VPCs Peer is Dead

[hauret]

I have a pair of Nexus 5596 switches that are configured to be VPC peers.  One of the switches is “dead” and I cannot get the active VPC member to restore the configured VPCs to resume operations.

The current config of the live switch (VPC peer) is:
VPC domain 1

role priority 1000

peer-keepalive destination 10.1.99.42 source 10.1.99.41

delay restore 150

peer-gateway

auto-recovery

[MHM Cisco World]

peer-keepalive destination 10.1.99.42 source 10.1.99.41 vrf default 

[hauret]

I’ve placed that command but the VPC are not recovering. Is there a subordinate command that is necessary to run after adding the vrf default?

[MHM Cisco World]

Peer keeplive is in global VRF or managment VRF ?
if it not in global VRF then same command above instead use 
vrf mgmt 

[hauret]

Peer keepalive is in global. However, to confirm, I changed it to mgmt, then in viewing the show VPC brief the keep-alive status shows “Suspended (Unusable VRF)”.

[MHM Cisco World]

peer-keepalive destination 10.1.99.42 source 10.1.99.41 vrf default <<-NSK-1

peer-keepalive destination 10.1.99.41 source 10.1.99.42 vrf default  <<-NSK-2

make Keep-alive in global and check above note 

NOTE:- vpc domain must be same 

[hauret]

Thank you MHM. I do have the surviving switch NSK-1 configured as suggested. I don’t have an NSK at current to modify setting since that switch did not survive a power event.  Hopefully there is a way to get the surviving switch to make the configured VPCs active.

[MHM Cisco World · Accepted solution]

You are So So welcome. 
if you have one NSK 
only run PO between NSK and Access SW without vPC.
thanks 

[hauret]

Removing the vPC does solve the current issue at hand.  Awesome suggestion. I was completely overlooking the easy answer there. Thank you MHM.

[balaji.bandi]

I followed the below steps - when we replaced failed one.

1. make sure both are the Nexus running the same version of the code.

2. make sure peer connected interface in shutdown (before you introduce the new failed switch in to network) like peer-link, and keepalive ports.

3. note down the config of STP, VTP, and VLAN this should match both the same and identical - configured on the new replaced switch as same as live one.

4. if you have any FEX configure the same config (for the fex to have dual home)

5.  disable auto-recovery, so there is no role changes occurring while the new switch added to vpc .

6.  role priority needs to be changed better than 1000 may be 1001 or more. (lower priority will be active all time)

7.  check the config on the new device offline, make sure config is identical except some parameters mentioned like priority

and peer-keepalive destination 10.1.99.41 source 10.1.99.42 ( we reverse this config on the new device)  - check is there any VRF configured for this, if so add vrf end of line.

8. Connect only keepalive link

9. bring up the interface up

and validated

show vpc

show vpc consistency checks

show vpc peer-keepalive

Note : make sure any features enabled - should be same as active one.

Once all working connect  peer-link and do the testing.

If all is good then you can connect end device Links _ bring up each service one at a time and do the testing.

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/nexus5000/sw/operations/interfaces/602_n1_1/N5K_Interfaces_Ops_Guide/vpc_ops.html

[hauret]

Thanks for the guide above, unfortunately, I don’t have a peer switch that I can use for the peer link.  For the meantime, all I have is the single switch so I am trying to get it to make active the configured VPC.  We had a power event that downed the second switch and am working towards obtaining a replacement. I’ll keep your instructions for when I do get to add a new peer.

[balaji.bandi]

You should not modify any configuration as I suggest, you should not change anything until any emergency requires.

if the devices are dual home they should work as expected.