vCenter HA configuration fails with the error, "vCenter HA has an invalid configuration. Remove vCenter HA to destroy the current cluster configuration and set up vCenter HA again"

Products

• VMware vCenter Server

Issue/Introduction

• The "Deploy a vCenter HA cluster" task fails with the following error status on the vCenter UI:
  
  A general system error occurred: Failed to ssh connect peer node [Passive node IP/Witness node IP]
  
  
• Active node is unable to establish an SSH handshake with Passive or Witness nodes despite sshd service being in a running state.
  
  
• In /var/log/vmware/vcha/sshconnect.log, the following error appears:
  
  WARNING sshConnect retry attempt #### failed [Errno None] Unable to connect to port 22 on [Passive node IP/Witness node IP]
  
  
• Connectivity is restored only when a network activity is triggered (ping to the gateway from the passive/witness node VM console or a vMotion is initiated)

Environment

VMware vCenter Server 8.x

Cause

• This issue occurs due to an ARP resolution delay on the physical switch environment.
  
  
• The environment exhibits ARP resolution failures where the Active node cannot map the hardware address of cloned peer nodes on the vCenter HA VLAN resulting in "Incomplete" ARP entries on the physical switch, preventing the routing of traffic over port 22.

Resolution

• Refer the below KB Article on requirements for VCHA and validate if all the requirements are met and if the the passive or witness nodes are still unreachable over network, engage internal network team to investigate why there are ARP delays in the physical environment and reattempt the VCHA workflow. Deploying vCenter High Availability with network addresses in separate subnets - vSphere 6.7, 7.0 and 8.0
  
  
• If the issue persists even after resolving ARP delays, Contact Support for further assistance.