SRM recovery fails with error: ‘Recovered Datastore Is Not Visible to Host’

Products

• VMware Live Recovery

Issue/Introduction

Symptom:

• The recovery plan is in “Incomplete Recovery” state, and the Re-protect option is greyed out.

• Within the Array Pair view, the status of the datastore group indicates that failover is complete, confirming that the storage array has successfully failed over the LUNs to the recovery site.

  

• In some cases, the ESXi host appears in a Responding state in vCenter, but recovery still fails on that host.

Validation:

• Review the history report of the recovery plan execution. It shows that the affected datastore is successfully recovered. However, the recovered datastore failed to mount on one of the ESXi hosts at the recovery site.

  

• You will notice that the virtual machines which failed to recover were mapped to the same host as their recovery host 

  

• Log into the vCenter Server at the recovery site and observe that the recovery host in question is in a “Not Responding” state.

• Log into DCUI and check if the ESXi has PSOD.
  
  
  
• Logging to the hardware interface of the host will report hardware failure.
  

Environment

VMware Live SIte Recovery 8.x

VMware Live Site Recovery 9.x

Cause

During a planned migration, the replica or snapshot LUN at the recovery site must be correctly mapped to all ESXi hosts within the recovery cluster where the placeholder virtual machines reside.

If one of the ESXi host in the recovery cluster is in a “Not Responding” or failure state during the migration process then we will observe:

 • The recovered datastore could not be mounted on that host.
 • The virtual machines mapped to the recovery host failed to recover, since the required datastore was not accessible on that host.

This led to a partial recovery with some VMs in a failed state and the recovery plan marked as Incomplete.

Resolution

The issue will be resolved by restoring connectivity between the unresponsive host and vCenter. If the host cannot be repaired immediately, temporarily remove it from the cluster. Once this is done, re-initiate the recovery process. 

Workaround

As this environment uses array-based replication, the recovery should be completed manually using the following steps:

• Manually recover the affected virtual machines by registering the respective .vmx files from the target datastore onto other available ESXi hosts in the recovery site where the datastore is accessible.

• Delete the existing Protection Group and Recovery Plan to clear the incomplete recovery state.

  Note: If any manual IP customization is configured within the recovery plan, please ensure to document and preserve this information before deletion, as it may be lost.

• Engage the storage vendor to reverse the direction of replication, ensuring the recovered VMs at the recovery site can be protected going forward.

• Recreate the Protection Group and Recovery Plan to reflect the updated environment and to enable future planned migrations or failovers.

 

Additional Information

If the host is responding but still preventing datastore mounting during SRM recovery, add the SRM tag SRM-autoPHVmDatastoreExclude to exclude the problematic host from participating in recovery.

After applying the tag and rerunning the recovery, the plan completes successfully without requiring manual VM registration.

Clarification that this is a temporary workaround, not a replacement for root cause remediation.

A note that once the host is remediated, the tag can be removed to restore full participation in recovery.