Taming Your File Server

You wouldn't scatter folders and important files all over your office, so why would you disperse them throughout your file server? Similar to the way you'd organize paper files -- neatly tucked away in a folder inside a file cabinet -- your server should be arranged in such a way that you can easily get at what you need without having to call in a search party.

File servers are often the underappreciated and neglected members of most office networks; even though they're typically the most commonly shared network resource. Too often, data goes in and never comes back out again, leading one CompuMentor consultant to refer to them as "LaBrea Tarpits of Data."

Sure, it can be painful to cull and organize a heavily used server containing a mess of folders and files scattered everywhere, but it's well worth the effort. Creating a file structure and folders that make sense to users will save time and headaches down the line.

Follow the steps outlined below to replace your existing file server with a smooth and secure system so you can turn that tarpit into a quick and strong "saber-toothed tiger of data." Note, though this article uses examples and terminology from Windows, the concepts and techniques are applicable to almost all flavors of operating systems.

The first step taming your server is planning. Sit down with your users to find out how they use the file server on a day-to-day basis, and more importantly, what sorts of things they'd like to do. Use this information, along with the pointers here, to plan out your implementation and to build and maintain a reliable and user-friendly server.

Physical Drives and Partition Organization

Most file servers have multiple physical drive partitions (parts or sections of a hard drive), depending on the number and configuration of disk drives. There are many different philosophies to hard drive organization, but creating two partitions, one for the operating system ( OS ) and applications and another one for data, usually works best. Creating separate OS and data partitions will keep most hard-drive issues -- things like capacity problems, file corruption, and drive failures -- limited to one partition, making support and recovery much easier. Please note that partitioning and formatting hard drives is a destructive process, and any existing data will be lost.

The rest of this article assumes that you've got multiple partitions, with the operating system and applications on the C: partition and data on the D: partition. For more information on partitioning, check out Microsoft's article .

Mirror, Mirror

Another consideration is whether to mirror your hard drives. Also referred to as RAID 1, a basic mirrored system has two hard drives, each containing a complete copy of all the data on the server. If one drive fails, a second one is ready to go with all the data intact. Though it's not a 100-percent reliable protection against drive failures, with new 120- GB drives costing as little as $85 and mirroring software built into most modern server operating systems, setting up mirroring is relatively cheap and easy.

For an explanation of RAID 1 and a list of requirements, check out PC Guide's RAID Level 1 requirements article. This PCTechGuide's article explains This . If you're using a Mac, read Apple's piece article.

Directory Organization on the Server

Your file server should bend to meet the needs of your organization, not the other way around. Most often, file servers contain a variety of different data folders at the root (or first) level of the hard drive, such as D:\staff, D:\programs, D:\admin, etc. (see figure 1 below). At first, this may seem like the simplest arrangement, but this setup makes maintenance difficult.

Instead, set up the different folders inside a single root level folder -- such as D:\data -- so that the different data folders end up something like this: D:\data\staff, D:\data\programs, or D:\data\admin (see figure 2 below). With this type of setup, you can maintain multiple folders for your staff to use and preserve a single container on the server for your data.

What follows is one example of a relatively simple file structure. The first level contains a single directory, named "data," which contains all of the different shared folders. Within that first-level folder are two primary folders: "users" and "shared." Within the users area are individual folders for everyone in your organization.

Inside the shared folder are directories for different program areas, such as "admin," "clients," "finance," etc. (see figure 3 below). The idea with this configuration is to provide two types of shares: individual shares for each user, and shares based around the different aspects of the organization's structure, such as finance and client services.

Again, the actual layout of the directory structure should take shape from your organization's arrangement. For instance, if your organization were a puppy rescue group, you'd create a top-level shared area for veterinary clinics with subfolders containing specific information for each dog, with spaying and neutering or vaccination records. When developing your file structure, it's always easier to add shares than to remove them, so start out with a simple setup and expand to a more complicated structure as needed.

Please note that shares are almost always folders, but folders are often not shares. A share (or fileshare, or shared folder) is a network address connected to a specific folder on a server's hard drive. A folder (or directory) is a container for files on a hard drive. Folders can be nested within other folders, creating a hierarchy, often referred to as a tree. Shares can point to any folder within the hierarchy. To make matters even more complicated, shares do not have to have the same name as the folder they point to (though we recommend using the same name for both).

Share and File Permissions

Perhaps the most important part of setting up a file server is creating appropriate folder and share permissions. Permissions are the fundamental mechanism used to secure computer files. They determine who has access to what file (or folder) and what they can do with the file. Each folder should have its own set of permissions, allowing different users and groups of users to have different permissions to access and modify the files in each folder as needed.

Sally is the accountant and needs access to the payroll area, but you don't want anyone else to see that information, so the payroll folder permissions should be set so only Sally and the director can access it. But the organization's phone list is in the staff folder and that folder can be accessed and read by anyone, while only the human resources director has write permissions. The details of setting up users and groups is outside the scope of this article, but Microsoft's article this article provides information on permission setup for Windows. For Macintosh computers, read Apple's article this .

When setting up permissions, you should ask yourself: How concerned is my organization with security and privacy? You probably don't want everyone on staff rooting through the contents of your server, nor do you want a complete lockdown of your data.

The best solution is something in between: an intermediate level of permissions, with users having read and write access to their individual folders. The different shared folders should have different permissions, depending on which users and groups need access to the files contained within. For the most part, the group, not the user would determine the permission level of the shared files.

For instance, anyone from accounting would have access to financial information, while customer service reps would be the only ones with access to customer records. It's much easier to add and remove users from groups than to add and remove users from individual shares.

If possible, you shouldn't give everyone permission to create folders within the shared areas because the more folders at each level, the easier it is to misplace files. A simplified set of permissions is shown with the sample file structure from the previous step.

Once you've set up everything, be sure communicate to staff how and why the shares have been configured. If your users know where files are supposed to go and why, they're much more likely to put stuff where it belongs, keeping clutter down. As part of this process, you should let your users know just what level of privacy and permissions they can expect when working on the server.

Maintaining Your File Server

Now that the server is configured, document the whole setup, including a listing of all shares, user accounts, groups, and relevant partitions. Use this document to train your users, and update it when you make any major changes to the setup. If more than one person is responsible for maintaining the server, be sure to make it clear who is responsible for what. If you find that users are having problems finding or accessing files, help them to create desktop shortcuts to often-used folders.

On a regular basis, really clean and scour that file server. Look for out-of-place files, check the permissions on new folders to make sure they match up with the overall scheme, and consolidate multiple folders from the same projects. You should focus this effort on the top levels of the file structure, as this is where misplaced and mislabeled files and folders can cause the most problems.

Once every few months, go through the file server and archive files and folders, clearing space on the drive for new files and folders. Remove ex-employee's folders, finished projects, and other material that your organization no longer uses.

You can use the operating system's search function to look for folders and files that haven't been accessed in a specific period of time (one year, three years, or whatever seems appropriate). Collect these "orphan" files and folders and put them in a specific unshared archive folder at first. If after a month or two (or however long you deem appropriate) no one has come to you looking for the archived files, archive them to tape, CD-R, or DVD-R. Make two copies and store one off-site, in case of disaster.

From time to time, sift out "junk" files. Massive MP3 collections, picture libraries, and copies of easily downloadable free applications have bogged down a number of file servers I've maintained. Use the operating system's search function to find these files and remove them, if appropriate. You can also provide a separate share or even a drive just for these "junk" files.

Back it Up

Sooner or later, a hard drive will fail, a staff member will delete an important spreadsheet, or a database will get corrupted. Running a daily or weekly backup of any file server is a must. To read more about backing up, check out this TechSoup's article Backing Up Your Data . Once you've configured a backup system, test it to make sure you can restore the data from it.

This process may seem like a daunting undertaking, but you can do it gradually over an extended period, if need be. Once you've organized, streamlined, backed up, and configured, the lean saber-toothed tiger hiding in the LaBrea Tarpit of lost files can finally be free.