[c-nsp] PoP(s) failure and WWW redirect

Rubens Kuhl Jr. rubensk at gmail.com
Tue Oct 4 13:27:02 EDT 2005


Assuming you are a transit provider:
1) Use the BGP conditional advertisement to announce a default route
if the default route from the upstream disappears, or some simpler
mechanism like a default rote with lesser priority than the one to the
upstream (provided a control channel like PPP or ATM OAM drops the
route if the circuit fails).
2) Drop all non-HTTP traffic on this sinkhole router;
3) Pass all remaining traffic to a *nix box. If it is a Linux 2.4 or
2.6 kernel, use iptables to DNAT all traffic to a local address.
4) On the *nix box, configure an Apache server with mod_rewrite to
ignore host-headers and make all requests (/whatever.html) to be
redirected (HTTP 302, temporarily moved) to a server on your network
(http://networkisdown.myisp). It must not be the same machine with
this trickery, or at least use another network interface and another
server instance.


Rubens

On 10/4/05, Mark Tohill <Mark at u.tv> wrote:
> Hi,
>
>
>
> Has any implemented or have thoughts on the following?
>
>
>
> If we loose our upstream provider (our only provider) connectivity or
> they have routing 'issues', we can redirect major customers to an
> off-site temporary holding page.
>
>
>
> I was pointed towards the following on CCO recently.
>
>
>
> http://www.cisco.com/en/US/products/sw/iosswrel/ps1839/products_feature_
> guide09186a00800b5d54.html
>
>
>
> Is this an option and are there any others? BTW, DNS has been ruled out.
>
>
>
>
>
> Thanks,
>
> Mark.
>
>
>
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>



More information about the cisco-nsp mailing list