I have a question with OSPF based on the attachment that I have posted.
1) Since there are 2 routers R1 and R4 that are connected to area 0, would this
cause routes to enter via one router and come back to area 1 via another router ?
Do I need to use prefix-list/distrubute-list to prevent a routing loop here ?
There would not be any routing loop and no need for any prefix/distribute lists
Here you will be receiving routes from both the routers and the area0 will choose the path based on best metric. If the cost of the links from R4 and R1 are same, your router will loadbalance the traffic via both links
Loadbalancing doesn't really mean that the traffic will go via one and return via another. By default the cisco routers run CEF which loadbalance on a per-destination basis. So you will find that a particular source-destination pair always take one path.
If you want only router to act as primary for all the traffic, you can manipulte the cost of one of the links of the routers between area0 and area1
I have a similar situation with the same question and was wondering if your answer would still hold true:
The difference is that all devices are on the same ethernet link.
I have an edge router that is getting a default route from the PE router via BGP and distibuting it into OSPF area 0
The default route is then advertised into OSPF.
I have two firewalls with their outside interfaces in the same OSPF area getting the defualt route.
I have a CSS that has one interface sitting in the area 0 and I want to make sure that the two firewalls will send all of their traffic through the CSS and then go out the Edge router to the Internet.
I have the CSS configured for OSPF in area 0 and have given the interface a cost of 1,
The firewalls and router have their interfaces with a cost of 10.
This seems to be working in a Lab with the firewalls prefering the path to the CSS then going to the edge router.
But, I was not sure from a design perspective if this was a proble to put into production.
The CSS will only allow me to configure it as a area-boudary device to do this.
Can you tell me your thougts?
Though u have given a very elaborate expanation of the network, it would be easier to visualize and suggest if you also post your network topology
Thanks for the reply.
The HQ CSS is in place and I have not put the DR CSS in place just yet.
The idea is to have all of the routed traffic from the HQ site go through the CSS first and then to the edge router.
I cannot use static routes because our failover scheme depends on the dynamic default route obtained from the BGP distribution from the PE router.
We are only getting the default route from them. When we loose the Internet connection from HQ, the default route flips to the DR site Internet gateway router.
The CSS is only advertising the default route.