NSX-T configured host encounters a PSOD when port used in Live Traffic Analysis session is disabled

Products

• VMware NSX

Issue/Introduction

Symptoms:

• You configured a Live Traffic Analysis (LTA) session with a source port.
• The port used in LTA session was disabled during the on-going session.
• The ESXi host where the port resides encounters a PSOD.
• In the ESXi host log vmkernel.log we see the following backtrace:

Panic Message: @BlueScreen: #PF Exception 14 in world 2098614:NSX_HELPER_L IP 0x420037ea2455 addr 0x18

Backtrace (top 8 lines):

0x453a27e1bed0:[0x420037ea2455]vmk_IOChainRemove@vmkernel#nover+0x19 stack: 0x40009b1, 0x431d7ae01c70, 0x10, 0x42003922c0d9, 0x430706002b80

0x453a27e1bf00: 0x42003922c0d8][email protected]#1.0.7.0.20115690+0xd1 stack: 0x0, 0x431bc48016e0, 0x453a27e1f000, 0x13c7b61d9cbed2, 0x4302e9801220

0x453a27e1bf40:[0x420037cda16f]HelperQueueFunc@vmkernel#nover+0x2d8 stack: 0x453a27e20b48, 0x431bc48016f8, 0x453a27e1f000, 0x0, 0x431bc4801780

0x453a27e1bfe0:[0x420037fb33e1]CpuSched_StartWorld@vmkernel#nover+0x86 stack: 0x0, 0x420037cc4b50, 0x0, 0x0, 0x0

0x453a27e1c000:[0x420037cc4b4f]Debug_IsInitialized@vmkernel#nover+0xc stack: 0x0, 0x0, 0x0, 0x0, 0x0

Environment

VMware NSX-T Data Center

Cause

The PSOD occurs when the port which is used in the LTA session is disabled, while the session is still in progress.
This can occur when the port is changes such as : mtu change; ring size change; queue changes or start/stop of pktgen (or any dpdk application) inside the guest.

Resolution

This issue is resolved in 3.2.3.1 Impactor(nsx-impactor-323-rel), nsx-impactor-323-telia-rel, 3.2.4.0 Impactor(nsx-impactor-ufo branch), NSX Transformers 4.1.1.0 Keeper, 4.0.1.57.0 VMC M20v9(nsx-jetfire-m20-rel), (VMC M22 / 4.1.0.x)(nsx-keeper-410-rel), available in the Broadcom Portal.

Workaround:
There are 3 workaround options: 

1. To avoid this issue, do not make any changes to the port being used in the LTA session, while the session is in progress.
2. Do not disable the port used in the LTA session
3. Do not use LTA(Live Traffic Analysis) feature if you need to disable the port frequently.