Virtualization

Hi,

This is a pretty common question to be honest - you're right that it's not really properly explained in the published documentation, but it is covered in the official (and our tailored) training (apologies for the not so subtle sales pitch!).

Anyway, in simple terms, there are a couple of things to know about the Q: drive:

• It doesn't have to be Q. It can be any available letter, but note that you will get much better results if the same letter is used across all of your client machines. It should not be a mapped or SUBST'd drive
  
• It is a 'real' drive (or partition) on your sequencer, but it is a 'virtual' drive on your client machine. It will be created by the app-v client installer and should NOT pre-exist on that machine.
  
• It does not exist on your server
  
• For best practice, you should install your application into the Q: drive, but not the root of that drive. It should be installed into something called the PKGROOT which is a directory which is unique to that application and should ideally conform to an 8.3 compliant naming convention. Note that your naming convention is one of the most important pieces to get right. In very simple terms, the way I tend to describe this it is that if your app's default install path is C:\Program Files\AppVendor\AppName\AppVersion\.... then you should change this while sequencing to Q:\UniquePKGROOT\AppVendor\AppName\AppVersion\..
  

Once you've absorbed that lot, have another crack at sequencing and perhaps review the best practices guides which will give you some direction with regards apps which must be installed to C:\ etc.

Hope that helps,
Danny

Thanks for your time Danny. Very very well explained. I thought it was a virtual Q drive on the client but thanks for clearing it up. It's a little bit like the M: drive on a Citrix session really.

So just to cement my knowledge a little more... when sequencing do I mount to Q and install to Q?

I'm not sure I entirely understand what you mean by 'mount to Q', but hope the following helps:

• The sequencing machine must have a Q (or equivalent) drive available before you start monitoring.
  
• You should ideally create your PKGROOT directory (the 8.3 compliant named, unique directory) on the root of the Q: before you start. For purposes of an example if you were sequencing Foxit PDF Reader, you could create Q:\Foxread.001
  
• When you start monitoring you will be asked to specify the 'installation directory' - I can't remember the exact dialogue but it's not technically correct since you MIGHT not install there, but nevertheless you should select the PKGROOT (Q:\Foxread.001) folder you created earlier
  
• When you install the application, if possible you should change the root of the installation path of the app to be the PKGROOT folder (referring again to my example earlier, instead of installing to C:\Program Files\Foxit Software\Foxit Reader for example, you might choose something like Q:\Foxread.001\Foxit Software\Foxit Reader)
  
• As far as your application is concerned, it is installed in a real directory of that name, and when executed on the client will run from that directory, blissfully unaware that the Q:\Foxread.001 directory is essentially a mount point in the virtual environment and not a 'real' directory structure on the local disk.
  

Regards,
Danny

You've answered it perfectly thank you.

What i meant was that
"When you start monitoring you will be asked to specify the 'installation directory' "

and

"if possible you should change the root of the installation path of the app to be the PKGROOT folder "

were they the same folder and you've answered that they are. Confusion over thanks very much [:D]

Cool, glad to have helped..

Good answers DannyC, thank you for sharing your knowledge.

On a side note I'd love to strangle the induviduals that developed this sort of confusing and non-intuitive system. Did they set out to create a system that is inherently confusing or what?

In terms of best practice I've yet to understand why I even need a Q: drive on my Sequencer machine. [Note that I am not experienced with APP-V, I've only done a few hours of packaging after a course]. Some applications cannot be installed to Q: and has to use the default C: path, and if this works just fine with APP-V why not just make the best practice of using the default install dir?!?!?! (possibly add a few more ?!'s). In my limited experience I've not been able to understand why that is a problem.

Older versions of App-V (aka the SoftGrid days) installing to C:\ created significant overhead to redirect the file system calls back to Q:\<8.3 folder>\VFS where the virtual file system actually lives. I talked casually to one MS consultant and he said that with more recent versions of App-V have about .004 of a second overhead per I/O operation but don't quote that as fact.

The reality is that performance isn't as much of a concern anymore but the best practice is to install do your virtual drive when possible because the virtualization supposedly works better in some instances (sorry I don't have a good technical explanation). I'm a bit of a cowboy and say installing to C:\ is just fine but in some cases where your application does not work Q:\ might be an option to try for the fun of it to see if your sequencing issue goes away. The reason I say C:\ isn't so bad is that I've packaged and sequenced enough applications to know that some installers just aren't designed to install an application successfully to a drive letter other than C:\.