High storage utilization after recovery and configuring replication in reverse direction using vSphere replication

Products

• VMware Live Recovery

Issue/Introduction

Symptoms:

• vSphere replication is in use without a site recovery manager.

• Since there is no site recovery manager, failover and failback of virtual machines needs to be performed manually using vSphere Replication

• Virtual machines are recovered using vSphere Replication using the steps documented in Recover Virtual Machines with vSphere Replication

• After recovering the virtual machine, the status of the virtual machine changes to recovered and replication is disabled

• The recovered virtual machine is removed from replication by selecting the option 'Retain replica seeds'.

  

• The replications are reconfigured in the opposite direction (DR to DC)

• While the replications start syncing, datastore usage increases on the target datastores

Environment

vSphere Replication 9.x

Cause

This issue occurs due to a combination of manual actions during the replication lifecycle of a virtual machine:

1. After recovery, the recovered virtual machine is removed from the replication configuration.
2. During this removal, the "Retain replica seeds" option is selected. This prevents vSphere Replication from automatically cleaning up the replica files on the old target datastore.
3. The replication is configured for the same VM, but this time a different datastore is selected as the target.
4. vSphere Replication creates new replica files on the newly selected datastore. The old replica files, which were deliberately retained as seeds but are no longer used by any replication job, become orphaned.

This results in the virtual machine files being duplicated across two different datastores, leading to an increase in overall storage consumption. 

Resolution

Remove the duplicate replication files from the old target datastores to free up the space