Redundancy Design

[WILLIAM STEGMAN]

                            Nexus2k---Nexus5k   

                           /          |

Server(NIC Team)             |(VPC)   

                           \          |    

                            Nexus2k---Nexus5k

My basic design for our new data center is illustrated above. I’m trying to determine how the default gateway for servers will behave when one of the links to either a 2k or 5k will fail. Where do I define that layer 3 gateway? I was hoping the VPC feature would allow me to define it once on either of the 5ks allowing that layer 3 interface to simultaneously exist on both 5ks, much like it would on a 6500 with dual supervisors, and avoiding the hassles associated with HSRP. Is that an option with the Nexus switches and does my design allow that option to work?

Thank you

[WILLIAM STEGMAN]

The formatting didn't come out right on the design illustration.  Basically, each 2k only connects to 1 5k and a VPC exists between the 5ks.  Those 2 vertical lines under and above the Nexus 2k should be under the Nexus5k, creating a line that connects the 2 Nexus5ks.

[Jerry Ye]

My question is what model is the N5K? The older 50x0? or 55xx?

Regards,

jerry

[WILLIAM STEGMAN]

They're the newer 55xx, 5548s.

[Jerry Ye · Accepted solution]

You will need to use HSRP as the FHRP to forward traffic out of the Nexus L2 network, there is no way around this.

HSRP in Nexus behaves a little bit different from IOS routers. When vPC forward traffic to the L2/L3 boundary Nexus, HSRP will forward the traffic even though it is not the HSRP primary router. Please see the following link on how vPC interact with HSRP

http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9402/white_paper_c11-516396.html

So for your 2nd question, not possible to have the Nexus behave like a C6K with dual SUP.

Regards,

jerry