Consider the following
I have a network with the following configuration that is used to redistribute static routes into an EIGRP network.
router eigrp 1 redistribute static metric 32 2000 255 1 1500 network 10.0.0.0 auto-summary ! ip route 30.0.0.0 255.255.255.0 10.0.0.2 ip route 40.0.0.0 255.255.255.0 10.0.0.2 !
We have a requirement to do some route slugging on a router in the high availability path. We need to lower the metric of the static route 40.0.0.0/24 this are being redistributed so that this becomes the preferred router for that route.
Question: What is the difference between the following two solutions:
Solution 1
router eigrp 1 redistribute static metric 32 2000 255 1 1500 route-map metricmod network 10.0.0.0 auto-summary ! ip route 30.0.0.0 255.255.255.0 10.0.0.2 ip route 40.0.0.0 255.255.255.0 10.0.0.2 ! logging alarm informational access-list 10 permit 40.0.0.0 0.0.0.255 ! route-map metricmod permit 10 match ip address 10 set metric 10000 10 255 1 1500 ! route-map metricmod permit 20
Solution 2
router eigrp 1 redistribute static route-map metricmod network 10.0.0.0 auto-summary ! ip route 30.0.0.0 255.255.255.0 10.0.0.2 ip route 40.0.0.0 255.255.255.0 10.0.0.2 ! logging alarm informational access-list 10 permit 40.0.0.0 0.0.0.255 ! route-map metricmod permit 10 match ip address 10 set metric 10000 10 255 1 1500 ! route-map metricmod permit 20 set metric 32 2000 255 1 1500
Hint
Would you like to a hint ? [slider title="Click Here"]Solution 1 is much better in a real network[/slider]
I will post the answer in a week or so. Answers in the comments and I will say why its the best solution.
Answer
Solution 2 will require you to issue the following:
router eigrp 1 no redistribute static
and then
redistribute static route-map metricmod
to make these changes. Of course this removes all the routes from the routing table and will cause an outage. Whereas Solution 1 will not cause a routing tables to change and thus not cause an impact to services.
Sermon
For anyone who is learning Cisco networking, this might help to understand why detailed knowledge, “more than one way” and memorisation is important. By knowing it backwards, you spend your time thinking about other things rather than thinking only about the redistribution. Typically, people with limited skills will be happy with ANY solution, and not attempt to think of other ways of doing something that might be a better choice.
For mendicants in CCIE study: this iss why you need to know four different ways to do same thing. You learn to keep thinking about options, and the best choice. It’s been a valuable lesson for me.
Postscript
In this particular case (as Ivan Pepelnjak points out in the comments) the router had a couple of hundred static routes. When you delete and add that many routes you create a CPU and bandwidth hit as the routes are deleted and propagated. †In networks with slow bandwidth, this can cause service impacts as the routing updates can suck up bandwidth.
It’s a big knock on effect. You need to think of these things in real life to keep the network up all the time.
