0

Our build process revolves around windows Deployment Workbench. Unfortunately this was setup by someone who is no longer with the company, and no-one has ever dared/needed to make any changes.

The other day it stopped working. It turns out that one of our build guys started thinking about changing some stuff in it, clicked something and now it no longer works (He is saying now that he right clicked on the 'LAB' entry in 'Deployment points' and hit 'Update', which took some time to run through apparently). The job has fallen on me to resolve and frankly I'm not sure what I'm doing.

I was wondering if someone with more experience than me can provide some pointers as to troubleshooting cos I'm feeling quite a lot in the dark here.

On the server I have Deployment Workbench up and running (MMC snapin) version 3.0. There is a WDS service that appears to be running ok, as does the tFTPd service. Nothing specific to this in event logs.

From the client side; PXE boot works and gets you to the Win PE launch, and it has the correct company logo as the background (proving to me that its loading win PE from the network).

WPEINIT runs, and asks for domain credentials, here the team simply put User/Pass/Domain in the boxes and click ok. Normally the build would kick off.

Instead they get an error message saying that the \NATBLU01\Distribution$ share isn't available.

Checking \NATBLU01\Distribution$ shows that its there and accessible over the network. Security/permissions seem ok, even 'ANONYMOUS LOGON' has read access to that share so I don't see that being a problem.

Digging the trace files from C:\MININT\SMSOSD\OSDLOGS\ after an attempt to run the build I can see an error saying much the same - 

<![LOG[Validating connection to \\NATBLU01\Distribution$]LOG]!><time="16:42:14.000+000" date="03-15-2012" component="LiteTouch" context="" type="1" thread="" file="LiteTouch">
<![LOG[FindFile: The file OSDConnectToUNC.exe could not be found in any standard locations.]LOG]!><time="16:42:14.000+000" date="03-15-2012" component="LiteTouch" context="" type="1" thread="" file="LiteTouch">
<![LOG[The network location cannot be reached. For information about network troubleshooting, see Windows Help.]LOG]!><time="16:42:24.000+000" date="03-15-2012" component="LiteTouch" context="" type="3" thread="" file="LiteTouch">
<![LOG[ERROR - Unable to map a network drive to \\NATBLU01\Distribution$.]LOG]!><time="16:42:24.000+000" date="03-15-2012" component="LiteTouch" context="" type="3" thread="" file="LiteTouch">

BDD.LOG shows much the same.

Full copies of the .LOG files can both files be found here : BDD.LOG LITETOUCH.LOG

I can get to a command prompt from the Win PE that boots from PXE, however there isn't any network stuff there. IPCONFIG returns nothing so none of the tests I would usually run resolve anything.

I'm at a loss frankly. I did wonder if I could perhaps start a new build process but if the change to the DeploymentWorkbench has knocked it offline I don't think I'm going to be able to create a new deployment.

Failing that; we do have a deployment point labeled type 'Media' which appears to be a DVD ISO image of one of the builds, but its dated 2008, is it possible to export the network build to .ISO and build from DVD? We are looking at new hardware to run this from anyway (for the impending Windows 7 roll out) so a temporary work round isn't going to be too much of a problem.

All assistance is appreciated!

EDIT : OK. Got it working again.

Solution was close to Newmanth's idea. The problem was that our PE image didn't appear to be connecting the network. I had an older copy of the PE boot.WIM on a stick that I had been using for other purposes. I booted that and correctly got a network connection. Showed a correct internal IP and could ping out etc etc.

However I was still getting the same errors in all the logs and in when wpeinit was running.

What I did seperately was to update the PE image that DeploymentWorkbench was pushing out to display a different back ground. I wanted to prove that I was working in the correct place. Turns out that I wasn't.

I went and looked at the other deployment stuff we had on this machine, Windows Deployment Services was installed and although all the install images are off line the boot image was online, so I uploaded the copy from my stick to that. Booted straight off. And fixed.

Working. Yay!

For anyone stumbling across this in the future you may find that although your deployment images are located in the DeploymentWorkbench, the Win PE boot image you are launching from is located in the associated Windows Deployment Services images.

Improve this question
1
  • On a final note, I now feel qualified for my next task which is to completely re-create this process on a clean VM and to document everything for my future self so I can easily trouble shoot next time something goes wrong.
    – Patrick
    Mar 21, 2012 at 16:36

1 Answer 1

Reset to default
2

The fact that ipconfig does not return any information indicates that the Windows PE image is missing the NIC driver for your particular model. It turns out that while Windows PE has many network drivers pre-installed, newer models will frequently use NICs that are not part of the base install. This is especially the case with large OEMs, such as Dell.

Technet has an article that explains how to add a driver to an offline Windows PE image using ImageX:

http://technet.microsoft.com/en-us/library/cc749350(WS.10).aspx

Caution! If your machine uses a multi-tiered NIC (such as Broadcom), you'll need to make sure you download the monolithic driver, which is usually made available from the OEM. Here is a blog that explains this issue:

http://blogs.technet.com/b/configurationmgr/archive/2010/02/09/nic-devices-that-require-a-special-driver-for-winpe-may-cause-a-configmgr-task-sequence-to-fail-if-a-vista-or-newer-os-is-being-deployed-via-an-operating-system-install-package.aspx

Improve this answer
4
  • 1
    Don't be confused by the fact that Windows PE pulls successfully from the network... this is done by the BIOS bootstrapper, NOT PE, so it (should) always work. The failure occurs once PE is loaded and online as that is the point when the workstation switches to trying to use the Windows NIC driver instead.
    – newmanth
    Mar 19, 2012 at 14:00
  • Thanks. My difficulty here is that this is a standard device, we have rolled out hundreds of this exact make/model with the current build. I wonder if this driver was dropped during the 'Update' process? I will have a read about adding it back in. And thanks for clearing up that I [i]should[/i] see some network information on the PE build. Not having a functional reference system is making it difficult to confirm whats what. And I would need to add the driver to the build on the WDS server, then PXE boot from the laptop at which point it should see the network (roughly).
    – Patrick
    Mar 19, 2012 at 14:21
  • My shop uses SCCM, instead of Deployment Workbench, so I am not familiar with how components are managed in the PE image with that tool. I'm guessing that when your guy updated the deployment points, the update (for whatever reason), did not include the NIC driver that was included in the prior version. It could be that he was working off an earlier version, or the workbench requires you to explicitly add the driver every time the image is updated... but it definitely looks like a missing driver issue.
    – newmanth
    Mar 19, 2012 at 14:24
  • really appreciate your advice, helped me narrow down where I was looking.
    – Patrick
    Mar 21, 2012 at 16:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .