cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
5869
Views
15
Helpful
7
Replies

Multi-Pod APIC Migration

I´m going to add a second POD to the Fabric and I have a question about one APIC.

Right now 3 APICS are on POD 1 . But once I configure Multi-Pod and both PODs are up, I would like to move one of the APICs to the second POD. To do so, I understand I need to wipe/reset the APIC and apply the script again.
My question is.... should the TEP pool be the one from Pod 1 or Pod 2 (during the basic configuration once you reset it)?
I have seen on documentations that all APICs have addressing from the POD 1 (even if they are on POD2) and this is confusing.

 

For example one of the white papers have this:

The final step is disconnecting one of the APIC nodes from Pod1 and connecting it to Pod2. The APIC node should be reset and the initial setup script should be run again to ensure it can get connected to the new Pod. It is worth noticing that an APIC node keeps the same IP address (assigned from the TEP Pool valid for the fabric it was initially connected).

 

So, when I reset the APIC and move it to the new POD, should I put the TEP pool from Pod 1 or 2? and what about the pod number as well? (Just to confirm this one as well)

1 Accepted Solution

Accepted Solutions

Francesco Molino
VIP Alumni
VIP Alumni

Hi

Apic on pod2 should have the same infra vlan and tep pool as pod1


Thanks
Francesco
PS: Please don't forget to rate and select as validated answer if this answered your question

View solution in original post

7 Replies 7

Francesco Molino
VIP Alumni
VIP Alumni

Hi

Apic on pod2 should have the same infra vlan and tep pool as pod1


Thanks
Francesco
PS: Please don't forget to rate and select as validated answer if this answered your question

In a Multi-Pod setup, the configuration of the APIC clusters should be specific to the pod to which they belong. Because the Tunnel Endpoint (TEP) pool for each pod must be unique, each pod has its own TEP address pool.

https://www.cisco.com/c/en/us/solutions/collateral/data-center-virtualization/application-centric-infrastructure/white-paper-c11-739714.html

 

Is there any explanation why Cisco doc calls for  TEP poll per pod? BTW that is the configuration we run for over 4 years in production fabric. 

I'm not sure i understand your question.
Devices on pod2 will get an ip within their own tep pool assigned.


Thanks
Francesco
PS: Please don't forget to rate and select as validated answer if this answered your question

Hi Francesco,

The question was, is there are any reason you recommendation contradicts Cisco documentation ?  

Here is your post :

"Apic on pod2 should have the same infra vlan and tep pool as pod1"

 

and here is Cisco Application Centric Infrastructure Multi-Pod Configuration White Paper :

"In a Multi-Pod setup, the configuration of the APIC clusters should be specific to the pod to which they belong. Because the Tunnel Endpoint (TEP) pool for each pod must be unique, each pod has its own TEP address pool. This setup enables east-west communication for endpoints connected to separate pods via VXLAN tunnels established between the specific leaf nodes VTEP addresses taken from those not overlapping TEP address pools."

 

So, APIC in POD1 /POD2 should have the same infra VLAN, Fabric ID, but own unique TEP address pool.  That is Cisco design and my own practical experience. 

I'm not contradicting Cisco documentation, this is the way it should be.

You can take a look on that documentation and you'll see all APICs are in the same subnet:

https://www.cisco.com/c/en/us/solutions/collateral/data-center-virtualization/application-centric-infrastructure/white-paper-c11-739714.html

 

For Spines and leaves, they will be part of the local tep pool.


Thanks
Francesco
PS: Please don't forget to rate and select as validated answer if this answered your question

This is a bit misleading in the documentation.

 

When asked for the TEP pool in the initial setup dialog on an APIC, you have to give the seed PODs TEP pool, which is always the TEP Pool of POD1!

This is because the APIC itself will get a TEP IP from that pool to build the cluster.

 

When creating POD2 in the Fabric Inventory, you'll have to give it a new TEP pool which is used to hand out TEP IPs for spines and leaves in POD2.

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:

Save 25% on Day-2 Operations Add-On License