ESXi Host lost Network Connectivity After 7.x to 8.0 Upgrade Due to Legacy vDS LACP Mismatch.

Products

• VMware vSphere ESXi
• VMware vCenter Server

Issue/Introduction

Upgrading an ESXi host from 7.x to ESXi 8.x while it remains attached to a vSphere Distributed Switch (VDS) 7.x branch configured with a legacy Link Aggregation Control Protocol (LACP) schema results in complete data-path network isolation.

The following symptoms can be observed across the management and production networking layers:

• The vCenter UI reports that newly upgraded ESXi 8.0 host network status as completely "In Sync" with the VDS container, while showing the virtual Link Aggregation Group (vLAG) set to "Active/passive" mode.
• Deployed virtual machines (VMs) and VMkernel adapters (vmks) experience a complete loss of network connectivity when this occurs. An inspection via esxtop (using the "n" view) reveals that the active uplink status displays as "Fallback", contrasting with the expected lag-# designation observed in previous versions.
• The systemMac of the proxy vDS on the ESXi host remains unchanged, which can be verified directly within the /var/log/lacp.log file.
• The local LACP daemon logs specific protocol validation timeouts and split-brain loop-prevention state (smVars 0x3) inside /var/log/lacp.log.
  • lacp[2098304]: 777, AddPort port(vmnic#/DR lag#-0) flag 0 systemMac 00:00:00:00:00:00 uplinkMac 00:00:00:00:00:00
lacp[2098304]: port for uplink vmnic# with mode 1 for lag DR lag#
lacp[2098304]: 777, UWSetPortInfo port(vmnic#/DR lag#-0) flag 0 systemMac ##:##:##:##:##:## uplinkMac ##:##:##:##:##:##
lacp[2098304]: 2993, Receive uplink vmnic# event (port connected) for lag DR lag#
lacp[2098304]: 2970, Add uplink vmnic# into lag DR lag#/########## of portset DvsPortset-0
lacp[2098304]: 777, AddPort port(vmnic#/DR lag#-1) flag 0 systemMac 00:00:00:00:00:00 uplinkMac 00:00:00:00:00:00
lacp[2098304]: port for uplink vmnic# with mode 1 for lag DR lag#
lacp[2098304]: 777, UWSetPortInfo port(vmnic#/DR lag#-1) flag 0 systemMac ##:##:##:##:##:## uplinkMac ##:##:##:##:##:##
lacp[2098304]: 2993, Receive uplink vmnic# event (port connected) for lag DR lag#
lacp[2098304]: 1398, 1 active port of group ID ##########
lacp[2098304]: 1314, Set host status property to green for portset DvsPortset-0
lacp[2098304]: 1544, No peer response: enable link vmnic# as individual
lacp[2098304]: 3602, Receive uplink vmnic# event (port blocked), smVars 0x3
lacp[2098304]: 2820, Remove uplink vmnic# from lag DR lag#(##########) of portset DvsPortset-0
lacp[2098304]: 1398, 2 active port of group ID ##########
lacp[2098304]: 1294, The host status property already set to green for portset DvsPortset-0
lacp[2098304]: 2329, Failed to send lacpdu to uplink vmnic#

• The upstream physical switch is misconfigured to static EtherChannel mode (Mode ON) or configured as a standard trunk/access port rather than using dynamic LACP negotiation.

Environment

VMware vSphere ESXi
VMware vCenter Server

Cause

In an upgraded vSphere 8 environment, the hypervisor strictly mandates the use of LACP V2.

When the uplinks fail to form a Link Aggregation Group (LAG) due to this updated protocol requirement, the active uplinks are immediately blocked and demoted to a fallback status.

In strict compliance with loop-prevention mechanics defined by the IEEE 802.1AX standard and enforced by the vSphere 8 kernel, the physical interfaces are placed into a blocked standalone state (State 0x03), cutting off the active data path(uplink) across all virtual networks attached to the vDS.

Resolution

To resolve this issue, implement one of the two administrative options detailed below. To maintain cluster availability, place a single ESXi host into Maintenance Mode at a time and execute these changes sequentially across the cluster.

• Option 1: Upstream Physical Switch Reconfiguration.
  • Reconfigure the respective peer ports on the upstream physical switch to LACP Active mode to align correctly with the existing Link Aggregation Group (LAG) configuration on the vSphere Distributed Switch (vDS).
• Option 2: Virtual Network Migration and Alignment
  • Migrate the network infrastructure to a temporary vSphere Standard Switch (vSS) configured with the correct load-balancing policy to match the physical switch topology, then transition the workloads back to the vSphere Distributed Switch (vDS) after provisioning a new distributed port group that mirrors the required load-balancing mechanism.
  • Reference KB for "Migrating ESXi host networking from Distributed Switch (vDS) to Standard Switch (vSS) via CLI".

Additional Information

• ESXi Uplinks in "Stand-alone" State Due to LACP Negotiation Failure
• Converting to Enhanced LACP Support on a vSphere Distributed Switch- "Source vCenter Server has instance(s) of Distributed Virtual Switch at unsupported lacpApiVersion"
• Migrating from Standard to Distributed vSwitch

To speak with a customer representative or a Support Engineer see Contact Support.