In my CCNP notes I usually have some questions about each topic to help me solidify the material (along with labs). Here are some EIGRP questions I would recommend trying to answer. If I were you, I would memorize these questions/answers. I promise they will help!
Questions:
- What is the administrative distance of internal EIGRP?
- What is the administrative distance of external EIGRP?
- How do you advertise an interface’s network into EIGRP?
- What are the default K values used to calculate the EIGRP metric? Also what are all of the 5 K values?
- What metrics does EIGRP use? Describe them.
- How do you advertise other routing protocols, static routes, or connected interfaces without the network command into EIGRP?
- How is the Router ID selected for EIGRP?
- How do you advertise a network (with the internal EIGRP AD), but suppress EIGRP hellos on an interface?
- What does EIGRP use for neighbor discovery? What address do most EIGRP packets go to?
- How does EIGRP know a neighbor is down/dead?
- What command can we use for unequal cost load balancing, how does it work?
- How do you form a neighborship over non broadcast media, what are the downsides of this?
- How does EIGRP prevent routing loops internally in the AS, explain in detail.
- How does EIGRP prevent external routing loops?
- Can two EIGRP routers with the same router ID become neighbors?
- What are EIGRP stub routers, how do you configure them? What are the defaults for it?
- List 5 ways to manipulate EIGRP route paths? Quickly describe them.
- List 4 ways to advertise a default route into EIGRP.
- Describe the SIA state, pre and post SIA queries.
- How do you change the EIGRP AD locally on the router? Set internal AD to 95, and external to 200.
- How do you change the EIGRP K values on a router?
- What is required to match for a neighborship to form?
Answers:
- The AD of internal EIGRP is 90.
- The AD of external EIGRP is 170.
- With the “network” command, e.g. “network <x.x.x.x> <wildcard bits optional>
- Bandwidth, and delay are the default K values used. All of the K values are Bandwidth, Delay, Reliability, Load, and MTU.
- EIGRP has feasible distance, and reported/advertised distance. Feasible distance is the distance to the prefix from the point-of-view of your local EIGRP router. Advertised/reported distance is the distance to the prefix from the router who sent you the route.
- To advertise those items, you need to redistribute. E.G. “redistribute static”, “redistribute connected” , “redistribute ospf <process ID> metric <k1 k2 k3 k4 k5” When redistributing you need to either set a “default metric” or set the metrics when redistributing.
- The router ID is selected when EIGRP first comes up. It is the highest loopback interface IP address. If a loopback interface does not exist EIGRP will choose the highest IP address of an interface.
- To advertise a network and suppress EIGRP hellos on the interface you need to set the “passive-interface <int>” command in EIGRP config.
- EIGRP used hello packets sent out every 5 seconds by default. Most EIGRP packets go to 224.0.0.10, some queries are sent via unicast.
- EIGRP has a hold/dead timer which is 15 seconds by default. If no hello is received in this time, the neighborship is taken down.
- Variance is used for unequal cost load balancing. Variance is a multiplier, which by default is set to 1. Variance multiplies the successors feasible distance (locally on the router) by X. Then any feasible successor that has a feasible distance less than that multiplied value will be put into the routing table. To view the traffic shares use “show ip route x.x.x.x”. Variance only works for routes that have passed the feasibility condition.
- To form a neighborship over non-broadcast media we need to use the “neighbor x.x.x.x int <int>” command in EIGRP config. This must be done on both ends of the link because it turns off multicast processing for that interface.
- EIGRP prevents internal routing loops by using the feasibility condition. The condition is this: To become a feasible successor, the routes advertised distance must be less than my successors feasible distance. This guarantees a loop-free path. This can however mean routes with very bad metrics do not pass the condition. This isn’t a problem because we can always go active for the route to get that path later if we lose the successor. We also have split-horizon. Split-horizon says you do not advertise a route on the same interface you learned it on. This helps filter what I like to call “route noise” from being sent to the test of the feasibility condition.
- Since EIGRP has a different AD for internal and external routes, this helps prevent routing loops. Also, when a network is redistributed into EIGRP the originating router ID is the router that redistributed it into EIGRP.
- Two routers with the same router ID can become neighbors, however they will not exchange internal or external routes. They will however pass routes. The behavior is the same if they are separated by other routers.
- EIGRP stub routers are routers that are not queried for routes that go active. This is because these routers are in a “stub” network, with only one path. To configure use “eigrp stub” in EIGRP config mode. The defaults for EIGRP stub are it will only advertise connected, and summary routes.
- Change active K values to make one route more or less preferred. Use variance to introduce unequal cost load balancing for feasible successors. Summarize routes to make them less preferred. Use offset lists to make a route less or more preferred. Or just filter the routes all together via a distribute-list.
- You can redistribute a static route to an IP address. You can use the “summary-address” in interface config to summarize a default route to a neighbor. You can use the “ip default-network <classfulnetwork here>” command (as long as you have a route to the network). You can use the “network 0.0.0.0” command which advertises a defualt route if you have a static route like this “ip route 0.0.0.0 0.0.0.0 <interface here>”.
- When we lose a neighbor/route and we don’t have a feasible successor for the route, the route goes into active mode. We are querying all of our EIGRP neighbors for the route. Since the whole chain needs to reply to queries it’s possible by congestion or packetloss we will lose a reply. This will cause the chain of EIGRP neighbors to go down after 180 seconds. After SIA queries were implemented, if we don’t receive a reply in 90 seconds, we send a SIA query to the neighbor to confirm that they are still up. This helps prevent re-convergence of the whole network.
- You need to use the “distance EIGRP” command in EIGRP config. “distance eigrp 95 200”
- Use the “metric weights 0 <k1 k2 k3 k4 k5>” command.
- Subnet, K values, AS.