Are MDT + SCCM task sequences worth the effort?

[ThatOneGuy]

What's so great about the MDT / SCCM integration, specifically about task sequences and OS deployment?

 

I understand that using MDT to build an image is preferable for a lot of reasons, and I'm okay with that part. It's the MDT-enabled task sequences that are bothering me. It seems like whenever I need to create an MDT task sequence, I have to bend over backwards, analyze everything in the task sequence created, tweak the defaults, pat my head, and rub my tummy before it will (somewhat) work correctly.

 

Here's my latest example: I needed a fairly simple refresh task sequence to be run on about fifteen computers of a single hardware model. I'm not concerned with preserving any data on them; I just need to format and install Windows, along with a set list of applications. I'd like to use copyprofile as well, so I turned to MDT instead of building my own unattend.xml from scratch. My MDT task sequence is now three times as long as my original "vanilla" task sequence was, contains unneeded variables for BitLocker, UEFI, and the UDI wizard, and it still fails! I get seemingly random 80040005 errors throughout the task sequence with no apparent rhyme or reason (they actually happen in different places in the same task sequence on the same machine) with no output in smsts.log. The SMSTS.log files are now stored in about four different places. In fact, when I create a new MDT task sequence without selecting the UDI wizard and try to run it from within Windows (using the Software Center), it doesn't even restart into Windows PE - I get an error when it tries to partition Windows from within Windows.

 

I'm fairly new to MDT, so it's very possible I'm doing this wrong...but from my perspective, integrating MDT with SCCM has been nothing but a headache. Could anyone explain some of the advantages of using these enhanced task sequences?

[Peter van der Woude]

Main advantages are the buildin scripts and monitoring of the deployment, but that doesn't add up for everybody. It really depends on the usage and the users. You can make it as difficult, or easy as you want, but in the end it all depends on if it's manageble and if it's what you or the customer wants.

[Apexes]

Example why i cannot live without MDT:

 

I work for a large financial organization, i'm heading up a migration project from WinXP to Windows 7 and am the sole contact for SCCM & Automation.

 

My MDT Task sequence contains 196 steps, and takes around an hour to install a fully new image, ready to go for users.

 

I've got 27 site DP's across 17 countries, so customization with minimal effort is a must.

 

MDT gives you the ability to have a variable image over multiple sites by running only one task sequence, sure - it takes some configuration in the first part, but in that MDT Database, i have settings which ensures the correct time zone is applied for each region, keyboard settings, input locale's etc. If i detect you connecting from a default gateway based in china, you get the chinese language packs for Office 2010, IE9 and Windows 7 by default. At the same time i've got a user in germany running the exact same sequence, getting only whats specific for his region

 

If i was to create a seperate TS for each region - it would have taken me probably 20x longer than it would have to of created an MDT integration.

 

I'm also able to monitor any OSD image at any time, no matter where i am - I injected DaRT 8.1 into MDT, which gives me remote control abilities, so i can troubleshoot clients remotely if they fail, even in WinPE stage.

 

Monitoring - i can see at any point what machine is imaging, how far it's into the image, and when completed.

 

Log file copying, all MDT logs go to a log share on my primary site.

 

Application & Package variation across locations - add in any package i want for a specific office.

 

Also by keeping just one task sequence, rather than 17 - means that if i need to update an application in my base image, i only do it once on my main task sequence, rather than 17 times.

 

I think MDT works for some people, but not for all. Certainly in my role i'd be lost without it, everything is so much easier with it in place. I agree with you that it can be fiddly at times, and it takes some getting used to, but i can reel off the log locations by heart now just by looking at what background i have in OSD

 

Stick at it, it's a great tool to use and once you've got your head around it you can take that on your CV to other job roles and it'll make your life alot easier and offer you perhaps a few more opportunities career wise if you can use it.

[topcal]

Microsoft Deployment Tool kit is great and works wonders after its initial setup. But again, its too complicated to get basic sequences done. I recommend using it for interactive sequences.

 

But for the scenario outlined in this post, a standard SCCM 2012 task sequence will suffice.

 

Use the following front-end to set the initial variables and then perform the normal SCCM 2012 task sequence steps (no dramas). This works for me all the time for small sequences.

 

http://gallery.technet.microsoft.com/SCCM-2012-OSD-Task-9c3f2be7

 

 

 

 

 

[JagoWu]

ThatOneGuy,

I went through the same delimma and realized that the default MDT TS in SCCM has bugs. However, the good news is there are small steps in the TS you can add to make it work correctly and you should treat those added steps in the sequence as part of the default sequence. For bitlocker I am still having a headache trying to get bitlocker to save the key as a .txt file to a network share because the laptop is apart of Workgroup and not domain.

 

Here are some things I found out that work for me in a MDT TS.

 

1. Right under the Execute Task Sequence (1st step) step you should add three steps in it.

 

SMSTSDownloadRetryCount = 5 <-- needed for downloading packages requests and such.

SMSTSDownloadRetryDelay= 15 <-- needed for downloading packages requests and such.

SMSTSRebootDelay=2 <-- this is the area where I do not want to wait 30 seconds for a reboot in the TS so I change it to 2 sec.

 

2. In the steps called Format and Partition Disk (UEFI) I deleted the first 3 partitons listed and the last partition i left there. The last partition is OS Disk (Primary)

 

3. In the step called OSDPreserveDriveLetter I set it to TRUE. If set to false then windows might install on some partition named E and when windows is installed in explorer you will have D windows instead of C windows.

 

4. In the Apply Windows Settings step make sure you set an admin password and the timezone should be the same as the sccm server or else it seems to bomb out for me during sysprep.

 

5. After Set Status 5 add new step called Request State Store.

 

6. After Restore User State step add a step called Release State Store.

 

Those added changes in the step should help you out.

NOTE: This has worked for me in a BareBones install and a Refresh install.

[Msquare]

On 5/9/2014 at 4:22 PM, Apexes said:

Example why i cannot live without MDT:

MDT gives you the ability to have a variable image over multiple sites by running only one task sequence, sure - it takes some configuration in the first part, but in that MDT Database, i have settings which ensures the correct time zone is applied for each region, keyboard settings, input locale's etc. If i detect you connecting from a default gateway based in china, you get the chinese language packs for Office 2010, IE9 and Windows 7 by default. At the same time i've got a user in germany running the exact same sequence, getting only whats specific for his region - How do I apply this configuration, is this documented somewhere?

 

Quote

I work for a large financial organization, i'm heading up a migration project from WinXP to Windows 7 and am the sole contact for SCCM & Automation.

 

My MDT Task sequence contains 196 steps, and takes around an hour to install a fully new image, ready to go for users.

 

I've got 27 site DP's across 17 countries, so customization with minimal effort is a must.

 

MDT gives you the ability to have a variable image over multiple sites by running only one task sequence, sure - it takes some configuration in the first part, but in that MDT Database, i have settings which ensures the correct time zone is applied for each region, keyboard settings, input locale's etc. If i detect you connecting from a default gateway based in china, you get the chinese language packs for Office 2010, IE9 and Windows 7 by default. At the same time i've got a user in germany running the exact same sequence, getting only whats specific for his region

 

If i was to create a seperate TS for each region - it would have taken me probably 20x longer than it would have to of created an MDT integration.

 

I'm also able to monitor any OSD image at any time, no matter where i am - I injected DaRT 8.1 into MDT, which gives me remote control abilities, so i can troubleshoot clients remotely if they fail, even in WinPE stage.

 

Monitoring - i can see at any point what machine is imaging, how far it's into the image, and when completed.

 

Log file copying, all MDT logs go to a log share on my primary site.

 

Application & Package variation across locations - add in any package i want for a specific office.

 

Also by keeping just one task sequence, rather than 17 - means that if i need to update an application in my base image, i only do it once on my main task sequence, rather than 17 times.

 

I think MDT works for some people, but not for all. Certainly in my role i'd be lost without it, everything is so much easier with it in place. I agree with you that it can be fiddly at times, and it takes some getting used to, but i can reel off the log locations by heart now just by looking at what background i have in OSD

 

Stick at it, it's a great tool to use and once you've got your head around it you can take that on your CV to other job roles and it'll make your life alot easier and offer you perhaps a few more opportunities career wise if you can use it.