Nexus 5500 VPC failover

[bikespace]

Hi All,

I have a Nexus 5K setup with two Nexus 5548 with layer 3 daughtercard. It’s not a particularly standard setup using them as a standard LAN core, with multiple VPC’s to various edge switches and HSRP on the Nexus'.

There’s also a few non VPC connections for which we have an entirely separate trunk. One VLAN providing layer 3 point to point connectivity between the Nexus, the other non VPC related VLANs trunked across to provide standard layer 2 connectivity.

Multiple peer links and separate keepalive link currently using directly connected management ports.

I believe my config would have worked on a 7K (I hope to test that in the next few days), but it’s eluding me on the 5K.

The problems have come when reloading the primary Nexus, the secondary also drops it’s VPC connections.

I had included the following commands:

vpc domain 11

peer-config-check-bypass

auto-recovery

The intention was that the first command allowed the VPC’s to stay up when the peer disappeared, and the second one allowed the VPC’s to come up after a reload if one of the switches doesn’t re-appear, but I may have misinterpreted those commands.

Running version 5.1(3)N1(1)

Can the "Nexii" work on their own? I'm sure I don't need both switches active to have the switches hanging off them active?

Thanks,

Bikespace

[Jerry Ye]

Can you provide the following command output with you reload the primary Nexus?

show vpc brief

Regards,

jerry

[bikespace]

I'm a little stuck for testing at the moment. Next week I have another similar install, and can test till my hearts content, but this was a little awkward that the install was a live cut over with almost zero testing time. I had 20 minutes to find out the failover doesn't work. For now they're making sure both Nexus stay up and we revisit after next weeks work.

I'll try to update when I find out what's happening with it.

[bikespace]

First Reload

CORE-B# 2009 Mar  5 12:53:37 CORE-B %$ VDC-1 %$ %VPC-2-VPC_SUSP_ALL_VPC: Peer-link going down, suspending all vPCs on secondary

2009 Mar  5 12:53:43 CORE-B %$ VDC-1 %$ %VPC-2-PEER_KEEP_ALIVE_RECV_FAIL: In domain 10, VPC peer keep-alive receive has failed

Second Reload

CORE-B# 2009 Mar  5 13:02:59 CORE-B %$ VDC-1 %$ %VPC-2-PEER_KEEP_ALIVE_RECV_FAIL: In domain 10, VPC peer keep-alive receive has failed

I’m consoled in to CORE-B, ssh’d to CORE-A and reloading CORE-A (The primary). No changes at all between the two reloads.

During the first one, it seems to lose the VPC peer link before it loses the keepalives, so it suspends all the VPC’s on B and I lose everything for a while.

During the second one, it loses the peer keepalive first and all is good.

[Jerry Ye]

Hm, interesting. Have you tried a different software like 5.0(3)N2(2c)?

Regards,

jerry

[relsethagen · Accepted solution]

Your fighting a bug that as per TAC many have fought, kind of wish TAC watched the boards and it wouldn't be so long for you to get answer

CSCtw76636

http://tools.cisco.com/Support/BugToolKit/search/getBugDetails.do?method=fetchBugDetails&bugId=CSCtw76636

The fix for the above bug is in 5.1.3.N1.1a.

[bikespace]

You little beauty. I should have found that, not sure how I missed it.

Thanks very much!

Sent from Cisco Technical Support iPhone App

[efrazee]

Hello,

I'm attempting a similar setup for the first time like the the one you described in this post. Would you be willing to share your 5548 configs with me?

Thank you

[canjohnson]

I still see this error message on a 5010 running 5.2.1.N1.