Routing

here is the questiondo we need to ping remote ce via the interface

direct route is missed

...............

Hi ,

You don't need CE routes in inet.0 table.

Please configure "vrf-table-label" inside the routing instance.

You should be able to ping from CE to CE

set logical-systems r2 routing-instances vpn-r1 vrf-table-label

set logical-systems r4 routing-instances vpn-r6 vrf-table-label

Regards

Moses N

-------------------------------------------------------

If this post answers your question, please mark it as "Accepted Solution".
Kudos are a nice way of expressing your gratitude

hi,Moses

it works now

jucao# run ping 6.6.6.6 source 1.1.1.1 logical-system r1    
PING 6.6.6.6 (6.6.6.6): 56 data bytes
64 bytes from 6.6.6.6: icmp_seq=0 ttl=63 time=5.256 ms
64 bytes from 6.6.6.6: icmp_seq=1 ttl=63 time=0.820 ms
^C
--- 6.6.6.6 ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.820/3.038/5.256/2.218 ms

[edit logical-systems r4 routing-instances vpn-r6]
jucao#

1:would u like to tell me why I need to add set vrf-table-label

what does this command mean?

2:another question is:

if I only configured rsvp in this topo. will it use label switching?

in some video,I saw tractouter some ip will use label switching

but in some other video,it tells me only nexthop of bgp will install entry in inet.3

Hi Robert,

Please check this forum topic about vrf-table-label.

http://forums.juniper.net/t5/Routing/with-and-without-vrf-table-label-enabled-on-PE-router/m-p/157218#M7163

Basically, this is necessary to configure this , if the CE-PE interface is a multi-access media ( eg- Ethernet).

2) Yes, RSVP in your case here is used for distributing the labels. RSVP will install routing information in inet.3 table and BGP will use this table for next-hop resolution.

Regards

Moses N

-------------------------------------------------------

If this post answers your question, please mark it as "Accepted Solution".
Kudos are a nice way of expressing your gratitude

hi,

still a little confused here

with VTL configured

Route                     next-hop      allocated label

10.0.1.0/24           10.10.10.1   16

192.168.1.0/24    10.10.10.1   16

192.168.2.0/24     20.20.20.1  16

even label is different ,

10.0.1.0/24           10.10.10.1   16

192.168.1.0/24    10.10.10.1   17

192.168.2.0/24     20.20.20.1  18

it doesn't matter per my understanding ,right?

the point is :VTL make the junos check the ip header and get layer 2 mac when compared wirh P2P link,right?

 P2P link also need lable pop and layer 3 lookup,but there is no layer 2 mac in p2p link

................................................

Can you please post the following out puts from r2,r4 and r5?

show mpls lsp extensive
show rsvp session extensive

Attach the output in a text file, that will make easy to read this thread.

Regards,

Moses N

............................

Hi Robert,

In L3 VPN case, MP-BGP advertises the VPN label to the remote PE , and RSVP distributes the label  to reach the next-hop of bgp routes.

In your case, if you have a direct LSP between r2 and r4, then r4 will advertise label 3 which instructs r2 to remove the outer label. ( The default PHP behavior).

When you have the LSP path r2-r5-r4, you can see the outer label pushed by r2.

You can check the labels  in mpls.0 table by  "show route table mpls.0".  ( check this in r5)

Yes. With VTL configured, VPN label will be popped, layer 3 lookup will be done, ARP resolution ,,, Layer 2 header will be constructed and forwarded out of the interface.

Regards,

Moses N

great thanks

you are the man!!

Not a problem Robert,

The command is correct. But you won't be able to ping since your VPN Export policy is not allowing directly connected routes to be sent to remote PE / CE.

You might need to add the following term inside your export policy.

<policy-options policy-statement export-vpn-r1>

term 2 {
    from protocol direct;
    then {
        community add vpn-r1;
        accept;
    }
}

simillarly, if you want to ping r1 (CE)  from r4 (remote PE) you need to add the same term in export-vpn-r6 policy.

Rgds,

Moses N

cool!!

Hi Robert,

In Juniper implementation, by default BGP only uses the MPLS paths for forwarding.

inet.3 table is not used for forwarding, but it is used by BGP to resolve the next-hop.

There are many ways to copy / move the routes in inet.3 table to main table inet.0, so that MPLS paths can be used to route IGP destinations.

Please read this document about routing tables related to MPLS

http://www.juniper.net/techpubs/software/junos/junos93/swconfig-mpls-apps/mpls-and-routing-tables.html

VPNs rely on the routes in the inet.3 routing table to function properly.

if you configure " set protocols mpls traffic-engineering bgp-igp" then all routes in inet.3 will be moved to inet.0 an hence VPNs will not work properly.

 traffic-engineering bgp-igp-both-ribs solves this issue by copying the routes from inet.3 to inet.0 ( both tables will have the routes)

Also, please check this link as well.

http://www.juniper.net/techpubs/en_US/junos9.5/information-products/topic-collections/config-guide-mpls-applications/mpls-configuring-traffic-engineering-for-lsps.html

Regards,

Moses N