Well, I've been working with Debian systems for over a year now and I've
been able to get it to dial into my ISP about three times. Each time, I
had to use "route" to make the default gateway to be the ip that I could
get with "ifconfig".
EVERY person I've talked to who has tried getting dial-up ppp going on
Debian has approached it like a heavyweight fighter preparing for a title
fight. They spend a few days just mentally preparing for the ordeal. Then,
when they finally DO get it working, I usually get some jubilant e-mail
from them... something along the lines of "Hey man!!! I actually got PPP
going!!!!!". One of these people was a coworker I see often. He got it all
working about two weeks ago. Then, his roommate had to use the modem so he
unhooked the connection for a while. To this day, the coworker has not been
able to reproduce the ONLY success he's ever had with ppp.
So, today, I resolved to put this nonsense to rest. Being the local Debian
guru, I told this coworker that I would figure out all of the changes that
needed to be done to the base distribution in order to do get everything
working smoothly. Now, having heard that you're supposed to be able to just
edit /etc/ppp.chatscript and make a "ppp_on_boot" file, I figured I'd try
that first.
This is my saga:
I editted /etc/ppp.chatscript to properly log into the dial-in server.
Since /etc/ppp.options_out made reference to /dev/modem, I went to /dev and
make a symlink from "modem" to "ttyS0". (I know I could have edited the
options file, but I wanted to leave the stock config files as pristine as
possible to illustrate how screwed up the whole setup is). I moved the
"no_ppp_on_boot" to "ppp_on_boot". Lastly, I edited "/etc/modules" to
include "serial" and "ppp" (although kerneld would probably load them on
demand anyway... again, I wanted to "roll out the red carpet" for ppp).
Then, I rebooted...
The system started up... started pppd and the modem began dialing. I watched
the whole show by periodically doing "tail /var/log/messages". Chatscript
logged in fine and started ppp on the other end. Then, the system hung up
the phone. /var/log/messages reported "Cannot determine ethernet address
for proxy ARP". This is because "proxyarp" is uncommented /etc/ppp/options.
Why? Beats me. So, to fix this error, I added "-proxyarp" to
/etc/ppp.options_out and gave it a go again....
This time, /var/log/messages recorded that pppd received the local and
remote IP addresses. I will refer to them as <remoteIP> and <localIP>.
(As well as showing that the AppleTalk and IPX drivers had become aware
of the connection, apparently). So, I tried pinging <remoteIP> and not a
single packet came back (although I could see them getting sent out on the
modem by watching the lights). Pinging <localIP> went fine, but didn't use
the modem. Pinging anywhere on the server's network other than <localIP>.
Doing an ifconfig (which paused for about 45 seconds... I've never had that
happen before) showed the proper local and remote IP's for interface ppp0.
Then, "route -n" reported:
(Keep in mind that I'm using "<remoteIP>" and "<localIP>" in place of the
real ones for easier reading...)
Destination Gateway Genmask Flags Met Ref Use Iface
<remoteIP> 0.0.0.0 255.255.255.255 UH 0 0 0 ppp0
127.0.0.1 0.0.0.0 255.0.0.0 U 0 0 1 lo
0.0.0.0 <remoteIP> 0.0.0.0 UG 0 0 0 ppp0
Which *seems* okay. I was a little concerned about not seeing a "default"
in there, but "route add default <remoteIP> metric 1", but that didn't
cause anything new to show up in "route -n". Lastly, after a couple minutes,
the connection will drop, with /var/log/messages reporting:
"Excessive lack of response to LCP echo frames".
Now, before you say that my dial-in server is messed up, let me point out that
I'm able to dial into it with Win95 without incident. The server is a Debian
box running a modem pool.
Also, I tried dialing into a Cisco terminal server and all I got was "Could
not determine local IP address". Again, I don't have this problem with Win95.
So, I have a few questions:
1 - Why is PPP this screwed up? Even if the "ppp_on_boot" thing *did* work,
why is there no mention of it in the instal program? There are a lot of
people out there who install Debian on their home systems and need to
use ppp in order to add/update packages via ftp. Shouldn't a little more
effort be made to make this a little simpler? I don't know of ANYONE
who looks forward to attempting ppp on Linux without a sense of dread.
2 - How can I fix it in the short term? Does anyone know what I can do to
be able to see the remote network?
3 - I think I'm resigned to the fact that this figgin' ppp catastrophe isn't
going to get fixed unless I do it myself. I'm tenatively planning on
writing a set of scripts and ppp.options files to allow people to
easily configure their system as a dial-in server or as a home machine
that dials into an ISP. Does anyone want to offer suggestions, help
code, or help test?
- Joe
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
