ESXi Host 'Not Responding' State Due to Upstream Network Congestion

Products

• VMware vSAN

Issue/Introduction

ESXi hosts may intermittently enter a "Not Responding" state in vCenter Server, characterized by management agent (hostd) failures, storage heartbeats timeouts, and vSAN component congestion. This article outlines the troubleshooting steps when standard host-level fixes do not resolve the issue, focusing on identifying underlying upstream network fabric errors.

Environment

VMware vSphere ESXi 8.x

VMware vSAN 8.x

Cause

The host disconnections are caused by the hostd service failing due to hanging storage and vSAN component congestion, which are secondary symptoms of severe, underlying network TCP/IP errors (such as packet drops, out-of-order packets, or congestion) affecting the host.

Resolution

If standard VMware troubleshooting steps—such as disabling Tx Pause Frames, swapping physical cables, and changing switchports—have been performed and the issue persists:

1. Analyze vSAN performance metrics for high rates of TCP RX duplicate packets and out-of-order packets.
2. Review ESXi vmkernel.log for extensive lock issues and LSOM component-release warnings.
3. Confirm if the issue persists after isolating the host in Maintenance Mode.
4. If network errors continue despite replacing local host hardware (NICs, cables), engage the network engineering team to investigate the upstream network fabric for persistent packet drops or congestion.

Additional Information

• Ensure all hardware is compliant with the VMware Compatibility Guide.
• For general troubleshooting of ESXi host disconnections, refer to: ESXi host disconnects intermittently from vCenter Server