Dup Pings

[kevinfowler]

I have just implemented a redundant network with redundant core and distribution nexus switches.   When I ping from "Core-1" ping runs normal.  However when I run a ping from any of the other 3 switches I get DUP ping replies.   Is this normal in a redundant network or do I have some kind of issue I'm not familiar with?

[MHM Cisco World]

This normal I think'

If you use nexus peer gateway under domain 

Then both NSK vPC reply to any ping to VLAN SVI or VIP of hsrp.

Do traceroute and see the IP it appear 

MHM

[kevinfowler]

I am using nexus peer gateways to both the core and the distro switches.   When I do a traceroute from the distro switches it just goes right to the IP.   All of the research online says if you have a DUP ping you have network issues but nobody is addressing the question if the network is redundant.  

[MHM Cisco World]

From SW you ping' show arp did you see two MAC or one (after traceroute)

MHM

[kevinfowler]

All of my ARP's done on the core's only show one mac address.

[MHM Cisco World]

Show mac in both NSK' do you see same MAC in both NSK ?

MHM

[kevinfowler]

Yes sir....All the MAC addresses are the same across both Core switches.  

[MHM Cisco World]

https://community.cisco.com/t5/switching/nexus-5500-duplicate-icmp-echo-replay/td-p/2083784

check this link 
as I guess this normal in vPC NSK with peer gateway under vpc domain 
the solution is no ip redirect in interface of NSK 
but to be sure share topology and the IP you ping TO 
thanks 

MHM

[kevinfowler]

Do I need to put that command on the port channel interfaces as well?   It's already on all of my interface vlans on both my core devices. 

[MHM Cisco World]

this command in for L3 interface so it need only under the VLAN SVI 
thanks 
MHM

[Ruben Cocheno]

@kevinfowler 

Normal behavior due VPC, disable the ip icmp redirects on that interface and that DUPs should not come up again

[kevinfowler]

Actually I have disable the icmp redirects and it's still happening.  

[Kevin15]

I'm still having this problem.   Where did you read this is normal behavior due to VPC being implemented?