We need this also. Right now Azure is like Hotel California - you can check in to Azure, but you can never leave.
We are accustomed to Hyper-V Replica where we can start with a machine in any location and fail it over to another location. Currently there is no way to initiate ASR from a machine that originated in Azure back to on-prem; and furthermore the button "Complete Migration" if accidentally pressed breaks the link to on-prem, and the only way to restart this is to copy the VHD down to on-prem, then re-setup the link.
206 votesplanned · AdminAzure IaaS Engineering Team (Azure IaaS Engineering Team, Microsoft, Microsoft Azure) responded
Currently you can increase the size of a disk when the VM is in the stop-deallocated state. We are also working on a solution where a disk can be increased in size while attached to a running VM.
This creates a problem when trying to retrieve a VHD back to on-prem - the fully-populated VHD size may exceed the disk space of the on-prem machine, even though the used data of the VHD is small.
This one isn't going to happen - sorry. The new portal is better in nearly every respect. Spend some time in it and you will find what you need.
341 votesunder review · AdminAzure IaaS Engineering Team (Azure IaaS Engineering Team, Microsoft, Microsoft Azure) responded
The status of this item has been moved back to Under Review. We initially planned to move to VHDX support as part of our support for HyperV Gen2 VMs, but we ended up using the VHD format for Gen2 VMs as well. Some aspects of the Azure Infrastructure do not cleanly support VHDX OS or data disks. So this feature is dependent on some of these internal services being updated which is an ongoing process.
Agreed - "Cloud First" is running on a platform 2 generations old. Let's go here folks!
We need VHDX support because of the time it takes to downconvert during an Azure Site Recovery failover operation. This adds 20 minutes to an hour to recovery time.