vSphere Replication Pairing Failure - 'existingSite' is null
search cancel

vSphere Replication Pairing Failure - 'existingSite' is null

book

Article ID: 441478

calendar_today

Updated On:

Products

VMware Live Recovery

Issue/Introduction

When attempting to pair sites or reconnect a broken site pair in VMware Live Site Recovery (VLR), the operation fails with the following error:

UI Error: Unable to connect or disconnect the vSphere Replication Management Server. Details: 'Cannot invoke "com.vmware.hms.remote.RemoteSiteInt._getRef()" because "existingSite" is null'

hms.log entries: err: Cannot invoke "com.vmware.hms.remote.RemoteSiteInt._getRef()" because "existingSite" is null; com.vmware.vim.binding.hms.remote.fault.SitePairingFault: Cannot invoke "com.vmware.hms.remote.RemoteSiteInt._getRef()" because "existingSite" is null.

Environment

VMware Live Site Recovery 9.X

Cause

This issue is caused by a state inconsistency or stale Site ID entries in the vSphere Replication Management Server (VRMS) database. Common triggers include:

  • One site's VR appliance was redeployed or changed without properly breaking the existing pair.
  • A UUID mismatch where the database at Site A expects a specific identifier for Site B that no longer exists or has changed.
  • Orphaned database entries in the hmsserverentity table preventing new registration. 

Resolution

To resolve this issue, you must establish a clean state for the vSphere Replication appliances by clearing stale identifiers and re-registering the services.

Prerequisites:
Take a virtual machine snapshot of both the Production and Disaster Recovery vSphere Replication appliances.
Ensure you have administrative credentials for both vCenter Servers.

Procedure:

  1. Clear Existing Replications:
  2. Remove any remaining replication jobs associated with the broken pair. You may choose to keep the replica disks as seeds to avoid a full initial sync later.
    Break Existing Pairs:
  3. Forcefully break the site pair from the site that still shows a connection (if applicable).
    Modify Site ID/Naming:
  4. Log into the VRMS VAMI (https://VR-IP:5480) or use the configuration CLI to ensure each site is using a unique and descriptive 'Site Name' label.
    Changing the Site ID forces the system to register as a new endpoint, bypassing the orphaned database entries.
  5. Reconfigure VRMS:
  6. Perform a RECONFIGURE on the vSphere Replication appliances to force a fresh login and registration with their respective vCenter Servers.
    Establish New Site Pair:
  7. Navigate to the Site Recovery UI and create a new site pair between the Production and DR sites.