Kppp / pppd

[Schotty]

Hello. I have found that some ISP's dont behave well with pppd, and do not get the default gateway set properly. Well, Red Hat did this somehow always thru their software. If I plan to go non-Red Hat, how can I overcome the issue of :

sudo route del default
sudo route
{[me=Schotty]looks for the new IP}[/me]
sudo route add default gw $NEW_IP
{[me=Schotty]hangs up}[/me]
sudo route del dfault
sudo route add default gw 192.168.0.2

Each time I dial up I gotta go thru this. My little brother used to have access and I had him using it to go online, but now -- I am not giving him root access to do route commands :

Any tips? I know Mandrake did the same that Red Hat did. SuSE 8 did not though. Then again it could have been a patch that the aforementioned 2 did have as opposed to SuSE.

Thanks.
Currently I am on a HD install of Knoppix, but Slack 9 did the same thing.

[Schotty]

update -- well I did get somewhere sofar :

[/b]route del default ; route add default ppp0[/b]

seems to work. Then at disconnect I can do a

route del default ; route add default gw 192.168.0.2

Of course I can script it or have kppp do it. I suppose I can see how well this works. If I suid the script (for connect/disconnect), that can take care of the permissions, right? If that is an idiotic approach, what is the correct way? I am good with *nix, but not perfect ;D

[Schotty]

Well I finally got this bad boy fixed. It turns out, after adding a new MAC address to my dhcp server (static IP assignment here), I saw a really nasty snafu.

Now it works fine.

[pbharris]

good job, i stared at that one for a while and had no idea

[Schotty]

Would definitely explain why Slack, Libranet, and Knoppix all didnt want to work right. I still may try to get slack to work, from within a VMWare session or something. I am not sure. Slack 7.1 treated my extremely well back in the day. It took some effort to make the dive to Red Hat. Once done though, I have rarely looked back ;D

As soon as I get some cash, I will be evaluating on MY box the latest Lindows. LindowsOS 3 was nice, and the inprovements in 4 seem worth it now to blow $149 bucks for the whole shebang (this is definitly convert stuff here).

But yeah, note this line in my /etc/dhcpd.conf (Red Hat 9)

subnet 192.168.0.0 netmask 255.255.255.0 {
        #option routers                  192.168.0.2;
        option subnet-mask              255.255.255.0;
        option domain-name              "schotty.net";
        option domain-name-servers       192.168.0.2, 209.153.128.4, 169.207.1.3;
        range 192.168.0.200 192.168.0.245;

And yes, there is a nameserver for the LAN ;D

Hopefully you see why this was a fatal error on my part. And why I asked (too bad nobody responded) for Mandrake and Red Hat dialup users to show me their routing tables -- both ran fine off my server with that option enabled. Them boys had some linux voodoo going on