BGP Free core with MPLS traceroute issue

Hi experts,

I'm currently testing out BGP free core using mpls and i'm missing some traces between endpoint. Here is topology

image

Goal is to tunnel traffic from AS4 to AS5 over the mpls core without running BGP on the P routers. I was able to get reachability but when i trace from one side to the other, the P router cant respond to the trace. Is this normal??

R5(config-router)#do trace 4.4.4.4 so lo 0
Type escape sequence to abort.
Tracing the route to 4.4.4.4
VRF info: (vrf in name/id, vrf out name/id)
  1 35.0.0.3 100 msec 56 msec 32 msec
  2  *  *  *  ----------------------------------------------------R2 cant respond to traceroute
  3 10.0.12.1 76 msec 96 msec 72 msec
  4 14.0.0.4 104 msec 140 msec 108 msec

Find configs below, can copy and paste directly to test.

R1
conf t
int f0/1
 ip add 10.0.12.1 255.255.255.0
 mpls ip
 no sh
int f0/0
 ip add 14.0.0.1 255.255.255.0
 no sh
int lo 0
 ip add 10.0.0.1 255.255.255.255

router ospf 1
 network 10.0.0.0 0.255.255.255 a 0

router bgp 123
 no bgp default ipv4
 nei 10.0.0.3 remote-as 123
 nei 14.0.0.4 remote-as 4
 nei 10.0.0.3 up lo 0

 address-family ipv4
  nei 14.0.0.4 activate
  nei 10.0.0.3 activate
  nei 10.0.0.3 next-hop-self

R2
conf t

int f0/0
 ip add 10.0.12.2 255.255.255.0
 mpls ip
 no shut
int f0/1
 ip add 10.0.23.2 255.255.255.0
 mpls ip
 no shut
int lo 0
 ip add 10.0.0.2 255.255.255.255
router ospf 1
 network 10.0.0.0 0.255.255.255 a 0

R3
conf t
int f0/0
 ip add 10.0.23.3 255.255.255.0
 mpls ip
 no sh
int f0/1
 ip add 35.0.0.3 255.255.255.0
 no shut
int lo 0
 ip add 10.0.0.3 255.255.255.255

router ospf 1
 network 10.0.0.0 0.255.255.255 a 0

router bgp 123
 no bgp default ipv4
 nei 10.0.0.1 remote-as 123
 nei 35.0.0.5 remote-as 5
 nei 10.0.0.1 up lo 0

 address-family ipv4
  nei 35.0.0.5 activate
  nei 10.0.0.1 activate
  nei 10.0.0.1 next-hop-self


R4
conf t
int f0/0
 ip add 14.0.0.4 255.255.255.0
 no sh
int lo 0
 ip add 4.4.4.4 255.255.255.0
router bgp 4
 nei 14.0.0.1 remote-as 123
 network 4.4.4.0 mask 255.255.255.0


R5
conf t
int f0/0
 ip add 35.0.0.5 255.255.255.0
 no sh
int lo 0
 ip add 5.5.5.5 255.255.255.0
router bgp 5
 nei 35.0.0.3 remote-as 123
 network 5.5.5.0 mask 255.255.255.0


Thanks.

Comments

  • The idea of a BGP free core is that the P routers don't run BGP but IGP and LDP so MPLS is needed on the P router. That is why you don't get any response for the second hop. You could tunnel BGP over GRE over the core. There is an ATC video that covers that specifically. The problem with a GRE tunnel and BGP is that there is no dynamic capability for reconvergence. I'm sure you could probably get MPLS l3 VPN up and running to some degree, I haven't played with that before. 


    On Monday, January 26, 2015 10:44 AM, Seyi <[email protected]> wrote:


    Hi experts,

    I'm currently testing out BGP free core using mpls and i'm missing some traces between endpoint. Here is topology


    Goal is to tunnel traffic from AS4 to AS5 over the mpls core without running MPLS on the P routers. I was able to get reachability but when i trace from one side to the other, the P router cant respond to the trace. Is this normal??

    R5(config-router)#do trace 4.4.4.4 so lo 0
    Type escape sequence to abort.
    Tracing the route to 4.4.4.4
    VRF info: (vrf in name/id, vrf out name/id)
      1 35.0.0.3 100 msec 56 msec 32 msec
      2  *  *  *  ----------------------------------------------------R2 cant respond to traceroute
      3 10.0.12.1 76 msec 96 msec 72 msec
      4 14.0.0.4 104 msec 140 msec 108 msec

    Find configs below, can copy and paste directly to test.

    R1
    conf t
    int f0/1
     ip add 10.0.12.1 255.255.255.0
     mpls ip
     no sh
    int f0/0
     ip add 14.0.0.1 255.255.255.0
     no sh
    int lo 0
     ip add 10.0.0.1 255.255.255.255

    router ospf 1
     network 10.0.0.0 0.255.255.255 a 0

    router bgp 123
     no bgp default ipv4
     nei 10.0.0.3 remote-as 123
     nei 14.0.0.4 remote-as 4
     nei 10.0.0.3 up lo 0

     address-family ipv4
      nei 14.0.0.4 activate
      nei 10.0.0.3 activate
      nei 10.0.0.3 next-hop-self

    R2
    conf t

    int f0/0
     ip add 10.0.12.2 255.255.255.0
     mpls ip
     no shut
    int f0/1
     ip add 10.0.23.2 255.255.255.0
     mpls ip
     no shut
    int lo 0
     ip add 10.0.0.2 255.255.255.255
    router ospf 1
     network 10.0.0.0 0.255.255.255 a 0

    R3
    conf t
    int f0/0
     ip add 10.0.23.3 255.255.255.0
     mpls ip
     no sh
    int f0/1
     ip add 35.0.0.3 255.255.255.0
     no shut
    int lo 0
     ip add 10.0.0.3 255.255.255.255

    router ospf 1
     network 10.0.0.0 0.255.255.255 a 0

    router bgp 123
     no bgp default ipv4
     nei 10.0.0.1 remote-as 123
     nei 35.0.0.5 remote-as 5
     nei 10.0.0.1 up lo 0

     address-family ipv4
      nei 35.0.0.5 activate
      nei 10.0.0.1 activate
      nei 10.0.0.1 next-hop-self


    R4
    conf t
    int f0/0
     ip add 14.0.0.4 255.255.255.0
     no sh
    int lo 0
     ip add 4.4.4.4 255.255.255.0
    router bgp 4
     nei 14.0.0.1 remote-as 123
     network 4.4.4.0 mask 255.255.255.0


    R5
    conf t
    int f0/0
     ip add 35.0.0.5 255.255.255.0
     no sh
    int lo 0
     ip add 5.5.5.5 255.255.255.0
    router bgp 5
     nei 35.0.0.3 remote-as 123
     network 5.5.5.0 mask 255.255.255.0


    Thanks.



    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx


  • Sorry, i meant BGP free core. MPLS is enabled on the P routers


    The idea of a BGP free core is that the P routers don't run BGP but IGP and LDP so MPLS is needed on the P router. That is why you don't get any response for the second hop. You could tunnel BGP over GRE over the core. There is an ATC video that covers that specifically. The problem with a GRE tunnel and BGP is that there is no dynamic capability for reconvergence. I'm sure you could probably get MPLS l3 VPN up and running to some degree, I haven't played with that before. 



  • OK, when you reference BGP free core my first thought is that only the PE devices will run BGP, any combination of internal or external BGP. The P routers won't run BGP just IGP and LDP. The PEs will form iBGP will need a loopback configured and advertised into IGP to form the iBGP peerings. 

    Not sure if your new to MPLS or not but check out the v5 MPLS ATC section. That'll help you out.

    Rob


    On Monday, January 26, 2015 3:01 PM, Seyi <[email protected]> wrote:


    Sorry, i meant BGP free core. MPLS is enabled on the P routers

    image rriker:

    The idea of a BGP free core is that the P routers don't run BGP but IGP and LDP so MPLS is needed on the P router. That is why you don't get any response for the second hop. You could tunnel BGP over GRE over the core. There is an ATC video that covers that specifically. The problem with a GRE tunnel and BGP is that there is no dynamic capability for reconvergence. I'm sure you could probably get MPLS l3 VPN up and running to some degree, I haven't played with that before. 






    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx



  • OK, when you reference BGP free core my first thought is that only the PE devices will run BGP, any combination of internal or external BGP. The P routers won't run BGP just IGP and LDP. The PEs will form iBGP will need a loopback configured and advertised into IGP to form the iBGP peerings. 



     

    thanks rob, i configured it just as you said above and i have full reachability. However the issue i'm facing here is why the P router can respond to the trace.

  • I ran into issues with GNS3 and MPLS in the past, a router not responding to a trace as well. I would make sure when you test reachability with labels as you trace from PE 1 loopback to PE 2 looback. There should be an LSP between them. Show mpls forwarding-table, sh mpls ldp nei. Blow away the IGP and rebuild it, try a reload or blow away the router altogether and rebuild from scratch. 

    For the reasons above is why I moved to the CSRs, but still have to reload a couple times per week. 

    HTH
    Rob


    On Monday, January 26, 2015 4:20 PM, Seyi <[email protected]> wrote:


    image rriker:

    OK, when you reference BGP free core my first thought is that only the PE devices will run BGP, any combination of internal or external BGP. The P routers won't run BGP just IGP and LDP. The PEs will form iBGP will need a loopback configured and advertised into IGP to form the iBGP peerings. 



     

    thanks rob, i configured it just as you said above and i have full reachability. However the issue i'm facing here is why the P router can respond to the trace.



    INE - The Industry Leader in CCIE Preparation

    http://www.INE.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx


  • image

     

     

    image

     

    image

     

     

    used your exact same config, mine worked.

     

    attached are images you can use to verify your label push and pop on r3 - r2 which is where your verification was different from mine. check the mpls forwarding table of r2 and 3 to confirm r2 is feeding r3 a label for r3's nex hop ip to the bgp route

     

    if exact same config worked on mine then problem is not with config, have u tried reloading ?

  • Interesting!!, what IOS version are you using?

    Mine is

    Cisco IOS Software, 7200 Software (C7200-ADVENTERPRISEK9-M), Version 15.2(4)S2, RELEASE SOFTWARE (fc1)

    used your exact same config, mine worked.

     

    attached are images you can use to verify your label push and pop on r3 - r2 which is where your verification was different from mine. check the mpls forwarding table of r2 and 3 to confirm r2 is feeding r3 a label for r3's nex hop ip to the bgp route

     

    if exact same config worked on mine then problem is not with config, have u tried reloading ?

Sign In or Register to comment.