We have hibernation enabled on our personal pools. We are having issues when users attempt to resume their hibernated VM but it fails because of capacity not being available on the Azure side. We configured Capacity Extender with additional backup SKUs hoping to have Nerdio start those VMs with a different SKU when the issue happens. When a user attempts to start the VM from the user self service portal, Nerdio still does not start it because hibernated VMs cannot be resized per MS KB article https://learn.microsoft.com/en-us/azure/virtual-machines/hibernate-resume#general-limitations .
We are asking for some additional controls to be implemented for admins to optionally use them if needed. When this issue happens the user can either wait and try later (which is not great), or deallocate the VM (by losing session state) and start it again (which normally succeeds). We need Nerdio to allow admins to have a configurable option that would change the flow of how the VM is started when the user attempts to start it from the user self service portal. 1. First it should attempt to normally start the hibernated VM in the default SKU. 2. If there is no capacity, then it should deallocate the VM (maybe a consent can be asked here to admins and/or users since the session state will be lost) and attempt to start it again on the default SKU (this should normally be enough). 3. If step 2 fails for capacity reasons again and if the admin has configured alternative SKUs in Capacity Extender, then Nerdio should attempt to start the VM using the alternative SKUs.
Comments (2 comments)