The Case of the Disappearing Philips Dictation Microphone in a XenApp-Delivered Application Running In A XenDesktop Virtual Desktop And How To Make It Work

A long title but with amazing results. What do you do when you are asked to fulfill these requirements?

• User will be using Epic Hyperspace delivered via XenApp 7.x
• User will be launching Hyperspace from their XenDesktop-delivered Windows 7 desktop
• User will be launching their VDI desktop from a Dell Wyse Xenith endpoint with a Phillips SpeechMike III attached
• User must be able to use their SpeechMike in Hyperspace for partial dictation with all buttons functioning

1. First, make sure you are running a recent Dell/Wyse firmware (I tested on ThinOS Lite 2.4_112 and ThinOS 8.4_112).
2. Next you will need to set a custom hardware string either in the .ini or your Dell Management Suite group policy->Advamced: Device=vusb ForceRedirect=0x0911,0x0c1c,0x01,0x01,0x00 Type=HDX
3. After logging in to your published desktop, you will see the Speechmike in audio devices but notice that the default audio device is Citrix HDX Audio for both speakers and microphone. In order for the SpeechMike to work in a XenApp session, the Speechmike must be default.
1. To accomplish this automatically, you can use Powershell and a 3rd-party utility, nircmdc, to make the change during login.
2. For the script to work, you must change permissions on
   HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\MMDevices by first taking ownership of the key and then giving write permissions for Authenticated Users and propagating them down.
3. Nircmdc uses the Windows generic name (eg. Microphone) and not the device name (Philips Speechmike III). In many cases there will be multiple devices with the same name (eg. Microphone or Speakers) so we need to run a Powershell script to find the Philips microphone and rename it's generic name to DocMic so we can use Nircmdc to set it as default.

   $regcheck = "HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\MMDevices\Audio\Capture"
   $soundcheck = Get-ChildItem -path Registry::$regcheck
   
   $enabledDevices = foreach($_ in $soundcheck)
   { 
       $path = $_.'name'
       Get-ItemProperty -path Registry::$path | ?{$_.DeviceState -eq "1"} | Select-Object pspath | Get-ChildItem | Get-ItemProperty | ?{$_ -match "SpeechMike III"}
   }
   if($enabledDevices)
   {
       $path =  $enabledDevices.pspath
       $key = ($enabledDevices.psobject.properties | ?{$_.Value -match "Microphone"}).name
       set-ItemProperty -path $path -name $key -value DocMic -type string -force
   }

4. You can modify the Powershell script to also kick off Nircmdc but I've left it separate:

   Nircmdc.exe -setdefaultsounddevice "DocMic"

4. Now just have these run at user logon. You may need to call a sleep function at the beginning of the script to allow time for Windows to complete the new hardware detection.
5. If you want to also use the speaker on the dictation mic, you will also need to run a slightly modified version of the above PS script:
   

   $regcheck = "HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\MMDevices\Audio\Render"
   $soundcheck = Get-ChildItem -path Registry::$regcheck
   

   
   

   $enabledDevices = foreach($_ in $soundcheck)

   {  

       $path = $_.'name'

       Get-ItemProperty -path Registry::$path | ?{$_.DeviceState -eq "1"} | Select-Object pspath | Get-ChildItem | Get-ItemProperty | ?{$_ -match "SpeechMike III"}

   }

   if($enabledDevices)

   {

       $path =  $enabledDevices.pspath

       $key = ($enabledDevices.psobject.properties | ?{$_.Value -Match "Speakers"}).name

       set-ItemProperty -path $path -name $key -value DocSpeaker -type string -force

   }

   And then modify the nircmdc command to reference "DocSpeaker".
   You can obtain the nircmdc utility here.

Powershell: Ensuring Values Are Returned As Arrays

It's Christmas Eve and I'm in the office catching up on some work.  This year has flown by pretty quickly and I have realized I have not added any content this year despite being very busy with VDI deployments, scripting, a zero-downtime SAN migration (which really deserved/deserves an article), and other miscellaneous tasks.
When scripting today, I came across an challenge that I remember solving earlier in the year. It took a bit of searching through previous scripts for the answer and I finally found it. I decided to turn it into a post for easier recollection later. I also hope it helps someone that runs across the same issue. Others will read this and probably just shake their head as common knowledge.

The Problem:
I have a function that returns multiple values into an array that I can then run a ForEach loop on.  However, it is possible that the function may only return a single value.  In that case, the variable is no longer an array, but a string or integer (depending on what is returned).
Example:

   $files = Get-ChildItem -Path E:\music –File | Select-Object Name
The example would return all the file names in E:\music as Strings and store it in the variable $files. If there were multiple values returned, the $files variable would be an array but let's say there is only one file in E:\music named mysong.mp3. The output of $files[0] would be:
   m
Not the entire file name. It makes sense when you think about it.  When you only assign one string value to a variable, the variable is going to be a string. Weird, eh?

The Solution:
When you are declaring a variable as an array, you can do something like: 
   $myArray = @( )
So when you are assigning the value of a function to a variable and want to ensure it is stored as an array, you would use:
    $files = @(Get-ChildItem -Path E:\music –File | Select-Object Name)
The result of running $files[0] using the same scenario as above would be:
   mysong.mp3
     

VDI: Xenith Pro Endpoints and Imprivata SSO High Availability

This is the first in, I imagine, a series of articles regarding VDI processes, lessons learned, tech issues overcome, etc. as it relates to my experience planning, configuring, deploying, and managing a Virtual Desktop Infrastructure roll out.
One issue that had cropped up after the pilot deployment was how fail over of single sign-on (SSO) authentication would work on the 0-client endpoints that use the ProveID Web API calls to the SSO infrastructure (2 Imprivata appliances, single site).  On the Windows side, HA was built in to the agent, whereas it downloads the SSO topology from an appliance after agent install.  Appliance failovers are seamless and Offline mode is configured to prevent authentication failure even if both appliances go offline.  However, all documentation on the ProveID side was sparse from Imprivata and further research did not reveal any answers to these questions:

Q. If the appliance configured in 'OneSignServer=https://<servername or IP> SignOn=yes' is down, will the endpoint go into an offline mode? Will it fail over to another appliance?
A. The answer is no to both. You will not be able to authenticate after the endpoint boots.  A visual cue is that the Imprivata logo will be missing from the logon box.

Q. Does the entry for OneSignServer accept multiple servers then?  If so, what is the syntax?
A. Yes, you can specify multiple servers.

OneSignServer=https://server1.domain.com,server2.domain.com SignOn=yes

Note: You can use a comma or semi-colon for separation.  Do not add the https:// to additional servers.