I am dealing with an issue for sometime now which is that Domain Controllers Virtual Machines that are replicated to Microsoft Azure using the Azure Site Recovery service, are failing to take Application-Consistent snapshots while they can take Crash-Consistent snapshots normally.
Of course prior setting everything up, I have ensured Microsoft supports replicating production Domain Controller VMs from premises to Azure.
All the environments in where the issue occurs consist of Windows Server 2012 R2 Hyper-V hosts and the guest VMs are Windows Server 2012 R2.
Please note all VMs are running only AD DS/DNS services as workloads thus the possibility of conflict with additional installed software nearly gets eliminated.
I have confirmed that there is a specific pattern followed when the issue occurs which is that NTDS writer is always in failed and non-retryable state and relevant failure events are always logged in Event Viewer.
Since this issue is recurring, I have logged a ticket with Microsoft's support but while it had been escalated numerous times the only answer I got was to uninstall the Hyper-V integration Tools from the affected VMs and see what happens.
Now what happens once Integration Tools get uninstalled from the affected VMs, the Application-Consistent snapshots are being taken successfully however once VMs install Windows Updates and restart, the Integration Tools will reinstall their selves once again resulting to the very same issue.
My question is why does this happen? Is it a known bug or something Microsoft is aware of?
Even if permanently uninstalling Integration Tools and ensuring they won't be installed again (if that's possible) resolves the issue I think the community has to get an official answer from Microsoft for the root cause of the issue as well as for its resolution.
Finally, despite I didn't find any relevant issue posted on the internet I have the sense that many others will be dealing with this issue so I hope if this get resolved to help others in the future as well.