NEXUS 5548 - HSRP strange behaviour

[oguarisco]

hello to all,

I'm experiencing a strange behaviour on NEXUS5548 with Release 6.0(2)N2(5)

We've a vPC domain ID 1 with 2x 5548 and vPC domain ID 2 with 2x other 5548

vPC domain ID 1 is connected to vPC domain ID 2 with a vPC configured as trunk 802.1q transporting different VLANs

HSRP configuration is the following:

ID 1 node1 - HSRP version 2 - role Active - Priority 200

ID 1 node2 - HSRP version 2 - role Listen - Priority 100

ID 2 node1 - HSRP version 2 - role Standby - Priority 150

ID 2 node2 - HSRP version 2 - role Listen - Priority 50

After enabling one VLAN for test on all nodes, We've seen HSRP working fine and each node is acting with the correct role as per HSRP configuration. Shutting down the VLAN on the Active cause the Standby to takes over and the Listen with the highest Priority become the new Standby... all is still working fine and when we shut down the VLAN on the new Active the Standby become the new Active but the traffic is not anymore forwarded outside the VLAN and the hosts directly connected to the vPC domain ID 1 node1 can't anymore reach the Virtual IP (VIP).

We've tried to simplify putting in shutdown state the test VLAN on both nodes in vPC domain ID 2. Correctly ID 1 node1 is Active and ID 1 node2 is Standby, traffic is routed fine and VIP is reachable via ping. Shutting down VLAN on the Active cause the Standby to be the new Active correctly but from the hosts the VIP is not anymore reachable and traffic is not routed... but Interface IP address is reachable.

Applying a no shut of the VLAN on the original Active cause all to be working again as expected and the Active which does not route/forward traffic outside the VLAN is becoming again Standby.

If we configure the ID 1 node2 as the only HSRP it become Active but nothing is working also in this situation... Very Very Strange is the fact that configuring the interface IP address in the VLAN with the IP of the VIP... all is working fine!!!! GULP!!!!! Configuring again this IP as as virtual IP then nothing won't work anymore.

Any ideas or suggestions are really appreciated since I'm running out of ideas

thanks a lot

Omar

[Reza Sharifi]

Hi,

Can you post the following?

relevant config from both 5548 for this test vlan

"sh hsrp group xx" from both 5548

"sh vpc" from both 5548

[oguarisco]

Hello Reza,

I've check the vPC configuration and on all the Nexus Domain ID1 and 2 states status UP, consistency "success" Reason success...I think that L2 connectivity is achieved between:

ID1 Node1-Node2 (vpc peerlink),

ID2 Node1-Node2 (vpc peerlink)

and on the vpc inter-Datacenter composed by the 2 following links:

ID1 Node1- ID2 Node1

ID1 Node2- ID2 Node2

In the attachement you'll find the HSRP info you asked

BTW Track 10 is used to track the L3 uplinks towards the Core to adapt HSRP Roles in certain situations.

BR

Omar

[Lee Yau Mun]

Omar,

i had the same issue on my nexus 3000 after perform upgrade to release 6.0.2 4(u)3. Do you managed to find any solution?

[oguarisco]

Hello Lee,

I've noticed this strange behaviour regarding HSRP during some tests routine...

The problem was related to the fact that VLANs on the vPC peers were not in same state (one peer UP/one peer ADMIN DOWN)

Omar