cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
3036
Views
5
Helpful
2
Replies

Nexus 9300 BFD to NSX-T tier 0 gateway

Patrick Colbeck
Level 3
Level 3

I have a fabric running VXLAN with DCNM 11.4.1 and the compute is running the latest VMware cloud foundation with NSX-T.

To get to the NSX-T world VMware runs a Tier0 gateway that is essentially a virtualised router running eBGP so far as the Cisco network is concerned.

So using the services dialogue I created route peerings to the Tier0 gateway from loopbacks on the leaf switches. This works well and the eBGP peers come up fine.

Now the customer wants BFD enabling for these eBGP connections to reduce failure detection time and its the VMware recommendation.

Its enabled in VMware and I added it to the peering on the 9300 leaf switches by modifying the eBGP peer template.

Unfortunately BFD does come up, its configured but doesnt detect the VMware end.

The 9300 leaf eBGP section looks like this:

 

vrf mdft
neighbor 10.16.160.1
bfd
remote-as 65258
password 3 2fcd630ce0b69502
update-source loopback3
ebgp-multihop 5
address-family ipv4 unicast
send-community
send-community extended
route-map extcon-rmap-filter out

The peering goes loopback -> VLAN -> VMware IP on the VLAN

The bfd looks like this:

# show bfd neighbors vrf mdft details




OurAddr         NeighAddr       LD/RD                 RH/RS           Holdown(mult)     State       Int                   Vrf                              Type

10.16.98.152    10.16.168.1     1090519041/0          Down            N/A(3)            Down        Vlan1401              mdft                             SH







Session state is Down and not using echo function




Session type: Singlehop

Local Diag: 0, Demand mode: 0, Poll bit: 0, Authentication: None

MinTxInt: 2000000 us, MinRxInt: 2000000 us, Multiplier: 3

Received MinRxInt: 0 us, Received Multiplier: 3

Holdown (hits): 0 ms (0), Hello (hits): 2000 ms (100)

Rx Count: 0, Rx Interval (ms) min/max/avg: 0/0/0 last: 174576 ms ago

Tx Count: 100, Tx Interval (ms) min/max/avg: 1689/1689/1689 last: 642 ms ago

Registered protocols:  bgp

Downtime: 0 days 0 hrs 3 mins 7 secs

Last packet: Version: 0                - Diagnostic: 0

             State bit: AdminDown      - Demand bit: 0

             Poll bit: 0               - Final bit: 0

             Multiplier: 3             - Length: 24

             My Discr.: 0              - Your Discr.: 0

             Min tx interval: 0        - Min rx interval: 0

             Min Echo interval: 0      - Authentication bit: 0

Hosting LC: 1, Down reason: No Diagnostic, Reason not-hosted: None

Any ideas on how to get BFD running?
2 Replies 2

Patrick Colbeck
Level 3
Level 3

The NSX guys say we are using bfd not bfd multihop from their fault finding. Can I force bfd multihop?

Patrick Colbeck
Level 3
Level 3

And I think I've found it. The BFD documentation for Nexus 9Ks is terrible but I found a reference to "bfd multihop" as a new command in 9.3.6 under eBGP peers. Tried it in the lab on 9000v and it seems to work (cant test fully as the 9000v lets you configure BFD but cant actually run it). We will upgrade the customers leaves to 9.3.6 in the new year when we can get change control approval.

The weird thing is that the bfd multihop timer commands are there in NXOS < 9.3.6 just not the plain bfd multihop command to switch it on.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: