VMware ESXi and vSphere Cluster Management

Configure Physical NIC Speed and Duplex on VMware ESXi

Learn how to view and change ESXi vmnic speed and duplex in the vSphere Client, when to use auto-negotiation, and how to verify or recover from link problems.

Physical NIC speed and duplex settings control how an ESXi host communicates with the Ethernet switch connected to each hardware network adapter. This guide covers viewing and changing those settings in the vSphere Client, using ESXi Shell commands for inspection and recovery, and validating the result safely.

Understand Speed, Duplex, and Link State

An ESXi host is the bare-metal VMware hypervisor system. A physical NIC is a hardware network adapter installed in that host. ESXi identifies each physical adapter with a name such as vmnic0 or vmnic1.

Link speed is the transmission rate used by the physical Ethernet connection. Common values include 100 Mbps, 1 Gbps, and 10 Gbps or higher. The speed may be negotiated automatically or selected as a fixed configuration.

Duplex describes how the interface sends and receives Ethernet traffic. Full duplex permits simultaneous transmission and reception. Half duplex alternates between transmitting and receiving and is generally limited to legacy Ethernet environments.

Link state indicates whether the physical connection is detected and operational. Link Up means the adapter has an active physical connection; an adapter can be administratively available while still showing Link Down because of a cable, transceiver, switch-port, or configuration problem.

The ESXi vmnic and the connected physical switch port must use compatible settings. A mismatch can cause poor throughput, packet loss, CRC or frame errors, collisions on legacy half-duplex links, and intermittent connectivity. A connection may remain down if the endpoints cannot agree on a compatible speed and duplex combination.

Plan the Change Before Editing a vmnic

Identify the correct ESXi host and physical adapter by vmnic name, MAC address, PCI or device details, and current link status.
Confirm that the adapter is connected to the intended physical switch port. Trace the cable or transceiver when the host has several similar adapters.
Check the NIC hardware capabilities, including supported speeds, duplex modes, transceiver requirements, and cable limitations.
Ask the network team to confirm the switch-port configuration whenever a fixed value is required.
Determine whether the vmnic carries management, vMotion, storage, fault tolerance, replication, VM, or other VMkernel traffic.
Review the active, standby, and unused uplinks assigned to the relevant standard vSwitch or vSphere Distributed Switch.

A standard vSwitch is a host-local virtual switch. A vSphere Distributed Switch is centrally managed across hosts. In either case, speed and duplex belong to the physical vmnic, not to a port-group VLAN setting. Changing the vmnic can therefore affect every traffic type mapped to that uplink.

View the Current Physical Adapter Configuration

Select the ESXi host in the vSphere Client inventory.
Open Manage, then open Networking.
Open Physical adapters.
Locate the target vmnic and review its adapter name, link status, speed, duplex mode, MAC address, and uplink context where available.

Use the MAC address and PCI or device information to distinguish similarly named adapters. Confirm that the displayed adapter is physically connected to the switch port you intend to change.

The Physical adapters view may show a configured or negotiated speed and duplex value. Always compare that information with the physical switch's operational status rather than relying on only one endpoint.

Change Speed and Duplex in the vSphere Client

Select the target ESXi host.
Open Manage and then Networking.
Open Physical adapters.
Select the required physical adapter and choose Edit.
Choose Auto negotiate or an explicitly supported speed and duplex combination.
Save the configuration.
Allow the link to renegotiate or reset as required, then verify the resulting state and values.

The edit operation can briefly interrupt the physical link. If the vmnic is part of NIC teaming, traffic may fail over to another uplink, but this depends on the configured failover order, link state, and available redundancy.

When to Use Auto-Negotiation

Auto-negotiation is the Ethernet mechanism that allows connected endpoints to select mutually supported link characteristics. It is the preferred default in normal modern environments.

For typical 1 GbE connections, leave the ESXi vmnic and switch port on auto-negotiation. Gigabit Ethernet commonly requires or depends on this behavior.
For modern 10 GbE and faster links, follow the NIC, transceiver, switch, and vendor requirements; automatic configuration is normally preferred unless a documented standard says otherwise.
Both ends should generally use auto-negotiation together.
Forcing one endpoint while leaving the other on auto-negotiation can produce an unintended duplex result, commonly a duplex mismatch.
Use a manual setting only for a legacy device, a documented network standard, a troubleshooting procedure, or a known switch-port requirement.

Speed and Duplex Configuration Decision Guide

Columns: Environment or condition; Recommended ESXi setting; Required switch-port setting; Reason; Notes.

Modern 1 GbE Ethernet connection: Auto negotiate; Auto negotiate; Allows compatible Gigabit operation and duplex selection; This is the normal recommendation.

Modern 10 GbE or faster connection: Auto negotiate or the vendor-required mode; Use the matching normal switch configuration; Preserves supported link behavior; Check optics, DACs, cabling, firmware, and hardware requirements.

Legacy device requiring a fixed value: The documented supported fixed speed and full duplex; The identical fixed speed and full duplex; Prevents an auto-versus-forced mismatch; Confirm support on both endpoints first.

Troubleshooting a suspected negotiation issue: Temporarily test auto-negotiation on both ends when appropriate; Set the switch port to the corresponding mode; Provides a consistent baseline; Record the original settings and use a maintenance window.

One endpoint forced and the other automatic: Do not leave this arrangement in production; Match both endpoints or return both to auto-negotiation; Avoids duplex mismatch; This commonly causes errors and poor performance.

Rules for Manual Configuration

Select only values offered by the ESXi adapter and supported by the switch port, transceiver, and cabling.
Configure identical fixed speed and duplex values on the ESXi vmnic and physical switch port.
Use full duplex for supported switched Ethernet links. Half duplex is generally legacy-only.
Do not select a lower or higher rate that the NIC, transceiver, cable plant, or switch port does not support.
Expect the link to go down briefly during renegotiation. It may remain down when the endpoints have incompatible configurations.

A fixed value is not automatically more stable than auto-negotiation. It is appropriate only when the entire path and the operational standard have been verified.

Inspect Settings from the ESXi Shell

The following commands help compare the vSphere Client view with the host's operational information. Replace vmnic0 with the target adapter.

esxcli network nic list
esxcli network nic get -n vmnic0

esxcli network nic list reports vmnic names, link state, speed, duplex, driver, and PCI location. esxcli network nic get -n vmnic0 provides detailed configuration, supported capabilities, auto-negotiation state, current link settings, and available diagnostic information.

To return a physical NIC to auto-negotiation:

esxcli network nic set -n vmnic0 -a

To set a fixed value, use only a speed supported by the adapter and switch, and configure the switch port identically:

esxcli network nic set -n vmnic0 -s 1000 -d full

Verify the Result

Confirm that the physical adapter reports Link Up.
Confirm that the displayed speed matches the intended configured or negotiated rate.
Confirm that the duplex mode is the intended mode, normally full duplex.
Review ESXi NIC statistics for receive and transmit errors, dropped packets, CRC errors, and other available counters.
Confirm that host management remains reachable if the modified vmnic is a management uplink.
Test VMkernel services such as management, vMotion, storage, or replication as applicable.
Test virtual machine traffic that relies on the affected uplink.
Ask the network team to confirm that the physical switch port reports the same expected speed and duplex.

Post-Change Validation Checklist

Columns: Validation item; Where to check; Expected result; Potential issue if unsuccessful.

Physical link state: vSphere Client, ESXi Shell, and switch port; Link Up; Cable, transceiver, port, or incompatible settings.

Configured or negotiated speed: Physical adapters view, esxcli network nic list, and switch; Intended rate on both endpoints; Port limitation, unsupported hardware, or cabling issue.

Duplex mode: Client and switch operational details; Matching full-duplex mode unless a documented legacy requirement exists; Duplex mismatch.

Host management reachability: vSphere Client, ping, or management tools; Host remains reachable; Management uplink interruption or failed failover.

VMkernel and workload connectivity: Relevant service and VM tests; Expected traffic succeeds; Affected uplink, VLAN, teaming, or physical-link problem.

ESXi NIC error counters: NIC statistics and host monitoring; No increasing abnormal errors or drops; Duplex, cable, optics, driver, or hardware problem.

Physical switch-port counters: Switch interface status and statistics; Matching operational values and clean counters; Switch configuration or physical-layer fault.

Understand Uplink and NIC Teaming Impact

An uplink is a physical NIC assigned to carry traffic for a standard vSwitch or vSphere Distributed Switch. NIC teaming uses multiple physical uplinks for redundancy or traffic distribution.

Before changing a vmnic, check whether it is active, standby, or unused for each relevant port group and VMkernel adapter. If one team member temporarily loses link, traffic may fail over, but only when another eligible uplink exists and the failover policy permits it.

A single vmnic change can affect management, vMotion, storage, fault tolerance, virtual machine, and replication traffic. Redundancy reduces the risk of an outage but does not eliminate the need for a maintenance plan and post-change testing.

Common Symptoms and Troubleshooting

Common Symptoms and Likely Causes

Columns: Symptom; Likely cause; Verification step; Corrective action.

Link remains down: Different fixed values, unsupported setting, or cable/transceiver/switch issue; Compare both endpoints, NIC capabilities, port status, and physical components; Restore a known-good matching configuration or use auto-negotiation on both ends.

Slow or inconsistent throughput: Duplex mismatch, damaged cabling, or unstable forced configuration; Compare speed and duplex and review error and discard counters; Correct the mismatch, remediate the physical layer, or return both ends to auto-negotiation.

CRC or frame errors: Cabling, optics, hardware, or link-configuration problem; Review ESXi and switch counters and inspect the physical path; Replace defective components and correct configuration.

Remote host management loss after edit: The modified vmnic carried management traffic and no usable failover path existed; Use console or out-of-band access and inspect management uplinks; Revert the setting and confirm management redundancy.

Unexpected negotiated speed: Switch limitation, insufficient cable or transceiver capability, driver or firmware issue, or manually constrained port; Compare supported capabilities and physical component requirements; Remove limitations and correct incompatible components or software.

Recover a Link That Stayed Down

Check the NIC details in the vSphere Client or with esxcli network nic get.
Check the physical switch port's administrative and operational status.
Confirm that speed and duplex match exactly when using fixed settings.
Inspect or replace the cable or transceiver and test another switch port if necessary.
Restore the previously documented matching fixed configuration or return both endpoints to auto-negotiation.

Recover Lost Remote Management

If vSphere access is lost, use the ESXi Direct Console User Interface, a remote console, or out-of-band host management. Check which uplinks serve the management network, restore the previous NIC setting or auto-negotiation, and verify the switch port before retrying.

Practical Configuration Examples

Keep a 1 GbE Uplink on Automatic Negotiation

An ESXi host connects through vmnic0 to a normal Gigabit Ethernet switch port. Leave both sides on auto-negotiation. The expected result is a compatible 1 Gbps full-duplex link.

Match a Legacy Fixed Switch Port

A documented legacy segment requires 100 Mbps full duplex. After confirming support, configure both the switch port and ESXi vmnic to 100 Mbps full duplex. Matching fixed values prevent an auto-versus-forced mismatch.

Protect Management Connectivity

If vmnic0 carries management traffic and vmnic1 is a functioning redundant management uplink, confirm that management is reachable through the alternate path before editing vmnic0. Retain console or out-of-band access in case failover does not occur.

Identify a Duplex Mismatch

If ESXi is forced to full duplex while the switch remains on auto-negotiation, compare both operational configurations and error counters. Configure both ends consistently or return both to auto-negotiation. Throughput and error counters should improve after the mismatch is removed.

Rollback and Recovery Procedure

Restore Auto negotiate when a forced value causes instability or no link.
If the original state was a documented fixed configuration, restore that matching speed and duplex on both the vmnic and switch port.
Use direct console or out-of-band access if remote vSphere administration is unavailable.
Reconfirm the switch-port configuration, cabling, transceiver, and NIC capabilities before repeating the change.
Document the final values, link status, counters, and affected uplink assignments.

Exam-Relevant Notes

Speed and duplex are physical NIC properties, not port-group VLAN settings.
Auto-negotiation is normally preferred, especially for Gigabit Ethernet and modern switched links.
When using fixed settings, both endpoints must use identical supported speed and duplex values.
Forcing one side while the other side remains automatic can create a duplex mismatch.
A vmnic change can affect every traffic type using that uplink, including management and VMkernel services.
Always verify link state, speed, duplex, error counters, management access, workload traffic, and switch-port status after the change.

For related host networking work, review Configure VLANs, VMware ESXi Online Course, and VMware ESXi.