VDI Hardware Templates
- Thread starter ArjenF1
- Start date
Thank you for the suggestion. I'm adding this to our backlog. Could you elaborate on the specific settings you would like to see in this hardware template?
Hi Freek,
We also have a ticket open regarding this (5401326).
Currently the VDI solution offered is based on a template which is also your golden image.
With this in mind the only way to provide multiple hardware specs is to create multiple images which means multiple images would need to be maintained.
What we would prefer is for the templates to become the sort of golden image where the resources assigned to that template wont matter to parallels.
Instead of taking the hardware preferences of this template a new option will be created (on a site basis) where you can create hardware templates.
For example:
Bronze = 2vcpu, 8gb ram
Silver = 2 vcpu, 16gb ram
gold = 4 vcpu, 16gb ram
platinum = 6vcpu, 32gb ram
Then on your host pool you would be able to select your current template (now golden image) and the required hardware template.
This would also mean you need to allow multiple host pools to select the same image as you would then create host pools for each of the required hardware templates. from there you would publish each of these pools and filter access to those based on a security group.
This way it would become a lot easier to rightsize a vdi based on user requirements instead of seeking a compromise where some users might have a oversized profile and some an undersized profile (which will generate compliants).
Also it would make the maintenance of the several vdi templates (golden images) a lot more friendly as you would now need only one template instead of multiple to provide the required solution.
We also have a ticket open regarding this (5401326).
Currently the VDI solution offered is based on a template which is also your golden image.
With this in mind the only way to provide multiple hardware specs is to create multiple images which means multiple images would need to be maintained.
What we would prefer is for the templates to become the sort of golden image where the resources assigned to that template wont matter to parallels.
Instead of taking the hardware preferences of this template a new option will be created (on a site basis) where you can create hardware templates.
For example:
Bronze = 2vcpu, 8gb ram
Silver = 2 vcpu, 16gb ram
gold = 4 vcpu, 16gb ram
platinum = 6vcpu, 32gb ram
Then on your host pool you would be able to select your current template (now golden image) and the required hardware template.
This would also mean you need to allow multiple host pools to select the same image as you would then create host pools for each of the required hardware templates. from there you would publish each of these pools and filter access to those based on a security group.
This way it would become a lot easier to rightsize a vdi based on user requirements instead of seeking a compromise where some users might have a oversized profile and some an undersized profile (which will generate compliants).
Also it would make the maintenance of the several vdi templates (golden images) a lot more friendly as you would now need only one template instead of multiple to provide the required solution.
Should be Super easy, barely an inconvenience..
Clones are generated from master template snapshots. "Clone to new vm".
From my guess the Parallels Ras provider connects to Nutanix cluster with Restapi to do this action.
Changing VM allocation of vCPU or RAM should be a small task to add.
My wish would be to either:
1. In host pool set the size of the provisioned machines, eg. not just take the Master Template hardware size
2. being able to create a hardware template (vCPU, Memor or maybe also CPU class) and select this in Host-Pool.
Changing the host-pool vm's could be done either on the fly (adding CPU/Memory as separate actions) or less complicated during rebuild (before powering on vm, running rasprep etc).
Clones are generated from master template snapshots. "Clone to new vm".
From my guess the Parallels Ras provider connects to Nutanix cluster with Restapi to do this action.
Changing VM allocation of vCPU or RAM should be a small task to add.
My wish would be to either:
1. In host pool set the size of the provisioned machines, eg. not just take the Master Template hardware size
2. being able to create a hardware template (vCPU, Memor or maybe also CPU class) and select this in Host-Pool.
Changing the host-pool vm's could be done either on the fly (adding CPU/Memory as separate actions) or less complicated during rebuild (before powering on vm, running rasprep etc).
We are working on supporting override capabilities to allow overriding the VM Size on a host pool level, providing more flexibility. More news to share soon, stay tuned!
Really cool. Love that Parallel listens to community.
On a side note, a golden image creation tool would also be great. This could create a golden image vm, that we could clone and import as template. Nice for us map's that have a lot of templates, and at least one template per customer.
On a side note, a golden image creation tool would also be great. This could create a golden image vm, that we could clone and import as template. Nice for us map's that have a lot of templates, and at least one template per customer.
As part of Parallels RAS v 21.2, we introduced Flexible VM sizing and Active Directory OU assignment for host pools. This allows you to configure VM sizing and Active Directory organizational unit placement at the host pool level, reducing template sprawl and simplifying large-scale session host deployments. We are now supporting this for Azure, AVD, AWS, and Nutanix providers. https://www.parallels.com/products/ras/whats-new
This is a great addition — template sprawl is one of those pain points that quietly grows over time, especially in larger environments where every VM size/OU combo used to need its own template. Moving that logic to the host pool level should cut down provisioning overhead significantly. Glad to see it rolled out across Azure, AVD, AWS, and Nutanix rather than limited to one provider — that consistency across hybrid/multi-cloud setups is exactly what admins managing mixed environments need. Looking forward to testing this out in our own host pool deployments."