With a data disk, you can provide data, applications, or extra storage for your end users. You can use data disks both with image VMs automatically created by the Import Virtual Machine from Marketplace wizard or image VMs that you have manually created and paired with your Horizon Cloud environment. The system supports using data disks with session-based desktops and remote applications from RDSH farms, floating VDI desktop assignments, and dedicated VDI desktop assignments. However, because of the different nature of the assignment types, the use cases vary for each type.
Data Disks and Dedicated VDI Desktop Assignments
The dedicated VDI desktop assignment is the most common use case for data disks. Initially, each desktop VM in the assignment's pool of VMs has the same data disk configuration and contents as for the original image VM on which the assignment is based. You might provide data and applications on the initial data disk that you want to give to all of the entitled end users. Each end user in a dedicated VDI desktop assignment is assigned a specific virtual desktop. The assigned end user returns to the same virtual desktop each time they launch the desktop and log in. Because the data disk persists with that virtual desktop, the assigned end user can make changes to the data on the data disk and all of the user's changes are preserved between sessions.
Data Disks and Floating VDI Desktop Assignments
In a floating VDI desktop assignment, each virtual desktop VM is reverted to the initial state of the original image VM when an end user logs out of the desktop. Like in the dedicated case, initially each desktop VM in the assignment's pool of VMs has the same data disk configuration and contents as for the original image VM on which the assignment is based. Also as in the dedicated case, you might provide data and applications on the initial data disk that you want to give to all of the entitled end users. Each time an end user connects to a desktop from the pool, that end user is connected to a desktop with any data disks in their initial state.
Unlike the dedicated case, when the end user logs out of the desktop, the virtual desktop's data disks are reverted to the initial data disk configuration and contents. Any files the end user might have saved to those disks are lost when the user logs out.
Data Disks and RDSH-Based Desktop and Application Assignments
The main use case for using data disks with RDSH VMs is to provide shared, read-only data or applications to all of the end users that you will entitle to use session-based desktops and remote applications provisioned from the RDSH farm. Any data disk that is attached to an RDSH VM is available to all end users that connect to that VM for their session-based desktops and remote applications. Also, because an end user might get connected to different VM instances each time the end user logs in to use their entitled desktop or application, there is no guarantee a particular end user would be able to access data they saved to the data disk during previous sessions. As a result, using data disks for personal data in this scenario is typically avoided.
To have data disks available to the pod-provisioned virtual desktops and remote applications, you use the Microsoft Azure portal to create the disks and attach them to a golden image VM before publishing the image. At a high level, you:
- Attach the created data disks to the VM.
- Initialize those data disks according to the steps in the Microsoft Azure documentation topic Attach a managed data disk to a Windows VM by using the Azure portal. Those steps include initializing the disks, defining volumes, and formatting partitions as appropriate for your needs.
- Add any initial contents you want on the data disks.
You must perform those steps before you convert the golden image to a published image. The system's image publishing process captures that initial state of the data disks as it seals the image. After publishing the image, you cannot add data disks to that sealed image. To update a sealed image for any reason, including adding a data disk, you update the image according to the information inManaging Published Images for Horizon Cloud Pods in Microsoft Azure and its subtopics.
For the detailed steps of preparing a data disk for a image VM used in Horizon Cloud, see Setting Up a Data Disk for an Image VM in Horizon Cloud.
Number of Data Disks per VM
The current recommendation on the number of data disks supported in Horizon Cloud on an image VM is up to five (5) data disks. Additional factors might restrict the number of data disks you can attach to a VM, such as Microsoft Azure policies about how many data disks can be attached to a particular VM size and the Microsoft Azure region into which your pod is deployed. See the Microsoft Azure documentation topic Sizes for Windows virtual machines in Azure and the pages for the various VM types for the charts that show the maximum numbers for each Microsoft Azure VM size.
Life Cycle of Data Disks
When you use the Horizon Cloud administrative console to delete a VM, the system looks for all of the resources associated with the VM and deletes those resources. Even though you manually created the data disks in the Microsoft Azure portal, when the data disks are attached to a VM in Horizon Cloud, the system will delete those data disks when it deletes the VM.
When farm RDSH instances and VDI desktop instances are created from images with attached data disks, the data disks for those instances are created and deleted automatically when the RDSH and desktop VMs are created and deleted, according to the system's standard behavior.