Using S2D Ready Nodes with Dell EMC servers with internal drives provides a software-defined storage solution at a fraction of the cost of traditional SAN or NAS arrays. Microsoft first released its software-defined storage solution, Storage Spaces Direct, with Windows Server 2016. In 2019, to reflect the integration of Storage Spaces Direct with Azure Stack, Microsoft changed the Storage Spaces Direct name to Azure Stack HCI. Azure Stack HCI is experiencing rapid growth in popularity, with more than 10,000 clusters worldwide running within the first 18 months of general availability.
The rapid growth of the platform has challenged many customers to find a protection solution that enables the IT organization to start with a right-sized solution and then grow incrementally. Using Dell EMC’s Microsoft Storage Spaces Direct Ready Nodes (now Azure Stack HCI Ready Nodes), our engineering lab ran a group of data protection tests to assess whether pairing the DP4400 appliance with Ready Nodes provides an optimal protection solution for VMs. Dell EMC S2D Ready Nodes offer the following benefits:
In our Dell EMC labs, we created a two-node S2D Ready Node cluster. We used the Microsoft Cluster Shared Volumes (CSV) file system, which enables multiple nodes to read and write to the same LUN (disk). CSV provides resiliency in the case of a node failure. The clustered roles can fail over quickly to a surviving node.
We also used Resilient Change Tracking (RCT), which increases backup reliability, scalability, and performance. RCT is a native Change Block Tracking (CBT) feature that was introduced in Windows Server 2016. RCT accelerates backups by providing a way to query a virtual hard disk for changes since the last backup. For every virtual hard disk, two files are created—an .rct file and an .mrt file. Backup solutions such as IDPA DP4400 with Avamar software can use these two files to determine the unique blocks to back up.
From the DP4400 appliance, we installed the Avamar Client plug-in for Windows on each of the two nodes and the Avamar Client plug-in for Hyper-V VSS on the management node. With these two Avamar plug-ins, we ran image-level backups and restores of VMs in the cluster to validate that the backup and restore of the VMs would work as expected. We then ran additional tests to validate file-level and folder-level restores to the backup cluster and VM as well as to a new cluster and VM. Across all seven tests, the DP4400 solution successfully backed up and restored VMs, files, and folders. The backup and restore procedures in our experience were intuitive, which is important when the backup administrator must recover from an unexpected data loss.
The DP4400 plug-in for Hyper-V streamlined our backup and restore procedures. Both activities required only two to four mouse clicks. We monitored progress across all the tests to validate success. The ability to monitor progress, particularly with restores, is important because the business can be promptly notified when a restored VM is available.