Having struggled to deploy Vivado 2018 through SCCM despite it working perfectly from the commandline i discovered the setting CreateProgramGroupShortcuts in the install_config.txt file which was generated previously by running bin\xsetup.bat -b ConfigGen. Setting CreateProgramGroupShortcuts=0 was the fix as for some reason this would fail when run as the SYSTEM account.
I ended up doing a command line install to get the shortcuts and copying them to the correct location as part of the install script.
Recently tasked with deploying Matlab Support Packages I went off to Google to find some instructions. Mathworks do provide instructions here and here, but I find them a little too wordy and I’m not a fan of the “read this guide” and then “read that guide” half way through kind of style.
Here are my simplified instructions:
- On a machine with Matlab installed (same version as you’re deploying to), open Matlab and open the Addon Manager
- Locate and download the addon you want to deploy
- Once downloaded you’ll have a folder here: C:\Users\<YourUser>\Downloads\MathWorks\SupportPackages\R2018b
- Copy the above folder, you need everything inside it. It includes a config file which tells the Support Package installer what to install: ssi_input.txt
- Install through SCCM with the following script:
“C:\Program Files\MATLAB\R2018b\bin\win64\SupportSoftwareInstaller.exe” -downloadfolder “%DownloadFolder%” -inputFile “%DownloadFolder%\ssi_input.txt”
Note . You can’t simply use %~dp0 for this as that adds a trailing slash which makes the installer fail.
Update: Microsoft suggest (despite the ambiguous wording) that the latest hotfix does resolve this issue. It may also be worth updating your boot images after applying the hotfix to ensure the fix applies.
This is a bug introduced in SCCM version 1702 where unknown machines can not be built as they cannot find a task sequence. The main symptom is unknown machines will reboot before you get to the SCCM password screen when network booting – you will see the background image.
- Run this query on the database to find computers matching the GUID: select * from System_DISC where SMS_Unique_Identifier0 like ‘%ce458683-a99a-4898-97dd-85e72e1478ab%’
- Remote onto the machine identified in step 1
- Stop the service SMS Agent Host on the machine
- Delete the machine from the SCCM console
- Delete C:\Windows\smscfg.ini on the machine
- Start the service SMS Agent Host on the machine
Step 1 assumes the Unknown Computer GUID. If that finds no results it may be the case that there are new Unknown Computer objects (this may happen when upgrading SCCM for example).
To find the Unknown Computer GUIDS, run the following query on the database: select * from System_DISC where SMS_Unique_Identifier0 like ‘%ce458683-a99a-4898-97dd-85e72e1478ab%’
How to run a query
- RDP to uiwdbsms03
- Launch SQL Server Management Studio
- Connect to UIWDBSMS03
- Expand Databases
- Right click on UN2 and select New Query
- Paste in the query and click Execute
- The results in this example show the offending computer that has stolen the GUID
I came across this problem which manifested itself differently in a few cases but the most common result is something like this:
Windows 10 1511 (build 10586) includes a new bitlocker encryption, XTS-AES encryption algorithm, which cannot be read by earlier versions of Windows including Windows 7, 8/8.1 and 10 older than v1511.
The fix is documented here: https://social.technet.microsoft.com/Forums/en-US/07c809fc-486b-49aa-8df8-70e374d90402/sccm-2012-r2-sp1-preprovision-bitlocker-windows-7-cannot-read-drive-after-reboot?forum=configmanagerosd and all credit goes to Anders Horgen for creating a PowerShell script to apply the fix. It may be a little overkill but it works so here it is: OSD_Set_BitLocker_Chiper-v1.0.0.ps1
Steps to implement this fix:
- Create a new package (or use an existing if you prefer)
- Add a new step to your TS just before Pre-provision BitLocker
- Fill it in something like this:
New issue here with deploying an automated MDT build of Windows 7.
I found after deploying that the clock was in US format, the keyboard – US, everything defaulted to US. I couldn’t find an obvious quick way of rectifying this through MDT, so here is the easiest solution, no rebuilds necessary. Not something I’d come across before using old skool hand made images.
When applying the image in an SCCM task sequence, specify an answer file. Here’s the answer file I created for Windows 7 SP1 x64: en-GB unattend
To create your own, follow the guide here: https://www.scconfigmgr.com/2014/01/30/create-an-answer-file-for-language-settings-during-osd-with-configmgr/
This is my script for installing Citrix Receiver. The reg changes are to stop it launching on startup and to stop the nag for a user e-mail address that it needs on first run. This still cannot be prevented (to my knowledge) if the user is logged on during install, so either make it available in SCCM Software Center or if you can garantee no user will be logged in during install (and won’t login half way through) then go for a required hidden install.
CitrixReceiver.exe /noreboot /silent
IF EXIST "%PROGRAMFILES(X86)%" (GOTO 64BIT) ELSE (GOTO 32BIT)
reg add HKLM\software\Wow6432Node\citrix\dazzle\ /v startmenudir /t reg_sz /d "\UoN Citrix" /f
REG DELETE HKLM\Software\Wow6432Node\Microsoft\Windows\CurrentVersion\Run /v CitrixReceiver /f
REG DELETE HKLM\Software\Wow6432Node\Microsoft\Windows\CurrentVersion\Run /v ConnectionCenter /f
REG DELETE HKLM\Software\Wow6432Node\Microsoft\Windows\CurrentVersion\Run /v Redirector /f
reg add HKLM\software\citrix\dazzle\ /v startmenudir /t reg_sz /d "\UoN Citrix" /f
REG DELETE HKLM\Software\Microsoft\Windows\CurrentVersion\Run /v CitrixReceiver /f
REG DELETE HKLM\Software\Microsoft\Windows\CurrentVersion\Run /v ConnectionCenter /f
REG DELETE HKLM\Software\Microsoft\Windows\CurrentVersion\Run /v Redirector /f
Key points of sequencing Anaconda:
use a PVAD
Install Anaconda to C:\Anaconda_x_versionhere\
After installing, in the sequence do the first run tasks, change preferences and change any settings to point the PVAD.
Screenshots planned.. one day.
Sorry for the briefness!
Just sequenced Android Studio in App-V 5.0 Sp3 successfully. Here’s how:
Use the default GUID based PVAD and install JDK and Android Studio in the same package, all to the PVAD.
Job done 🙂
This lync may be of use too if Java has any issues trying to update itself and other bits and pieces: https://packageology.com/2014/02/sequencing-java-the-definitive-guide-part-1/
Here are 3 handy resources to have a read through before going ahead with the upgrade.
Well worth a read through the full section for the SP2 upgrade as it applies to R2 Sp1 too. So the check list, considerations and planning to upgrade to Sp1 right at the end of the article.
Test the database prior to upgrading
Expand the “To test….” section for better instructions on how.
And finally, a great Step-by-Step SCCM 2012 R2 Sp1 Upgrade Guide in pictures.
So MDT won’t capture. Towards the end of BDD.LOG I get this:
Error creating an image of drive C:, rc = 2
ZTIBackup COMPLETED. Return Value = 2
ZTI ERROR - Non-zero return code by ZTIBackup, rc = 2
Command completed, return code = -2147467259
Litetouch deployment failed, Return Code = -2147467259 0x80004005
The problem here became pretty much obvious on going just a few lines up:
About to run command:
"D:\Deploy\Tools\X64\imagex.exe" /append C: "\\PUIP02932\DeploymentShare$\Captures\WIN7X64X32_6-3-2015.wim" "WIN7X64X32CDrive" /flags Enterprise
Due to the default naming of the captured WIMs, you cannot capture the same task sequence twice in a single day without deleting/moving/renaming the first capture.