ACI Multi-POD HA Design

[a.azambuja]

What is the best design with Multi-POD when using 04 APIC Controllers?

02 Active controllers at DC01 and 01 Active controller at  DC02 + 01 controller in standby mode? 

[RedNectar]

Hi @a.azambuja,

You indeed answered your own question:

What is the best design with Multi-POD when using 04 APIC Controllers?

02 Active controllers at DC01 and 01 Active controller at  DC02 + 01 controller in standby mode? 

---

You assign each APIC Node a unique OOB Management IP address, but assign Node IDs 1 & 2 to the APICs in DC 1 and node ID 3 to the active APIC in DC2. The standby APIC is assigned Node ID 2, it's own OOB management IP and everything else the same as APIC 3. And then you power-down the Standby APIC so that you don't get error messages about there being two APICs with ID=2. It's a good idea to have the CIMC configured on the Standby APIC so you can power it up remotely if needed.

But you need to understand that should there be a disaster that separates the two DCs, DC1 will continue in R/W mode, while DC2 is in read only mode, and will continue that way until the connectivity is restored. The standby APIC can't keep the second DC in RW mode, but policy will still be applied and traffic will still flow.

If you wish to bring the Standby APIC online (because DC1 has burned down or some other disaster), there is a special procedure to go through which is described in the ACI Stretched Fabric Design paper.

Links:

https://www.cisco.com/c/en/us/solutions/collateral/data-center-virtualization/application-centric-infrastructure/white-paper-c11-737855.html#_Toc521671382

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/kb/b_kb-aci-stretched-fabric.html#concept_524263C54D8749F2AD248FAEBA7DAD78

I hope this helps

---

Don't forget to mark answers as correct if it solves your problem. This helps others find the correct answer if they search for the same problem

---

[richmond]

With later releases the standby APIC is numbered from node ID 20-29. Typically 21 is used for the first standby, 22 for the second and so-on.

ACI_Best_Practices Reference:

"When APIC redundancy is configured, you should use IDs 1 to 19 for active APICs and IDs 20 to 29 for standby APICs. This allows for expansion of the fabric."

It's kept powered on and will receive software updates but will not receive any configuration. When you need to activate it you replace an existing controller with the standby. The procedure is documented here:

APIC_Getting_Started_Guide_Rel_3_x

When you replace an APIC it wipes the replaced APIC if reachable. If not reachable (for example because of a failure) then you can run into some interesting scenarios where both APICs think they are node 2 for example. This is resolved by wiping and re-joining the APIC to the cluster. I'd recommend opening a TAC case if you are in this position though.