Linux Filesystem Structure Advance Draft Daniel Quinlan, quinlan@bucknell.edu Filesystem Standard Group ABSTRACT This document is an extensive undertaking to correct outstanding problems with the Linux filesystem structures in use by developers, programmers, administrators, and users. Our purpose and goal is to describe a filesystem structure of exceptional quality that developers and others will voluntarily adopt to solve well-acknowledged problems. The FSSTND group hopes that this draft will be eventually adopted as a better standard than the de-facto standard produced by the current disarray of ideas. 1. Foreword 1.1. Status of this Draft This draft is being distributed to members of the Linux community in order to solicit their reactions to the series of ideas, concepts, and proposals included within it. While the entire content of this draft may not conform to what every individual desires, it should prove to be a good start to solving many problems. This draft is a working document of the Filesystem Standard (FSSTND) mailing list, the author, and many others who have worked together to make this draft possible. 1.2. Legalese Copyright (C) 1994 Daniel Quinlan (quinlan@bucknell.edu) Permission is granted to copy and distribute verbatim copies of this document provided the copyright and this permission notice are preserved on all copies. - 1 - Linux Filesystem Structure January 15, 1994 Permission is granted to copy and distribute modified versions of this document under the conditions for verbatim copying, provided that no text from this document is removed from the derived work, that modifications are clearly marked, and that the entire resulting derived work is distributed under terms of a permission notice identical to this one. UNIX is a Registered Trademark of X/Open Co. Ltd. All other copyrights are owned by their owners, unless specifically noted otherwise. Use of a term in this document should not be regarded as affecting the validity of any trademark or service mark. - 2 - Linux Filesystem Structure January 15, 1994 2. Introduction We originally felt that it was desirable to first call attention to some of the fundamental problems with the current filesystem situation: o There was no single well-accepted Linux directory structure. Instead, there are many different ones, each being incompatible with each other. This is a problem that justifies our effort and should overshadow any differences in opinion, the same differences in opinion that make Linux filesystems an utter mess. o In the most widely used filesystem hierarchies, the directories were not well structured and differ gratuitously from more modern standards (such as POSIX, System V, BSD, and others). o The filesystem was disturbing to experienced UNIX** users and administrators who have experience on more mainstream UNIX systems. o The current layout was confusing for Linux newcomers, especially those coming from a non-UNIX background. o Any incompatibilities between primary installation packages and other software packages are typically solved by methods of a less than appealing nature. o Overall, symbolic links are used too often within the filesystem to fix problems. (However, there are times when symbolic links need to be used to ensure backward compatibility or to allow specific systems to have an individual filesystem structure.) The FSSTND group seeks to correct these problems by proposing a good filesystem structure that the Linux community may voluntarily follow. While developing this draft, approval and input was received from a number of Linux developers, noted Linux programmers, many system administrators, and both experienced and novice users. For this reason, we think that following our recommendations is a good thing. If you feel that there is a problem with this effort or the substance of the draft, feel free to first contact the draft coordinator, Daniel Quinlan , with your comments. 2.1. Specific Problems Naturally, while defining a Linux filesystem structure, there were some specific problems that we sought to correct. Here are some of the major and well-accepted ones: _________________________ ** UNIX is a Registered Trademark of X/Open Co. Ltd. - 3 - Linux Filesystem Structure January 15, 1994 o The primary binary directories, /bin and /usr/bin, do not have well defined divisions between them. As a result, the binaries that are in found in each directory greatly differ between various Linux distributions. o Including both binaries and configuration files in /etc makes it more confusing and harder to maintain for inexperienced users or system administrators with especially large systems. o The divisions between /usr/etc and /etc as a location for configuration is muddled at best. This is because the boundary between what is a site-wide configuration file and what is a machine-local configuration file is difficult to establish. o The current implementation of /usr cannot be mounted read-only because it contains variable files and directories that need to be written to. o In a networked environment it is desirable to serve software to workstations via a NFS mounted filesystem. Such filesystems need to be mounted read-only so that accidents or malice on one workstation cannot damage the files on the server. This requires identification and separation of files that a machine must write to and separation of files that are specific to a single machine. o Linux is not well prepared for a network installation including the possibility of a read-only /usr partition and diskless (or small local disk) workstations. While these are some of the major problems we addressed, there were numerous additional problems that needed to be solved. This draft attempts to address many of those other problems, but there may be something we missed. If you wish to bring something to our attention, please note there are some things we have discussed at length, but did not include in this draft (for good reasons). 2.2. Objectives In trying to solve the above problems, we saw several objectives that needed to be accomplished in addition to the more technical matters. These goals comprise the correction of outstanding problems as well as the validation of our discussion and work. o Solve the above problems while also limiting the possible transition difficulties resulting from moving away from the former de-facto standards. o Gain approval of distributors, developers, and other important people in the Linux movement, as well as their suggestions. - 4 - Linux Filesystem Structure January 15, 1994 o Provide a standard that all of the Linux community would choose to follow because it will solve the above problems as well as provide the most sensible structure for Linux's filesystem. o While conforming to this or any other standard in Linux is obviously completely voluntary, we wanted to impress upon developers that this organization is a very sensible way to lay out a Linux filesystem. If you, as a developer, wish to suggest any improvements, we are quite willing to listen. 2.3. History and Progress The original post that motivated this effort to restructure the Linux filesystem was written by Olaf Kirsh on August 2, 1993 to the NORMAL channel of the Linux activists mailing list. Soon thereafter, it was decided that the best way to accomplish such a restructuring of the Linux filesystem would be to create a mailing list for the purpose of trying to develop a consensus standard. After a comprehensive discussion, with surprisingly few flames, a preliminary draft was written. Then, with the help of several dedicated people, the draft was finished and that resulting draft submitted to the FSSTND channel for more discussion. The first draft was submitted to the channel on September 18, 1993 by Daniel Quinlan. As the discussion continued and this draft of FSSTND recommendations was developed further, contacts were established with accessible Linux distributors who then offered their input and support to our effort. Many Linux developers already agree that what we are doing is a good thing. These are some of the developers who are trying to follow the recommendations of the FSSTND draft [alphabetical order]: o Debian Linux Ian A. Murdock o Linux/PRO Fred N. van Kempen o Slackware Linux Patrick J. Volkerding o TAMU Linux Dave Safford o util-linux packages Rik Faith - 5 - Linux Filesystem Structure January 15, 1994 The draft, although much more stable and comprehensive than it once was, will probably never be truly finished -- the needs of the Linux community will continually change in relation to emerging technology. It is also possible that better solutions to filesystem problems could be discovered. We are not above making mistakes nor are we unwilling to admit them. Because of these reasons, the FSSTND group plans to release supplementary drafts in addition to periodic updates to the main draft. -- Daniel Quinlan [93/12/21] - 6 - Linux Filesystem Structure January 15, 1994 3. The Filesystem The UNIX filesystem is characterized by: o a hierarchical structure o consistent treatment of file data o protection of file data Our recommendations on the Linux filesystem follow the same basic principles as the UNIX filesystem. However, we did not try to conform in every possible respect to the UNIX (SVR4) filesystem structure. This is after careful consideration of other factors, including: o common (good) practices in the Linux community o the existence of other filesystem structures (such as BSD) o applicable standards (such as POSIX) The directory hierarchy is separated into unsharable data, "/" (root), and sharable data, "/usr". These two filesystems are divided according to the following two data types: static data and variable data. Static data includes binaries, libraries, documentation, and anything that does not change without system administrator intervention. Variable data is anything that does change without the system administrator's intervention. (In this document, the term "filesystem" is used to refer to either a single formatted partition for data, or to refer to the entire directory hierarchy.) We have defined 2 orthogonal categories of file data: sharable- unsharable and variable-static. We have defined /usr as sharable data and / as unsharable data. Each hierarchy, / and /usr, divides data into static and variable types. Throughout this document, and in any well planned filesystem, the acceptance of these facts will help guide the structure and lend it additional consistency. The distinction between sharable and unsharable data is needed for several reasons: o In a networked environment, certain filesystems contain information specific to a single machine. Therefore these filesystems cannot be shared (with NFS). o The current implementation of /usr cannot be mounted read-only because it contains variable files and directories that need to be written to. This is a factor that must be addressed when /usr is - 7 - Linux Filesystem Structure January 15, 1994 shared on a network or mounted read-only because of other considerations (safety). The "sharable" factor can be extended in two directions: o A /usr mounted (read-only) through the network (using NFS). o A /usr mounted from read-only media. (You can think of that CD- ROM drive as a networked, using postal mail, filesystem that you are sharing with other Linux systems.) The "static" vs. "variable" factor dramatically affects the filesystem in 2 major ways: o Since / contains both variable and static data, it needs to be mounted read-write o Since /usr contains both variable and static data, and since we want to mount it read-only (see above), it is necessary to provide a method to have /usr mounted read-only. This is done through the creation of a /var hierarchy which is mounted read-write (or is a part of another read-write partition, such as /), taking over much of the /usr partition's traditional functionality. - 8 - Linux Filesystem Structure January 15, 1994 4. The root directory This is the root directory structure. In general, enough data should be contained in the root partition to boot, restore, recover, and/or repair the system: (1) To boot a system, enough must be present to mount /usr. This includes utilities, configuration, boot loader information, and other essential start-up data. (2) To recover and/or repair a system, those utilities needed by an experienced user to diagnose and reconstruct a damaged system should be present on root. (3) To restore a system, those utilities needed to restore a system from backups (on floppy, tape, etc.) should be present on root. The primary concern used to balance these desires (placing many things in root) is the goal of keeping root as small as reasonably possible. It is desirable to keep root small in terms of number of directories, files, and disk space for several reasons: (1) The root is often mounted from very small media. For example, most people using Linux install and do recovery by mounting root off of a RAM disk which is copied from a single 1.44M or 1.2M floppy disk. (2) Root has many system-specific configuration files in it, a kernel that is specific to the system, a different hostname, etc. This means that root isn't always sharable between networked systems. Keeping root small on networked systems minimizes the amount of space lost on servers to unsharable files. It also allows workstations with smaller local hard drives. However, with diskless clients, this does not have to be entirely the case, unless each client has a different root image. (3) While you may have a large root partition, and may be able to fill it to your heart's content, there will be people with smaller partitions. If you have more files installed, you may find incompatibilities with other systems using limited root partitions. If you are a developer then you may be sharing this problem with a large number of users. (4) Disk errors on the root partition are a greater problem than errors on any other partition. A small root partition is less prone to corruption as the result of a system crash. Since root is small and host-specific (due to the division between / and /usr), this scheme necessitates a writeable root. However, this does not necessitate a fully locally stored root. The root partition doesn't - 9 - Linux Filesystem Structure January 15, 1994 have to be locally stored just to be system specific (i.e., root mounted from a NFS root server.) Software should never create or require special files or subdirectories in the root directory. This structure provides more than enough flexibility for any package. Any package which does occupy a directory under root suffers from sheer arrogance. / : the root directory | |- bin : essential command binaries |- boot : static files of the boot loader |- dev : device files |- etc : machine-local system configuration |- home : user home directories |- lib : shared libraries (libc.so.*, libm.so.*, and ld.so) |- mnt : mount point of temporary partitions |- proc : process information pseudo-filesystem |- root : home directory for root |- sbin : essential system binaries |- tmp : temporary files |- usr : second major permanent mount point |- var : files that tend to grow or vary in size +- {kernel image} Following this section, each directory is explained in full. /usr and /var are discussed in separate sections of this draft. The root directory normally contains the current kernel image. The kernel image name is locally configurable, but the name we suggest (that has been used in recent Linux kernel sources) is `vmlinux' which may or may not be a (symbolic-)link to the actual file, possibly depending on the system distribution used. More information on kernel placement is located in the /boot section of the draft. 4.1. /bin : essential command binaries (for use by all users) There should be no subdirectories within /bin. The commands (static data) that are needed in single user mode by the super-user (root) are stored in /bin. However, the commands in /bin are for use by both root and other users. On the same note, the /bin directory should not contain anything that is to be used only by root for system administration. All root-only binaries such as standard daemons, `init', `getty', `mkfs', et al. (previously found in /etc), shall now be placed in /sbin or /usr/sbin depending on the necessity of the command. For discussion and our definition of essential (necessity and related concepts) please read the issues and rationale section towards the end of this draft. - 10 - Linux Filesystem Structure January 15, 1994 Command binaries that are not essential enough to place into /bin should be placed into /usr/bin, instead. Items which are only used by non-root users are not essential (interactive shells, pagers, `passwd', et al.) and should be placed elsewhere. REQUIRED files for /bin: o general commands: The following commands have been added because they are essential. A few have been added because of their traditional placement in /bin. { arch, cat, chgrp, chmod, chown, cp, date, dd, df, echo, ed, false, free, kill, ln, loadkeys, login, ls, mkdir, mknod, mv, ps, pwd, rm, rmdir, setserial, sh, stty, su, sync, true, uname } If /bin/sh is BASH, then /bin/sh should be a symbolic or hard link to /bin/bash since bash behaves differently when called as `sh' or `bash'. The same goes for `pdksh', which is often the /bin/sh on install disks and such (link from /bin/sh to /bin/ksh). I personally prefer the advantage of using a symbolic link in these cases, because it allows users to easily see that /bin/sh is not a true Bourne shell. Both `[' and `test' are built into BASH, pdksh, zsh, recent Korn shells, ... essentially every Bourne shell replacement there is for Linux. They should be placed into /usr/bin. (That is, they must be included with any Linux system, attempting to comply with the POSIX.2 standard.) /bin/arch should produce the same output as "uname -m", specifically "i386" or "i486". o restoration commands: These commands have been added to make restoration of a system possible (provided that / is intact). { tar, gzip, gunzip (link to gzip), zcat (link to gzip) } If system backups are made with a package other than `gzip' and `tar', then that administrator should include the minimal necessary restoration components in the root partition. For instance, many systems should include `cpio' as it is the next most commonly used backup utility after tar. Conversely, if no restoration from the root partition is ever expected, then these binaries may be omitted (e.g., a ROM chip root, mounting /usr through NFS). If restoration of a system is planned through FTP, then ftp or tftp (along with everything necessary to get an ftp connection) should be available on the root partition. - 11 - Linux Filesystem Structure January 15, 1994 o networking commands: These are deemed the only necessary networking binaries that both root and users will want or need to execute other than the ones in /usr/bin or /usr/local/bin. { domainname (link to hostname), hostname, netstat, ping } 4.2. /boot : static files of the boot loader This directory contains everything for boot except configuration files and the map installer. This includes saved master boot sectors, sector map files, and anything else that is not directly edited by hand. The boot loader program should be placed into /sbin and configuration files for boot loaders into /etc. For LILO: Old location New location ------------------------ ----------------- /etc/lilo/config.defines /etc/lilo.defines /etc/lilo/config /etc/lilo.conf /etc/lilo/disktab /etc/disktab /etc/lilo/lilo /sbin/lilo /etc/lilo/boot.NNNN /boot/boot.NNNN /etc/lilo/part.NNNN /boot/part.NNNN /etc/lilo/map /boot/map /etc/lilo/*.b /boot/*.b *.b are the first and second stage boot loader, plus all those chain loaders. `QuickInst' (if it is included) should be placed into /usr/sbin. (The `activate' command is left out of this scheme because its future is uncertain at this time.) Extra kernel images may be stored in /boot. The main kernel can either be placed in / or in /boot according to preference of the administrator. If placed in /, the kernel may also possibly be a symlink to a kernel image in /boot. Note that the standard location for the kernel is still in /. 4.3. /dev : Device files /dev usually also contains a file, MAKEDEV, a shell script designed to create devices as needed. It also often contains a MAKEDEV.local for any local-only devices. Symbolic links in /dev should not be distributed with Linux systems, as the local setup will often differ from that on the distributor's development machine. Also, if a distribution install script configures - 12 - Linux Filesystem Structure January 15, 1994 the symlinks at install time, these symlinks will often not get updated if local changes are made in hardware. When used responsibly at a local level, however, they can be put to good use. This standard incorporates the the Linux Device List which is maintained by Rick Miller , the Linux Device Registrar. This is the current standard that we strongly recommend using. It is available by anonymous ftp at sunsite.unc.edu as /pub/Linux/docs/hardware/DEVICES. 4.4. /etc : Machine-local system configuration No binaries should go directly into /etc. Binaries which would have in the past been found in /etc should now be placed in /sbin. This includes such files as init, getty, and update. Binaries such as hostname which are used by users as well as root should not be placed in /sbin, but in /bin. REQUIRED files for /etc: o general files: These files are needed on most Linux systems. { adjtime, csh.login, fdprm, fstab, gettydefs, group, inittab, issue, ld.so.conf, magic, motd, mtab, mtools, passwd, printcap, profile, securetty, shells, termcap, ttytype, utmp } The `securetty' file is not truly required because the "getty_ps" package uses a list in it's `login.defs' by default. o networking files: These are only needed if networking is installed. { ftpusers, hosts, host.conf, hosts.equiv, networks, protocols, resolv.conf, services } There are two models for setup of the "rc" command scripts which are invoked by init(8) at boot time, the /etc/rc.* (BSD model) and the /etc/rc.d/* (System V model). Either model is acceptable at this time. This is a local issue and the implementation that is used should be determined by the system administrator or the developer. The `wtmp' log file belongs in /var/adm because it can grow in size without bound. Systems which choose to use the shadow password suite will have additional files in /etc (/etc/shadow and whatever else) and /usr/sbin (useradd, usermod, et cetera). - 13 - Linux Filesystem Structure January 15, 1994 If you use getty_ps, there might also be a couple of other files: login.defs, default/getty.ttyS*, etc. [I am not especially thrilled about the "default" subdirectory of /etc, but that is another matter -djq] 4.5. /home : User home directories /home is a fairly standard concept, but it is clearly a site-specific filesystem. The setup will differ from machine to machine. This is only a suggested placement for user home directories, but we recommend that all Linux distributions utilize this as the default home directory. On small systems, each user's directory is typically one of the many subdirectories of /home such as /home/smith, /home/linus, /home/operator, etc. On large systems (especially when the /home directories are mounted across a number of machines using NFS) it is a good idea to subdivide user home directories. Subdivision can be accomplished by using subdirectories such as /home/staff, /home/guests, /home/students, etc. Different people prefer to place user accounts in a variety of places and because of this reason, no programming should rely on this location. If you want to find out a user's home directory, you should use the field in /etc/passwd or another reliable method (if a program is being run by a user, the $HOME variable is a fairly reliable method). 4.6. /lib : Shared libraries (needed to run dynamically linked binaries) Only the shared library images necessary to boot the system should be placed in /lib. The shared library images are "/lib/libc.so.*", "/lib/libm.so.*", and "/lib/ld.so" (and not the actual ".a" files). XFree86 and other libraries do not belong in /lib. Essentially, only the dynamic shared libraries needed to run programs in /bin and /sbin should be here. A single symbolic link for the C preprocessor currently exists in /lib pointing /lib/cpp to /usr/bin/cpp which should be either be the actual `cpp' binary or a link to /usr/lib/gcc-lib/i?86-*-linux/2.5.?/cpp. No commands should be added to /lib in addition to `cpp'. /lib/cpp is where XFree86 and some other packages expect to find the C preprocessor. 4.7. /mnt : Mount point for temporarily mounted filesystems This is the location where the system administrator may temporarily mount filesystems as needed. The setup of this directory is a local issue and should not affect the manner in which any program is run. - 14 - Linux Filesystem Structure January 15, 1994 This directory should probably not be used by installation programs due to the site-specific nature of it. A temporary directory not in use by the system should be used instead. 4.8. /proc : Proc based process system The procps filesystem is becoming the standard Linux method for handling process information rather than /dev/kmem and other nasty methods. This is only recommended, but may in time become the standard for the storage and retrieval of process information as well as other kernel and memory information. 4.9. /root: home directory for root / is traditionally the home directory of the root account, however on most Linux systems this is found in /root. One thing that is certain is that the root account's home directory must be stored on the root partition. It is preferred practice not to use the root account for mundane things such as mail and news, but solely for systems administration. For this reason, subdirectories such as "Mail" and "News" should not appear in the root account's home directory. Mail is usually forwarded to a more appropriate account. 4.10. /sbin : System binaries (binaries once kept in /etc) Utilities used for system administration (and other root-only commands) are stored in /sbin, /usr/sbin, and /usr/local/sbin. /sbin typically contains files essential for the booting phase of starting the system up. Anything executed after /usr is known to be mounted (when there are no problems) should be placed into /usr/sbin. Local-only system administration stuff should now be placed into /usr/local/sbin. The concept of what goes into "sbin" directories is simple. If a user will need to run it, then it should go somewhere else. If it will only be run by root (i.e., system administration commands, networking daemons, system startup), then it should go in /sbin (or in /usr/sbin if the item is not essential to system operation). Files such as `chfn' and `ac' which users only occasionally use should still be placed in /usr/bin. `ping', although it is absolutely necessary for root (network recovery and diagnosis) is often used by users and should live in /bin for that reason. If there is any chance that a user should need to run it, do not put it here! Users should never have to place /sbin (or any of the "sbin" directories) in their path. It is true that they should probably not even be able to execute anything dangerous in /sbin if the job has been - 15 - Linux Filesystem Structure January 15, 1994 done correctly. It is reasonable to let them see what files are in /sbin. Therefore, don't make the directory totally unreadable unless you must. /sbin was not created to protect users or to prevent them from seeing the OS, but to provide a good division between binaries that everyone uses and the commands that are to be used only by administrators. REQUIRED files for /sbin: o general commands: { clock, getty, init, update, mkswap, swapon, swapoff } o shutdown commands: { halt, reboot, shutdown } o filesystem commands: { fdisk, fsck, fsck.*, tunefs (Ext2 only), mkfs, mkfs.*, mount, umount } "*" = (ext, ext2, minix, msdos, xiafs) o LILO commands: { lilo } o networking commands: { arp, ifconfig, ifsetup, route } OPTIONAL files for /sbin: o static binaries: Static ln and static sync are useful when things go wrong. The primary use of sln (to repair hosed symlinks in /lib after a poorly orchestrated upgrade) is no longer a major factor now that the `ldconfig' program exists and can act as a guiding hand in upgrading the dynamic libraries. Static sync is useful in some emergency situations. Note that these are not just statically compiled versions of `ln' and `sync'. The `ldconfig' binary is optional for /sbin as well. Basically, if you use it properly when upgrading the shared libraries in /lib, you won't need it in root. { ldconfig, sln, ssync } - 16 - Linux Filesystem Structure January 15, 1994 4.11. /tmp : temporary files /tmp is used for temporary files, preferably on a fast device (a memory based filesystem, for instance). The "persistence" of the data that is stored in /tmp is different than that which is stored in /usr/tmp. /tmp is usually cleaned out at boot time (or at relatively frequent intervals). Because of this, data stored in /tmp should not be expected to remain for any long period of time on the system... it is frequently deleted. Programs should use /tmp or /usr/tmp (which is usually symlinked to /var/tmp) according to the expected persistence of the data, but should not rely on any particular persistence for any tmp directories The precise arrangement of /tmp and /usr/tmp is a local issue. If there are distinct /usr/tmp and /tmp directories, then the persistence of /usr/tmp files should be at least as long as for /tmp, but beyond that, a site administrator or distributor can arrange these any way they want, over RAMdisk, symlinks or whatever. - 17 - Linux Filesystem Structure January 15, 1994 5. The /usr directory /usr is the second major division of the filesystem. /usr is sharable data. That means that /usr should be sharable between various machines running Linux. Because it is sharable between machines, any information that is machine-local must be stored elsewhere, hence /var enters the picture. /usr : Second major mount point (permanent) | |- X386 : X Window system on x86 platforms |- bin : Most user commands |- dict : Word lists |- doc : Miscellaneous documentation |- etc : Site-wide system configuration |- g++-include : GNU C++ include files |- games : Games and educational binaries |- include : Header files included by C programs |- info : GNU Info system's primary directory |- lib : Libraries |- local : Local directory (empty after main installation) |- man : Online manuals |- sbin : Non-vital system administration binaries |- share : Architecture-independent data |- src : Source code +- tmp : Temporary files, used to keep /tmp small. The following list of directory symbolic links need to be added. This only needs to be done until compatibility with the /var scheme can be assumed to exist. /usr/adm -> /var/adm /usr/preserve -> /var/preserve /usr/spool -> /var/spool /usr/tmp -> /var/tmp /var/spool/locks -> /var/lock Most of the above symlinks should in time become unneeded as packages are changed to support /var in addition to /usr. The exception, the directory that should be referenced in /usr rather than /var is /usr/tmp because some systems link it to different directories within /var. The GNU Emacs lock file directory, if Emacs is installed, should be a symlink pointing to /var/lock/emacs if you want to be able to mount /usr read-only. It is usually found in /usr/emacs, /usr/lib/emacs, or /usr/local/lib/emacs (preferably not the first). 5.1. /usr/X386 : X Window system on x86 platforms This hierarchy is reserved for the use of XFree86 X11 releases. - 18 - Linux Filesystem Structure January 15, 1994 /usr/X386 : XFree86 installation directory |- bin |- doc |- include |- lib +- man To simplify matters and make X386 more compatible with other X11 packages, our recommendation is to place the following symbolic links on your system: /usr/bin/X11 -> /usr/X386/bin /usr/lib/X11 -> /usr/X386/lib /usr/include/X11 -> /usr/X386/include/X11 5.2. /usr/bin : Most user commands /usr/bin : Binaries which are not needed in single-user mode. |- mh : commands for the MH mail handling system +- X11 : symlink to /usr/X386/bin This is the primary directory of executable commands on the system. 5.3. /usr/dict : Word lists REQUIRED files for /usr/dict: { words } Traditionally this directory contains only the English `words' file, which is used by look(1) and various spelling programs. `words' may use either American or British spelling. Sites which need both around may link `words' to /usr/dict/american-english or /usr/dict/british-english. Another possible solution goes something like this: $ cat american-english british-english | sort | uniq > words Word lists for other languages may be added using the English name for that language, e.g., /usr/dict/french, /usr/dict/danish, etc. These should use the ISO 8859 character set appropriate for the language in question; if possible the Latin1 (ISO 8859-1) character set should be used. The rationale behind having only word lists here is that they are the only files used by all spell checkers. For example, ispell(1) uses a complicated format for its "hashed dictionaries" that is only useful to ispell and should thus go in /usr/lib/ispell. - 19 - Linux Filesystem Structure January 15, 1994 5.4. /usr/etc : site-wide system configuration Storing configuration in /usr/etc for the software found in /usr/bin and /usr/sbin is a problem. It makes the read-only mounting of /usr through CD-ROM or NFS delivery very difficult at best. This is true for many of the same reasons that /usr/tmp or /usr/spool/cron cause problems. One possible solution that we considered was to completely eliminate /usr/etc and specify that all configuration be stored in /etc. A problem with this approach is that it does not properly anticipate the possibility that many sites may want to have some configuration files which are not machine-local. We eventually decided that /etc should be the only directory which is actually referenced by programs (that is, everything should look for configuration in /etc and not in /usr/etc). Any configuration files which need to be site-wide and are not needed either before /usr is mounted (or in an emergency situation) should then be placed in /usr/etc. Then, specific files (in /etc) on specific machines may or may not be symbolically linked to appropriate configuration files located in /usr/etc. This also means that /usr/etc is technically an optional directory in the strictest sense, but we still expect all Linux systems to incorporate it. It is not recommended that symbolic links exist in /usr/etc that point to files in /etc as is unnecessary and would interfere with local control over machines which share a /usr directory. 5.5. /usr/include: Directory for standard #include files. This is where all of the system's general-use include files should be placed. - 20 - Linux Filesystem Structure January 15, 1994 /usr/include: include files. |- X11 : Symlink to /usr/X386/include/X11 |- arpa : ARPAnet defined protocol definitions |- asm : Symlink to /usr/src/linux/include/asm |- bsd : BSD compatibility include files |- gnu : GNU include files |- linux : Symlink to /usr/src/linux/include/linux |- net : Generic network-related definitions |- netax25 : +AX25 (ARRL AX.25) specific definition |- netinet : TCP/IP specific networking stuff |- netipx : +IPX (Novell IPX/SPX) specific definitions |- protocols : Protocol definitions (mostly INET-based) |- readline : The GNU readline library |- rpc : Sun Microsystems RCP definitions |- rpcsvc : Sun Microsystems RCP service definitions +- sys : System generation include files 5.6. /usr/lib : Libraries for programming and packages /usr/lib : Libraries for programming and packages |- X11 : Symbolic link to /usr/X386/lib |- emacs : Support files for the GNU Emacs editor |- groff : Libraries/directories for GNU groff |- gcc-lib : System specific files/directories for `gcc' |- mh : Libraries for the MH mail handling system |- mf : Meta-Font data |- news : Cnews/INN |- smail : Smail |- terminfo : Directories for terminfo database |- tex : TeX (and LaTeX) data libraries |- uucp : Commands for uucp +- zoneinfo : Timezone information and configuration The directory /usr/lib includes static data files and some internal binaries like `sendmail'. `smail', if used, should be stored in /usr/bin because both normal users and administrators will often want to use it from the command line (/usr/lib/sendmail should also be linked to /usr/bin/smail, of course). 5.7. /usr/local : Local directory - 21 - Linux Filesystem Structure January 15, 1994 /usr/local : Local directory |- bin : Local only binaries |- doc : Local Documentation |- etc : Configuration for local only binaries |- games : Locally installed games |- lib : Libraries for /usr/local |- info : Local info pages |- man : Man page hierarchy for /usr/local |- sbin : Local only system administration +- src : Local source code The /usr/local directory is used by the system administrator to install software that needs to be safe from overwriting when the system software is updated. It is used to store anything that is sharable among a group of machines, but not found in /usr. This directory should always be empty after first installing Linux, no exceptions to this rule should be made other than the listed directory stubs. Locally installed software should be placed within /usr/local rather than /usr unless it is being installed to replace or upgrade software in /usr or it is felt that the installed software is "important enough" to place in /usr or in /. Note that software placed in / or /usr will often be overwritten with complete upgrades (unless proper backups are made) therefore local software should not be placed in /usr/local without good reason. 5.8. /usr/man : Manual pages /usr/man : Manual page hierarchy |- man1 : User programs |- man2 : System calls |- man3 : Library functions and subroutines |- man4 : Devices |- man5 : File formats |- man6 : Games |- man7 : Miscellaneous |- man8 : System Administration +- man9 : Kernel internal variables and functions The cat page sections (cat[1-9]) containing formatted manual page entries are also found within subdirectories of /usr/man, but are not required nor should they be distributed in lieu of nroff source manual pages. Local and X Windows manual pages (if present) should be stored in /usr/local/man and /usr/X386/man, respectively. These directories have a similar structure to /usr/man (man[1-8], cat[1-8], empty subdirectories being omitted). - 22 - Linux Filesystem Structure January 15, 1994 The MH mail handling system manual pages should have "mh" affixed to all manual page filenames. X11 manual pages usually have "x" affixed to all X11 manual pages. As Linux (and UNIX) is further utilized in non-English speaking countries and manual pages are translated to non-English versions, there is the impending problem that these manual pages will have to be stored somewhere else. Some German releases of Linux have already created a manual page system that is placed in /usr/man with the suffix "g". This is a poor solution and will cause further problems in the long run as other languages appear, especially other languages starting with the same letter (Greek, Gaelic, whatever). Therefore, all non-English manual pages sections should be stored in subdirectories within /usr/man named according to the language that the the contained manual pages are written in (lowercase characters), hence, for the German manual pages: /usr/man/german/man[1-9] and possibly /usr/man/german/cat[1-9] Then, German-speaking Linux users can add /usr/man/german to their MANPATH before /usr/man so that /usr/man/german manual pages are referenced first. If a German manual page is not found for a given command then the English version may be referenced. This setup will be needed as the number of foreign (non-English) manual pages increases. German is the language mentioned here since it is the only non-English manual page system distributed with any Linux system at this time. Other languages will probably follow and they should follow this scheme as well. We only use German as our example because it was the first non-English manual system to be completed. The practice of placing non-English in subdirectories of /usr/man should be followed as well for other manual page hierarchies, such as /usr/local/man and /usr/X11/man. Note: Using the language itself (/usr/man/deutsch) rather than the English (/usr/man/german) was considered, but this was met with disapproval from many people, including those who do not speak English as a first language. The reasons include: simplicity, the difficulty in displaying many languages' names in ASCII characters, and the fact that everyone should be able to recognize their language name in English. A description of each section follows: o man1: User programs Manual pages that describe publicly accessible commands are contained in this chapter. Most program documentation that a user will need to use is located here. o man2: System calls This section describes all of the system calls which are entries - 23 - Linux Filesystem Structure January 15, 1994 to the Linux kernel (operating system). This section can be very useful to programmers, but users have little need of the items in section 2. o man3: Library functions and subroutines Section 3 describes user-level library routines. This is another chapter that is only really of interest to programmers. o man4: Special files Section 4 describes the special files, related driver functions, and networking support available in the system. Typically, the device files found in /dev. o man5: File formats The formats for many nonintuitive data files are documented in the section 5. This includes various include files, program output files, and system files. o man6: Games This chapter documents games, demos, and generally trivial programs. Different people have various notions about how essential this is. o man7: Miscellaneous Manual pages that are difficult to classify are designated as being section 7. The *roff and other text processing macro packages are found here. o man8: System administration Documentation for programs used by system administrators for system operation and maintenance are documented here. Some of these programs are also occasionally useful for normal users. o man9: Kernel internal variables and functions This appears on Linux systems to document the kernel source code. 5.9. /usr/sbin : Non-vital standard system binaries This directory contains any non-vital binaries used exclusively by the system administrator (as root) which are not required for system repair and recovery, mounting /usr, or other essential operations. This includes networking daemons, large system administration tools, or anything that really isn't needed. Local system binaries and local administration shell scripts belong in /usr/local/sbin. - 24 - Linux Filesystem Structure January 15, 1994 5.10. /usr/share : Architecture-independent data Any specifications for /usr/share will be included in a supplementary draft to the main FSSTND draft. 5.11. /usr/src : Source code /usr/src : Source code | +- linux : Source code for Linus' kernel Any non-local source code should be placed in this subdirectory. The only source code that should always be placed in a specific location is the kernel source (when present or linked in part to the /usr/include structure). Subdirectories may be used here if desired. The source code for the kernel should always be in place or at least the include files from the kernel source. Those files are located in these directories: /usr/src/linux/include/asm /usr/src/linux/include/linux /usr/include usually contains links to `asm' and `linux' in the source directory. Since they are needed by the C compiler, at least those include files should always be distributed with installations. They should also be distributed in the /usr/src/linux directory so there are no problems when system administrators upgrade their kernel version for the first time. 5.12. /usr/tmp : temporary files, used to keep /tmp small Files in /usr/tmp are stored for an unspecified duration (please remember that system temporary directories are not guaranteed to hold data for any particular duration). Data stored in /usr/tmp is typically cleaned out "in a site-specific manner", but usually at less frequent intervals than /tmp. More information on temporary directories is in the section of the draft devoted to /tmp (above). /usr/tmp is usually symlinked to /var/tmp. - 25 - Linux Filesystem Structure January 15, 1994 6. The /var directory /var : Directories of files that tend to grow or vary in size |- adm : System logging and accounting files |- lock : Lock files |- named : DNS files (for `named'), networking only |- preserve : Saved text edited by `vi' after crash or hang-up |- spool : Directories for queue work to be performed later +- tmp : (symlink from /usr/tmp) The directory for variable length files. This includes spools, administrative files, logging files, transient files, and temporary files. A good reason to use /var is to make it possible to mount /usr read- only. Everything that once went into /usr that is written to on a temporary basis, now goes into /var. The aforementioned symbolic links, also mentioned below in the issues and rationale section, should be added to /usr for compatibility. This is very helpful if you are mounting /usr through NFS or if you want a read-only /usr. 6.1. /var/adm : System logging and accounting files All system logging should be written to this subdirectory and not to /var/log. `wtmp' and `lastlog' should be stored here. 6.2. /var/lock : Lock files Lock files should be stored within the /var/lock directory structure. /var/lock : Lock files +- emacs : Emacs lock files To preserve the ability to mount /usr read-only, no lock files or lock file systems should be stored on the /usr partition. Device lock files, such as the serial device lock files which are found in /usr/spool/locks or /usr/spool/uucp, should be stored in /var/lock. The naming convention which should be followed is `LCK..{device}'. For instance, if you are using a program which uses /dev/cua0, then it should create a lock file, `/var/lock/LCK..cua0'. The contents of these lock files should be the pid (process ID) of the process which is using the device (stored in 7 bit ASCII), followed by a newline character. Then, anything wishing to use /dev/cua0 can observe the lock file and act accordingly. Programs with large lock file systems should utilize a subdirectory of /var/lock. For instance, GNU Emacs normally stores Emacs lock files in /usr/lib/emacs/lock -- this poses a serious problem if you are trying to - 26 - Linux Filesystem Structure January 15, 1994 mount /usr read-only. Emacs lock files should instead be placed in /var/lock/emacs. If you are a developer/maintainer of an application which uses lock files in any manner, it is a good idea to contact the FSSTND mailing list or me to discuss the arrangement of your lock files. At this time, lock files which are stored in /etc, such as pid lock files, should remain in /etc. 6.3. /var/named : DNS and `named' stuff This is only needed for systems running a DNS name server (networking protocol for name servers). 6.4. /var/spool : Spool directories The disputed definition of this directory is "queue work or work to be done later". I suppose that more appropriately, it would be that plus anything that wouldn't fit in /usr or seemed related to the original stuff in here, but not always. Of course, that doesn't sound so great. So... who has a real definition? /var/spool : Spool directories |- at : at jobs |- cron : Cron jobs |- lpd : Printer spool directory |- mail : Directory for user mailboxes |- mqueue : Outgoing mail queue |- news : News spool directory +- uucp : Spool directory for uucp /var/spool/lpd : Printer spool directory |- {printer} : Spools for a specific printer +- {printer}.LOCK : Lock file for a specific printer - 27 - Linux Filesystem Structure January 15, 1994 7. Issues and Additional Rationale This section discusses several areas that may require further explanation. 7.1. What is Essential? The answer is: essential to clean, create, prepare, check, find and mount other filesystems (possibly on remote machines). There are other definitions, but this is a general definition that most people will at least incorporate into their own. 7.2. Networking Networking presented an interesting dilemma. Some people wanted to separate networking binaries and configuration from other binaries and configuration. However, we disagree. We feel that networking is not a "package", but an integral part of most UNIX (and UNIX-like) machines. Because of this networking should not be placed into a single directory, but systematically placed in the appropriate directories. o /bin: anything a user will want to use that is also considered vital {hostname, netstat, ping, ...} o /sbin: anything only root needs and is considered vital {arp, ifconfig, ifsetup, route} o /usr/bin: any binaries a user will want to use, but are not vital {finger, rcp, rlogin, telnet, ...} o /usr/sbin: any root only networking binaries that are not vital {ftpd, inetd, lpd, telnetd, ...} While this may seem confusing at first (and it does take a moment to digest), it does make sense. If you can only mount root for some reason and you need access to networking to repair your system, you don't need the files to be off in /usr/etc (as they often are). Files that are needed to mount /usr in normal (and emergency situations) are placed on the root subtree and any others are placed in /usr in order to keep the size of root small. Configuration files for networking belong in /etc. 7.3. Architecture-independent Structures Many people have noted that in this draft, that there was no /usr/share. There is now. The structure, /usr/share, typically contains - 28 - Linux Filesystem Structure January 15, 1994 architecture-independent files such as man-pages, timezone, terminfo information, et al. As of this time, there are no different architectures for Linux, but with the passage of time we should see Linux include other architectures and other UNIX-like systems. One note, no program should ever reference anything in /usr/share. For instance, a manual page program should never directly look in /usr/share/man/man1/ls.1, but it should refer to /usr/man/man1/ls.1 at all times. Anything in /usr/share will be "pointed-to" by the use of symlinks from other areas in the filesystem, such as /usr/man, /usr/lib/{something}, etc. The specifications for /usr/share are still be worked out and discussion within the FSSTND channel or with me is encouraged (by me). 7.4. Symbolic Links There are a wide range of uses for symbolic links (symlinks) in every filesystem. While symlinks are not encouraged for default setup (found after installing Linux) in a standard such as this, they are often used with good purpose on different systems. The point is that symlinks should be there to keep everything where everyone else expects find it Be prepared to accept that certain directories, even those contained on the root directory, are still going to be symlinks. For instance, on some systems /home will not be on the root, but symlinked to a /var directory, or to somewhere else. /home could also have its own physical partition, of course, and be mounted on its own. Similarly, because /usr might be on a central file server mounted via NFS, /usr/local could be symlinked to /var/local. Like /usr/emacs/lock, this change can be justified by recalling one definition of /var: "directories of files that vary on different systems and machines". Sometimes systems will also link /tmp to /var/{something} if the root partition becomes too small (or starts out too small). There are more examples of "good" uses of symbolic links, but the entire issue boils down to two things: packages should be able to find things where they expect them (within reason) and symbolic links can be used to solve the problem in many cases. However, problems also can arise from using too many symbolic links. These problems include over-reliance on symbolic links to solve problems, confusion resulting from overuse of symbolic links, and the aesthetic preferences of different people. 7.5. Statically linked binaries Linux is currently running on a wide variety of systems, some single user with small disks, some as servers in large networked environments. Because of this variety, this standard sets no rule regarding what - 29 - Linux Filesystem Structure January 15, 1994 binaries are static or dynamic with the following two exceptions. Both `ln' and `sync' should exist in /bin; any statically linked versions may be placed in /sbin, or replace those in /bin. Large Linux systems may wish to include other statically linked binaries (sh, init, mkfs, fsck, tunefs, mount, umount, swapon, swapoff, getty, login, etc.). The developers and/or system administrators are free to statically/dynamically link these and other binaries as they see fit, as long as the location of the binaries doesn't change. Networked systems (especially those of the future which may lack floppy drives), may want to make ifconfig, route, hostname, and other networking utilities static as well. This is usually not needed. - 30 - Linux Filesystem Structure January 15, 1994 The FSSTND mailing list The FSSTND mailing list is located at linux-fsstnd@ucsd.edu. This list was originally located on the linux-activists@Niksula.hut.fi "Mail-Net" as the FSSTND channel. (To subscribe to the list send mail to listserv@ucsd.edu with the body of "ADD linux-fsstnd".) Thanks to the people at the University of California at San Diego who allowed us to use their excellent mailing list server. Acknowledgments Credit for this text should be given to the FSSTND activists, developers, system administrators, and users whose input was essential to this standard. I also wish to thank each of the contributors who helped me to write, compile, and compose this, a consensus standard. I also wish to give real credit to those Linux developers who have seen that giving Linux a common filesystem layout is something that will further the development of the Linux operating system. Original contributors Drew Eckhardt Ian Jackson Ian McCloghrie Daniel J. Quinlan Mike Sangrey David H. Silber Theodore Ts'o Stephen Tweedie - 31 -