broadcast 10.255.255.255 netmask\
255.0.0.0
/sbin/route add -net 10.0.0.0 device eth0
-----------------------------

(these go in /etc/rc.d/rc.inet1 under Slackware, your mileage may vary
with other distributions or Unices).

Note that I am *not* setting a default gateway for this machine.  When
I start pppd, it uses the -defaultroute option, making it the default
gateway.

A bit of theory: the routing table (type route -n to see yours)
provides a list of 'destination' addresses, usually discrete networks,
and instructions to get to each of those destinations.  For example,
my own routing table has 10.0.0.0 as the first entry, with a listed
Iface of eth0.  The next line has a destination of 127.0.0.0 and an
interface of lo0 (the virtual loopback device).  This means every time
I try to access an IP address beginning with a 10, the system knows to
use ethernet to get to the destination; likewise if I ping or telnet
to 127.0.0.1, the routing table shows that's really just another way
of addressing the same machine (localhost), so it won't waste any time
looking on outside networks.

Most routing tables also have a 'default' entry, which just means "try
this way if you can't find an address explicitly listed elsewhere in
the table'.  Your default route is usually the gateway machine on your
LAN (typically yourip.yourip.yourip.1), but if you're intermittently
running PPP, you probably have a dynamically assigned IP every time
you connect, so you can't predict this...  but you don't have to,
since pppd will add the default route when you connect.  *Unless*
you've already set the eth0 interface as the default route, in which
case outgoing packets will always try (unsuccessfully) to find their
destination on the ethernet segment, and ignore ppp altogether.

Anyway, I'm not sure if I'm doing a good job of explaining this or
not, but it's all pretty easy once you've gotten it to work once or
twice.  :-)  You may also want to read the NET-3 Howto for more
information.

 
 
 

1. "eth0" vs. "eth0:1"

Hello, all.

After several years running a Linux server here at my home office you'd
think I wouldn't be a "newbie," but the fact is that once I get it doing
what I need doing, I never really have to think about it much any more.

So when I come across a basic problem, sometimes I'm just at a loss what to
do, my "knowledge gap" is so wide.

I have an AMD Athlon single-processor box. Eventually I plan to use it to
route, but I haven't gotten there yet. Meanwhile, I have two Ethernet
"adapters," one on the mainboard and one an actual adapter card.

For some reason, one day when I was upgrading the system from Red Hat 8 to
Red Hat 9, I found I had to fiddle somehow with the network configuration.
Not recalling which card was "live," I somehow "created" a "new" device
called "eth0:1", which is LIKE the old "eth0" device I used to use,
but...er...worse.

Now, if the machine reboots, I have to MANUALLY "activate" the ethernet
card. The system is blind to the network until I do this. I cannot figure
out (a) what I did to screw this up, and (b) how to get things back right
again.

Any insights would help.

2. Video Drives IBM 9504 21" Monochrome

3. Linux vs OS2 vs NT vs Win95 vs Multics vs PDP11 vs BSD geeks

4. exportfs: Function not implemented?

5. PPP vs SLIP vs CSLIP

6. make error

7. PPP vs. SLIP vs. PLIP

8. Problems Under Linux - limit and ulimit are ignored

9. PLIP vs. PPP vs. (C)SLIP

10. Perfomance: tar vs ftp vs rsync vs cp vs ?

11. Slackware vs SuSE vs Debian vs Redhat vs ....

12. DOS vs. Windows vs. Mac vs. Unix vs. NS

13. KDE vs. Openlook vs. Xfree86 vs. MetroX vs. CDE