[c-nsp] Mpls with layer-2 in between

[Scott Miller]

On the 3850, can you do something like:

mpls ldp router-id Loopback0

would probably take bouncing all the ldp neighbors for it to take effect.

On Mon, Oct 12, 2020 at 12:38 PM Shawn L <shawn at rmrf.us> wrote:

> I have kind of a strange situation.  Trying to figure out what to do moving
> forward, and not sure what the best (or quickest) way to solve the issue.
>
> I have a mpls cross-connect (pseudowire) between 2 sites.  The link comes
> up, but I can't pass any traffic.
>
> sh mpls l2transport vc - shows the connection is up
> ping mpls pseudowire x.x.x.x YY reply mode ipv4 - gives me a time-out.
>
> The link goes from an ASR920 - Cisco 3850 ------- Cisco 3850 - ASR920
>
> The long dashed line is a 40-gig layer-2 link between the 2 Cisco 3850
> switches.  Both switches run eigrp and terminate mpls links just fine; ie
>
> interface TenGigabitEthernet1/0/4
>  no switchport
>  ip address 10.10.0.73 255.255.255.252
>  mpls ip
>  mpls label protocol ldp
>
> ASR920 ---- 3850 ---- ASR920 pseudowires work fine.  It's an issue when the
> 2nd 3850 comes into the mix.
>
> When I do a traceroute from end-to-end, I see the vlan 1 interface of the
> 3850, not the loopback interface.  So, I'm guessing it's taking the layer-2
> path instead of using the loopback interfaces and routing it.
>
> The best way to fix it is to remove the 3850s from the equation and put in
> real routers.  Unfortunately that means I can't utilize the 40 gig link to
> the other site.
>
> Is it possible to have a layer-2 connection in-between 2 mpls speaking
> routers?  ie
>
> ASR920 --- 'new' router --- 3850--------3850 -- 'new' router ---- ASR920
>
> Or, is there a way to tell the 3850s to route the connection instead of
> switch it (and not break switching on them)?
>
> Thanks
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>