Stale ARP Proxy Entries on Tier-0 Gateway After Tier-1 Deletion

Products

• VMware NSX

Issue/Introduction

When managing NSX-T gateways, the following symptoms may be observed:

• After deleting a Tier-1 (T1) Gateway that had active SNAT rules and Route Advertisement enabled (specifically "All NAT IP's"), the corresponding ARP Proxy entries remain on the Tier-0 (T0) Gateway external interface.
• Manual verification on the Edge node (e.g., get arp-proxies or checking edge/arp-entry) confirms the presence of these stale entries.
  
  

  "table": [
    {
      "port_uuid": "[masked]",
      "ip_address": "[masked]",
      "mac_address": "[masked]"
    }
  ],

Cause

The issue is caused by a failure in the L3 trigger mechanism. While the trigger to clean up ARP proxy entries executes correctly during a "detach" operation (unlinking T1 from T0), it fails to fire when the T1 Gateway is deleted directly. This results in the T0 Gateway retaining stale ARP information in its database and datapath.

Resolution

This issue is scheduled to be resolved in upcoming releases. Please subscribe this article to get the latest infomation.

To clean up the stale entries, a Corfu database dump is required to provide specific removal instructions.
Please contact to Broadcom support.