[mdt] Clients lose multicast connection after downloading image...HELP!

[scottpowers82]

Hey Everyone!

 

Hopefully someone will be able to help me on this as I have been tearing my hair out the last few days trying to nail down what is causing my issues.

 

OK, so I have had an MDT 2012 Update 1 server (Server 2008 R2 with WDS) up and running for quite some time now with (usually) no issues. I have gone ahead and started making my new image for my school system (Windows 7 x64 w/SP1) and am using a thick image approach so that everything is in the image and we can hopefully get things deployed faster. However, when I go to deploy my new image, the client makes the connection and goes through the process of downloading the image, restarts (as part of the normal standard task sequence) and then presents the attached error shown below in PrimaryImageError2.PNG. After click OK on that error, I am then presented with the deployment summary screen with a listing of errors (as shown in PrimaryImageError.PNG).

 

The odd thing is that if I go into my Deployment Workbench, right-click on my Deployment Share and turn off multicasting, the image loads perfectly every time. Seeing as how I am trying to apply this to several hundred machines this Summer, using unicast sessions is far from ideal.

 

I have Googled up and down the internet for possible fixes. I have re-captured my image (made in VMWare Workstation) several times with different options in sysprep thinking that may have caused issues, but no luck.

 

We were hoping to start imaging machines almost 2 full days ago now and I can only stay up till 1-2AM so many times in a row trying to figure this out.

 

Anybody out there have any ideas or advise?

 

Was debating upgrading to MDT 2013 to see if that fixes any issues, but I'm trying not to cause other problems/conflicts as well this close to my deployment time . However, if anyone thinks that might work, I'll definitely try it...

 

Thanks in advance for any suggestions and your time

 

Scott

 

[scottpowers82]

A couple more notes that I have found out (some new as of this morning).

 

1. Yes, I can still image a different physical machine with an older image. No issues there. Therefore, Multicast across our network should be working correctly.

 

2. If I start deploying my new image via multicast, then tell it to “bypass” multicast from the WDS server (hence, changing it over to unicast mid-download), it works.

a. If I let that same session continue with the multicast, it fails.

 

3. As of this morning, I am not longer able to multicast into my copy of VMWare Workstation. Before, my tests would fail (with the same errors I get on the physical machines), but they would at least connect to the multicast session and download the image. Now, both my vms are stuck at “Attempting multicast transfer”. If I tell it to bypass the multicast (again, from the WDS server), the connection works right away and things work just fine. One thing to note though is that when I look at my multicast transmissions, it shows my VM status as waiting until I tell it to bypass…then it goes through as normal.

 

4. I have also tried manually injecting my network drivers into my main image WIM file using DISM, but same results.

 

Am about to try upgrading my copy of MDT 2012 Update 1 to MDT 2013 here in a few minutes (making a backup now) to see if that has any effect, but I have to admit I have my doubts that will actually fix anything, but here’s hoping!

 

Thanks for the advice so far!

 

Scott

[SYNACK]

Sounds networky or drivery, perhaps older nic drivers, rebooting wds, looking for any changed switch config, splitting wds multicast into speed bands and finally a wireshark to look for multicast flow problems.

[scottpowers82]

I was thinking about the same thing with the possiblitity of the drivers. So, I downloaded fresh drivers from Lenovo's website (it's a Lenovo X130e Netbook) and used the DISM tools to inject the driver directly into the image. I have rebooted the server several times as well as changing the WDS multicast speeds around, but no luck. As far as I know, no changes have been made to our switches in quite some time and, as I stated earlier, I am still able to image a physical machine (on Multicast) without issue using the old image. But if I image a X130e with the new image on multicast, it drops out. Change it to unicast and it works just fine.

 

Just upgraded to MDT 2013 a few minutes ago and am regenerating my boot images now. Hopefully, that will make a difference...

 

A damn strange problem...

[scottpowers82]

So, just wanted to update this in case anyone else is having similar issue.

 

After updating to MDT 2013 (twice), my issue still persisted. Therefore, I took my same WIM file that I was having problems with and have now switched over to SCCM for my deployments. Had been wanting to switch over anyway, just wasn't planning on doing it right now. Anyway, I have pretty much given up on the issue and am delving full force into SCCM (finally!).