Recently I was pinged by mail buy a guy who asked me if it is possible to create the “Publishing MSI” for a bunch of App-V Packages. In the past, that organization only used the App-V Management Server, so their Packagaing Guideline did not include the creation of that MSI. About at the same time I realized that I have “a million” of old SoftGrid and App-V Packages that still would do their job, but their inner structure is quite old so the App-V Client would apply some “conversion magic” whenever certain functions of those packages would be executed. Some packages had that “old” Publishing MSI that was not compatible with newer Client versions. And of course non of my packages was compressed before.

So, the guy asked me: “Yeah, you told me that it can be done in Powershell using the Command Line Sequencer… but how?”

Hm, I am not really what you would call a Powershell Expert. This is why I asked my colleague Matthias to “help” me with the App-V Analyzer tool before. But with that sample and some internet search I actually created my first “real” PoSH scripty and here we go:

When you upgrade your App-V Client to a somewhat newer version (like from 4.2 to 4.5 / 4.6 / 4.6 SP1), Microsoft recommends to open and save all packages with the new Sequencer. MS says, that by resaving the packages their “inner structure” gets converted to the new version, resulting in an increased stability and performance. When “legacy” packages are executed on newer clients, some “tweaks” might be applied to let the old packages behave like new ones.

Feature

• Convert all packages under a folder (subfolders are parsed recursively) to the actual Seqeuncer’s version. Network Shares work as well
• Create an Publishing MSI for every package
• Compress every package
• Keep FB1/FB2 separation

Script

##########################################################
##
##    Resave App-V Packages
##
##    2012-26-01 Falko Gräfe
##

##########################################################
## Purpose: Open all App-V Packages of a certain folder recursively and Save them right away
## Create MSI and Compress automatically
## Not implemented: Error Handling and Reporting
## Adjust $StartFolder !!
cls
$StartFolder = 'C:\Data Folders\Packages'
$sprjFileName = ''
# Get all folder names (except for "... ICONS" folders) into one list
# Because we want to process the .sprj AND want to "clean" the old sft(s), just getting the .SPRJs is not enough
$AppVPackagesFolders1 =  Get-Childitem $StartFolder -Recurse | Where {($_.psIsContainer -AND ! ($_.name -match "Icons") )}
#Process each folder
FOREACH ($AppVPackagesFolder in $AppVPackagesFolders1){   # parse each list item
    # Now get the "full name" of the current folder. If the folder doesn't contain an sprj, there is no App-V packagee in there so wie skip it
    $CurrentFolderContent = Get-Childitem $AppVPackagesFolder.FullName | Where {($_.name -match ".sprj")}    
    # Console output
    "Processing "+$AppVPackagesFolder.FullName
    If ($CurrentFolderContent -ne $null){
        # get the SPRJ File Name
        $sprjFileName = Get-Childitem $CurrentFolderContent.FullName | where {$_.name -match ".sprj"}
        # get the "old" SFT File Names (for later deletion)
        $sftFiles = Get-Childitem $AppVPackagesFolder.FullName | where {$_.name -match ".sft"}             
        # specify the Sequencer location
        [string]$ExeFile = '"'+$env:ProgramFiles+"\Microsoft Application Virtualization Sequencer\sftsequencer.exe"+'"'  
        # Compose the command line
        [string]$CmdlineArgument = ' /open:"'+$sprjFileName.FullName+'" /MSI /COMPRESS'
        # Launch the Sequencer, pass the .sprj file name
        Start-Process $ExeFile $CmdlineArgument -wait 
        # Delete every old SFT file in the current folder
        FOREACH ($sftFile in $sftFiles){                       
            remove-item $sftFile.FullName  
        }  #FOREACH 
    }  #IF 
    ELSE{
       # Console output
        "          ... previous folder did not contain any Package files (but may contain folders)"
    }
}   
"Finished"

Preparation and Execution

Before using it, check and adjust the PoSH script. You may save it as “Resave_App-V_Packages.ps1″ if you want to use it regulary.

Required adjustments

The value fo $StartFolder. This is the main folder were all your packages are located.

The value for $ExeFile: This is the location of SFTSEQUENCER.EXE. In my example, it’s the default location on a 32bit machine.

The value for $CmdlineArgument: currently includes MSI creation and compression. If you don’t want to have that, delete /MSI and/or /COMPRESS

Execution

Impotant: Do not run against your “original” App-V Content folder. The Sequencer will override each input package!

Make a Backup Copy before!

The Script needs elevated admin permissions to execute (because the Sequencer requires that) and the user needs write access to the Content-Copy folder.

I’m using the Powershell ISE (launched as Administrator) on the Sequencer machin, but you also could use a little CMD to launch it:

set Ps1Path= C:\Data Folders
powershell set-executionpolicy unrestricted
powershell "& '%PS1Path%\Resave_App-V_Packages.ps1'"

Here, you’d have to adjust the path to the Powershell script that you might have saved.

Remember to run the CMD with “run as adminisstrator”.

Limitations

This release has not been tested thoroughly. You use it at your own risk!

There is absolutely no error handling in it. If anything goes wrong, it may delete all “old” SFT files without creating new ones. The Sequencer may crash.

To check for issues, use the Sequencer’s error log (sft-seq-log.txt in the Sequencer’s Program / Logs folder).

It’s my first PoSH script. Remember that. I know it’s not that smart at all (no GUI, no Command Line parameters…, no PoSH specific extra tweaks) Just Quick and Dirty.

Apparently, I am not that Server Hardware, Network Equipment nor Storage guy, so I was quite sceptic about the value HP Discover could give to me.

So I was seeking for at least ‘sort of’ interesting stuff amongst the 650 breakout sessions and technical demos. Of course there were tons of Storage and Server related sessions, but also a lot of ‘cloud’ stuff I actually started with.

To start with the surrounding stuff: I was there together with some other App-V MVPs: Ruben (@rspruijt) , Ment (@mentvanderplas), Nicke (@znackattack) and Kalle (@gridmetric). Our stay was organized by ivy worldwide.

The event in short: There were about 7000 attendees and about 650 sessions and technical presentations. Mostly there were 40 sessions in parallel (making a choice quite difficult).

The over all organization during the event – and Ivy’s dedicated services – were quite perfect.

So, let’s get to the content.

HP’s vision – while focusing on the layers ‘up to the hypervisor’  – is quite broad. Yes, they have their own equipment for public, private and hybrid clouds, but with solutions like HP CloudSystem, different approaches, technologies and vendors can be addressed.

They seriously try to integrated the variety of best-of-breed solutions (like the most relevant commercial and Open Source hypervisors) into unified Management Consoles. So perhaps the Operations Team just doesn’t notice that you migrated to another vendor.
Especially on the Network layer, HP already has their Network Management offering that manages literally ‘all’ possible network devices you can think of. More than 5000 device classes from 140 vendors. In one single console.That best-of-breed approach obviously spans most of their software and services offering, that leads me to the first conclusion.

HP really is about to integrate solutions, across various layers of the technology stack as well as across different solutions and approaches within a certain layer.

In alignment with that aim, HP’s Application Lifecycle Management and Application Transformation Solutions offering should definitively be mentioned. Again, the area of interest is quite broad. Talking about Mainframe applications transforming to Windows apps or enabling transitions into (public or private) clouds: HP is offering pieces to this puzzle as well. Within their IT Performance Suite, there are products that analyze applications regarding their ability to be transformed to another architecture. HP even tries to identify the purpose of code (like User Interface, Backend Communication or Core Processing) in order to define higher and lower re-development priorities or recommend something like ‘don’t recode, but run it on VDI’ – based on both, technical and business driven parameters. HP has an impressive portfolio to address the transitions processes we all will face in the future. So, here is the second one:

HP is perhaps the one company with the most comprehensive approach to address Data and Application readiness for various paradigms, including own and 3rd party solutions.

While investigating about that, we heard about Application Transformation Experience workshop – a service offered by HP. In fact I knew that HP is offering various services like Support or DataCenter outsourcing. But that Transformation Experience Workshop (PDF) is really far beyond those task oriented activities.
HP High Level Consultants are moderating these workshops and guide technical executives through a vast variety of aspects they have to consider to make their IT organization agile and cost effective throughout the next years. Strongly highlighting non-technical aspects like Process Implications, Business Impact or Human aspects, this Workshop aim to enable customer’s decision makers to understand the complexity of Application Transformation- and agree on the high level targets right away. Of course such a one-day workshop can’t dig into technical details. Instead, general directions for future projects are defined here. And again – HP doesn’t really care if the customer wants to realize those projects alone, with HP or even with other System Integrators. And the news for me?

HP is an IT Professional Services provider with a holistic view on technical and non-technical dimensions of IT, including very professional, high level Consulting offerings.

There are a bunch of other things to tell, but I’d like to overall conclude my take-away:

HP is way more than just a hardware vendor. Their Services and Solutions are focusing on integration and flexibility. But it is exactly that flexibility that makes it difficult to pin HP somewhere. They do very good things, but because they do such a lot they struggle to claim there space outside the hardware market. They do interesting things beyond that but its really hard to learn about that.

Would I attend a next HP Discover? Probably yes!

There will be about half of all world wide App-V MVPs on site (most of them even presenting), so this is probably a unique oppurtunity to meet with them and other community leader.

While the day is fully packed with awesome presentations, time is planned for questions and answers. And as Community events should be: it’s free!

You can register at www.appvug.com.

If you’d like to tweet about it: #AppVUG is the tag of choice.

Hope to see you there!

Falko

Forced to by some projects I tried to collect all the Citrix Hotfixes that
are:

• Designed for Citrix Provisioning Services 5.6 SP1
• Are public available
• Have been released until Sep. 6^th, 2011

Please refer to the PDF for more details on the hotfixes: Citrix-PVS-Hotfixes_2011-09-06

[Update 2011-09-12] Hotfix clarification by Citrix Support about E028 and E006 applied: E006 replaces E028. The updated PDF is Citrix-PVS-Hotfixes_2011-09-12.

Recommended Microsoft Hotfixes for Provisioned Systems

Citrix maintains a list of recommended MS Updates that should be applied to
provisioned systems: http://support.citrix.com/article/CTX128058.

Also, there is a list of recommended Hotfixes for XenApp 6, regardless of
their deployment method: http://support.citrix.com/article/CTX129229.

So far, I didn’t come across a similar list for XenDesktop.

Installation Method

Did you ever “install” PVS hotfixes? Then you perfectly know how file copy
operations are used within your Software Deployment Solution or Scripting
Language: PVS Hotfix “installations” mostly are readme files that explain where
the few files should be copied to. No MSI, MSU, MSP, EXE or other nice method.

From time to time it even happens that two “current” Hotfixes ask to
replace the same file. Can’t believe it? Check the readme for “E028” and
“E006”!

Due to the short life time of some of the hotfixes it’s almost useless to
publish installation scripts – they would expire to quick.

Citrix Provisioning Services 5.6 SP1 Hotfix List

Note: the Citrix-PVS-Hotfixes_2011-09-06 PDF contains more Details, like fixed issues or the bug IDs

Hotfix Name	PVS Server	PVS Target	CTX Article
			CTX125361
CPVS56SP1E014			CTX129104 superseded by E022
CPV56SP1E003_CPVS56E019			CTX127528 included in E036
CPVS56SP1E001			CTX127189 superseded by E002
CPVS56SP1E002			CTX127502 superseded by E008
CPVS56SP1E004			CTX127565
CPVS56SP1E005			CTX127558
CPVS56SP1E007			CTX127813 superseded by E026
CPVS56SP1E008			CTX127937 included in E040
CPVS56SP1E009			CTX128602 superseded by E020
CPVS56SP1E011			CTX128160
CPVS56SP1E013			CTX128245
CPVS56SP1E014			CTX129104
CPVS56SP1E015			CTX128591
CPVS56SP1E016			CTX128726 superseded by E029
CPVS56SP1E017Filename E036			CTX128770 included in E036
CPVS56SP1E018			CTX128603
CPVS56SP1E019 Filename E036			CTX127528 included in E036
CPVS56SP1E020			CTX129474
CPVS56SP1E022			CTX129123
CPVS56SP1E023			CTX128727 superseded by E028
CPVS56SP1E024			CTX129471
CPVS56SP1E026			CTX128987
CPVS56SP1E027			CTX128989 superseded by E041
CPVS56SP1E028			CTX129021 superseded by E006
CPVS56SP1E029			CTX129381
CPVS56SP1E036			CTX130273
CPVS56SP1E040			CTX130275
CPVS56SP1J003			CTX127375 Japanese
CPVS56SP1E038			CTX130117
CPVS56SP1E006			CTX130664
CPVS56SP1E033			CTX130661
CPVS56SP1E041			CTX130662

Summary and Recommendation

Well, I won’t summarize my thoughts about Citrix’ Patch “Management” for
PVS that I had several times during the last few weeks. If I would, this
article would be marked as “verbal abuse”.

However, there are some recommendations:

• Do check for PVS Updates on a regular basis. Usually, new patches
  fix really important issues. I tend to check once per month or at least once
  every three months.
• Do automate the PVS Hotfix installation – at least by using CMD
  scripts. The effort of the initial creation pays off with the second server/device
  that you install.
• Read the Readme: Check for requirements, HF installation order and
  replaced files. Some HFs require other HFs, some replace the same files like
  others and some should be installed even before the actual PVS Software.
• Reboot the Server / (Master)Target after you applied Hotfixes. Some
  replace drivers and I doubt they are replaced on the fly.
• Verify file replacements. If have seen scripts that replaced/added
  files only after the second time they run. Do not rely on the “.md5” checksum
  files. They are not usable as an indicator whether a HF has been installed or
  not. Compare file dates!

Actually, the recommendation is: Please, Citrix, redesign the Hotfix Management and Release process for Provisioning Services. Build out-of-the-box installers. Urgently!

Do you want to verify quickly, if Naming Conflicts between virtual packages, 8+3 folders and other items occur?

Then, get Login Consultant’s “App-V Analyzer” to get all the infos usefully presented!

Today, Login Consultants released the first, free version of their “Application Virtualization Analyzer” tool.

It has been developed by Matthias Wessner and me, because we thought that no commercial nor free product actually compiles some critical information about App-V package management information.

The original aim was to develop a “conflict checker”, but it turned out that – as a side effect – one also would get a comprehensive overview of App-V packages.

Download

You can download AVA from the Login Consultants Homepage or directly browse to the Tools Section. A free Registration is required.

Credits

Credits go out to Community Contributors Jurjen van Leeuwen and Thomas Øyen  as well as to our Login Colleauges Volker Kleiner and Holger Brink for testing some earlier versions of App-V Analyzer and providing very valuable feedback.

Feature

Virtual Application Inventory

AVA generates an Excel Report about critical properties of virtual applications and packages. At a glance you can inspect all the package names, primary directroy names, application names and others, with direct links to the corresponding OSD files.

App-V Conflict Check

AVA parses all virtual applications to identify potential conflicts on App-V specific properties, namely the Primary Directory (aka 8+3 folder), Package Name, or Application Name+Version. Those conflicts are highlighted.

Szenario Agnostic

AVA works with any App-V deployment scenario, like SCCM integrated, App-V Native, 3rd-Party (Altiris, Frontrange, Matrix42..) deployment solutions and even in stand-alone mode.

Lightweight

AVA does not require any database nor backend service to operate. It is developed in Powershell 2 and requires MS Excel. It simply parses the local or remote file share, where the App-V Packages are stored.

Requirements

• Powershell 2
• MS Excel (installed)
• Access to a lokal folder or remote share that contains the App-V Packages
• A given subfolder should only contain package files that belong to the same logical package. If you mix files from different packages into one subfolder, AVA may respond with inaccurate results. A recursive folder structure is supported.
• Patience

How to use it?

Well, that’s quite easy: Install it on a machine that has Powershell 2 and MS Excel installed, then open it from the Start Menu.

Browse (or enter) a location that contains all the OSD files an hit “Analyze”. Depending on the amount of App-V Applications (and the network speed), AVA may require several minutes to complete. During this period, the main window appears to be not responsive or “stuck”. Just give AVA its time. In some tests, AVA took as long as 30-45 minutes to complete.

Feedback

To provide feedback, you may use Login Consultant’s Online Forum (SoftGrid Tooling section) - or you can contact me directly or via this Blog’s “Comment” feature.

According to Sebastian, this patch addresses the following issues:

• Applications can’t be launched if the SFT file is not located at the right location on an HTTP server – although RequireAuthorizationIfCached is set to 0 and the package is fully cached
• The Sequencer’s “Save as” feature is avaiable again
• Applications failed to launch from a Network Share on WinXP and W2003 compute
• Applications, that call the LoadLibrary Windows API during sequencing may fail
• Running applications from a network share that call DLLs may cause error 04-00000020
• The App-V Client Service takes a long time to launch and decreases perfomance (also boot performance)

Treat the topics with caution: they are a “free translation” from German!

Some of the topics have been addressed with HF1 and HF2 already.

If you encounter any of the issues described above, you may contact Microsoft Support to request the Hotfix.

The KB article ID will be 2571168.

What is the perfomance impact of Application Virtualization?

Now, Project Virtual Reality Check (http://www.projectvrc.com) has released a Whitepaper that provides an idea of that. ProjectVRC compared Microsoft App-V, VMware ThinApp and Citrix XenApp Streaming running in Virtual Windows 7 machines.

While the impact is significant, there are some things to note…

Announcement

Read the official announcement on BrianMadden.com:

http://www.brianmadden.com/blogs/rubenspruijt/archive/2011/06/15/the-impact-of-application-virtualization-solutions-in-vdi-the-reality-check.aspx (short: http://bit.ly/ii0pCY)

Measurement Methodology

The way ProjectVRC measures is somewhat different from most classical “Performance Test” tools. Most of these tools process loops of well-defined tasks. The results then directly reflect the execution speed (the faster the tasks are processed, the better the result).

ProjectVRC (actually a joint venture of PQR and “my” Login Consultants) uses an index that reflects the maximum amount of simultaneous user sessions on a given piece of hardware.

Essentally, more and more user sessions are created on that hardware, and when the user sessions start to get slow or inresponsive, the number of users is counted.

For the Application Virtualization test the decision was made to measure against a VDI environment. (how many users, each having an own VM, can work with an acceptable response time on the hardware). Another approach could have been: “How many Remote Desktop Services / Terminal Server User Sessions can be hosted”, but this was not tested.

For the AppVirt impact tests, ProjectVRC took “localy/classicaly” installed applications as a baseline (or 100%). Values below 100 mean that – with a given Application Virtualization solution – less users can work on the same “iron” with an acceptable performance (measured by the application’s response times).

Overall Results

Well.. no. I won’t disclose any details here. Check the Whitepaper, it’s free. Anyway there are some things that I want to share with you:

Max User Sessions Impact

As you can see, each Application Virtualization solutions only allows a reduced number of users compared to VMs with locally installed applications. The result shows that – using a given deployment scenario - there are at least 20% less sessions possible by having the core business applications virtualized. Microsoft Application Virtualization, Citrix Streaming and VMware Thinapp have different impacts, while two of them are quite close together and the third one seems to have some difficulties. The grey bar at the top marks 100% for locally installed applications.

This is an example for a “streaming” scenario with a fully virtualized Office suite. If the Virtual Applications are pre-cached on the Clients or some of the Office applications are installed locally, the results are different.

The ProjectVRC Whitepaper includes quite detailed results from all AppVirt solutions, like the difference between pre-cached and streamed delivery, the usage of App-V’s Share Cache or differences in the transfer protocol. It includes various charts like the one above.

Results for Selected Resources

Beyond that overall results, ProjectVRC also investigated the impact on Memory or the application launch time.

It has put an focus onto Disk I/O, because esp. in VDI environments Disk I/O (that is Storage I/O) is a sginifcant cist drive and perfromance bottleneck.

It shows that Application Virtualization can decrease the Read I/Os, but probably will increase Write I/Os from the VM’s (or VM Host’s) perspective.

Conclusion

I think the answer is clear: Application Virtualization solutions, namely App-V, XenApp Streaming and ThinApp do have a significant impact on the amount of users that can work “acceptably” fast on a given hardware. While introducing AppVirt to VDI environments, Storage I/O requires additional attention.

Considerations

While the key message is clear, keep in mind:

• The results are not 1:1 applicable for your environment. Project VRC:
• Run all core applications, including MS Office, in a virtualized manner. This is not a Best Practice
• Used a very well-defined workload/taskset that may/may not reflect the majority of your users
• Evaluated VDI VMs, that are different from an RDS/TS/SBC environment. On Terminal Serves, you may seem significant different result.
• The applications were not specially “tuned” for Application Virtualization. It is sometimes recommended to avoid installation of certain features
• Your environment can be faster or slower with different resources that impacts the performance.

Unluckwise, ProjectVRC did not (yet?)

• Compare App-V applications that were installed to C: instead of Q: (VFS vs. MNT install) (Applications have been sequenced to Q: what is considered to be faster)
• Compare the impact on SBC hosts
• Check the impact on the “network”, namely the required bandwidth and time to download the packages, the impact on CPU and RAM of the streaming server…

But hey, there is always room for improvements, right?

Download

You can download the Whitepaper for free directly from the Project VRC WebSite. Go to http://www.projectvrc.com and look for the “Phase IV” whitepaper.

Or check the announcement at http://bit.ly/ii0pCY

After working with the new Sequencer for a while now I tried to figure out which wizard allows (or prevents) certain activities.

The following table compares the wizards based on the main Sequencing tasks:

Task description

Install / Monitoring

This wizard allows to set the Sequencing machine into „monitoring“  mode that allows you to install applications, apply updates or perform other system changes that should become part of the App-V package.

Configure Applications

This wizard allows to launch the application after the actual Monitoring to pre-configure the application for „most users“. Typical tasks include confirming License Agreements, disabling Splash Screens, disabling Updates and pre-configure the most common application settings.

Edit Shortcuts

This wizard allows to adjust the name of the application shortcut as it appears in the user session, modify the shortcut locations and configure File Type Associations (FTAs). Adding or removing „shortcuts“ causes OSD files to be created or removed.

Optimize for Streaming  (FB1)

This wizard allows to launch the application(s) in order to determine the blocks that are important for
launching the application(s). Blocks that were read during this wizard are marked as „Feature Block 1 (FB1)“. Feature Block 1 components are the first ones that are streamed down to the client.

Specify Operating System

This wizard allows to select a list of Operating Systems that should be compatible with the application. Also, it allows to select the application to be offered on „any“ Operating Systems. This task also can be performed using the Sequencer Editor.

Sequencer Editor (also called „Tabbed Editor“ or „Tab View“)

This wizard allows to modify items like Package Name, Virtual Registry, Files within the package (including adding new files to the VFS section), modifying Virtual Services‘ settings as well as modifying the OSD file(s).

Recommendation

If you want to create a very new package, you should start with the „Create Package / Standard Application“ wizard. It includes the tasks that are required to create new Standard Applications, Plug-Ins and Middleware App-V Packages.

If you want to modify an existing package you should determine if you just need to make some smaller modifications or if you need to launch the Monitoring again. For smaller modifications, „Edit Package“ is sufficient. For extended modifications, „Add new Applications“ would be the choice with a higher flexibility (that also allows to „update“ applications).

When using Package Accelerators, basically there is no choice because that method uses a very specific wizard. For certain tasks (like Shortcut modification, Registry manipulation or others) you may need to re-open that „accelerated“ package again.

Also, if you are somewhat experienced with older Sequencers, you may find the following choices useful:

In the „Customize“ window, select „Customize“. This allows you to access the Edit Shortcuts, Optimize for Streaming and the Select Operating System tasks.

In the „Document the package…“ step, select to „Continue to modify package“. This allows using the Sequencer Editor:

Additional information

For a detailed flow chart of the new App-V 4.6 SP1 Sequencer, check out Tim Mangan’s eBook: http://www.tmurgent.com/Books/Default.aspx

While monitoring an application setup with the App-V Sequencer, you will have to decide if you install the application to the C: drive (the installer’s default location) or to the Q: drive. Microsoft strongly recommends to always install to the Q: drive (an “MNT install”). However, many
applications proved to work fine after they were installed to C: (a “VFS install”).
This article discusses some of the myths and facts around that “C: vs. Q:” or “VFS vs. MNT” installations.

Note: To limit the size of that post, there is an “Extended Version” availabe as PDF at http://bit.ly/mam6oO. This PDF contains more detailed information, esp. the details to reason the table’s content.

[Update 2011-05-23: Tim Mangan blogged about an issue when Sequencing to Q: in the App-V 4.6 SP1 Sequencer. http://www.tmurgent.com/TmBlog/?p=402]

What’s up with the Q drive?

Microsoft App-V is the only application virtualization solution I know that incorporates the concept of a virtual drive letter. On the client, the App-V Client software creates a new (virtual) drive that can be considered as the “mount point” for all virtual applications. Whenever a virtual
application requests a file from the package, it will get this file finally from Q – even if the app requested the file from C.

(more details can be found in the PDF)

Each Virtual Package (like the OpenOffice Suite, certain Tools or Line-of-Business-Suites) has its own, dedicated “Primary Directory”
that is located directly under Q:. Each of these folders is individual per-package: two packages cannot share the same Primary Directory. The App-V Clients throws an error if two different packages try that. In the “ancient days” (actually before SP1 for App-V 4.6)) this Primary Directory had to follow the “8+3″ naming convention, but that’s another story..

Install Directory and Primary Directory

During Sequencing, one or the other Sequencing Wizard Screen will ask you for the “Primary Directory”. This is where the Sequencer will collect all files that have been added or modified during Sequencing, also it will use this directory for its own, internal files (like osguard.cp) and folders (like SoftgridUserSettings).

Depending on the Sequencer version, the Screen in question will prompt for “The primary directory the application was/will be installed to”
before or after the actual monitoring.

The Primary Directory has to be on the Q: drive. There is no discussion about that.

The Install Directory, as referred to by this article, is the folder where you actually install the application to during Sequencing. You may be
prompted by the application’s MSI or Setup.exe or you may manually copy files to the Install Directory. Whether or not this will be a folder on the Q: or C: drive – this is what we’re discussing here. If you install an application to C:, every file later on will be copied into the “VFS” (Virtual File System) folder under the Primary Directory. This sometimes is called a “VFS install”.
If you install the application to a folder within the Primary Directory, this is going to be a “MNT install”, because the app is installed into the “Mount Point” (actually the Q: drive).

MNT vs. VFS: An Overview

Let’s start with a given statement: Microsoft strongly recommends installing applications to the Q: drive (hence perform an MNT install).

VFS vs. MNT – Some Details

Please refer to the attached PDF for more details on

• Official Statement
• Sequencing Performance
• Execution Performance
• Installer Compatibility
• Execution Compatibility
• Platform Compatibility
• Application Update Compatibility
• Package Branching Support
• DSC Conflict Prevention

Summary and Recommendations

As you can see, there is no easy answer to the question whether an application
should be installed to the C: or the Q: drive during Sequencing. At the end,
you will be required to try-end-error for some applications.

The recommendation is

• If you intent to “branch” an application a lot, sequence it to  C:
• Sequence Dynamic Suite Composition Components to Q:
• Sequence all other applications to Q, except if
  • They cant’s be installed/run from Q:
  • You have lots of “sophisticated” Unattended Installations that point
    to C: (then Sequencing it might be much easier and faster)

[Update 2011-05-23: Tim Mangan blogged about an issue when Sequencing to Q: in the App-V 4.6 SP1 Sequencer. http://www.tmurgent.com/TmBlog/?p=402]

Details PDF

In the attached PDF you will find the content of this post plus

• More details on the motivation
• More details on the different aspects that have been examined for the summary (VFS vs. MNT details)

–> To-Q-or-not-to-Q_That-is-the-Question

Sebastian namely thanks Jan Pfisterer and Marius Gawenda, who pointed him to that issue.

Because the new Sequencer does not require a separate hard disk partion for the Q: drive (but uses a Symbolic Link), some components (like OCX controls) may not register correctly. In their registration, they may point to the Symbolic Link’s target (on the C: drive) instead.

The Issue

The new App-V Sequencer 4.6 SP1 does not require a separate drive or partition for the Q: drive letter any longer. During installation, a Symbolic Link is created that is labeled “Q:” and that points to “C:\ProgramData\Microsoft\Application Virtualization Sequencer\Package Root”.

All files that are installed to the Q: drive directly during monitoring and all files that the Sequencer creates or collects in Q are stored in that folder on C:

Obviously it now may happen that some activities do recognize the Symbolic Link’s target location (instead of the “original” label location) for certain reasons.

It Sebastian’s post (and in a German discussion at dsgug.de), namely OCX registration via Regsvr32.exe seems to struggle:

The component “Foo123.ocx” is located in Q:\PrimaryDirectory\InstallationDestinationDir and has to begistered with Regsvr32.

“Local” Registry Value would look like:

[Standard] (Reg_SZ)
Q:\PrimaryDirectory\InstallationDestinationDir\Foo123.ocx

Expected Virtual Registry:

[Standard] (Reg_SZ)
%SFT_MNT%\PrimaryDirectory\InstallationDestinationDir\Foo123.ocx

Real Virtual Registry:

[Standard] (Reg_SZ)
%CSIDL_COMMON_APPDATA%\Microsoft\Application Virtualization Sequencer\Package Root\PrimaryDirectory\InstallationDestinationDir\Foo123.ocx

Because on the client there is no “C:\ProgramData\Microsoft\Application Virtualization Sequencer\Package Root”, also not in the Virtuel Environment, the component can’t be loaded: The application (probably) fails.

The Solutions (well: Workarounds)

1. On the Sequencer Machine, attach a secondary drive or create a second partition and label it Q: before installing the Sequencer 4.6 SP1 Software. The installer then does not create a Symbolic Link to C:\… but uses the Q: drive directly. (actually, use the same like for pre-SP1 Sequencers).

2. When registering the OCX, don’t run regsvr32.exe from “somewhere” and point to the full path of the OCX. Instead, change to the OCX’ folder and run Regsvr32 without specifying a path. According to Sebastian’s Blog, this should work.

So:

Do not

C:\> Regsvr32.exe "Q:\PrimaryDirectory\InstallationDestinationDir\Foo123.ocx"

Do

Q:\PrimaryDirectory\InstallationDestinationDir>  Regsvr32.exe Foo123.ocx

3. Don’t install the application to the Q: drive, but install it to C: instead. While Microsoft strongly recommends to install to the Q: drive, there are various reasons why not to follow that and install to C:.

Future

This issue has been reported to Microsoft and they may fix it in a future version of the App-V Sequencer

Remark:

Again, tributes go to Jan Pfisterer and Marius Gawenda who reported that issue and to Sebastian Gernert who analyzed and blogged about it.