Linux Server Hacks, Volume Two

Hack #54. Make Files Easier to Find with Extended Attributes

Define file- and directory-specific metadata to make it easier to find critical data.

Most projects and users organize their files by taking advantage of the inherently hierarchical nature of the Linux filesystem. Conceptually related items are given meaningful names and stored in hierarchical directories with equally memorable or evocative names. But unfortunately, file and directory names and structures that were memorable at the time of creation are not always equally memorable a month or two later, when you're desperately looking for a specific file.

Extended attributes are name/value pairs that you can associate with any file or directory in a Linux filesystem. These are a special type of metadata, which is the term for data about data, such as modification and access times, user and group ownership, protections, and so on. Extended attributes can be associated with any object in a Linux filesystem that has an inode. The names of extended attributes can be up to 256 bytes long, are usually standard ASCII text, and (like standard Linux strings) are terminated by the first NULL byte. The value of an attribute can be up to 64 KB of arbitrary data in any format.

Extended attributes are often used for system purposes, by tagging files with metadata about who can access them, and under what circumstances. This hack discusses how extended attributes can be extremely useful for any user or system administrator who wants to ...