Hi Malik,
1-In JNCIP Book case study there is task to increase the level 1 prefernce to 155 which in turn requires us to keep the
OSPF external preference below 160 i.e.159.
Why is it neccassary to keep the preference below 160 how will it impact 192 subnets will they be automatically
advertised to backbone.
Referring to the case-study topology, let's assume that you have set the ospf external-perference as "165" and R7 gets the OSPF 192.168.x prefixes first compared to R6.
Now since ISIS L1 adjaceny is formed between R6 and R7 over direct connected interface and the isis export policy on R7 is defined as below, R6 gets the 192.168.x prefixes as ISIS Level 1 external routes from R7, who's preference is 160.
show policy-options
policy-statement ospf-isis {
term 1 {
from {
route-filter 192.168.0.0/22 longer;
}
then accept;
}
term 2 {
from {
route-filter 0.0.0.0/0 exact;
}
then reject;
}
}
In meantime, R6 gets the same 192.168.x prefixes from OSPF neighbor as extrenal routes who's preference is 165 (as defined).
WIth this, R6's routing table will have 192.168.x installed as :
suryak@R6# run show route 192.168/22
inet.0: 52 destinations, 57 routes (51 active, 0 holddown, 1 hidden)
+ = Active Route, - = Last Active, * = Both
192.168.0.0/24 *[IS-IS/160] 00:00:25, metric 10
> to 10.0.8.2 via ge-1/0/0.0
[OSPF/165] 00:00:03, metric 0, tag 0
> to 172.16.40.1 via ge-2/1/1.0
192.168.1.0/24 *[IS-IS/160] 00:00:25, metric 10
> to 10.0.8.2 via ge-1/0/0.0
[OSPF/165] 00:00:03, metric 0, tag 0
> to 172.16.40.1 via ge-2/1/1.0
192.168.2.0/24 *[IS-IS/160] 00:00:25, metric 10
> to 10.0.8.2 via ge-1/0/0.0
[OSPF/165] 00:00:03, metric 0, tag 0
> to 172.16.40.1 via ge-2/1/1.0
192.168.3.0/24 *[IS-IS/160] 00:00:25, metric 10
> to 10.0.8.2 via ge-1/0/0.0
[OSPF/165] 00:00:03, metric 0, tag 0
> to 172.16.40.1 via ge-2/1/1.0
This clearly indicates R6 will take sub-optimal path via R7 to reach 192.168.x prefixes.
Hence to avoid this situation you set the ospf external preference less than 160.
2- In case study when i use rip instead of ospf there is strange behaviour i.e. the rip router can reach R6 10.0.9.6
directly where as the R7 10.0.9.7 is reachable through R6 is it a normal behaviour just becoz of rip can not load balance automatically however it installs both R6 and R7 as equal cost next hops.
This is because of the route metric behaviour associated with RIP. By default, the routes that are imported from RIP neighbor are incremented by 1. While redistributing, the routes are advertised with metric 1. Hence the RIP router see same metric value for a given prefix from R6 and R7, which ends up as ECMP paths.
3-What if i keep the link between the R3 & R4 in level 1 and their loopbacks in both levels i.e. l1/l2 will it cause some serious issue like RR suboptimal routing or something like ospf when it is neccassary to keep the link b/w r3 and r4
in backbone i.e. area 0.
For the given case in JNCIP, this will not cause any sub-optimal routing. But to meet the specified criteria in case-study, you need to have loopbacks in L2 only.
However there can be some cases where incluing loopbacks in both L1 and L2 can cause suboptimal cases. Let's take an example as below:
L1
|-------R2-------BGP1
| |
| |
R1 | L2
| |
| |
|-------R3--------BGP2
L1
In this case, if R2 and R3 loopback are included in L1 and L2, then to reach any prefixes of BGP2 (with next-hop as R3's loopback) from BGP1 via R2, the path would be BGP1-R2-R1-R3-BGP2 as L1 routes (preference 15) are preferred over L2 routes (preference 18)
Hope this helps...
Regards
Surya prakash
If you like this, kudos would be appreciated.