Hi team,
We currently have a Host Pool that contains an equal number of Session Hosts in both Primary and Secondary Regions. We also have enabled the 're-image' on logoff setting. As the reimage process takes some time, Autoscale kicks in and adds a new Session Host, before waiting for the re-image to complete, in order to fulfil the 'Min Active Session Hosts' option.
This autoscale logic consistently adds to the primary region, as there is only 1 add host job, it doesn't 'need' to balance them across both regions. Is it possible to base the autoscale logic off the current host pool allocation?
The scale in task then removes VMs from the secondary region...
Solution would be something like checking Host Pool count of both primary and secondary regions. If there is an imbalance, it adds the VM to the region that has the imbalance?
The impact of this is as follows;
We pay for reserved instances, equally spread across both Primary and Secondary regions, so we're paying double. Firstly for the reserved instance, secondly for a new VM in the other region, at the full cost.
Hope that all makes sense. Thanks.
Comments (3 comments)