Hi popvm,
Random-ish logon delays in non-persistent scenarios are typically caused by SMB clients not having had the time to correctly "sign out" from the SMB server before the image was refreshed, causing oplock-related delays when the SMB client on another VM subsequently tries to access configuration files. Please see Re: UEM hanging/delaying randomly. for a workaround.
Having said that, something else might be at play here (as well), as that would not really explain why your 10 async drive mappings take 12 seconds to be processed. That's just a matter of reading 10 tiny files from the config share... Does that takes as long if you do not refresh at logoff? Anyway, let's first try shutting down the workstation service, to see the impact of that.
