Virtual machine never finished full synchronization

Products

• VMware Live Recovery

Issue/Introduction

In VMware Live Recovery, a virtual machine (VM) may enter a continuous loop where it restarts a "full sync" . This behavior typically occurs immediately following the completion of an incremental backup task performed by third-party software like Veeam .You may observe the following event in the ESXi hostd logs at the time of the backup snapshot consolidation:

[Originator@6876 sub=Hbrsvc] ReplicationGroup: received DISK_ATTACHED event for the last disk, reinitializing (groupID=GID-...) (numDisks=2)
[Originator@6876 sub=Vimsvc.ha-eventmgr] Event 1472104 : Full sync started by System for virtual machine...

 

Environment

 

• Product: VMware Live Recovery (formerly vSphere Replication)

• VDDK based 3rd party backup software

 

Cause

This issue is caused by a race condition between the replication's initial synchronization and the snapshot consolidation process of the backup. When a backup finishes, the snapshot is removed and the disk is re-attached to the VM, triggering a DISK_ATTACHED event. If the VM is still in the middle of its initial sync, the replication filter (LWD) re-initializes and loses its current progress, forcing the system to start the full sync over from the beginning.

Resolution

To resolve this issue, you must ensure the initial synchronization completes without being interrupted by disk-attach events:

1. Identify the virtual machines currently stuck in the full sync loop.

2. Pause the Veeam incremental backup schedule for the affected VMs.

3. Allow the initial synchronization to reach 100% and show a status of "OK" or "Success".

4. Resume the daily backup tasks once the initial sync is finished.

Additional Information

Once the VM has moved past the initial sync phase and is performing standard delta replications, subsequent backup tasks and their associated DISK_ATTACHED events will no longer trigger a full sync reset.