[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Automatic ISP failover if connection fails



Hello,

Albert Ulmer:
> On Fri, 08 Oct 2004 14:41:00 +0200, Richard Verwayen <holle@ackw.de> wrote:
> > Hello list,
> > 
> > I've been looking after an ISP failover solution for my sarge box for a
> > while. My Internet sonnection is a ADSL one (pppoe), but it goes down
> > every now and then so I would like to set up a failover solution to my
> > ISDN-Card.
> > Googling doesn't helped me...
> > I there a working configuration or howto out there?
> 
> I've been looking for a proper solution to this problem for quite a
> while now, but I have not found anything viable as of yet.
> 
> So, I'd really be interested in hearing if there is a good solution.
> I'm convinced that it can be done!

The trick is to have two routes with different Metric.  If for example
ippp0 is your ISDN-device and ppp0 is your DSL connection, you could
have two default routes: one with metric == 0 on ppp0 and another one
with metric == 1 on ippp0.  If both pppd-Demons are configured to 
"dial" on demand, the ISDN connection will be established automatically,
if the pppoed dies for some reason.

Since my setup is little bit different (ISDN goes directly to a computer
in our company network and is a fallback for the VPN) I can't provide
a working configuration file.  
Just a guess: you have to insert or modify some 'up' and 'down' commands
in your /etc/network/interfaces file.  Like so 
...
iface ppp0 inet ... # DSL (pppoed)
...
	up route add default gw XXX.XXX.XXX.XXX metric 0 
        down route del default gw XXX.XXX.XXX.XXX
...
iface ippp0 inet ... # ISDN
...
	up route add default gw XXX.XXX.XXX.XXX metric 1 
        down route del default gw XXX.XXX.XXX.XXX

Read /usr/share/doc/ifupdown/examples/network-interfaces.gz
for more information.

Regards, Peter
-- 
Peter Funk, Oldenburger Str.86, D-27777 Ganderkesee, Germany
office: +49 421 20419-0 (ArtCom GmbH, Lise-Meitner-Str. 5, D-28359 Bremen)
http://wecanstopspam.org



Reply to: