VM boot cycles on ESXi hosts due to upstream physical switch port failure

Products

• VMware vSphere ESX 7.x
• VMware vSphere ESX 8.x

Issue/Introduction

• Virtual Machines (VMs) experience abrupt reboots or continuous boot cycles. Some specific types of VMs (e.g. VNF VMs) don't show console messages, and they are stuck in a "boot loop" if no network connectivity is present.
• Network traffic disruption is observed on specific gateway nodes or ESXi hosts in a cluster, not all of them.
• Impact may be localized to a specific enclosure or set of ESXi hosts sharing a physical switch fabric.

Environment

• VMware ESXi 7.x
• VMware ESXi 8.x
• Physical network infrastructure using TOR switches.

Cause

A physical hardware fault on an upstream Top-of-Rack (TOR) switch port or slot. This disruption causes intermittent or total path failure for the affected vmnic affected  that is one be the uplinks for the DVS where the virtual nics of the VM are connected 

Resolution

1. Identify Faulty Interface: Review Link Layer Discovery Protocol (LLDP) or CDP (Cisco Discovery Protocol) telemetry to identify the specific physical switch port associated with the impacted vmnic.
2. Apply Workaround: Administratively shut down the faulty vmnic on all affected ESXi hosts or the switch ports connected to them. This forces traffic to redundant, healthy uplinks and stabilizes the host environment.
3. Physical Remediation: Engage the hardware or network administration team to investigate and remediate the physical switch (e.g., port replacement, slot inspection, or switch reload).
4. Verification: Confirm VM stability and successful boot cycles once the faulty network path is bypassed.

Additional Information

• For guidance on speaking with a representative, see Contact Support.
• Subscribe to this article to receive updates on hardware compatibility and related networking alerts.