ESXi 6.0 host fails with a purple diagnostic screen after upgrading the Mellanox ConnectX3 Ethernet driver to Async version 3.2.0.7

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• ESXi 6.0 host fails with a purple diagnostic screen after upgrading the Mellanox ConnectX3 Ethernet driver to Async version 3.2.0.7.
• After upgrading from nmlx4_en 3.0.0.0 Inbox driver to nmlx4_en 3.2.0.7 Async driver, the ESXi host fails with a purple diagnostic screen.

Cause

This issue occurs due to an additional module in the Inbox driver that is not available in the Async driver. This causes an error in the upgrade process from Inbox driver to Async driver, resulting in a purple diagnostic screen failure.

The inbox rdma module must be removed before updating from the nmlx4_en 3.0.0.0 Inbox driver to the nmlx4_en 3.2.0.7 Async driver on ESXi 6.0/6.0U1. Upgrading without removing the rdma module can trigger the purple screen failure.

Resolution

To prevent this issue, ensure you remove the Inbox RDMA module before upgrading from the nmlx4_en 3.0.0.0 Inbox driver to the nmlx4_en 3.2.0.7 or later Async driver on ESXi 6.0/6.0 Update 1.

To remove the Inbox driver RDMA module and install the new Async driver:

1. Connect to the ESXi 6.0 console as root. For more information, see Using ESXi Shell in ESXi 5.x and 6.0 (2004746).
2. Run this command to remove the driver module:
   
   # esxcli software vib remove -n nmlx4-rdma
   
   
3. To confirm that the driver module is no longer present, run this command:
   
   # esxcli software vib list | grep nmlx4-rdma
   
   The expected response is no output.
   
   
4. Install the new Async driver. For more information, see Installing async drivers on VMware ESXi 5.x and ESXi 6.0.x (2005205).