cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
334
Views
0
Helpful
3
Replies

Connectivy issue on ACI L3out connected to CATO Socket cluster

lorenzo_grusi
Level 1
Level 1

Hello,

on a customer tenant we configured an L3Out (SVI mode) with 2 ports connected to 2  Cato Socket devices in HA

During the failover test we had the following issue:

If we shut the leaf port where the cato1 is connected the cato2 become master and everything works correctly

When we reactivate the port previously shut, the cato1 become active and the cato2 standby, but on the arp of the nodes in L3Out we still see the virtual mac on the port where cato2 is connected and this cause connectivity issue

The only way to solve the issue is shut the leaf port where the cato2 is connected.

Any idea what can be the problem?

3 Replies 3

AshSe
VIP
VIP

Dear @lorenzo_grusi The issue you're describing seems to be related to ARP (Address Resolution Protocol) and MAC address handling during the failover and failback process between the two Cato Socket devices. When Cato1 becomes active again, the ARP tables on the nodes in the L3Out are not being updated correctly, causing connectivity issues.

Let me know, if you need to deeper understanding of potential causes and solutions for the same.

Have a good one!!!

Many thx for your reply.
We understood that the issue is with the ARP, we cannot understand why the leaf didn't consider the GARP sent by the CATO devices and didn't update the ARP table
Have you any idea to find the error taken by the leaf?
Thx


@lorenzo_grusi wrote:
Many thx for your reply.
We understood that the issue is with the ARP, we cannot understand why the leaf didn't consider the GARP sent by the CATO devices and didn't update the ARP table
Have you any idea to find the error taken by the leaf?
Thx

To diagnose why the leaf switch is not updating its ARP table in response to the gratuitous ARP (GARP) sent by the Cato devices, you can follow these steps:

1. Verify GARP Transmission

Ensure that the Cato devices are indeed sending GARP packets when the failover and failback occur. You can use a packet capture tool like Wireshark on the network to confirm this.

2. Check Leaf Switch Configuration

Review the configuration of the leaf switch to ensure it is set to accept and process GARP packets. Some switches have security features or specific settings that might block or ignore GARP packets.

3. Examine ARP and MAC Table Entries

Check the ARP and MAC address tables on the leaf switch before and after the failover and failback events to see if the entries are being updated correctly.

4. Review Logs and Debug Output

Enable logging and debugging on the leaf switch to capture any errors or warnings related to ARP processing. This can provide insights into why the switch is not updating its ARP table.

 

Happy Troubleshooting !!!

Review Cisco Networking for a $25 gift card

Save 25% on Day-2 Operations Add-On License