It’s really the “unicorn” of mac imaging in my opinion. I’ve used DeployStudio for years and been really happy with it but have never been able to get multicast working until today. I could have sworn I did these exact steps in the past attempts, but I was never able to get it fully operational. Here is what I did. I’m not walking through the standard setup. I’m assuming you already have that going.
1. Setup DeployStudio Admin and the various Netboot, DHCP and NFS services on your OS X Server. I am going to document how to do it on an OS X client soon when I give that shot, but I have a few OS X 10.6 servers now that I have run DS on for years, so going with that for the time being.
2. Create a new workflow in DS admin to master the image you want to multicast restore to the Macs. Here is a screenshot of my create workflow.
3. Boot your golden mac computer into DeployStudio and run this workflow on your golden master drive. It will generate a big image. This was 100+ GB for me in this test today. DeployStudio will grind away a bit. I let the imaging run overnight and dealt with it this AM.
4. Create a new workflow in DS admin for restore function. Here is mine. This step is critical. You cannot run a multicast restore properly if you don’t have a specific workflow for it. This is not something you do can with the settings you get on the fly when you netboot your Mac. The workflow is a two-step: one for partitioning and one for restoring.
I set mine to my large master image and had the multicast trigger at 4 clients. You should point the restore step to your image created in step 3.
5. Flip-on the multicast on the master via DS Admin by going to the “Masters” area in DS admin and clicking the airport-looking radar icon.
NOTE: This does not seem to be necessary anymore with DS v1.0rc135 but if you are running a version prior to that, do it.
This puts the muticast as an active process in DS
6. The DS server is now ready to dish the large, reliable master image if you use the workflow. Boot your macs on the network and use the multicast restore workflow.
My testing to confirm I was getting multicase going was to take 4 iMacs and run them quickly as separate streams from the DeployStudio server and get a network activity graph from the OS X server. Then, reboot the restore and start it again but via the multicast workflow. You can see here in the graph that it went from getting close to 160MB on the 4 separate streams before or around 8am and then solid 40MB-ish later when I was running the multicast workflow.