Troubleshooting “No Connection to VR Server: Unknown” Errors During Bulk Migrations with VMware HCX

Products

• VMware HCX

Issue/Introduction

During bulk migrations in VMware HCX, administrators may encounter a “No Connection to VR Server: Unknown” error. This error often appears during the switchover phase, causing the source VM to power off before the destination VM can power on, leading to downtime or necessitating a rollback. This article provides high-level troubleshooting steps to resolve the issue and ensure your environment is operating at the recommended VMware and HCX versions.

For further details on vSphere Replication prerequisites—such as network connectivity, NTP settings, and Enhanced Replication—please see Broadcom KB Article 379468.

Environment

VMware HCX

Cause

Partial Connectivity to the VR Server(IX)

Only some ESXi hosts in the cluster have connectivity to the IP Address of the IX vSphere replication on port 31031. 

Transient Connectivity or VR Server Issues

Temporary network glitches or vSphere Replication server hiccups can trigger this error during switchover.

HCX Site Pairing or Authentication Problems

Incorrect credentials or incomplete pairing between the source and target vCenter/HCX instances.

Version Incompatibility

Outdated ESXi, vCenter, or HCX versions that do not meet the recommended requirements for stable replication and Enhanced Replication.

Miscellaneous Factors

VM guest OS or VMware Tools issues are preventing a clean power-off/on during the final switchover.

HCX Switchover Delays and Broken Pipe

The vSphere Replication service is not explicitly enabled on the dedicated replication VMkernel interface (vmkX), causing the Host-Based Replication (HBR) daemon to default to the non-optimized management interface (vmk0) for all outgoing replication traffic to the HCX IX appliance on TCP port 31031 which can cause HCX Switchover Delays and Broken Pipe

Resolution

Resolve any connectivity issue between the ESXi and the VR server(IP Address of the IX vSphere replication)

• Make sure necessary ports are opened between the ESXi servers and the VR server(IP Address of the IX vSphere replication) - especially port 31031 - Refer to VMware Ports and Protocols
• From the ESXi host
  

  nc -zv <IP Address of the IX vSphere replication> 31031

Stop and Retry the Migration

• If you see a “No Connection to VR Server: Unknown” error during a bulk migration, cancel/stop the current migration task and retry.
• A second attempt often resolves transient network or replication server issues.

Clone the VM and Test Migrate

• If downtime is a concern (e.g., the source VM powers off and the destination does not power on), create a clone of the problematic VM.
• Test the migration on the clone to see if the error persists.
  • If the cloned VM succeeds: Retry the migration on the original VM.
  • If the clone also fails: Proceed to updates and further troubleshooting.

Confirm Environment is Up to Date

• HCX: Upgrade to 4.11.3 (or newer).
• vCenter Server: Recommended upgrade to 8.0u3 (or newer).
• ESXi Hosts: Recommended upgrade to 8.0u3 (or newer).
• Make sure VM is on the latest version of VMware Tools
• For additional details on network configurations, NTP synchronization, and replication testing, refer to Broadcom KB Article 379468.

Contact Broadcom Support if Issues Persist

• If errors continue after retries and environment upgrades, open a support case with Broadcom.
  • Provide
    • Logs (vCenter, HCX, ESXi, vSphere Replication)
    • Error Messages with timestamps
    • Environment Details (versions, deployment details)
  • Broadcom Support can perform more advanced log analysis and guide additional remediation steps.