NSX-T Load Balancer crashes and generates nginx coredump when L4/L7 LB pool neither contains Pool member port nor Monitoring port

Products

• VMware NSX Networking

Issue/Introduction

This article is published to describe a known issue affecting all currently available versions of NSX-T.

Symptoms:

1. You are using NSX-T Load Balancer with L4 or L7 Virtual Servers

2. In the pool associated with the virtual server, none of the pool member is configured with pool member port (The field is left empty) and also the monitoring profile associated with the pool doesn't contains any port.

3. As soon as such configuration is save from NSX-T UI, we could see similar crash log in syslog of edge node,

/var/log/syslog
edge01.corp.local kernel - - - [310227.187941] nginx[1628139]: segfault at 0 ip 000061e0682426e5 sp 000071743fda1d48 error 4 in libc-2.31.so[61e0680dc000+178000]
 
2023-06-19T06:12:02.747Z edge01.corp.local NSX 1631076 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="INFO"] Core dump generation received by process: None [nginx]
 
2023-06-19T06:12:02.751Z edge01.corp.local NSX 1631076 - [nsx@6876 comp="nsx-edge" subcomp="node-mgmt" username="root" level="WARNING"] Core file generated: /var/log/core/core.nginx.1687155122.1628139.134.11.gz

4. We can also see a coredump generated with ngnix process inside /var/log/core directory of the edge node,

-rw-r--r-- 1 root root 1262050 Jun 19 06:12 core.nginx.1687155122.1628139.134.11.gz
root@edge01:/var/log/core# pwd
/var/log/core

Environment

VMware NSX-T Data Center

Cause

Port in pool member or monitoring profile can be empty. But it cannot be empty in pool member and monitoring profile at the same time.as soon as such configuration is pushed, the LB config process crashes and a core dump will gets generated

Resolution

Currently there is no resolution on this and this will be fixed in a future releases

Workaround:
To workaround from this issue, please follow the below workaround steps,

1. Configure all the pools such a way that either the pool members or the monitoring profile associated with the pool contains required port. Both of them can't be empty.

2. Once such config is saved, please restart the LB services. The easiest way to perform this is to put the Edge nodes in NSX Maintenance Mode and put it out again. Please note, this needs to be performed in both Active and Standby Edge nodes one by one.

Additional Information

Impact/Risks:
NSX-T load balancer service unavailable, impact on traffic flowing via the NSX-T Load Balancer for all the hosted VIPs, Pools.