Support and News Forum

Hello.

I experience the following behaviour of a private cloud network:
I have two different nodes connected to the same private cloud network in two labs each.
The communication between the nodes accross the two labs is working well.

But as soon as I disable the interface connected to this private cloud on only one of the four nodes, the link on all other thee nodes is also going down.
It looks like that the private cloud singals link UP to the connected nodes only when all connections (4 out of 4) are up, but as soon as even one connection is down all others are also singaled "no link".

Is this correct?
Is there a way to overcome this?

Thank you very much,
Karl

Hi,

first question are you using L3 devices/routers interconnect labs? The behave which you wrote was observed with vEOS Arista switches, when switchport connected to private cloud send to suspend link, it disable all interfaces which are connected to private cloud.
Recommendation is to use L3 device connect labs via private cloud.
Also, please check what MAC address you have on all interfaces connected to private cloud, MAC address must be different! Same MAC will make only collisions in your internconnection.
Example I made 4x labs connected to Private cloud test.
Device connected to private cloud: vIOS router. Suspending some lab connection does not affect others.
MAC address I changed manually on the connected interfaces:
1. add node on topology
2. right click edit
3. Set different MAC for interface
in the pic below: Lab1, router has MAC 50:00:00:01:00:00
in Lab2 router interface has MAC 50:00:00:03:00:00

Hi.

Thank you for your answer.

I exactly use Arista VEOS-LAB switches in my scenario.
The interfaces are configured as "no switcheport" (= L3 routed port) but it matches your description.

I experience it a little different than you:
-- If I disable the interface on the vEOS node, than all other interfaces conneted to the same private cloud go down.

-- If I just suspend the link by eve than the other links stay up.
Doing so, I can achive my goal for redundancy tests.

But good to know, that it is a well know behaviour.

Thank you very much for your support.

Sad to say, but it is just Arista hawing such issue, even L3 port