Package Software While You Sleep Using the SCCM Application Packager Script – Part 2

, , , , , , , , ,

Welcome to Part 2 of the 2-part series about the SCCM Application Packager Script! If you have not yet read Part 1, I highly recommend doing so before continuing.

The SCCM Packager Tool works by parsing and processing XML files known as “recipes”. Recipes provide the instructions to the SCCM Application Packager Script that allow it to create applications in SCCM. Recipes are broken down into 5 main nodes. This post will cover each of the nodes, many of the tags and properties found in each node, and their purpose. Once a general understanding of the recipes and their properties are established, we will dissect the 7-Zip recipe for further insights. Please note that there are more properties available than the ones explained here. All available properties and options are documented in the Template.xml file located in the “Recipes” folder of the SCCM Application Packager Script.

Application Node

The Application Node contains information for the application that will be visible to the end user when the it is deployed. Essentially, the Application Node defines what will be seen in Software Center for each application. The Application Node does not have many options overall, so they are all broken down below:

  • Name – The name of the application as shown in SCCM and the Software Center
  • Description – The description of the application in Software Center. Also sets the “Administrative Comments”
  • Publisher – The application publisher as shown in Software Center
  • AutoInstall – Whether to allow the application to install during a task sequence or not
  • UserDocumentation – Link for the application to provide the end user with more information
  • Icon – The filename of the application icon in the icon repository folder

Downloads Node

The Downloads Node is responsible for downloading and copying the application installers and any associated files to the application share. Multiple downloads can be specified (for things like 32 and 64-bit downloads), and each download should be named to correspond with a “DeploymentType” in the Deployment Types Node. Different destinations can also be specified for each download (see the Template.xml file for more information). The Downloads Node also has just a few options, the most common are as follows:

  • PrefetchScript – The prefetch script can be used to download the software or find the URL of the download for the application if a direct link to the application download isn’t available.
  • URL – The link to download the latest version of the application.
  • DownloadFileName – Specify the name of the file when it is downloaded
  • Version and FullVersion – These can be blank, but MUST be present in each download definition
  • DownloadVersionCheck – This can be used to check the version of the application that was downloaded. If the version can’t be checked directly, additional processing can be done here as well
  • AppRepoFolder – Allows different folders to be specified on the application share per download
  • ExtraCopyFunctions – If additional files like configurations, installation scripts, or “.mst” files are needed, they can be copied to the content share here

Deployment Types Node

The Deployment Types Node is where all the magic really happens. In the Deployment Types Node, multiple deployment types can be specified for a single application. At this time, Script and MSI deployment types are supported. Each deployment type supports things like: Custom installation and uninstall commands, Custom and multiple detection methods including script detection methods, Installation behavior, and dependencies. The number of options available are too great to list for the deployment types, so it is recommended to check out the Template.xml to see the options available.

Distribution Node

The Distribution Node is responsible for distributing the content to SCCM distribution points. Generally, only one option will need to be specified here, and that is whether or not to distribute the content. If the “DistributeContent” property is set to “True”, the content for the application will automatically be distributed to the distribution point group specified in SCCMPackager.prefs. The following options are also available in the Distribution Node:

  • DistributeContent – Set to True to distribute content, Set to false to skip distributing content
  • DistributeToGroup – Distribution Point group to distribute content to
  • DistributeToDPs – Distribution Point to distribute content to

Deployment Node

The Deployment Node is similar to the Distribution Node. It has very few options and, generally, the only option that needs to be set is “DeploySoftware”. If the “DeploySoftware” property is set to “True”, the software will automatically be deployed as available to the default collection specified in SCCMPackager.prefs. Additionally, a collection other than the default can be specified for deployment. Other options available in the Deployment Node are as follows:

  • DeploySoftware – Set to True to deploy the software, Set to false to skip deploying software
  • DeploymentCollection – Specify the name of a collection to deploy to other than the default


Recipe Breakdown – 7-Zip.xml

Now that we have a general idea of the purpose of recipes and the features available to them, let’s break down a simple one to demonstrate a real life example. The recipe for 7-Zip is relatively simple, containing two downloads and two MSI deployment types. One of these deployment types also contains a requirements rule. Below, each node of the 7-Zip recipe will be dissected and discussed.

7-Zip Application Node

    <Description>7-Zip is a file archiver with a high compression ratio. Free and Open Source Software</Description>
    <Publisher>Igor Pavlov</Publisher>

As described earlier, the Application Node is fairly simple overall. The 7-Zip application will be packaged with the name “7-Zip <Version>”, the Description as shown in Software Center will be: “7-Zip is a file archiver with a high compression ratio. Free and Open Source Software”. The Publisher and Link provided in Software Center will be “Igor Pavlov” and “” respectively. The SCCM Application Packager Script will also mark the option to allow the packaged application to be installed during a Task Sequence, and will choose the icon file “7Zip.ico” from the icon repository specified in SCCMPackager.prefs.

7-Zip Downloads Node

    <Download DeploymentType="DeploymentType1">
        <PrefetchScript>$URL = "$((Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*-x64.msi")[0].href)"</PrefetchScript>
        <DownloadVersionCheck>[String]$Version = ([String](Get-MSIInfo -Path $DownloadFile -Property ProductVersion)).TrimStart().TrimEnd()
        $Version = ($version.Split('.'))[0..1] -join "."</DownloadVersionCheck>
    <Download DeploymentType="DeploymentType2">
        <PrefetchScript>$URL = "$((Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*.msi")[0].href)"</PrefetchScript>
        <DownloadVersionCheck>[String]$Version = ([String](Get-MSIInfo -Path $DownloadFile -Property ProductVersion)).TrimStart().TrimEnd()
        $Version = ($version.Split('.'))[0..1] -join "."</DownloadVersionCheck>

The Downloads Node for 7-Zip has a bit more going on than the Application Node. Upon initial inspection, we can see that there are 2 “DeploymentType(s)” specified, one is tied to “DeploymentType1”, and the other is tied to “DeploymentType2”. We can also tell pretty quickly based on the “DownloadFileName(s)” properties of these downloads that they download the 64-bit and 32-bit msi installers of 7-Zip respectively.

You will notice that a URL is not provided in either of the downloads, however, each download has a “PrefetchScript”, that sets a variable called “$URL”. Let’s look at what the “PrefetchScript” in the “DeploymentType1” download does:

$URL = "$((Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*-x64.msi")[0].href)"

We can see that this script is setting a $URL variable, it just so happens that the $URL variable is tied to the URL tag in the recipe, which is what the SCCM Application Packager Script will attempt to download. That means that whatever we set $URL to in the prefetch script will be downloaded by the Packager Script. If we break down the command a bit more, and run the command found in the inner-most parenthesis in PowerShell:

PS C:\Users\Andrew> (Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*-x64.msi")

innerHTML : Download
innerText : Download
outerHTML : <A href="a/7z1900-x64.msi">Download</A>
outerText : Download
tagName   : A
href      : a/7z1900-x64.msi
innerHTML : Download
innerText : Download
outerHTML : <A href="a/7z1604-x64.msi">Download</A>
outerText : Download
tagName   : A
href      : a/7z1604-x64.msi
innerHTML : Download
innerText : Download
outerHTML : <A href="a/7z920-x64.msi">Download</A>
outerText : Download
tagName   : A
href      : a/7z920-x64.msi

we will see that it returns 3 possible download links to 7-Zip MSIs. If we expand a bit and run everything found in the outer-most parenthesis:

PS C:\Users\Andrew> $((Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*-x64.msi")[0].href) 

we are greeted with a file path for the latest 7-Zip MSI file. Finally, if we run the entire “PrefetchScript” script block:

PS C:\Users\Andrew> $URL = "$((Invoke-WebRequest |Select -ExpandProperty Links |where -Property href -like "*-x64.msi")[0].href)"
PS C:\Users\Andrew> Write-Output $URL

we can see that it returns a full URL. If we go ahead and point our browser to that $URL, we can see that it is a direct download of the latest 7-Zip MSI file. This clarifies the purpose of the “PrefetchScript” Property, it allows the Packager to find the latest download link, even when a hotlink to the latest version is not provided. Crafting a good “PrefetchScript” is one of the harder things to do when creating a recipe, it can take much trial and error. A suggestion for creating a “PrefetchScript” is to build it outside of the packager, and only add it in once it is perfected as it’s own little script.

Moving on, the final thing we will look at in the Downloads Node is the “DownloadVersionCheck”. This property is also just a PowerShell script block. The goal of the “DownloadVersionCheck” is to provide the proper version of the application to the Packager so that it can determine if this download is a new version, and package it if it is new. The “DownloadVersionCheck” script for this recipe is fairly simple:

[String]$Version = ([String](Get-MSIInfo -Path $DownloadFile -Property ProductVersion)).TrimStart().TrimEnd()
$Version = ($version.Split('.'))[0..1] -join "."

First, the “ProductVersion” is grabbed by the “Get-MSIInfo” function, any leading or following white space is trimmed, and the string is saved to the “$Version” variable. Now, this version string will be something like “” which is great and all, but 7-Zip does not even use the last two 0’s on their version string. So, for the sake of consistency with the vendor, we split the “” version string by the “.” and only keep the first two array values (“19” and “0”), then we join those back together with a “.” and assign the joined string back to “$Version”.

Now you may notice, just like there is a $URL that corresponds to the URL tag in the XML, there is also a “Version” tag in the XML; if you were thinking that the “$Version” variable ties to the “Version” tag in the XML, you would be exactly right! It is also important to note that having a Version assigned to every application packaged is a very important aspect to the packaging process. The “Version” property plays an important role, not only in the naming of the application, but in the folder structure on the application share.

7-Zip Deployment Types Node

    <DeploymentType Name="DeploymentType1">
        <DeploymentTypeName>7Zip Silent Install x64</DeploymentTypeName>
        <Comments>Silent Installer for 7Zip</Comments>
            <RuleName>Existential of AutoPackage - OSArchitecture x64 Not Equal to 0</RuleName>
    <DeploymentType Name="DeploymentType2">
        <DeploymentTypeName>7Zip Silent Install x86</DeploymentTypeName>
        <Comments>Silent Installer for 7Zip</Comments>

Alright here we are, the meat of the Recipe. The Deployment Types Node usually ends up being the largest part of any Recipe (depending of course on how tough it is to scrape the latest download for an application).

Looking over the Deployment Types Node, we can see that there are two “DeploymentType” objects with the names “DeploymentType1” and “DeploymentType2”, which correspond with the downloads we saw earlier with the same names. There is also a “DeploymentTypeName” specified in each deployment type, but don’t let that fool you, the “DeploymentTypeName” tag only specifies the name of the deployment type in SCCM. Looking over the rest of the deployment type properties, most of the options should look familiar if you have ever created an application in SCCM before.

We can also see that the 64-bit deployment type has a “Requirements” tag with an associated “RuleName”. These properties specify that a requirements rule with the name “Existential of AutoPackage – OSArchitecture x64 Not Equal to 0” should be added to the application. By default, on it’s first run, the SCCM Application Packager Script will create a few Global Conditions for use when packaging; one of which is “Existential of AutoPackage – OSArchitecture x64” which returns a “1” on a 64-bit operating system. Therefore, this rule will ensure that the 64-bit version will only install on 64-bit computers, while the 32-bit install will run on anything else.

Some other options of note for this recipe are:

  • InstallationType – This sets the type of installation for the deployment type. The MSI “DeploymentType” is nice because if you provide an MSI, SCCM will automatically provide things like Install/Uninstall commands and a detection method. You will notice that specific install and uninstall commands are not specified in this recipe, however they can be added to MSI deployment types in the Packager
  • InstallationMSI – This is the MSI provided to SCCM for it to generate the Install/Uninstall commands and the detection method
  • DetectionMethodType – This specifies that the MSI file should be used to generate the detection method automatically, instead of using a script, file, or registry key for the detection method

7-Zip Distribution Node


Now that we have covered all of the complex parts of the recipe, we can end on a much more gentle note. The Distribution Node for the 7-Zip application only contains a single property. The “DistributeContent” property for 7-Zip is set to “True” meaning it will be distributed to the default distribution point group specified in the preferences file.

7-Zip Deployment Node


Last but not least, the Deployment Node for the 7-Zip recipe also contains just a single Property. The “DeploySoftware” property here is also set to “True” meaning the software will be deployed as available to the default collection specified in the preferences file. I will note that this is the default for all of the recipes provided, but can be changed per recipe depending on your specific needs.

7-Zip Application End Result

This post went a bit long, so I am going to let the results speak for themselves here:

This slideshow requires JavaScript.

The SCCM Application Packager Script is a very powerful tool. Its recipes were designed for ultimate flexibility with the intention of automating the packaging of just about any application in SCCM. Creating your own recipes can be a time consuming process, however, the end result will end up saving a ton of time. In my own production environment, I have saved an estimated 220 hours a year with this tool, and that number will only grow as more recipes are created, and more application updates are released per month.

Thank you for taking the time to read this post and your interest in the SCCM Packager Script! I created this tool with the intention of saving people time and energy, and was able to deepen my knowledge in SCCM and PowerShell as a result. If you are interested in contributing your own recipes to the project, feel free to submit pull requests to the GitHub repository.

Andrew Jimenez

Twitter: @AndrewJimenez_

In the MEMCM Lab…

, , , ,

In the MEMCM Lab

With a pen and a pad, trying to get this deployment off

Updated: 7/8/2020 on

My first lab was Johan’s hydration kit. It’s incredibly powerful, customizable, and educational. Unfortunately it takes a little more time and know-how than a novice like myself was initially prepared for.

However, at MMS Steve Jesok pointed out that Microsoft provides an all-in-one solution: the Windows and Office Deployment Lab Kit. Within minutes, we can have a fully functional domain controller and MEMCM server.

The Requirements

  1. Set up a host device – For this lab, I’m using a Windows 10 Pro workstation with an old i7 CPU, 16GB of RAM, and a secondary 500GB hard drive.
  2. Enable Hyper-V – If Hyper-V is not yet installed, open Turn Windows features on or off, check Hyper-V, and click OK. Reboot.
  3. Configure networking – Launch Hyper-V as administrator, and open the Virtual Switch Manager. Under Virtual Switches, select New virtual network switch. Select External and click Create Virtual Switch. Name it Lab and and leave everything else default. Click OK, and if you are prompted with a warning, click OK again.

The Setup

  1. Download the kit from the link above. It has the virtual machines and step-by-step documentation on how to configure services. This is the only thing we need to download.
  2. Extract the lab zip file, preferably to a drive that is large and fast.
  3. Install – Right click Setup.exe and run as administrator. If prompted by SmartScreen, click More Info and then click Run anyway.
  4. Setup Wizard – Click Next all the way through to the end. It will import all the VMs into Hyper-V.
  5. Configure VM Settings – You should see HYD-DC1 and HYD-GW1 already running. Shut them down. We won’t be using HYD-GW1 again.
  6. Domain Controller – Right click HYD-DC1 and select Settings. Set Maximum Memory to 2048MB and leave Enable Dynamic Memory checked. Set CPU to one virtual processor.
  7. MEMCM Server – Right click HYD-CM1 and select Settings. Leave memory settings at default. Set CPU to two virtual processors.

NAT Networking

Note: We will NOT be using the external virtual switch called Lab from Step 3 of the Requirements section. It was only necessary so that Setup.exe from the Setup section would run.
  1. NAT Networking – We’ll use Ami Arwidmark’s NAT network script instead of the Internet Gateway (HYD-GW1) to make the lab simpler. You can learn more about NAT networking here.
  2. Prepare the Virtual Switch – The Deployment Lab Kit creates it’s own private network switch, so we need to make it an Internal one to work with Ami’s script. In Hyper-V click Virtual Switch Manager. Click on HYD-CorpNet. Select Internal network and click OK.
  3. Customize the script – On the host system, launch Windows Powershell ISE as Administrator. Copy and paste the following code into the top script pane. This is an edited version of Ami’s code customized for our Microsoft lab. Hit F5 to run it.
     New-NetIPAddress IPAddress -PrefixLength 24 -InterfaceAlias "vEthernet (HYD-CorpNet)" 
     New-NetNat Name HYD-CorpNetNATNetwork InternalIPInterfaceAddressPrefix
  4. We now have our host Windows 10 OS performing NAT on the internal virtual switch HYD-CorpNet. Our VMs are already pointing to it as the default gateway.

The Test

  1. Power on HYD-DC1 and wait for the log on screen. This is so our servers and workstations can talk to Active Directory.
  2. Power on HYD-CM1 and log in. The passwords for the local administrator accounts and for CORP\LabAdmin is P@ssw0rd
  3. Confirm the MEMCM server has internet access by launching command prompt and pinging
  4. Give HYD-CM1 another moment for services to start up. Launch the Microsoft Endpoint Manager Configuration Manager Console and confirm that it loads successfully.

Refer to the troubleshooting section at the end if anything isn’t working at this point

The Finishing Touch

Now that we’ve got a functioning domain, MEMCM server, and internet access, it’s time to update.
  1. In the MEMCM console, navigate to the Administration node and select Updates and Servicing. Click on Check for updates.
  2. If the latest version hasn’t already started downloading, select it (in this case 1902), right click and choose Download.
  3. Once it is downloading, on the bottom pane click on the Show Status link.
  4. On the Updates and Servicing Status page for our chosen update, right click update package and choose Show Status again.
  5. From here, we can follow the download AND installation statuses of the latest MEMCM upgrade.
  6. Once the download is complete, go back to the Administration node and click on Updates and Servicing again. The update we downloaded should now say Ready to install.
  7. Right click the update and select Install Update Pack. Check Ignore any prerequisite check warnings… and click Next until we reach the License Terms. Check the box, and keep clicking Next until the wizard completes successfully. Click Close.
  8. Repeat steps 3 and 4 and watch the update installation progress. Refresh until the Update Wizard is complete and click Close.
  9. Close the MEMCM console and relaunch it. We may be prompted to upgrade the console to the new version. Click OK, and if prompted for elevation click OK again.

Congratulations! We now have a functional MEMCM environment we can configure and customize.


If there are any obstacles during set up, we can try some of these troubleshooting tips
  1. Firewall – If you cannot ping, we don’t have access to the internet. From CM1, try pinging DC1 at If that works, try pinging the NAT gateway at If that doesn’t work, try temporarily disabling the firewall as that might be blocking access.
    You may need to remove and redo the NAT networking as well, so run the following command in an elevated Powershell terminal:

     Remove-NetIPAddress -IPAddress
     New-NetIPAddress IPAddress -PrefixLength 24 -InterfaceAlias "vEthernet (HYD-CorpNet)" 
     New-NetNat Name HYD-CorpNetNATNetwork InternalIPInterfaceAddressPrefix
  2. NAT – If you can ping but STILL can’t ping, make sure HYD-GW1 is powered off. If it is, the issue is on the host system. From the host system, ping CM1 at to confirm NAT is working. If NAT is working, from CM1 ping the host IP of the physical adapter.
  3. Subnet – The lab network is If our home network is also on we’ll have trouble getting out. We will either need to ditch the NAT and rely on GW1, or re-IP DC1 and CM1 and our NAT configuration on a different network. Just keep in mind in subsequent blog posts we’ll need to adjust networking respectively.
    For example if you wanted to change the lab from the default network to a network, change the CM1 IP to and the DC1 IP to Remove the NAT config and make a new one on that network like so:

     Remove-NetIPAddress -IPAddress
     New-NetIPAddress IPAddress -PrefixLength 24 -InterfaceAlias "vEthernet (HYD-CorpNet)" 
     New-NetNat Name HYD-CorpNetNATNetwork InternalIPInterfaceAddressPrefix