Narbik Gap from CCNP to CCIE BGP lab 10

CCIEWANNABE

ok, quick question on this lab. i am doing the last bgp configuration in lab 10. I set bgp best path med missing-as-worst on R1 so that all med values of routes in the BGP topology table with no med are given the worst possible med (4294967295) and am wondering why R1 does not see R2's locally generated routes as having the worst possible med. all other routes do (routes from R3 and R4), but R2 does not. Anybody else out there that has this workbook who can help out. There are no in/out filters on either R1 or R2 towards eachother. So why aren't these routes being advertised into R1 with the worst possible med. Very strange. Thanks for the help.

CCIEWANNABE

i think i figured it out. narbiks lab is about manipulating med, however in this scenario he does not use bgp always-compare-med and bgp bestpath as-path ignore, which in this scenario gives R1 the best path to R2's locally generated networks via the BGP AS_Path attribute, not MED (b/c of the BGP Best path selection). See this is confusing because he doesn't explain this at the end of the lab, and what further confuses things is that the worst MED value is not given to these routes (Routes from R2), maybe b/c he didn't do a clear ip bgp on R2. So what you end up with in R1's BGP topology table is 2 paths to the locally generated routes behind R2 and they are the following as shown in his workbook:

R1's BGP topology table:

_____________NEXT HOP_______MED___AS_PATH
*>2.0.0.0____131.1.12.2 (R2)_____0______200 i
*___________131.1.14.4 (R4)____100_____400 300 200 i

*>22.0.0.0___131.1.12.2 (R2)______0_____200 i
*___________131.1.14.4 (R4)____100_____400 300 200 i

*>3.0.0.0___131.1.14.4 (R4)____100______400 300 i <-- SAME AS_PATH DISTANCE
*__________131.1.12.2 (R2)_4294967295__400 300 i <-- SAME AS_PATH DISTANCE


which is confusing b/c it looks like R2 wins b/c of MED, however AS_PATH is the reason why it wins.

And the only routes that match the bgp bestpath med missing-as-worst and win this process are the routes locally generated by R3 because they have an equal hop path to R1 via R2 and R4, which makes MED the next BGP best path selection :]

I forgot to mention but R4 is also connected to R1 with a route map specifing a metric of 100 for all routes from R4, that is why the MED is 100.

I'm just glad i finally figured it out :] I wonder if anyone else has caught this before. I was starting to go crazy. Oh, by the way, BGP is awesome.