In two earlier posts, I showed you how you can use Azure File Sync to significantly reduce the your on premises storage of files to free up capacity and extend the life of your existing storage infrastructure.
To recap, I walked through the process on a Windows Server Fileserver VM hosted on VMware which had a 1 TB thick provisionsed VMDK file – and reduced it by more than 90%! The high-level steps we followed were:
- Install / configure Azure File Sync with cloud tiering enabled
- Tiered all the files to Azure (freeing up space on the local server)
- Zeroed out the disk using Sdelete
- Compressed the VMDK by cloning the VM using a different storage policy (thin)
The only real downtime we took was during the cloning process, which took several hours to complete – but there are options you can use to all but eliminate down time.
Reducing / Eliminating Downtime
Sometimes the objective in reducing the size of a VM is other than simply to save space – it could be to make migration to a new platform or data center simpler / shorter.
Often replication technology (like Azure Site Recovery or VMware HCX) can replicated storage to a new infrastructure (hopefully Azure!) without incurring the downtime we saw with cloning, since it can move most of the data while the VM is running.
With the help of one of my co-workers, we used VMware’s HCX to replicate the same VM (with the 1TB thick provisioned VMDK) to Azure VMware Solutions, which lets you run your VMware workloads natively on Azure.
The total process to replicate to an AVS cluster in Azure (using vSAN) took about 4 hours to complete:
…and the space was reclaimed with minimal downtime: