Hi, any idea how other providers offer such redundancy to end customers (if they do at all) ? We have a mass of customers with /29 or /28 networks and losing IPs isn't an option in such cases imo. Using bigger networks would require giving up vlan separation each customer, no option either.
regards Rolf > On Thu, 2012-03-01 at 16:30 +0100, "Rolf HanÃen" wrote: >> Is there a way to configure virtual IPs that do not belong to the >> "hard-coded" network (ip address x.x.x.x y.y.y.y) of the interface ? >> I see that it is possible to configure other IPs, but this results in a >> warning and there is no possibility to set the netmask at all. > > I was wondering the same some years ago. Take a look at this thread: > > http://puck.nether.net/pipermail/cisco-nsp/2007-November/045409.html > > We never got it to work. ARP requests are sourced from the real address, > and you cannot add a "connected static" route for a VRF enabled > interface, i.e. "ip route vrf A 192.168.1.0 255.255.255.0 Vlan50" fails. > > Also keep in mind that TTL exceeded replies (traceroute) would source > from the "real" interface address. > >> Is there a possibility to have static routes that are only active if the >> node has enabled the virtual IP ? > > This in itself would be possible with an EEM script that follows the > HSRP log messages and adjusts the configuration. It would trigger a > configuration change, so Rancid or whatever you might use would log a > change every time the HSRP state changes. > >> Is there anything else to take care of ? >> Any limitations except the 4096 HSRP-IDs ? > > That's 256 for HSRPv1 by the way. > > -- > Peter > > > _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/