Solved: BGP : route advertised but not received

Jad KAROUT · ‎04-04-2011

Hi everyone,

I'm running into a bgp problem that is really confusing to me. Here's the topology :

So R2 advertises a connected prefix (10.212.254.0/24) to R0. This prefix is then advertised to R1, R3 and R4. So far so good.

And then i have R1 advertising a connected prefix (10.212.255.0/24) to R0 as well. This prefix is then advertised by R0 to R3 and R4 but it doesn't make it to R2. The strange thing is that R0 does advertise the prefix to R2 (i checked with a

show ip bgp vpnv4 vrf ce R0 neighbor 10.245.32.1 advertised).

However R2 does not receive this route

 (checked with show ip bgp vpnv4 vrf ce R0 neighbor 10.245.32.2  received-routes)

Does someone have any idea how this prefix gets lost on its way from R0 to R2 ??? There's no configured filtering of any kind on R0. I checked with multiple

debug

commands (for bgp and routing) by that wasn't very helpful.

Finally R1 and R2 are two VRF running on the same 3750, same for R3 and R4, and R0 is on its own on yet another 3750.

Thx for any help.

jitendra22703 · ‎04-05-2011

Hi,

Its the default behaviour of local-as command that changes the AS for remote peers not for local router AS. R2 is not accepting the routes because AS-PATH Splithorizon rule. R2 look its own 65001 AS in AS-PATH hence its not installing the routes in routing table to avoid loop. where in R1 looks prefix originated in R2 from AS 65002 hence its installed in routing table.

You have two workaround to avoid this issue.

1. On R2:- neighbor R0 allowas-in

2. On R0:- neighbor R2 as-override

Regards,

Jitendra

CCIE 22703#

View solution in original post

Richard Burts · ‎04-04-2011

It might be helpful if you can show us configuration details from R1 and R2 - especially the configuration of BGP.

Your drawing seems to indicate that R1 has BGP process with AS number 65001 and that R2 has 65002. If I understand your post correctly then both are running on the same 3750. I am curious how you get 2 different BGP processes on the same 3750. If we understood this then perhaps we might have better answers about the missing route.

HTH

Rick

HTH

Rick

Jad KAROUT · ‎04-05-2011

Hi Rick,

thanks for your input.

Indeed there's a limitation on the 3750 that won't allow more than one BGP process to run. In order to circumvent this, i had to resort to using the replace-as option of the neighbor command.

Here are some outputs that may be of some help :

BGP config on 3750 running R1 and R2

interface loopback 0

ip vrf forwarding ce_R1

ip address 1.1.1.1 255.255.255.255

interface loopback 1

ip vrf forwarding ce_R2

ip address 2.2.2.2 255.255.255.255

router bgp 65001

no synchronization

bgp router-id 1.1.1.1

bgp log-neighbor-changes

no auto-summary

address-family ipv4 vrf ce_R2

  redistribute connected

  redistribute static

  neighbor 10.245.32.2 remote-as 65100

  neighbor 10.245.32.2 local-as 65002 no-prepend replace-as

  neighbor 10.245.32.2 activate

  neighbor 10.245.32.2 soft-reconfiguration inbound

  no synchronization

  bgp router-id 2.2.2.2

exit-address-family

!

address-family ipv4 vrf ce_R1

  redistribute connected

  redistribute static

  neighbor 10.245.31.2 remote-as 65100

  neighbor 10.245.31.2 activate

  neighbor 10.245.31.2 soft-reconfiguration inbound

  neighbor 10.245.31.2 prefix-list filter-ce1 out

  no synchronization

  bgp router-id 1.1.1.1

exit-address-family

BGP config of 3750 running R0

router bgp 65100

  no synchronization

  bgp router-id 3.3.3.3

  bgp log-neighbor-changes

  no auto-summary

  address-family ipv4 vrf core_R0

   neighbor 10.245.31.1 remote-as 65001

   neighbor 10.245.31.1 activate

   neighbor 10.245.31.1 soft-reconfiguration inbound

   neighbor 10.245.32.1 remote-as 65002

   neighbor 10.245.32.1 activate

   neighbor 10.245.32.1 soft-reconfiguration inbound

   neighbor 10.245.33.1 remote-as 65000

   neighbor 10.245.33.1 activate

   neighbor 10.245.33.1 soft-reconfiguration inbound

   neighbor 10.245.34.1 remote-as 65000

   neighbor 10.245.34.1 activate

   neighbor 10.245.34.1 soft-reconfiguration inbound

   no synchronization

  exit-address-family




Below are the peering state of the three routers R0, R1 and R2 :




Switch_sites#sh ip bgp vpnv4 vrf ce_R1 sum                       
BGP router identifier 1.1.1.1, local AS number 65001
BGP table version is 539, main routing table version 539
17 network entries using 2397 bytes of memory
17 path entries using 1156 bytes of memory
4/3 BGP path/bestpath attribute entries using 560 bytes of memory
2 BGP AS-PATH entries using 48 bytes of memory
0 BGP route-map cache entries using 0 bytes of memory
0 BGP filter-list cache entries using 0 bytes of memory
BGP using 4161 total bytes of memory
BGP activity 205/178 prefixes, 263/236 paths, scan interval 15 secs




Neighbor        V    AS     MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd
10.245.31.2    4  65100   16706     16690        539      0    0       17:34:23       10







Switch_sites#sh ip bgp vpnv4 vrf ce_R2 sum                                    
BGP router identifier 2.2.2.2, local AS number 65001
BGP table version is 539, main routing table version 539
10 network entries using 1410 bytes of memory
10 path entries using 680 bytes of memory
4/3 BGP path/bestpath attribute entries using 560 bytes of memory
2 BGP AS-PATH entries using 48 bytes of memory
0 BGP route-map cache entries using 0 bytes of memory
0 BGP filter-list cache entries using 0 bytes of memory
BGP using 2698 total bytes of memory
BGP activity 205/178 prefixes, 263/236 paths, scan interval 15 secs




Neighbor        V    AS       MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd
10.245.32.2    4    65100   15683     15660        539      0    0       18:47:57        7




Interestingly R2 sees itself as 65001 although the other routers see it as 65002.




Switch_CORE#sh ip bgp vpnv4 vrf core_R0 summary 
BGP router identifier 3.3.3.3, local AS number 65100
BGP table version is 198, main routing table version 198
14 network entries using 1974 bytes of memory
17 path entries using 1156 bytes of memory
4/3 BGP path/bestpath attribute entries using 560 bytes of memory
3 BGP AS-PATH entries using 72 bytes of memory
0 BGP route-map cache entries using 0 bytes of memory
0 BGP filter-list cache entries using 0 bytes of memory
BGP using 3762 total bytes of memory
BGP activity 68/54 prefixes, 103/86 paths, scan interval 15 secs




Neighbor        V    AS       MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd
10.245.31.1    4    65001   18416     18436        198       0    0      17:42:29        4
10.245.32.1    4    65002   15664     15689        198       0    0      18:52:32        3
10.245.33.1    4    65000   15620     15649        198       0    0      1w3d             5
10.245.34.1    4    65000   15638     15668        198       0    0      19:36:30        5




Here, as you can see in red, R0 sees R2 as part of AS 65002.







Now  you can see R2 advertises its 10.212.254.0/24 prefix to R0:




Switch_sites#sh ip bgp vpnv4 vrf ce_R2 neighbors 10.245.32.2 advertised-routes 
BGP table version is 539, local router ID is 2.2.2.2
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
              r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 4:4 (default for vrf ce_R2) VRF Router ID 2.2.2.2
*> 2.2.2.2/32           0.0.0.0                  0         32768 ?
*> 10.212.254.0/24  0.0.0.0                  0         32768 ?
*> 10.245.32.0        0.0.0.0                  0         32768 ?







R0 receives the 10.212.254.0/24 prefix sent  by R2 and it's seen as originating from AS 65002:




Switch_CORE#sh ip bgp vpnv4 vrf core_R0 neighbor 10.245.32.1 received-routes 
BGP table version is 198, local router ID is 3.3.3.3
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 3:3 (default for vrf core_R0)
*> 2.2.2.2/32           10.245.32.1             0             0 65002 ?
*> 10.212.254.0/24  10.245.32.1             0             0 65002 ?
r> 10.245.32.0         10.245.32.1             0             0 65002 ?







R0 then advertises it to R1 :




Switch_CORE#sh ip bgp vpnv4 vrf core_R0 nei 10.245.31.1 advertised-routes 
BGP table version is 198, local router ID is 3.3.3.3
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 3:3 (default for vrf core_R0)
*> 2.2.2.2/32          10.245.32.1             0             0 65002 ?
*> 4.4.4.4/32          10.245.33.1             0             0 65000 ?
*> 5.5.5.5/32          10.245.34.1             0             0 65000 ?
*> 10.200.190.0/24  10.245.33.1             0             0 65000 ?
*> 10.202.211.0/24  10.245.33.1             0             0 65000 ?
*> 10.205.0.0/24      10.245.33.1             0             0 65000 ?
*> 10.212.254.0/24  10.245.32.1             0             0 65002 ?
r> 10.245.32.0         10.245.32.1             0             0 65002 ?
r> 10.245.33.0         10.245.33.1             0             0 65000 ?
r> 10.245.34.0         10.245.34.1             0             0 65000 ?




And R1 receives it correctly and install it in its RIB :




Switch_sites#sh ip bgp vpnv4 vrf ce_R1 neighbors 10.245.31.2 received-routes 
BGP table version is 539, local router ID is 1.1.1.1
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 3:3 (default for vrf ce_R1) VRF Router ID 1.1.1.1
*> 2.2.2.2/32           10.245.31.2                           0 65100 65002 ?
*> 4.4.4.4/32           10.245.31.2                           0 65100 65000 ?
*> 5.5.5.5/32           10.245.31.2                           0 65100 65000 ?
*> 10.200.190.0/24  10.245.31.2                           0 65100 65000 ?
*> 10.202.211.0/24  10.245.31.2                           0 65100 65000 ?
*> 10.205.0.0/24     10.245.31.2                           0 65100 65000 ?
*> 10.212.254.0/24  10.245.31.2                           0 65100 65002 ?
*> 10.245.32.0        10.245.31.2                           0 65100 65002 ?
*> 10.245.33.0        10.245.31.2                           0 65100 65000 ?
*> 10.245.34.0        10.245.31.2                           0 65100 65000 ?




Total number of prefixes 10




So this works fine.




Now if i look the other way around from R1 to R2, R1 advertises its 10.212.255.0/24 prefix to R0 :




witch_sites#sh ip bgp vpnv4 vrf ce_R1 neighbors 10.245.31.2 advertised-routes 
BGP table version is 539, local router ID is 1.1.1.1
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 3:3 (default for vrf ce_R1) VRF Router ID 1.1.1.1
*> 1.1.1.1/32           0.0.0.0                     0         32768 ?
*> 10.3.10.0/24       10.212.255.4             0         32768 ?
*> 10.199.228.0/24  10.212.255.4             0         32768 ?
*> 10.212.255.0/24   0.0.0.0                     0         32768 ?




Total number of prefixes 4




R0 then advertises to  R2 the routes it received from R1 correctly :




Switch_CORE#sh ip bgp vpnv4 vrf core_R0 nei 10.245.32.1 advertised-routes 
BGP table version is 198, local router ID is 3.3.3.3
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 3:3 (default for vrf core_R0)
*> 1.1.1.1/32          10.245.31.1             0             0 65001 ?
*> 4.4.4.4/32          10.245.33.1             0             0 65000 ?
*> 5.5.5.5/32          10.245.34.1             0             0 65000 ?
*> 10.3.10.0/24       10.245.31.1             0             0 65001 ?
*> 10.199.228.0/24  10.245.31.1             0             0 65001 ?
*> 10.200.190.0/24  10.245.33.1             0             0 65000 ?
*> 10.202.211.0/24  10.245.33.1             0             0 65000 ?
*> 10.205.0.0/24      10.245.33.1             0             0 65000 ?
*> 10.212.255.0/24  10.245.31.1             0             0 65001 ?
r> 10.245.33.0         10.245.33.1             0             0 65000 ?
r> 10.245.34.0         10.245.34.1             0             0 65000 ?




and here finally is my problem as R2 doesn't receive the routes R0 sends it originating from R1 (AS 65001) :




Switch_sites#sh ip bgp vpnv4 vrf ce_R2 neighbors 10.245.32.2 received-routes 
BGP table version is 539, local router ID is 2.2.2.2
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
               r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete




   Network               Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 4:4 (default for vrf ce_R2) VRF Router ID 2.2.2.2
*> 4.4.4.4/32           10.245.32.2                           0 65100 65000 ?
*> 5.5.5.5/32           10.245.32.2                           0 65100 65000 ?
*> 10.200.190.0/24  10.245.32.2                           0 65100 65000 ?
*> 10.202.211.0/24  10.245.32.2                           0 65100 65000 ?
*> 10.205.0.0/24      10.245.32.2                           0 65100 65000 ?
*> 10.245.33.0         10.245.32.2                           0 65100 65000 ?
*> 10.245.34.0         10.245.32.2                           0 65100 65000 ?

Total number of prefixes 7

I thought maybe the issue has something to do with the 3750 limitation as R2 sees itself part of AS 65001 although the other routers see it as 65002. But if that were the reason, then why are its routes learned by R1 but not the other way around ?

Can someone shed any light on this ?

Thx for your help.

jitendra22703 · ‎04-05-2011

Hi,

Its the default behaviour of local-as command that changes the AS for remote peers not for local router AS. R2 is not accepting the routes because AS-PATH Splithorizon rule. R2 look its own 65001 AS in AS-PATH hence its not installing the routes in routing table to avoid loop. where in R1 looks prefix originated in R2 from AS 65002 hence its installed in routing table.

You have two workaround to avoid this issue.

1. On R2:- neighbor R0 allowas-in

2. On R0:- neighbor R2 as-override

Regards,

Jitendra

CCIE 22703#

Jad KAROUT · ‎04-05-2011

Hi Jitendra,

i had considered that R2 saw itself par of AS 65001 as the reason for the problem but was left a bit confused by the various ouputs of the show commands. Indeed while the

show ip bgp summary

command that i posted above indicates R2 as part of AS 65001, the

show ip bgp neighbor

command indicates that it's part of AS 65002 :

Switch_sites>sh ip bgp vpnv4 vrf ce_R2 neighbors 10.245.32.2
BGP neighbor is 10.245.32.2,  vrf ce_R2,  remote AS 6475,  local AS 65002 no-prepend replace-as, external link
  BGP version 4, remote router ID 3.3.3.3
  BGP state = Established, up for 1d00h
  Last read 00:00:30, last write 00:00:06, hold time is 180, keepalive interval is 60 seconds

I didn't know which one had precedence but it seems logical as you explained that it's the former.

I will try your two suggestions and post a feedback.

Thanks,

Jad.

Jad KAROUT · ‎04-08-2011

Just tried it and it works thanks !