Prob & Solution: Unable to make RDP connection using newly added Microsoft Account to win10 tech preview hosted on AzureVM

Question

Problem description:
After VM deployment is finished, remote desktop into the Windows 10 client OS with the created admin credentials
Open PC Settings -> Users and Accounts -> manage other users (this is the new UX not the old classic add user dialog)
Add a Microsoft account to the desktop, leave it as a standard user
Log Out
Make RDP connection using login of DNSName\[email protected]
Note: DNSName = connection string you are using in RDP
[email protected] is whatever liveID you are using.
Problem: If it was working you will get a message that this user does not have access rights to remote desktop to the computer
Error: Unknown user or bad password is what I would get
Solution:
Log back on as admin user with RDP
Using the new UX from " PC Settings -> Users and Accounts -> "  delete the recently created microsoft account
Create a local (not microsoft account) account where with the name that would be before the @ (i.e. user77)
Go to the Old user management interface (either the computer management MMC or the "System Properties->Remote" tab and give the created local user remote desktop access rights.
Now log off as admin
Remote desktop and use username of DNSName\User77  (note no @outlook.com)
Let the login process finish
Make sure any feedback questions are answered
Now navigate to the new UX using " PC Settings -> Users and Accounts -> " and link the local account with your Microsoft account.
You will have to authorize the microsoft account, typically with a text message + verification code
Wait 2 minutes (using a A1 size VM there seemed to be some background stuff for account linking)
Log out
Now log in using DNSName\[email protected]
Success!
I hope this helps anyone else who is using Azure VM's to test out Windows 10 client technical preview.
-N
neilgo

Answer 1

Hi
We use a custom imagine and experience this quite alot.  Usually host name is missing and we have to redeploy.
Seems to have gotten worse as we add more vms.
Should we not sysprep when capturing a vm?
Orginal vm came from the gallary and we've simply resized os to 30gb.
Main role for our vms is job processing so stop deallocate when finish to keep costs down buts prooving to be more trooblesome that expected due vms dieing.
We have a number which we don't shut down which are problem free.  so I suspect the stop deallocate and self heal are still early days.
My monitor has reported issues with 40+ vms this evening.  Are there issues or updates going on in western europe?