The NSX-T Edge Node dataplane service may crash when it attempts to clear an existing IGMP snooping entry

Products

• VMware NSX

Issue/Introduction

Symptoms:

• You are running or have recently upgraded to NSX-T version 4.0.0.1 or 4.0.1.1.
• Edge Nodes may process IGMP packets, although multicast is disabled on Tier-1 or Tier-0 gateway.
• Core dumps may be observed on the NSX-T Edge Node at /var/log/core:

core.dp-ipc35.xxxxx.gz

• The dataplane service will restart and this can lead disruption for North-South traffic.
• You may encounter similar entries below in the Edge node log file  /var/log/syslog:

202x-xx-27Txx:xx:xx.799Z edge01 NSX 5869 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="INFO"] Service datapathd coredump at 2022-xx-27 xx:xx:31 file /var/log/core/core.dp-ipc35.xxxxx.gz
202x-xx-27T04:21:07.994Z edge01 NSX 5869 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="WARNING" eventFeatureName="infrastructure_service" eventType="edge_service_status_changed" eventSev="warning" eventState="On"] The service dataplane changed from STARTED to CRASHED.
202x-xx-xxTxx:xx:11.221Z edge01 NSX 10052 ROUTING [nsx@6876 comp="nsx-edge" subcomp="datapathd" s2comp="igmp-learning" tname="dp-ipc35" level="INFO"] Removing IGMP (*,x.y.z.a) in vrfid X from port XXXXXXXXXXXX-XXXX

Environment

VMware NSX-T Data Center
VMware NSX-T Data Center 4.x

Cause

During the IGMP snooping entry cleanup process, the NSX-T Edge Node references the masked interface id to access the interface record.
This results in a segmentation fault, which leads to the NSX-T Edge Node dataplane service crash.

Resolution

This issue is resolved in NSX-T version 4.1 available at VMware Downloads

Workaround:
Two option available:

• As the Gateway is not configured for IGMP, on the physical network, prevent/block all the IGMP packets reaching the NSX-T Edge Node.
• Configure IGMP snooping on Physical switches, this should prevent IGMP Report packets reaching the NSX-T Edge Node.