Routing
Highlighted
Routing

EVPN-VXLAN without LEAF?

‎03-31-2020 09:33 AM

Hi Expert,

 

I'm just found the article for this url https://www.juniper.net/documentation/en_US/release-independent/nce/topics/concept/nce-178-collapsed...

 

Based on the url it's have a detail explanation. But i have a question for scalablity if i'm using this kind of setup. Below is my question:

 

1.) Lets say i start with just two spine for first deployment and lets say after one year i have a budget to purchase LEAF. So is it too much revamp / total revamp on current setting if i add LEAF?

 

2.) If it too much redesign back due add LEAF so is it just small change need if i add more SPINE instead of add LEAF?

 

My purpose is the pay as per growth. I will try to simulate the design on my EVE-NG base on the url.

 

Thanks and appreciate any feedback

 

3 REPLIES 3
Highlighted
Routing

Re: EVPN-VXLAN without LEAF?

‎03-31-2020 09:57 AM

Hi,

 

This should be possible. If you only have one layer, you just do a full-mesh between them. When you add another layer, just do a full mesh between spine and leaf layer, but not between spine and spine, and not between leaf and leaf 


Mengzhe Hu
JNCIE x 3 (SP DC ENT)
Highlighted
Routing

Re: EVPN-VXLAN without LEAF?

‎03-31-2020 09:17 PM

Hi Mhu,

 

So u means if i add new LEAF in the future, then i need disconnect back-to-back link between SPINE1 and SPINE2 right? Also i need to disable the bgp configuration between SPINE1 & SPINE2 right?

 

Please correct me if i wrong.

 

Thanks

 

Highlighted
Routing

Re: EVPN-VXLAN without LEAF?

‎04-02-2020 11:07 AM

Hi all,

 

Based on the url given previously, i read paragraph like below it have statement "choose one spine switch to be the standby switch". May i know how to choose spine switch as standby? Is it any hidden knob?

 

Appreciate any feedback

 

 

Split-Brain State

How to Prevent a Split-Brain State

Problem

 

If the links between the spine switches are down, causing the BGP peering to go down, both spine switches are active and forwarding. The downstream aggregated Ethernet interfaces are active and forwarding. This scenario is known as a split-brain state and can cause multiple problems.

 

Solution

To prevent this issue from occurring, choose one spine switch to be the standby switch. Use an event script on the standby switch that is triggered by a BGP peer down event. When the BGP peering goes down, the script checks if the other spine switch is still up using the out-of-band interface. If the other switch is reachable, the script disables the aggregated Ethernet interfaces on the standby switch. The downstream device then only uses the other spine switch to forward traffic.