Create a Virtual Workstation
Get started with creating a Virtual Workstation
A Virtual Workstation is a specialized use of a CoreWeave Virtual Server.
In this tutorial, several Virtual Workstations are created using the CoreWeave Cloud UI. The first machines demonstrate different configurations for artist use for both Windows and CentOS 7. Then, a few lighter-weight machines are deployed for administrator use.
Network Policies are deployed to provide strong security barriers between these different types of users.
Prerequisites
This guide assumes that the user already has an active CoreWeave Cloud account.
Create a Virtual Workstation
Log in to your CoreWeave Cloud account, then navigate to the Virtual Servers configuration screen either by clicking the Deploy Now button on the Deploy Virtual Server card, or by clicking the Virtual Servers link on the left-hand navigation.
 (1).6c6fdc3.1200.png)
Create additional machines for artists
Simple Virtual Workstations for artist use are easily created from the Virtual Server deployment form. In these examples, either Teradici or Parsec is installed as an Operating System add-on to provide remote access to the machines.
- You may choose to create your own Parsec Teams account for management capabilities, or you can license and administer Parsec through CoreWeave. To do this, please contact your CoreWeave support specialist.
- V100s are not compatible with Teradici. The best solution when using Teradici for remote access is to use a Quadro GPU, such as an A6000, A5000, A40, RTX6000, RTX5000 or RTX4000. For more information, see Node Types.
Begin by repeating the process for creating a new Virtual Workstation: navigate to the Virtual Server creation screen and begin configuring the machine.
For this tutorial, the first artist machine will run Windows 10 Professional. Select Windows 10 Professional from the Operating System configuration screen:
.2c1a230.1200.png)
Continue to configure the artist machine.
If you are utilizing Active Directory, you will have to put in a temporary user before you connect the storage volume.
Windows with Parsec
To deploy a Windows 10 Workstation with Parsec, click the Parsec Remote Desktop toggle under Operating System Additions.
.ea0c54f.1200.png)
Next, click the Deploy button to launch the Virtual Workstation in your namespace. Once the Virtual Workstation is running, log into it. Start Parsec, then enter your credentials.
Windows with Teradici
To deploy a Windows 10 Workstation with Teradici instead of Parsec, click the Teradici toggle under Operating System Additions.
.9595e8c.1200.png)
Deploy the Virtual Server by clicking the Deploy button. Once the Virtual Server is done initializing, the credentials specified in the creation interface may be used to connect to it.
Next, click the Deploy button to launch the Virtual Workstation in your namespace. Once the Virtual Workstation is running, log into it. Start Teradici, then enter your credentials.
To add additional applications upon initialization, the Windows package manager Chocolatey is recommended.
To use Chocolatey, add a list of application names to the choco_install parameter in the Workstation YAML manifest in the cloudInit block.
For example:
cloudInit:
choco_install: [googlechrome,firefox,vlc]
For a list of valid application names, refer to the Chocolatey community packages directory.
Linux Workstations
At this time, Parsec is not supported on CentOS 7. Teradici requires NVIDIA Drivers to be installed.
Storage in Linux Workstations is mounted in via Virtiofs. The process is the same as the other workstations: select the desired Linux distribution and version, then toggle the desired Operating System Additions.
.de141b4.1200.png)
Workstation YAML manifests
If you are comfortable with Kubernetes, you may view the Workstation configurations as YAML manifests. Click the EDIT YAML tab on the right-hand side to expand the YAML editor.
Workstation configurations will look similar to this example manifest:
Click to expand - Example Workstation manifest
apiVersion: virtualservers.coreweave.com/v1alpha1
kind: VirtualServer
metadata:
namespace: tenant-coreweave-name
name: new-1252
annotations:
external-dns.alpha.kubernetes.io/hostname: new-1252.tenant-coreweave-name.coreweave.cloud
labels: {}
spec:
region: LGA1
resources:
definition: a
gpu:
type: Quadro_RTX_4000
count: 1
cpu:
count: 4
memory: 8Gi
os:
definition: a
enableUEFIBoot: false
type: windows
storage:
root:
storageClassName: block-nvme-lga1
volumeMode: Block
accessMode: ReadWriteOnce
size: 79Gi
source:
pvc:
namespace: vd-images
name: win10-master-20230331-lga1
filesystems: []
additionalDisks: []
network:
public: true
directAttachLoadBalancerIP: true
users:
-
username: ''
password: ''
cloudInit: |
autologon: true
parsec: true
edid: true
readinessProbe:
failureThreshold: 30
initialDelaySeconds: 10
periodSeconds: 20
tcpSocket:
port: 1337
runStrategy: RerunOnFailure
initializeRunning: true
With the exception of the source image (root.source), which specifies the Operating System image, all other values will remain the same for each workstation.
You may decide at this point to load up a Virtual Workstation with the applications you'd like to be incorporated in a base image. For more information on how to do this, review our Root Disk Lifecycle Management guide.
Configuring administrators
Typically, administrators and support staff do not need the same resources as artists for their work. Setting up an administrator machine can be done in a few simple steps.
In this example, a new Virtual Machine is configured running Windows 10.
.e3efb35.1200.png)
When artist machines are powered off, only the use of persistent storage and public IP addresses are billed.