Unexpected Issue

RDS 2012 R2. We have the following environment:

• 1 - rds gateway (licensing and rd web)
• 2 - connection brokers
• 2 - session hosts (load balancing)

All these are hyper-v 2012 R2 VM's. We also use UPD and we use remoteapp publishing. So, session host 1 has never worked correctly since deployment. The problem with SH 1 is users would always get temp profiles at login. So i deleted the SH 1 VM using the following command:

Get-VM "VM Name" | %{ Stop-VM -VM $_ -Force; Remove-VM -vm $_ -Force ; Remove-Item -Path $_.Path -Recurse -Force}

This removed the vm, hard disk and xml file. Then i reinstalled 2012 R2 and deployed the session host role to the server. Logged in with domain admin and was given a temp profile. SH 2 was working fine so i left this project and moved on to some other things while i thought about what to do next. I picked the project back up a couple of days ago and didn't have anything to do but delete the VM again in the hopes this delete would somehow delete all traces of the VM. So i used the same command from above and then verified the VM, hard drive and xml file were gone, they were. So i created another VM and installed 2012 R2 again. Did windows updates and logged in as domain admin and was given a temp profile!

You can see why i am thinking some configuration trace is left somewhere! The question is where? Is it possible the connection brokers are holding something related to session host 1? When the SH 1 was removed, you would get "server SH 1 is no longer in the pool..." when you tried to manage the deployment in server manager. I found:

https://social.technet.microsoft.com/Forums/windowsserver/en-US/cee15edd-d1f5-4393-aafa-93e8e4f8e519/remove-already-deleted-session-host-server-from-deployment-configuration?forum=winserverTS

in regards to that issue and wonder if the connection broker is holding some info that keeps giving the temp profile issue. Anyway, i am at a loss on how 2 installs can give this same issue every time. Any ideas?