Any suggestions/limitations to hosting FMS 15 in a virtualized environment?  Currently, we host a 30-40 user, 10 gig FM db on a physical server, however, p to v is being suggested by others in our organization (They're also inclined to config pooled versus static resource settings which would seem to affect the manner FMS functions) 

Elaborating on specific VM configurations in a Windows environment would be greatly appreciated!

There is no difference in the resource specs for a VM vs. a physical server.  With one caveat: you have to make sure that the host machine is not constrained for resources by the other virtual instances that it runs.

Go by your current FMS stats.log numbers (you do have the 'usage statistics' log toggled in the FMS admin console hopefully, on the 'logging' tab).

What's is the 'pooled' vs. 'static' about?  How do they see this happening?  What would not be configured statically?

"Pooled" meaning that the FM server would use resources "as needed" (depending on server workload) versus a static configuration which would "lock-in" resources regardless of server workload. 

Is there a minimum assignment to the machine?  I don't think I would be happy with a pooled setup for a database server (of any kind).  

Processor cores are usually assigned on-demand, but RAM should definitely be statically (not dynamically) assigned.  Dyanmically assigned RAM is a bad idea no matter what you're running on a VM server.

- John

Coming from experience migrating to MS Hyper-V, I agree with John in stating that having the RAM statically assigned is a very wise idea.  Especially in the way FMServer uses RAM cache.  Having personal remote desktops be dynamically managed is OK, but not a VM running FMServer.  

For what it's worth, I've found that the typical bottleneck in a VM environment is the disk(s) read/write speeds.  Make sure to not only set aside a good portion of budget for fast drives/interface, but that whomever is setting this environment up knows what they are doing.  If you don't have experience, then paying for a consultant might make the difference in a system that migrates smoothly and a failed project.