[VCDR] Failback plan may fail with one or more VMs failing to restore.

Products

• VMware Live Recovery

Issue/Introduction

To understand the cause of the Failback failure.

 

 

 

 

Symptoms:

Failback plan fails from VMC to On-premises with one or more VM failing to successfully restore.

Cause

This may happen due to a snapshot task running at the time when failover is initiated. A new snapshot ID is updated and conflicts with the failover task. 

This is a race between failover and PG snapshot creation. The VM snapshot are discarded because the PG snap could not be created (due to schedule disable during failover) and the latest VM snaps will not exist, so this check will fail and we will fall back to full restore.

 

Resolution

VCDR engineering has analyzed and plan to fix the orphaned snapshot id in future release. 

Starting with V27, following enhancements are available for more visibility into the restore tasks.

a. If a VM falls back to a full restore, there will be an event raised for it in the event tab in the UI. It will also have some information about why the VM fell back to full restore.

b. The percentage progress bar in the restore task will help keep a check on the progress. 

Workaround:

As a workaround, please disable the schedules of PGs manually before starting a DR operation. 

Additional Information

Impact/Risks:
Customer unable to failback the VMs back to source on-prem datacenter.