CCIE R&S IGP notes (RIP, EIGRP & OSPF)

RIPv2:

  • ip rip triggered is an interface command to only send rip updates when there is a change. Only works on P2P links
  • passive interface in RIP only stops the router sending updates out, it’ll still receive updates in from neighbours. If you have configured a static neighbour it’ll send unicast updates even if passive-interface has been configured. i.e. if you want to limit the neighbour relationship out an interface, you can use passive-interface as well as the neighbour command
  • ip rip v2-broadcast is an interface command to allow you to broadcast RIPv2 updates out
  • Use the no validate update-source interface command if the neighbour is speaking to the router using an IP not on the local subnet (secondary address is an example)

EIGRP:

  • k1 = bandwidth
  • k2 = load
  • k3 = delay
  • k4 = reliability
  • k5 = MTU
  • By default, only bandwidth and delay are used
  • K values need to match in EIGRP AS domain in order for neighbours to form
  • Many ways to allow EIGRP to use equal or unequal paths. Can use offset-lists to increase metric, change bandwidth/delay, add additional K values into metric, increase K multiplier and so on
  • IOS allows up to 16 paths to be used, but only 4 by default. Changed using the maximum-paths eigrp process command
  • eigrp stub receive-only tells the local router to receive eigrp routes, but don’t send anything

OSPF:

  • FLOOD-WAR means that 2 routers (not directly connected) share the same router-id. If they were directly connected the neighbour relationship would not form
  • Router-id’s can change depending on the lab, this is important as virtual-links and certain other filtering mechanisms are configured using the RID. If the RID changes, your configuration needs to change
  • Virtual-links are in Area 0, hence if you need to authenticate all Area 0 links, you need to authenticate the virtual-link
  • Authentication is configured on the interface, however virtual-link authentication is configured under the ospf process itself using area (x) virtual-link (x.x.x.x) (message-digest-key|authentication-key) (…)
  • If you need to configure an interface in area 0, but not allowed to use area 0 command, you can always use area 0.0.0.0 – More info on this 32bit area number here: http://mellowd.co.uk/ccie/?p=910
  • Virtual links are on-demand links. Hence if you don’t do authentication on one side, you’ll not notice until packets actually needs to go down the link.
  • When router A sends router B an LSA, it includes router B’s cost to the destination. It does NOT include the local shared link. Router A will add the local link cost itself. This is the same as spanning-tree costs
  • max-lsa (options) – configured the maximum amount of non self generated lsa’s the local router can have. Can drop peer or warn when lsa amount is breached
  • ip ospf flood-reduction – stops an OSPF speaker from updating LSA’s every 30 minutes. If configured you’ll need to do it on all OSPF speakers
  • show ip ospf border-routers is a handy command to use when checking costs to border routers (perhaps for load-balancing to external destinations)
  • Watch out for strange frame-relay set ups. You may need to have frame-relay maps to certain other spokes, depending on the ospf network type
  • OSPF routes are always chosen in the order: O; O IA; E1; E2; N1; and N2 regardless of metric. You can however tell ospf to use different ADs for each of these types with the distance ospf [options] command which is handy for complex redistribution labs

Redistribution:

  • distribute-list out on DV protocols will of course affect what DV routes go INTO another protocol
  • Use debug ip routing
  • Use tags wherever possible
  • When redistributing from OSPF into BGP, it will only redistribute internal OSPF routes by default. You need to specify the external routes if you want them redistributed
  • When reditributing through a route-map, you can specify in the route-map that certain routes will be E1 and others E2. Can also specify the metric itself

Policy-based routing:

  • ip policy route-map will be configured on the interface in which traffic is coming in on
  • ip local policy route-map is used for policy routing locally generated traffic by the router

Changing AD:

  • router (eigrp|rip|ospf)
    • distance (#) x.x.x.x x.x.x.x (ACL)
  • RIP – x.x.x.x = advertising neighbours IP address
  • EIGRP x.x.x.x = advertising neighbours router-id
  • OSPF x.x.x.x = router-id of router originating the LSA into the area
  • If you use 0.0.0.0 255.255.255.255 then you’re telling the router not to care WHERE the route came from

© 2009-2019 Darren O'Connor All Rights Reserved -- Copyright notice by Blog Copyright