VM Network Disconnection on NSX-T Enhanced Data Path (EDP)
search cancel

VM Network Disconnection on NSX-T Enhanced Data Path (EDP)

book

Article ID: 448201

calendar_today

Updated On:

Products

VMware NSX

Issue/Introduction

  • A virtual machine  experiences sudden or frequent network disconnection issues.

  • Below logs are seen in the var/run/log/vmkernel.log
2026-05-25T11:22:49.751Z cpu26:39817519)ENS: Ens_DoDetachPort:5327: Detach and free port 0x80002d0 resources (interrupt) from ENS handle: 0x#########
2026-05-25T11:22:49.751Z cpu26:39817519)Vmxnet3: 12912: Detached port 0x######### from ENS
2026-05-25T11:22:50.053Z cpu78:39817494)Net: 3608: dissociate dvPort #########-#########-#########-######### from port 0x#########

Environment

VMware NSX

Cause

The issue is related to the Enhanced Data Path (EDP) feature enabled on the NSX-T Distributed Virtual Switch (DVS).

Because no infrastructure- or hypervisor-level packet-dropping errors are present in the logs, it is highly suspected that a specific Guest OS security feature or driver interaction causes the vmxnet3 virtual network adapter driver inside the VM to become unresponsive when running on an Enhanced Data Path-enabled switch

Resolution

The issue is resolved in VMware ESXi 8.0.3, available at Broadcom downloads.
If you are having difficulty finding and downloading software, review the Download Broadcom products and software KB.

To workaround this issue there are two possible options:


Workaround 1: Revert Enhanced Data Path to Standard (Recommended for VMXNET3)

If you must preserve the performance capabilities of the vmxnet3 adapter, you can transition the host configuration from Enhanced Data Path back to Standard mode.

  1. Modify the Host Settings: Change the Enhanced Data Path setting to Standard on a per-ESXi host basis for each NSX-T prepared host in the cluster.

  2. Update the Transport Node Profile (TNP): * Create  a Transport Node Profile representing the cluster.

    • Under the Advanced Configuration section, set the mode to Standard.

    • Apply this profile to the cluster to enforce the standard mode globally across hosts.

  3. Revert VM Adapters: Once the host switch is set back to Standard mode, configure the virtual machine network adapter(s) from E1000 back to VMXNET3


or

Workaround 2: Change Virtual Network Adapter to E1000

This workaround bypasses the VMware-specific vmxnet3 driver stack by leveraging the built-in, native driver inside the Guest OS.

  1. Power off the affected virtual machine.

  2. Edit the VM settings and change the network adapter type from VMXNET3 to E1000.

  3. Power on the virtual machine.