Client Management Suite

Hello all,

I experienced a strange thing today when trying to capture an image on a machine. Maybe someone can tell me what I did wrong..

Using an earlier image from December, three machines were deployed last week. After imaging a forth today, I ran Windows upates on it and decided to pull a new image so that new deployments would already have the new Windows updates. Using the same job  as used three months ago, containing the "Prepare for XP image capture" task and the "Capture XP Image" task, I ran the job on this forth PC and moments later we get three phone calls from the users stating that thier PC's just shut down and rebooted. Upon investigating, the machines had also ran this prepare for image capture task. Although not actually selected in the Management console, some link between these machines and Prepare for image task still exists. They all ran a sysprep before being joined to the domain and all had the resetguid command run on the agents after deployment. We're using Altiris 7.0.

I'm obviously reluctant to do anymore imaging until this is resolved. Can anyone suggest a fix?

Thanks!

Mike

This should not happen...

Can you do a quick test, run a Run Script task, and type notepad.exe in the script area. Now run, this task on foruth machine and confirm on other 3 machines that this task is executed.

Also, have you created your own collection/groups?

which option are you using to execute task, Quick Run, New schedule or drag n drop from Silverlight page?

where the tech ran an image job on 3 computers and later added a fourth. Then he added the fourth all four machines ran the task\job, instead of just the new system he added.

Can you provide some screenshots? Are the computers still in the distribution list for the task?

Hi guys,

Thank you for taking the time to reply. I will try to run the script task when I get the office and ask the  users to notify me if the event does occur on all the machines involved. The machines are not in the list anymore as the three were imaged in december. I should mention that all the machines were booted to PXE so they were identified to Altiris with a computer name similar to "MININT-XXXXXX". Also, all three were run from the Job itself as a "quick Run". Another interesting thing that happened was when the machine I imaged yesterday completed, it prompted for a computer name when it rebooted after the image deployment. Thats how I setup the sysprep file and I accepted the default name which was generated randomly by setup.

Later when the issue happened, I noticed all three machines were renamed to this forth PC's randomly generated name, they were disjoined from my domain and added to a local workgroup. I resolved it by simply renaming them back to thier original names and then rejoining them to my domain.

Why Altiris decided to name all these machines to to the forth PC's random name, I don't know.

https://www-secure.symantec.com/connect/forums/ns7-run-script-task-still-executing-even-though-no-longer-scheduled

It appears that the client cache is not being purged. It offers a workaround, using deployment server, although you should be able to do the same with a task.

Okay folks,

For anyone who's been wondering about this issue, I've worked with Symantec support and here's what we found out.

After looking at the machines that were involved with this, we discovered that all 4 machines had the same GUID. Obviously, this isn't normal or correct. My understanding is that the GUID is created on the built out machine prior to capturing the image. When a "Prepare for Image Capture" task is performed on the box, the GUID is indeed copied but it is marked by the task to let the system know that once deployed, the client will need to generate a new GUID or that the existing GUID shouldn't be reused. We weren't able to determine why it wasn't doing this but we were able to confirm it was only behaving this way with that single image file. A test client was created, the GUID recorded, the image was captured and then deployed to two other test machines and all three machines generated unique GUID's.

I again deployed the faulty image to a test box and verified it issued the same GUID again. I ran the Aexagentutil /resetguid command off of that client but Altiris simply re-assigned the same duplicate GUID to the client again. It wasn't until I deleted the computer from the management console, ran the /resetguid command and rebooted the client that it finally received a new and unique GUID.

In the end, it comes down to having a "defective" image.. but why or how it got created in this state, no one seems to know.

Thanks for your suggestions and help!

Kind regards,

Mike

Thanks for stopping back with the solution.  It makes sense that a duplicate GUID would cause this issue, though as you said it makes no sense why that would have happened.  I'm glad you were able to get this resolved.