ESXi host suffers a PSOD with error: "nmlx5_QueryNicVportContext:188 command failed: IO was aborted"

Products

• VMware vSphere ESXi

Issue/Introduction

Symptom: ESXi hosts experience a sudden link status down on physical NICs (vmnic). In some cases, this may lead to an ESXi host crash or Purple Diagnostic Screen (PSOD).

• The error "<NMLX_ERR> nmlx5_QueryNicVportContext" may occur at 5-minute intervals.
  
  
• In some instances, the ESXi host may crash, resulting in a Purple Diagnostic Screen (PSOD).
  
  
• Specific error messages such as nmlx5_QueryNicVportContext:188 command failed: IO was aborted may also be observed in /var/run/log/hostd.log on the ESXi hosts:

 

Log lines

2026-04-06T16:22:31.786Z Al(177) vmkalert: cpu4:24068897)ALERT: <NMLX_ERR> nmlx5_QueryNicVportContext:188 command failed: IO was aborted

2026-04-06T16:22:31.786Z Wa(180) vmkwarning: cpu4:24068897)WARNING: <NMLX_WRN> nmlx5_QueryPortRoce failed to query qkey violations counter

2026-04-06T16:22:31.952Z Al(177) vmkalert: cpu61:24068899)ALERT: <NMLX_ERR> nmlx5_QueryNicVportContext:188 command failed: IO was aborted

2026-04-06T16:22:31.952Z Wa(180) vmkwarning: cpu61:24068899)WARNING: <NMLX_WRN> nmlx5_QueryPortRoce failed to query qkey violations counter

2026-04-06T16:22:32.165Z Al(177) vmkalert: cpu3:24068900)ALERT: <NMLX_ERR> nmlx5_QueryNicVportContext:188 command failed: IO was aborted

2026-04-06T16:22:32.165Z Wa(180) vmkwarning: cpu3:24068900)WARNING: <NMLX_WRN> nmlx5_QueryPortRoce failed to query qkey violations counter

2026-04-06T16:22:33.732Z Al(177) vmkalert: cpu17:24068910)ALERT: <NMLX_ERR> nmlx5_QueryNicVportContext:188 command failed: IO was aborted

 

In the case of (PSOD) the display will look like the screenshot below.

Environment

VMware vSphere ESXi 8.0.3
VMware vSphere ESXi 8.0.2

Cause

 A defect in the nmlx5-core driver's health-monitoring logic causes it to incorrectly identify functional Mellanox network cards as faulty, triggering a link-down event or system crash.

Resolution

Upgrade the affected ESXi host to the following version or newer:

• ESXi Build: 8.0 Update 3e (Build 24674464)
• Fixed Driver Version: nmlx5-core 4.23.6.5-1vmw

Verification Steps

1. Log in to the ESXi shell and verify the driver version: esxcli network nic get -n vmnicX (Replace X with the affected NIC number).
2. Monitor the vmkernel.log for 24–48 hours to ensure the IO was aborted alert no longer appears.
3. Confirm physical link stability via esxcli network nic list.

Related Articles:

• ESXi Network Adapter Shows Repeated "IO was aborted" Errors in VMkernel Logs (381949)

Refer to the following KB article for FAQ: Recommendation for drivers/firmware for ESXi hosts

 

Additional Information

• The issue is resolved starting with ESXi 8.0 Update 3e (Build 24674464) using the nmlx5-core VIB version 4.23.6.5-1vmw.
  
  
• For VCF 9.0 environments, the fix is included in the inbox driver with nmlx5-core version 4.24.0.7-16vmw.
  
  
• Refer VMware ESXi 8.0 Update 3e Release Notes
  
  

 "Health: Miss counters detected" alerts for Mellanox driver on ESXi 8.0.2 and 8.0.3

Japanese KB: PSOD: nmlx5_QueryNicVportContext:188 command failed: IO was aborted