Title: Linux in the Enterprise Subject: Have no idea what this is for Ideally, your system would mark the failing disk read-only, block additional writes to the disk, and start to write the data to another drive, while copying the failing disk blocks to a new drive as well. When the old disk is empty of any data files, you should be able to remove it from the file system, bring that disk offline, replace it with another disk, bring that disk online, bring it into the existing file system where the old disk was attached, and have the disk take its place along with all the other disks in that file system. The file systems should be able to handle reasonably sized disks and disk farms (in the case of Intel Linux, over 2GB in size) and should be able to have journaling, and not have to depend on fsck to check out a really large file system. An enterprise OS should be able to scale fairly well, so when you buy your next two or three CPUs and put them into your server system, you can take advantage of those CPUs' power. These features (and many others) are the ones D.H. Brown examines in its study.