Image Management

Overview

Citrix virtualization on Google Cloud includes built in features which are designed to simplify VDA provisioning and image management at scale. These features are often referred to as “Machine Creation Services”, or MCS for short. MCS uses IAM service accounts on Google Cloud to facilitate VDA management on GCP.

Tip:

Before you dive into setting up and using MCS on GCP, review the Citrix documentation on setting up and using MCS on Google Cloud Platform virtualization environments. This document walks you through enabling Google Cloud APIs, creating and configuring the IAM Service Account, creating Hosting connections and resources, and much more.

MCS Issues and Limitations

MCS has a couple of important limitations you need to know about before deploying a Citrix virtualization system on GCP today. As the MCS feature set is delivered as a cloud service, you can expect this list to change over time as the service evolves. In the interim, you need to know about them so you can adjust your design and execution plan accordingly.

The current known issues/limitations are:

  • Provisioning of Linux VDAs. MCS on GCP has not been fully tested for provisioning of Linux VDAs, so this option is not currently supported. Running Linux VDAs on Google Cloud, even with GPUs attached, is. To work around this issue: provision Linux VM instances outside of MCS using Google Deployment Manager templates or another mechanism, then add pre-provisioned instances to a power managed machine catalog for assignment and power management.

Importing VDA Images from On-Premises

Custom VDA images are something many customers may already have and want to use on GCP. While deploying a fresh new instance and configuring that from scratch is preferable, there are times where it may not be feasible. For example, there can be a base image that has been configured on-premises where application owners or application dependencies are not known but are relied upon for critical business operations. Fortunately, on GCP you can import your on-premises image into GCP. Importing is also necessary when deploying Windows client OS variants (i.e. Windows 10) as Windows client operating systems are not natively available in Google Cloud’s image catalog. For more information, see importing a virtual disk.

Note:

Before importing an existing disk, be sure to read and understand the differences of importing a virtual disk from your on-premises environment. Where possible, it’s preferable to deploy new instances from the public image library and create your master images from scratch.

Using Sole-Tenant Nodes on GCP

Google Cloud has a feature called sole-tenant nodes which are useful for various use cases, including bringing your own license for Windows OS, and deploying Windows client VDAs on GCP. When configuring sole-tenant nodes, you can configure them in one or more zones within a GCP region. Citrix MCS fully supports provisioning VDAs on sole-tenant nodes, but is not automatically aware of which GCP zones your sole-tenant nodes are deployed in. If you intend to deploy VDAs to sole-tenant nodes, be sure to check out the GCP Zone Selection documentation for details.

Note:

For a good tutorial on deploying Windows 10 on GCP, see the GCP Windows 10 Sole Tenant with Optional Shared VPC Catalog Creation POC guide. The article covers both custom image import and sole-tenant node usage on GCP.

Image Management

In this article