I have been kicking around a few ideas about this and I was wondering if anyone else has got this working. To start with, I am using 9.5 SP1 and been using LANDesk since 8.8 SP4 and still haven't figured out a better mousetrap on this.
As you may already know, true one to many multicasting is not a feature in LANDesk. Also, multicasting and OSD doesn't really work nor is it built to be able to image many computers simultaneously. At least in a way that it won't take hours to complete.
So I had a thought of changing the OSD script through advanced to somehow point to a Ghostcast server session like the old days of Ghost. So I am starting to look around and the possibility of having a tech at the remote site start the Ghostsrv.exe to begin a Symantec Ghostcast Server, create a session name, set it for restore an image and tell it the image name. They can then just hit accept clients and then pxe boot the PCs to an OSD script that instead of pointing to an image name, would have the line modified to point to the session. Then when they are all booted, the tech can go to the GhostCast server and hit send.
I know, this is like going back to the old sneaker net days, except instead of a boot disk you are using LANDesk to PXE boot. Of course this could also be centralized a bit more by creating the OSD as a scheduled task and then put all the devices in that task from the lab and run that. Then once they all show up in the Ghostcast server session, you hit send and then you have a faster version of imaging to multipe devices.
I did talk to LANDesk about this and of course I was told they don't test for this and don't support this method. But I can't go back to the staff and say, just image and wait when you want to do 30 PCs that need to be imaged. Come back the next day and they will all be done.
So the question is, has anyone else worked out a method like this so I don't have to go crazy with research and reinvent the wheel?
On another note. Has anyone tried modifying OSD to use Windows Deployment Services instead for true multicasting of an image?
Oh, and is there a better method that you have succesfully used to expedite the imaging of the lab that I may not have brought up?
Thanks in advance,
-Jonathan