Thanks for clarifying Quinn - So on a pair of 3Ks, a "typical" VPC setup would be 2 x 10Gb links + "a" link(i.e. Management ports) for the keepalives?
And on a pair of 9Ks, 2 x 40Gb links, plus management port link? Cheers ________________________________________ From: quinn snyder <[email protected]> Sent: Friday, 20 November 2015 8:13 AM To: CiscoNSP List Cc: cisco-nsp Subject: Re: [c-nsp] Nexus / VPC - Management port "needed" in VPC? > On Nov 19, 2015, at 14:07, CiscoNSP List <[email protected]> wrote: > > We have a customer that is wanting to do VPC on some N9Ks and also N3Ks - I > "thought" VPC would be similar to VSS...i.e. dual link between the > switches...but my (brief) reading up on the setup, I see some setup guides > where there are dual links(2 x 10Gb, or 2 x 40Gb), plus the use of the > management port for vPC peer keepalives? > > > http://www.cisco.com/c/en/us/products/collateral/switches/nexus-3000-series-switches/white_paper_c11-685753.html > > Any info on the "correct" method to setup VPC on the Nexus would be greatly > appreciated the above is correct. vpc requires the “data plane” (vpc peer-link) that performs synchronization using cfsoe between vpc domain peers. it also *can* be used to forward actual data-plane traffic under failure scenarios. its important to understand the baked-in vpc drop conditions that exist to provide loop prevention under steady-state. the management (or some other set of layer-3 adjacencies within an isolated vrf) are used for simple heartbeats between the devices. failure of this link does not mean catastrophic failure of the domain. this is similar to something like ‘fast-hellos’ using an oob link when dealing with vss. q. -- quinn snyder | [email protected] _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
