Mac OS X,  Mac OS X Server,  Mass Deployment

Imaging Maturity for Mac OS X environments

Seems like most companies start out manually installing their computers.  This works for awhile but then they run into a situation where they’re manually installing too many and realize that for the most part they’re all pretty much the same.  That’s when products like Carbon Copy Cloner and Ghost (for those pesky Windows workstations) come into play.  Suddenly they’re saving time by mounting up an image and restoring it.  Then there is a collection of post-flight tasks they perform, like setting up user environments, settings and directory services bindings.  Then, there are just too many for this, so people start doing network based imaging.  Something like NetInstall, NetRestore, ASR or Ghost.  Some do unicast and then move into multicast but this seems to be the next step.  At this step people are typically still doing monolithic imaging, or one image as one big file.  

The next step is where things start to get really interesting.  And NetInstall, NetRestore and Ghost Enterprise still have options for it.  But in some cases, not a lot.  This is where specialty options come into play.  This is Jamf software’s Casper Suite, LANrev, etc.  At this point, people are typically building a package based set of installer files and distributing them along with a base image, which is basically just a bare metal image.  At this point many environments start to save a lot of time by doing what, in essence, is object-oriented imaging.

Once people start building out all these packages, some start to realize that they can then start to automate more and more.  Suddenly you’re writing shell scripts for everything, including getting you coffee and donuts and even putting down small revolts in the Balkans.  But there’s still one more step.  And this is where things get really interesting.  Regression testing.  The software isn’t cheap, but in a Mac environment you can leverage Eggplant from Redstone Software to do it.  Eggplant uses VNC to perform image recognition and uses those images to process scripts.  Eggplant takes a few days to get used to, but once you do you will start finding that if you want to open Office and verify there are no errors on 15 different image sets that you can do so easily and effectively without manually going through your images.