[f-nsp] 3-node MPLS with MLXs

Adam E Barratt (Network Operations) abarratt at socket.net
Mon Jun 28 12:50:22 EDT 2010


Grady,

Do you have an underlying IGP that you're running for the LDP?  (ISIS or 
OSPF)?  I'm thinking your LDP is just using the connected routes between 
the MLXs and when you fail a link, it has no IP route between them so it 
doesn't have a path.  The BGP will be used to transfer customer routes 
between PEs... if the MLXs are simply P (core) devices and not facing 
customers, then there is no need for BGP on them (unless you're using 
BGP for the IGP, but I don't like the hold-down time for it to react to 
a link/route failure).

-Adam

-- 
Adam E Barratt
Project Engineer
SOCKET
(573) 817-0000 ext. 223
abarratt at socket.net
1-800-SOCKET-3  www.socket.net



> Message: 4
> Date: Mon, 28 Jun 2010 08:36:21 -0500
> From: Grady Neely<grady.neely at adtran.com>
> To: foundry-nsp<foundry-nsp at puck.nether.net>
> Subject: [f-nsp] 3-node MPLS with MLXs
> Message-ID:<21AC538E-761B-49CB-89F7-B363F1DD429A at adtran.com>
> Content-Type: text/plain; charset=us-ascii
>
> Hi all,
>
> I have 3 MLX nodes connected in a ring configuration via 10gig links, call them R1, R2, R3.  I also have MPLS up and running via BGP between the three.
>
> All works well as long as all the phy links are up.  But if I fail one of the 10gig links, i.e. the link between R2 and R3, the MPLS tunnel between them goes down and does not reroute around the other side of the ring, via R1.  Is this normal?
>
> I am new to MPLS, but being a routed protocol, I was hoping it would see the route along R1 and re-establish the MPLS tunnel.
>
> Here are the MPLS configs of each router:
>
> R1 (10.27.250.1)
>
> router bgp
>   local-as 1027
>   neighbor 10.27.250.2 remote-as 1027
>   neighbor 10.27.250.2 update-source loopback 1
>   neighbor 10.27.250.3 remote-as 1027
>   neighbor 10.27.250.3 update-source loopback 1
>
>   address-family ipv4 unicast
>   redistribute connected
>   redistribute static
>   redistribute ospf match internal
>   redistribute ospf match external2
>   exit-address-family
>
>   address-family ipv4 multicast
>   exit-address-family
>
>   address-family ipv6 unicast
>   exit-address-family
>
>   address-family ipv6 multicast
>   exit-address-family
>
>   address-family vpnv4 unicast
>   exit-address-family
>
>   address-family l2vpn vpls
>   neighbor 10.27.250.2 activate
>   neighbor 10.27.250.2 send-community extended
>   neighbor 10.27.250.3 activate
>   neighbor 10.27.250.3 send-community extended
>   exit-address-family
> end of BGP configuration
>
> router mpls
>
>
>    mpls-interface ve8
>    ldp-enable
>
>    mpls-interface ve9
>    ldp-enable
>
>    mpls-interface ve10
>    ldp-enable
>
>   vpls MGMT 3
>    auto-discovery
>    vlan 3
>     tagged ethe 1/1 eth 1/4 eth 1/7
>
>
> R2 (10.27.250.2):
>
> router bgp
>   local-as 1027
>   neighbor 10.27.250.1 remote-as 1027
>   neighbor 10.27.250.1 update-source loopback 1
>   neighbor 10.27.250.3 remote-as 1027
>   neighbor 10.27.250.3 update-source loopback 1
>
>   address-family ipv4 unicast
>   redistribute connected
>   exit-address-family
>
>   address-family ipv4 multicast
>   exit-address-family
>
>   address-family ipv6 unicast
>   exit-address-family
>
>   address-family ipv6 multicast
>   exit-address-family
>
>   address-family vpnv4 unicast
>   exit-address-family
>
>   address-family l2vpn vpls
>   neighbor 10.27.250.1 activate
>   neighbor 10.27.250.1 send-community extended
>   neighbor 10.27.250.3 activate
>   neighbor 10.27.250.3 send-community extended
>   exit-address-family
> end of BGP configuration
>
> router mpls
>
>
>    mpls-interface ve3
>    ldp-enable
>
>    mpls-interface ve4
>    ldp-enable
>
>   vpls MGMT 3
>    auto-discovery
>    vlan 3
>     tagged ethe 4/3 to 4/5 ethe 4/7 ethe 4/13 ethe 4/19
>
> R3 (10.27.250.3):
>
> router bgp
>   local-as 1027
>   neighbor 10.27.250.1 remote-as 1027
>   neighbor 10.27.250.1 update-source loopback 1
>   neighbor 10.27.250.2 remote-as 1027
>   neighbor 10.27.250.2 update-source loopback 1
>
>   address-family ipv4 unicast
>   redistribute connected
>   exit-address-family
>
>   address-family ipv4 multicast
>   exit-address-family
>
>   address-family ipv6 unicast
>   exit-address-family
>
>   address-family ipv6 multicast
>   exit-address-family
>
>   address-family vpnv4 unicast
>   exit-address-family
>
>   address-family l2vpn vpls
>   neighbor 10.27.250.1 activate
>   neighbor 10.27.250.1 send-community extended
>   neighbor 10.27.250.2 activate
>   neighbor 10.27.250.2 send-community extended
>   exit-address-family
> end of BGP configuration
>
>
> router mpls
>
>
>    mpls-interface ve2
>    ldp-enable
>
>    mpls-interface ve3
>    ldp-enable
>
>   vpls MGMT 3
>    auto-discovery
>    vlan 3
>     tagged ethe 4/1 to 4/4 ethe 4/7 ethe 4/13 ethe 4/19
>
>
> ------------------------------
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp at puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>
> End of foundry-nsp Digest, Vol 89, Issue 22
> *******************************************
>
>    




More information about the foundry-nsp mailing list