Monthly Archives: September 2013

Using Microsoft Deployment Toolkit to start building a pristine infrastructure

Building Private Clouds is becoming more and more prominent in today’s IT world. But what if you have to start from scratch?

Imagine yourself, you need to build a Private Cloud infrastructure but you have to start from scratch? Ok, so you have purchased all the hardware needed to build the infrastructure but you have nothing. I guess you want to have something that will get you started quickly.

Microsoft Deployment Toolkit allows you start building quickly. You can easily build a machine (just a workgroup machine), install roles such as WDS, WSUS, DHCP (if not available) and DNS (nice to have) and install and configure MDT. This allows you to quickly create some sort of automation in deploying the base foundation of your Private Cloud.

Even though I integrate MDT in almost every Configuration Manager 2012 SP1 deployment, I never used it in a stand-alone setup. While I had some time I decided to have a look. What surprised me a lot is that many people wrote blogs about MDT. I remember I attended a session at MMS 2013 where MDT was explained extensively, the session was hosted by Mikael Nyström and Johan Arwidmark which I considered very valuable (and it was a lot of fun too)…

So, with all the stuff available on the Internet regarding MDT, I got started…

I used a machine that was eating dust but is still able to run Windows Server 2012 rather well. It has enough storage to store a bunch of Windows images and all related stuff I might want to deploy.

Once the basic stuff is there, I remove that machine from the infrastructure. It makes me think that a VM on my laptop could do the same thing, food for thoughts…

I used MDT 2012 U1 as a reference because it is the current version, I went to to download the installer. I am aware release of MDT 2013’s release is pending (at time of writing) but I expect I can reuse everything already used in MDT 2012 U1.

MDT comes with quite a bunch of scripts that should get you started pretty quickly. You can also adopt your own scripts and you can even install some applications if you like. It might prevent you from using standard Windows media and do everything manually…

Oh, and don’t forget. You can even build a .vhd that you can use to get started right away with deploying VM’s once your core infrastructure is running with something.




ConfigMgr 2012 SP1: an error 80091007 – hash values couldn’t be matched fairytale, or nightmare…

As we all know, Configuration Manager has a mechanism which checks if content that is stored on a distribution point is the same when being deployed to a ConfigMgr 2012 SP1 client machine. It compares the hash value that ConfigMgr knows with the hash value of the content on the client machine. If they’re the same, then everything is fine. If not, then you receive the error message displayed above…

While ConfigMgr 2012 is good at verifying content distribution and alerting users and admins when something, it is not able to diagnose or explain why this happens. After all, content distribution is nothing more than a ‘dumb’ stream of data from source to destination. The destination will tell if the content arrived correctly (sounds like ‘basic’ TCP to me) or not.

Recently, one of our customers was confronted with this issue. He was facing this problem during OSD. It happened on different machines using different Task Sequences on different steps. What became particularly frustrating was that all lights in the ConfigMgr console are green. Content verification on all content was successful and all distribution points are working as they should. I also found it very hard to believe that the large number of test machines used had faulty RAM or disks.

Analyzing the smsts.log file of a failed deployment doesn’t provide any information that might ring some bells. All I see is the error message displayed above.

At that point the only thing I could conclude was that something went wrong during transport since all content storage was working as it should. I could also conclude that the error message cannot really help me because the possible cause is outside ConfigMgr’s scope, more digging was required. ConfigMgr is working as it should so something else caused the issue.

After a lengthy investigation I discovered that I had the same problem as some guys had in this forum thread:

I discovered that the customer had the same ‘cocktail’ of components which consists the following:

  • ESXi 5.1 hypervisor
  • Windows Server 2012 VM
  • ConfigMgr 2012 SP1
  • VM with a E1000 NIC
  • Unfortunately, I don’t know which network switches are used (in this case I considered this less relevant)

Unfortunately, the customer did modify the NIC used from E1000 to vmxnet3 using PowerCLI. However, he modified the NIC instead of replacing it with a new one. This happens unintentionally so that was OK. After deleting the existing NIC and adding a new one we could try again.

Personally, I try to treat a VM the same way as a physical server. Changing the hardware configuration means that components are being replaced. In this very example you take out the old NIC and put a new one in.

The error message went away the very first time after changing the NIC, so the nightmare was over…

I have to admit, this one was hard to find. My main focus is on ConfigMgr so I had to think outside the box on components where my knowledge is somewhat limited.

Steve Thompson [MVP]

The automation specialist

Boudewijn Plomp

Cloud and related stuff...

Anything about IT

by Alex Verboon

Deployment Made Simple

Modern Workplace

Azure, Hybrid Identity & Enterprise Mobility + Security

Daan Weda

This site is all about System Center and PowerShell

IT And Management by Abheek

Microsoft certified Trainer -Abheek

Heading To The Clouds

by Marthijn van Rheenen