Radmind – Unix Workstation Management
During my time at the University of Oregon I spent a large number of hours maintaining our computer lab. The Honors College maintains 8 iMacs with dual-boot 10.5 and XP along with 7 other Mac computers that run 10.4. Maintaining dual-boot lab machines has proven to be fairly time consuming. Once a term I would apply the most current patches to the OS and software, and once per year I would update our image and completely reimage the lab. This proved to be quite time consuming, but necessary due to our use of roaming profiles. For management of the 7 Macs that did not dual-boot we used Radmind. In ths case I would update the Radmind image on the server once per term to patch the OS and software and then simply reboot the client computers. The updates were then applied and all was well. In the last three terms I have only had to solve network connectivity issues with those computers. Otherwise, they consistently work as needed and maintain a steady, stable OS that is not bogged down by the roaming profiles.
Radmind is designed for Unix workstation managment and is very similar in function to Deep Freeze (a Windows workstation management software). When run, Radmind will completely examine the client file system and return the workstation to its original state. This is incredibly useful in a Lab environment. At the same time it allows you to pick and choose which files you want to be managed by Radmind, thus allowing you to enable users to have selected preferences.
Configuration of Radmind is by no means simple. In fact, it takes hours and the process is poorly documented. You first begin by installing the Radmind Server software on whatever machine is going to act as your server. The box you install it on should be able to handle (at worst case) all of your client machines connecting simultaneously. Therefore if you are only going to use it for a few computers you can use just about any computer as a Radmind server. Best practices of, course, would be to install it on server grade hardware to prevent failure of the system. The next step is to load the Radmind client software on the client machines and create the main transcript files. There are two types of transcript files – positive and negative. The positive transcript contains the information for all of the files on the client machines that should always be equivalent to what is stored on the server. The negative transcript lists the files that should never be touched by the Radmind software. For instance, you want Radmind to keep your main OS files the same regardless of what a user may do to them, but at the same time you may want the list of printers to be managed by each user. Once you have created the transcripts you then create a command file that tells which transcripts to apply and to which machines. This allows you to manage multiple computers and have different software configurations on different groups of computers. Thus if you have a computer lab that is supposed to have a certain software set and a group of office computers that are supposed to have a different software set you can set Radmind to differentiate the two groups and maintain their software accordingly. Finally you install the Radmind client side software on the client machines and connect them to the server. The machine will then be updated to the configuration pushed down by the server.
I really like Radmind. It is an effective solution to mass computer deployment, but it requires some serious tech chops. If you decide to use Radmind be prepared for some long days of frustration and suffering as you learn how it works. I have been working on a wiki posting concerning Radmind, but that is slow going and done only in my free time. If you have questions feel free to post and I will do my best to help you out.

Tim Seymour lives in Washington, DC. He works as a Systems Administrator for