Azure Site Recovery (VMware) Failback for Modernised – Does reprotection replicate full disks or only deltas, and what happens to the original on-prem disks?

Hi @Anandha Chandrasekaran Thank you for the detailed scenario and and it's important to understand how Azure Site Recovery (ASR) handles reprotection and failback before planning a production recovery workflow.

Assuming your original on-premises VMware VM still exists, its VMDKs remain available, and the datastore is accessible, the failback process uses Original Location Recovery (OLR).

1. Does ASR transfer the full 2 TB again or only the changes?

In an Original Location Recovery (OLR) scenario, Azure Site Recovery replicates only the changes back to the original on-premises VM rather than performing a full data download. Microsoft explicitly states.

"If you failed over a VMware virtual machine, you can fail back to the same source on-premises virtual machine if it still exists. In this scenario, only the changes are replicated back."

Therefore, if only a small amount of data changed while the VM was running in Azure, ASR is expected to transfer those changes rather than retransferring the full 2 TB of disk data.

2. How does ASR determine what needs to be replicated?

Azure Site Recovery uses block-level replication, and Microsoft documents that resynchronization operations minimize data transfer by computing checksums and transferring only delta changes when appropriate. After the initial replication process, Azure Site Recovery continues with delta replication rather than repeatedly sending the entire disk contents.

Because the exact amount of transferred data depends on the changed blocks and the synchronization state, Microsoft does not document an exact transfer size calculation for every failback scenario. However, the documented OLR behavior is that only changes are replicated back.

3. What happens to the original VMDKs?

In an OLR scenario, ASR is designed to identify the original VMDKs and write the pending changes back to them. Microsoft specifically recommends enabling:

disk.enableUUID = TRUE

so that Site Recovery can correctly identify the original VMDK during failback.

Microsoft further notes that if the original VMDK cannot be identified, ASR may create an additional disk and write the data there instead.

For this reason, ensuring that the original VM, datastore, and VMDKs remain available is an important prerequisite for a successful OLR failback.

4. What if the original on-premises VM no longer exists?

If the original VM is unavailable, ASR uses Alternate Location Recovery (ALR).

In this case, Microsoft documents that:

• A new on-premises VM is created.
• A full data download is performed.
• New VMDKs are created on the selected datastore.

5. Are newly added Azure disks replicated back during failback?

No. Microsoft's failover/failback documentation states that only disks that were originally replicated from on-premises to Azure are replicated back during reprotection. Disks added to the Azure VM after failover are not replicated back to the on-premises machine.

Please find below references for more information:

• Failback of VMware virtual machines after disaster recovery to Azure
• About failover and failback in Azure Site Recovery - Modernized
• VMware to Azure disaster recovery architecture - Modernized
• Prepare for reprotection and failback of VMware VMs

If you find the answer helpful, please click "upvote" and accept it. This will help others in the community with similar questions easily find the solution.

Welcome to Microsoft Q&A

Hello @Anandha Chandrasekaran I hope you are doing well,

Here is exactly what happens

• What gets transferred? ASR will only send the 5 GB of changed data back to your local network. When you click "Reprotect," Azure simply scans your local 2 TB disks, compares them to the cloud, finds the 5 GB difference, and only sends those specific changes.
• What happens to the original disks? Your original local disks are kept completely safe while you are running in Azure. ASR reuses them. It essentially patches your original VMDK files with the 5 GB of new data. It does not create brand new 2 TB disks.
• What happens after failback? When you fail the server back to your local environment, it turns on using your original disks. Because Azure updated them directly, there is no leftover "junk" data or old disks that you need to manually delete.
• https://learn.microsoft.com/en-us/azure/site-recovery/failover-failback-overview-modernized
• https://learn.microsoft.com/en-us/azure/site-recovery/vmware-azure-architecture-modernized

😊 If my answer helped you resolve your issue, please consider marking it as the correct answer. This helps others in the community find solutions more easily. Thanks!

During reprotection from Azure back to VMware, ASR does not transfer the entire contents of both disks again. If the original on-premises VM and its disks are still available, ASR will transfer only the changed blocks (the 5 GB delta). It accomplishes this by performing a full content checksum analysis across the full 2 TB of data to match block hashes between Azure and your local storage. Once this comparison phase finishes, ASR switches to delta/block-level replication, transferring only the identified changes over the network.

The original on-premises VMware disks and VMDKs that existed before the failover are reused and updated. ASR does not overwrite the files with a completely new blank copy - instead, it overwrites the specific blocks inside the existing VMDKs with the changes replicated down from Azure. ASR does not create new replica VMDKs for this process unless the original on-premises VM or its datastore files were deleted or became inaccessible during the failover period.

After a successful failback, the recovered VM uses the original VMDKs that existed before the failover occurred. Because ASR directly updated those existing files during the reprotection phase, they are already up to date and ready for production. During the Azure failover period, your original on-premises disks remain intact and are neither automatically deleted nor replaced. They are preserved by ASR to act as the seed baseline, which eliminates the need for any manual storage cleanup after your failback is complete.

More at https://learn.microsoft.com/en-us/azure/site-recovery/failover-failback-overview-modernized

If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin