A new feature in Windows Server 2016 is an expanded virtual machine start order for clustered virtual machines. Previously in Hyper-V you were able to assign high, medium and low priorities to VM’s. For example, you could have a group of DC’s that were considered high priority and would start first, a group of, say DB servers deemed medium that would start 2nd and a group of low priority web/app servers that would start last.
It offered administrators flexibility, but there were only three buckets in which to place your servers. While this may be okay for small VM deployments, it was lacking for larger deployments that housed one or more complex, multi-tiered applications.
The new start order priority, for clustered virtual machines, gives admins the ability to define sets, assign VM’s to those sets and also specify dependencies between sets and VM’s. It expands your ability as an admin to define when VMs (that provide critical infrastructure services like DC’s, DNS, virtual network appliances) start in the correct sequence, or when the various tiers of your application start. This is all powered via PowerShell cmdlets at the moment so there is no GUI. You can get a list of cmdlets by running the following cmd on your server: “Get-Command *ClusterGroup*.
This new feature will provide some enhanced tools for a more organized and automated recovery in the event of a planned or unplanned event. In the past, such events were done by spreadsheet, script or manually powering on VM’s based on naming convention. With this new toolset, admins can breathe a little easier when a cluster node fails or their generator didn’t quite come online before the UPS dropped off. It won’t protect you from data loss, but it will make powering up your VM’s (in the correct order) significantly easier. Such operational ease will remain true when the necessary work is put in up front.