vMotion Fails with "Migration was cancelled" Due to Low Bandwidth

Products

VMware vCenter Server

Issue/Introduction

When attempting to migrate a virtual machine using vMotion, the operation fails. You will typically observe the following symptoms:

- Stalled Progress: The vMotion task stalls at 20% to 30% completion for approximately 30 minutes before eventually failing.
- vCenter Error Message:
  "The migration was cancelled because the amount of changing memory for the VM was greater than the available network bandwidth, meaning the migration was not making forward progress. Please attempt the migration again when the VM is not as busy or more network bandwidth is available."
- Slow Baseline Performance: Even an empty, newly created VM (powered on, but lacking a Guest OS) migrates successfully but takes an abnormally long time.
- VMkernel Log Errors: The var/run/log/vmkernel.log on either the source or target ESXi host registers an extremely low estimated bandwidth prior to the failure:
  [YYYY-MM-DD] [Time] In(###) vmkernel: cpu40:####### opID=########)VMotion: ####: ################### <D/S>: Estimated network bandwidth 0.7## MB/s before failure

Environment

VMware vCenter Server 8.0.x
VMware vSphere ESXi 8.0.x

Cause

This failure occurs when the vMotion network cannot sustain the throughput required to transfer the VM's active memory pages. Common culprits for restricted network throughput include:

Physical switch misconfigurations(e.g., MTU mismatches or port speed limitations).
Outdated or incompatible ESXi network adapter(NIC) drivers and firmware.
Misconfigured Traffic Shaping policies on the vSphere Standard/Distributed Switch.

Resolution

Before escalating to hardware vendors, verify the configuration and capabilities of your vMotion network using the following steps.

1. Verify Physical NIC Link Speeds:

Ensure the physical uplinks (vmnics) dedicated to vMotion are negotiating at the correct speed (10Gbps or higher is strongly recommended).

Connect to the ESXi host via SSH.
Run the following command to verify link speed and duplex settings:

esxcli network nic list
If the speed is lower than expected (e.g., 1000Mbps instead of 10000Mbps), the issue is likely physical cabling, transceivers, or switch port configurations.

2. Check for Traffic Shaping Restrictions:

Ensure that vSphere isn't artificially throttling your vMotion traffic.

Navigate to your vCenter Server > Networking.
Select the Port Group used for vMotion.
Go to Edit Settings > Traffic Shaping.
Verify that Average Bandwidth, Peak Bandwidth, and Burst Size are not restricted to values that would bottleneck a migration.

3. Test Network Throughput and MTU (Jumbo Frames):

If you are using Jumbo Frames (MTU 9000) for vMotion, an MTU mismatch along the physical network path will cause severe packet fragmentation and bandwidth drops.

Test the connection between the source and destination vMotion IP addresses using vmkping, forcing a large packet size and preventing fragmentation:

vmkping -I vmkX <destination_vMotion_IP> -s 8972 -d
(Replace vmkX with your vMotion VMkernel port number).
If this ping fails but a standard ping succeeds, there is a Jumbo Frame configuration error on the physical switch.

4. Engage Vendor Support:

If the internal vSphere configurations are correct and network tests pass, the bottleneck is occurring at the physical hardware layer:

Physical Network: Engage your networking team or physical switch vendor to inspect the switch logs for dropped packets, port errors, or spanning-tree topology changes.
Host Hardware: Verify your NIC firmware and drivers match the VMware Compatibility Guide. If you need assistance investigating driver compatibility, open a support case with Broadcom Support.