Gentoo Linux amd64 Handbook: Working with Gentoo

From Gentoo Wiki
Jump to:navigation Jump to:search



Welcome to Portage

Portage is one of Gentoo's most notable innovations in software management. With its high flexibility and enormous amount of features it is frequently seen as the best software management tool available for Linux.

Portage is completely written in Python and Bash and therefore fully visible to the users as both are scripting languages.

Most users will work with Portage through the emerge tool. This chapter is not meant to duplicate the information available from the emerge man page. For a complete rundown of emerge's options, please consult the man page:

user $man emerge

Gentoo repository

Ebuilds

When Gentoo's documentation talks about packages, it means software titles that are available to the Gentoo users through the Gentoo repository. This repository is a collection of ebuilds, files that contain all information Portage needs to maintain software (install, search, query, etc.). These ebuilds reside in /var/db/repos/gentoo by default.

Whenever someone asks Portage to perform some action regarding software titles, it will use the ebuilds on the system as a base. It is therefore important to regularly update the ebuilds on the system so Portage knows about new software, security updates, etc.

Updating the Gentoo repository

The Gentoo repository is usually updated with rsync, a fast incremental file transfer utility. Updating is fairly simple as the emerge command provides a front-end for rsync:

root #emerge --sync

Sometimes firewall restrictions apply that prevent rsync from contacting the mirrors. In this case, the Gentoo repository can be updated via daily generated snapshots. The emerge-webrsync tool automatically fetches and installs the latest snapshot on the system:

root #emerge-webrsync

Maintaining software

Searching for software

There are multiple ways to search through the Gentoo repository for software. One way is through emerge itself. By default, emerge --search returns the names of packages whose title matches (either fully or partially) the given search term.

For instance, to search for all packages who have "pdf" in their name:

user $emerge --search pdf

To search through the descriptions as well, use the --searchdesc (or -S) option:

user $emerge --searchdesc pdf

Notice that the output returns a lot of information. The fields are clearly labelled so we won't go further into their meanings:

CODE Example output for a search command
*  net-print/cups-pdf
      Latest version available: 1.5.2
      Latest version installed: [ Not Installed ]
      Size of downloaded files: 15 kB
      Homepage:    http://cip.physik.uni-wuerzburg.de/~vrbehr/cups-pdf/
      Description: Provides a virtual printer for CUPS to produce PDF files.
      License:     GPL-2

Installing software

When a software title has been found, then the installation is just one emerge command away. For instance, to install gnumeric:

root #emerge --ask app-office/gnumeric

Since many applications depend on each other, any attempt to install a certain software package might result in the installation of several dependencies as well. Don't worry, Portage handles dependencies well. To find out what Portage would install, add the --pretend option. For instance:

root #emerge --pretend gnumeric

To do the same, but interactively choose whether or not to proceed with the installation, add the --ask flag:

root #emerge --ask gnumeric

During the installation of a package, Portage will download the necessary source code from the Internet (if necessary) and store it by default in /var/cache/distfiles/. After this it will unpack, compile and install the package. To tell Portage to only download the sources without installing them, add the --fetchonly option to the emerge command:

root #emerge --fetchonly gnumeric

Finding installed package documentation

Many packages come with their own documentation. Sometimes, the doc USE flag determines whether the package documentation should be installed or not. To see if the doc USE flag is used by a package, use emerge -vp category/package:

root #emerge -vp media-libs/alsa-lib
These are the packages that would be merged, in order:
 
Calculating dependencies... done!
[ebuild   R    ] media-libs/alsa-lib-1.1.3::gentoo  USE="python -alisp -debug -doc" ABI_X86="(64) -32 (-x32)" PYTHON_TARGETS="python2_7" 0 KiB

The best way of enabling the doc USE flag is doing it on a per-package basis via /etc/portage/package.use, so that only the documentation for the wanted packages is installed. For more information read the USE flags section.

Once the package installed, its documentation is generally found in a subdirectory named after the package in the /usr/share/doc/ directory:

user $ls -l /usr/share/doc/alsa-lib-1.1.3
total 16
-rw-r--r-- 1 root root 3098 Mar  9 15:36 asoundrc.txt.bz2
-rw-r--r-- 1 root root  672 Mar  9 15:36 ChangeLog.bz2
-rw-r--r-- 1 root root 1083 Mar  9 15:36 NOTES.bz2
-rw-r--r-- 1 root root  220 Mar  9 15:36 TODO.bz2

A more sure way to list installed documentation files is to use equery's --filter option. equery is used to query Portage's database and comes as part of the app-portage/gentoolkit package:

user $equery files --filter=doc alsa-lib
 * Searching for alsa-lib in media-libs ...
 * Contents of media-libs/alsa-lib-1.1.3:
/usr/share/doc/alsa-lib-1.1.3/ChangeLog.bz2
/usr/share/doc/alsa-lib-1.1.3/NOTES.bz2
/usr/share/doc/alsa-lib-1.1.3/TODO.bz2
/usr/share/doc/alsa-lib-1.1.3/asoundrc.txt.bz2

The --filter option can be used with other rules to view the install locations for many other types of files. Additional functionality can be reviewed in equery's man page: man 1 equery.

Removing software

To safely remove software from a system, use emerge --deselect. This will tell Portage a package is no longer required and it is eligible for cleaning through --depclean.

root #emerge --deselect gnumeric

When a package is no longer selected, the package and its dependencies that were installed automatically when it was installed are still left on the system. To have Portage locate all dependencies that can now be removed, use emerge's --depclean functionality, which is documented later.

Updating the system

To keep the system in perfect shape (and not to mention install the latest security updates) it is necessary to update the system regularly. Since Portage only checks the ebuilds in the Gentoo repository, the first thing to do is to update this repository using emerge --sync. Then the system can be updated using emerge --deep --update @world.

Portage will, with --deep, search for newer version of the applications that are installed. Without --deep, it will only verify the versions for the applications that are explicitly installed (the applications listed in /var/lib/portage/world) - it does not thoroughly check their dependencies. This option should almost always therefore be used:

root #emerge --update --deep @world

The standard upgrade command should include --changed-use or --newuse because of possible changes within the repository's profiles, or if the USE settings of the system have been altered. Portage will then verify if the change requires the installation of new packages or recompilation of existing ones:

root #emerge --update --deep --newuse @world

Metapackages

Some packages in the Gentoo repository don't have any real content but are used to install a collection of packages. For instance, the kde-plasma/plasma-meta package will install the KDE Plasma desktop on the system by pulling in various Plasma-related packages as dependencies.

To remove such a package from the system, running emerge --deselect on the package will not have much effect since the dependencies for the package remain on the system.

Portage has the functionality to remove orphaned dependencies as well, but since the availability of software is dynamically dependent it is important to first update the entire system fully, including the new changes applied when changing USE flags. After this one can run emerge --depclean to remove the orphaned dependencies. When this is done, it might be necessary to rebuild the applications that were dynamically linked to the now-removed software titles but don't require them anymore, although recently support for this has been added to Portage.

All this is handled with the following two commands:

root #emerge --update --deep --newuse @world
root #emerge --ask --depclean

Licenses

Beginning with Portage version 2.1.7, it is possible to accept or reject software installation based on its license. All packages in the tree contain a LICENSE entry in their ebuilds. Running emerge --search category/package will show the package's license.

Important
As a disclaimer and limitation of liability, the LICENSE variable in an ebuild is merely a guideline for Gentoo developers and users. It is not a legal statement or a guarantee that it will reflect the license of every file installed by an ebuild. It should not be relied upon for a completely accurate legal representation of all files provided by a package. To gain assurance, system administrators should perform an in-depth check of each file installed by a package for proper licensing alignment and/or compliance. If a discrepancies is found in the ebuild, please file a bug to suggest a change to the value(s) assigned to the ebuild's LICENSE variable.

By default, Portage permits licenses that are explicitly approved by the Free Software Foundation, the Open Source Initiative, or that follow the Free Software Definition.

The variable that controls permitted licenses is called ACCEPT_LICENSE, which can be set in the /etc/portage/make.conf file. In the next example, this default value is shown:

FILE /etc/portage/make.confThe default ACCEPT_LICENSE setting
ACCEPT_LICENSE="-* @FREE"

With this configuration, packages with a free software or documentation license will be installable. Non-free software will not be installable.

It is possible to set ACCEPT_LICENSE globally in /etc/portage/make.conf, or to specify it on a per-package basis in the /etc/portage/package.license file.

For example, to allow the google-chrome license for the www-client/google-chrome package, add the following to /etc/portage/package.license:

FILE /etc/portage/package.licenseAccepting the google-chrome license for the google-chrome package
www-client/google-chrome google-chrome

This permits the installation of the www-client/google-chrome package, but prohibits the installation of the www-plugins/chrome-binary-plugins package, even though it has the same license.

Or to allow the often-needed sys-kernel/linux-firmware:

FILE /etc/portage/package.licenseAccepting the licenses for the linux-firmware package
# Accepting the license for linux-firmware
sys-kernel/linux-firmware linux-fw-redistributable

# Accepting any license that permits redistribution
sys-kernel/linux-firmware @BINARY-REDISTRIBUTABLE
Important
Licenses are stored in /var/db/repos/gentoo/licenses/ directory, and license groups are kept in /var/db/repos/gentoo/profiles/license_groups file. The first entry of each line in CAPITAL letters is the name of the license group, and every entry after that is an individual license.

License groups defined in the ACCEPT_LICENSE variable are prefixed with an @ sign. A possible setting (which was the previous Portage default) is to allow all licenses, except End User License Agreements (EULAs) that require reading and signing an acceptance agreement. To accomplish this, accept all licenses (using *) and then remove the licenses in the EULA group as follows:

FILE /etc/portage/make.confAccept all licenses except EULAs
ACCEPT_LICENSE="* -@EULA"

Note that this setting will also accept non-free software and documentation.

When Portage is complaining

Terminology

As stated before, Portage is extremely powerful and supports many features that other software management tools lack. To understand this, we explain a few aspects of Portage without going into too much detail.

With Portage different versions of a single package can coexist on a system. While other distributions tend to name their package to those versions (like gtk+2 and gtk+3) Portage uses a technology called SLOTs. An ebuild declares a certain SLOT for its version. Ebuilds with different SLOTs can coexist on the same system. For instance, the gtk+ package has ebuilds with SLOT="2" and SLOT="3".

There are also packages that provide the same functionality but are implemented differently. For instance, metalogd, sysklogd, and syslog-ng are all system loggers. Applications that rely on the availability of "a system logger" cannot depend on, for instance, metalogd, as the other system loggers are as good a choice as any. Portage allows for virtuals: each system logger is listed as an "exclusive" dependency of the logging service in the logger virtual package of the virtual category, so that applications can depend on the virtual/logger package. When installed, the package will pull in the first logging package mentioned in the package, unless a logging package was already installed (in which case the virtual is satisfied).

Software in the Gentoo repository can reside in different branches. By default the system only accepts packages that Gentoo deems stable. Most new software titles, when committed, are added to the testing branch, meaning more testing needs to be done before it is marked as stable. Although the ebuilds for those software are in the Gentoo repository, Portage will not update them before they are placed in the stable branch.

Some software is only available for a few architectures. Or the software doesn't work on the other architectures, or it needs more testing, or the developer that committed the software to the Gentoo repository is unable to verify if the package works on different architectures.

Each Gentoo installation also adheres to a certain profile which contains, amongst other information, the list of packages that are required for a system to function normally.

Blocked packages

CODE Portage warning about blocked packages
[ebuild  N     ] x11-wm/i3-4.20.1  USE="-doc -test"
[blocks B      ] x11-wm/i3 ("x11-wm/i3" is soft blocking x11-wm/i3-gaps-4.20.1)

* Error: The above package list contains packages which cannot be
 * installed at the same time on the same system.

(x11-wm/i3-4.20.1:0/0::gentoo, ebuild scheduled for merge) pulled in by
    x11-wm/i3

(x11-wm/i3-gaps-4.20.1-1:0/0::gentoo, installed) pulled in by
    x11-wm/i3-gaps required by @selected

Ebuilds contain specific fields that inform Portage about its dependencies. There are two possible dependencies: build dependencies, declared in the DEPEND variable and run-time dependencies, likewise declared in RDEPEND. When one of these dependencies explicitly marks a package or virtual as being not compatible, it triggers a blockage.

While recent versions of Portage are smart enough to work around minor blockages without user intervention, occasionally such blockages need to be resolved manually.

To fix a blockage, users can choose to not install the package or unmerge the conflicting package first. In the given example, one can opt not to install x11-wm/i3 or to remove x11-wm/i3-gaps first. It is usually best to simply tell Portage the package is no longer desired, with emerge --deselect x11-wm/i3-gaps, for example, to remove it from the world file rather than removing the package itself forcefully.

Sometimes there are also blocking packages with specific atoms, such as <media-video/mplayer-1.0_rc1-r2. In this case, updating to a more recent version of the blocking package could remove the block.

It is also possible that two packages that are yet to be installed are blocking each other. In this rare case, try to find out why both would need to be installed. In most cases it is sufficient to do with one of the packages alone. If not, please file a bug on Gentoo's bug tracking system.

Masked packages

CODE Portage warning about masked packages
!!! all ebuilds that could satisfy "bootsplash" have been masked.
CODE Portage warning about masked packages - reason
!!! possible candidates are:
  
- gnome-base/gnome-2.8.0_pre1 (masked by: ~x86 keyword)
- lm-sensors/lm-sensors-2.8.7 (masked by: -sparc keyword)
- sys-libs/glibc-2.3.4.20040808 (masked by: -* keyword)
- dev-util/cvsd-1.0.2 (masked by: missing keyword)
- games-fps/unreal-tournament-451 (masked by: package.mask)
- sys-libs/glibc-2.3.2-r11 (masked by: profile)
- net-im/skype-2.1.0.81 (masked by: skype-eula license(s))

When trying to install a package that isn't available for the system, this masking error occurs. Users should try installing a different application that is available for the system or wait until the package is marked as available. There is always a reason why a package is masked:

Reason for mask Description
~arch keyword The application is not tested sufficiently to be put in the stable branch. Wait a few days or weeks and try again.
-arch keyword or -* keyword The application does not work on the target architecture. If this is not the case, then please file a bug.
missing keyword The application has not yet been tested on the target architecture. Ask the architecture porting team to test the package or test it for them and report the findings on Gentoo's Bugzilla website. See /etc/portage/package.accept_keywords and Accepting a keyword for a single package.
package.mask The package has been found corrupt, unstable or worse and has been deliberately marked as do-not-use.
profile The package has been found not suitable for the current profile. The application might break the system if it is installed or is just not compatible with the profile currently in use.
license The package's license is not compatible with the ACCEPT_LICENSE value. Permit its license or the right license group by setting it in /etc/portage/make.conf or in /etc/portage/package.license.

Necessary USE flag changes

CODE Portage warning about USE flag change requirement
The following USE changes are necessary to proceed:
#required by app-text/happypackage-2.0, required by happypackage (argument)
>=app-text/feelings-1.0.0 test

The error message might also be displayed as follows, if --autounmask isn't set:

CODE Portage error about USE flag change requirement
emerge: there are no ebuilds built with USE flags to satisfy "app-text/feelings[test]".
!!! One of the following packages is required to complete your request:
- app-text/feelings-1.0.0 (Change USE: +test)
(dependency required by "app-text/happypackage-2.0" [ebuild])
(dependency required by "happypackage" [argument])

Such warning or error occurs when a package is requested for installation which not only depends on another package, but also requires that that package is built with a particular USE flag (or set of USE flags). In the given example, the package app-text/feelings needs to be built with USE="test", but this USE flag is not set on the system.

To resolve this, either add the requested USE flag to the global USE flags in /etc/portage/make.conf, or set it for the specific package in /etc/portage/package.use.

Missing dependencies

CODE Portage warning about missing dependency
emerge: there are no ebuilds to satisfy ">=sys-devel/gcc-3.4.2-r4".
  
!!! Problem with ebuild sys-devel/gcc-3.4.2-r2
!!! Possibly a DEPEND/*DEPEND problem.

The application to install depends on another package that is not available for the system. Please check Bugzilla if the issue is known and if not, please report it. Unless the system is configured to mix branches, this should not occur and is therefore a bug.

Ambiguous ebuild name

CODE Portage warning about ambiguous ebuild names
[ Results for search key : listen ]
[ Applications found : 2 ]
  
*  dev-tinyos/listen [ Masked ]
      Latest version available: 1.1.15
      Latest version installed: [ Not Installed ]
      Size of files: 10,032 kB
      Homepage:      http://www.tinyos.net/
      Description:   Raw listen for TinyOS
      License:       BSD
  
*  media-sound/listen [ Masked ]
      Latest version available: 0.6.3
      Latest version installed: [ Not Installed ]
      Size of files: 859 kB
      Homepage:      http://www.listen-project.org
      Description:   A Music player and management for GNOME
      License:       GPL-2
  
!!! The short ebuild name "listen" is ambiguous. Please specify
!!! one of the above fully-qualified ebuild names instead.

The application that is selected for installation has a name that corresponds with more than one package. Supply the category name as well to resolve this. Portage will inform the user about possible matches to choose from.

Circular dependencies

CODE Portage warning about circular dependencies
!!! Error: circular dependencies: 
  
ebuild / net-print/cups-1.1.15-r2 depends on ebuild / app-text/ghostscript-7.05.3-r1
ebuild / app-text/ghostscript-7.05.3-r1 depends on ebuild / net-print/cups-1.1.15-r2

Two (or more) packages to install depend on each other and can therefore not be installed. This is most likely a bug in one of the packages in the Gentoo repository. Please re-sync after a while and try again. It might also be beneficial to check Bugzilla to see if the issue is known and if not, report it.

Fetch failed

CODE Portage warning about fetch failed
!!! Fetch failed for sys-libs/ncurses-5.4-r5, continuing...
(...)
!!! Some fetch errors were encountered.  Please see above for details.

Portage was unable to download the sources for the given application and will try to continue installing the other applications (if applicable). This failure can be due to a mirror that has not synchronized correctly or because the ebuild points to an incorrect location. The server where the sources reside can also be down for some reason.

Retry after one hour to see if the issue still persists.

System profile protection

CODE Portage warning about profile-protected package
!!! Trying to unmerge package(s) in system profile. 'sys-apps/portage'
!!! This could be damaging to your system.

The user has asked to remove a package that is part of the system's core packages. It is listed in the profile as required and should therefore not be removed from the system.

Digest verification failure

CODE Digest verification failure
>>> checking ebuild checksums
!!! Digest verification failed:

This is a sign that something is wrong with the Gentoo repository - often, caused by a mistake made when committing an ebuild to the Gentoo ebuild repository.

When the digest verification fails, do not try to re-digest the package personally. Running ebuild foo manifest will not fix the problem; it quite possibly could make it worse.

Instead, wait an hour or two for the repository to settle down. It is likely that the error was noticed right away, but it can take a little time for the fix to trickle down the rsync mirrors. Check Bugzilla and see if anyone has reported the problem yet or ask around on #gentoo (webchat) (IRC). If not, go ahead and file a bug for the broken ebuild.

Once the bug has been fixed, re-sync the Gentoo ebuild repository to pick up the fixed digest.

Important
Be careful to not sync the Gentoo ebuild repository more than once a day. As stated in the official Gentoo netiquette policy (as well as when running emerge --sync), users who sync too often will be soft-banned from additional syncs for a time. Abusers who repeatedly fail to follow this policy may be hard-banned. Unless absolutely necessary it is often best to wait for a 24 hours period to sync so that re-synchronization does not overload Gentoo's rsync mirrors.





What are USE flags

The idea behind USE flags

When installing Gentoo, users make choices depending on the environment they are working with. A setup for a server differs from a setup for a workstation. A gaming workstation differs from a 3D rendering workstation.

This is not only true for choosing what packages to install, but also what features a certain package should support. If there is no need for OpenGL, why would someone bother to install and maintain OpenGL and build OpenGL support in most of the packages? If someone doesn't want to use KDE, why would they bother compiling packages with KDE support if those packages work flawlessly without?

To help users in deciding what to install/activate and what not, Gentoo wanted the user to specify his/her environment in an easy way. This forces the user into deciding what they really want and eases the process for Portage to make useful decisions.

Definition of a USE flag

Enter USE flags. Such a flag is a keyword that embodies support and dependency-information for a certain concept. If a certain USE flag is set to enabled, then Portage will know the system administrator desires support for the chosen keyword. Of course this may alter the dependency information for a package. Depending on the USE flag, this may require pulling in many more dependencies in order to fulfill the requested dependency changes.

Take a look at a specific example: the kde USE flag. If this flag is not set in the USE variable (or if the value is prefixed with a minus sign: -kde), then all packages that have optional KDE support will be compiled without KDE support. All packages that have an optional KDE dependency will be installed without installing the KDE libraries (as dependency).

When the kde flag is set to enabled, then those packages will be compiled with KDE support, and the KDE libraries will be installed as dependency.

By correctly defining USE flags, the system will be tailored specifically to the needs of the system administrator.

Using USE flags

Declare permanent USE flags

All USE flags are declared inside the USE variable. To make it easy for users to search and pick USE flags, we already provide a default USE setting. This setting is a collection of USE flags we think are commonly used by the Gentoo users. This default setting is declared in the make.defaults files that are part of the selected profile.

The profile the system listens to is pointed to by the /etc/portage/make.profile symlink. Each profile works on top of other profiles, and the end result is therefore the sum of all profiles. The top profile is the base profile (/var/db/repos/gentoo/profiles/base).

To view the currently active USE flags (completely), use emerge --info:

root #emerge --info | grep ^USE
USE="a52 aac acpi alsa branding cairo cdr dbus dts ..."

This variable already contains quite a lot of keywords. Do not alter any make.defaults file to tailor the USE variable to personal needs though: changes in these files will be undone when the Gentoo repository is updated!

To change this default setting, add or remove keywords to/from the USE variable. This is done globally by defining the USE variable in /etc/portage/make.conf. In this variable one can add the extra USE flags required, or remove the USE flags that are no longer needed. This latter is done by prefixing the keyword with the minus-sign (-).

For instance, to remove support for KDE and Qt but add support for LDAP, the following USE can be defined in /etc/portage/make.conf:

FILE /etc/portage/make.confUpdating USE in make.conf
USE="-kde -qt5 ldap"

Declaring USE flags for individual packages

Sometimes users want to declare a certain USE flag for one (or a couple) of applications but not system-wide. To accomplish this, edit /etc/portage/package.use. package.use is typically a single file, however it can also be a directory filled with children files; see the tip below and then man 5 portage for more information on how to use this convention. The following examples assume package.use is a single file.

For instance, to only have Blu-ray support for the VLC media player package:

FILE /etc/portage/package.useEnabling Blu-ray support for VLC
media-video/vlc bluray
Tip
If package.use is pre-existing as a directory (opposed to a single file), packages can have their USE flags modified by simply creating files under the package.use/ directory. Any file naming convention can work, however it is wise to implement a coherent naming scheme. One convention is to simply use the package name as the title for the child file. For example, setting the bluray USE flag for the media-video/vlc package can be performed as follows:

root #echo "media-video/vlc bluray" >> /etc/portage/package.use/vlc

Similarly it is possible to explicitly disable USE flags for a certain application. For instance, to disable bzip2 support in PHP (but have it for all other packages through the USE flag declaration in make.conf):

FILE /etc/portage/package.useDisable bzip2 support for PHP
dev-lang/php -bzip2

Declaring temporary USE flags

Sometimes users need to set a USE flag for a brief moment. Instead of editing /etc/portage/make.conf twice (to do and undo the USE changes) just declare the USE variable as an environment variable. Remember that this setting only applies for the command entered; re-emerging or updating this application (either explicitly or as part of a system update) will undo the changes that were triggered through the (temporary) USE flag definition.

The following example temporarily removes the pulseaudio value from the USE variable during the installation of SeaMonkey:

root #USE="-pulseaudio" emerge www-client/seamonkey

Precedence

Of course there is a certain precedence on what setting has priority over the USE setting. The precedence for the USE setting is, ordered by priority (first has lowest priority):

  1. Default USE setting declared in the make.defaults files part of your profile
  2. User-defined USE setting in /etc/portage/make.conf
  3. User-defined USE setting in /etc/portage/package.use
  4. User-defined USE setting as environment variable

To view the final USE setting as seen by Portage, run emerge --info. This will list all relevant variables (including the USE variable) with their current definition as known to Portage.

root #emerge --info

Adapting the entire system to the new USE flags

After having altered USE flags, the system should be updated to reflect the necessary changes. To do so, use the --newuse option with emerge:

root #emerge --update --deep --newuse @world

Next, run Portage's depclean to remove the conditional dependencies that were emerged on the "old" system but that have been obsoleted by the new USE flags.

Important
Double-check the provided list of "obsoleted" packages to make sure it does not remove packages that are needed. In the following example the --pretend (-p) switch to have depclean only list the packages without removing them:
root #emerge --pretend --depclean

When depclean has finished, emerge may prompt to rebuild the applications that are dynamically linked against shared objects provided by possibly removed packages. Portage will preserve necessary libraries until this action is done to prevent breaking applications. It stores what needs to be rebuilt in the preserved-rebuild set. To rebuild the necessary packages, run:

root #emerge @preserved-rebuild

When all this is accomplished, the system is using the new USE flag settings.

Package specific USE flags

Viewing available USE flags

Let's take the example of seamonkey: what USE flags does it listen to? To find out, we use emerge with the --pretend and --verbose options:

root #emerge --pretend --verbose www-client/seamonkey
These are the packages that would be merged, in order:
 
Calculating dependencies... done!
[ebuild  N     ] www-client/seamonkey-2.48_beta1::gentoo  USE="calendar chatzilla crypt dbus gmp-autoupdate ipc jemalloc pulseaudio roaming skia startup-notification -custom-cflags -custom-optimization -debug -gtk3 -jack -minimal (-neon) (-selinux) (-system-cairo) -system-harfbuzz -system-icu -system-jpeg -system-libevent -system-libvpx -system-sqlite {-test} -wifi" L10N="-ca -cs -de -en-GB -es-AR -es-ES -fi -fr -gl -hu -it -ja -lt -nb -nl -pl -pt-PT -ru -sk -sv -tr -uk -zh-CN -zh-TW" 216,860 KiB
 
Total: 1 package (1 new), Size of downloads: 216,860 KiB

emerge isn't the only tool for this job. In fact, there is a tool dedicated to package information called equery which resides in the app-portage/gentoolkit package

root #emerge --ask app-portage/gentoolkit

Now run equery with the uses argument to view the USE flags of a certain package. For instance, for the app-portage/portage-utils package:

user $equery --nocolor uses =app-portage/portage-utils-0.93.3
[ Legend : U - final flag setting for installation]
[        : I - package is installed with flag     ]
[ Colors : set, unset                             ]
 * Found these USE flags for app-portage/portage-utils-0.93.3:
 U I
 + + nls       : Add Native Language Support (using gettext - GNU locale utilities)
 + + openmp    : Build support for the OpenMP (support parallel computing), requires >=sys-devel/gcc-4.2 built with USE="openmp"
 + + qmanifest : Build qmanifest applet, this adds additional dependencies for GPG, OpenSSL and BLAKE2B hashing
 + + qtegrity  : Build qtegrity applet, this adds additional dependencies for OpenSSL
 - - static    : !!do not set this during bootstrap!! Causes binaries to be statically linked instead of dynamically

Satisfying REQUIRED_USE conditions

Some ebuilds require or forbid certain combinations of USE flags in order to work properly. This is expressed via a set of conditions placed in a REQUIRED_USE expression. This conditions ensure that all features and dependencies are complete and that the build will succeed and perform as expected. If any of these are not met, emerge will alert you and ask you to fix the issue.

Example Description
REQUIRED_USE="foo? ( bar )" If foo is set, bar must be set.
REQUIRED_USE="foo? ( !bar )" If foo is set, bar must not be set.
REQUIRED_USE="foo? ( || ( bar baz ) )" If foo is set, bar or baz must be set.
REQUIRED_USE="^^ ( foo bar baz )" Exactly one of foo bar or baz must be set.
REQUIRED_USE="|| ( foo bar baz )" At least one of foo bar or baz must be set.
REQUIRED_USE="?? ( foo bar baz )" No more than one of foo bar or baz may be set.




Portage features

Portage has several additional features that make the Gentoo experience even better. Many of these features rely on certain software tools that improve performance, reliability, security, ...

To enable or disable certain Portage features, edit /etc/portage/make.conf and update or set the FEATURES variable which contains the various feature keywords, separated by white space. In several cases it will also be necessary to install the additional tool on which the feature relies.

Not all features that Portage supports are listed here. For a full overview, please consult the make.conf man page:

user $man make.conf

To find out what FEATURES are set by default, run emerge --info and search for the FEATURES variable or grep it out:

user $emerge --info | grep ^FEATURES=

Distributed compiling

Using distcc

distcc is a program to distribute compilations across several, not necessarily identical, machines on a network. The distcc client sends all necessary information to the available distcc servers (running distccd) so they can compile pieces of source code for the client. The net result is a faster compilation time.

More information about distcc (and how to have it work with Gentoo) can be found in the Distcc article.

Installing distcc

Distcc ships with a graphical monitor to monitor tasks that the computer is sending away for compilation. This tool is automatically installed if USE="gtk" is set.

root #emerge --ask sys-devel/distcc

Activating Portage distcc support

Add distcc to the FEATURES variable inside /etc/portage/make.conf. Next, edit the MAKEOPTS variable and increase the number of parallel build jobs that the system allows. A known guideline is to fill in -jN where N is the number of CPUs that run distccd (including the current host) plus one, but that is just a guideline.

Now run distcc-config and enter the list of available distcc servers. For a simple example assume that the available DistCC servers are 192.168.1.102 (the current host), 192.168.1.103 and 192.168.1.104 (two "remote" hosts):

root #distcc-config --set-hosts "192.168.1.102 192.168.1.103 192.168.1.104"

Don't forget to run the distccd daemon as well:

root #rc-update add distccd default
root #/etc/init.d/distccd start

Caching compilation objects

About ccache

ccache is a fast compiler cache. Whenever an application is compiled, it will cache intermediate results so that, whenever the same program and version is recompiled, the compilation time is greatly reduced. The first time ccache is run, it will be much slower than a normal compilation. Subsequent recompiles however should be faster. ccache is only helpful if the same application version will be recompiled many times; thus it is mostly only useful for software developers.

For more information about ccache, please visit its homepage.

Warning
ccache is known to cause numerous compilation failures. Sometimes ccache will retain stale code objects or corrupted files, which can lead to packages that cannot be emerged. If this happens (errors like "File not recognized: File truncated" come up in build logs), try recompiling the application with ccache disabled (FEATURES="-ccache" in /etc/portage/make.conf or one-shot from the commandline with the following) before reporting a bug:


root #FEATURES="-ccache" emerge --oneshot <category/package>

Installing ccache

To install ccache run the following command:

root #emerge --ask dev-util/ccache

Activating Portage ccache support

Open /etc/portage/make.conf and add ccache to any values defined in the FEATURES variable. If FEATURES does not exist, then create it. Next, add a new variable called CCACHE_SIZE and set it to 2G:

FILE /etc/portage/make.confEnabling Portage ccache support
FEATURES="ccache"
CCACHE_SIZE="2G"

To check if ccache functions, ask ccache to provide its statistics. Because Portage uses a different ccache home directory, it is necessary to temporarily set the CCACHE_DIR variable:

root #CCACHE_DIR="/var/tmp/ccache" ccache -s

The /var/tmp/ccache/ location is Portage' default ccache home directory; it can be changed by setting the CCACHE_DIR variable in /etc/portage/make.conf.

When running ccache standalone, it would use the default location of ${HOME}/.ccache/, which is why the CCACHE_DIR variable needs to be set when asking for the (Portage) ccache statistics.

Using ccache outside Portage

To use ccache for non-Portage compilations, add /usr/lib/ccache/bin/ to the beginning of the PATH variable (before /usr/bin). This can be accomplished by editing ~/.bash_profile in the user's home directory. Using ~/.bash_profile is one way to define PATH variables.

FILE ~/.bash_profileSetting the ccache location before any other PATH
PATH="/usr/lib/ccache/bin:${PATH}"

Binary package support

Creating prebuilt packages

Portage supports the installation of prebuilt packages. Even though Gentoo does not provide prebuilt packages by itself Portage can be made fully aware of prebuilt packages.

To create a prebuilt package use the quickpkg command if the package is already installed on the system, or emerge with the --buildpkg or --buildpkgonly options.

To have Portage create prebuilt packages of every single package that gets installed, add buildpkg to the FEATURES variable.

More extended support for creating prebuilt package sets can be obtained with catalyst. For more information on catalyst please read the Catalyst FAQ.

Installing prebuilt packages

Although Gentoo doesn't provide one, it is possible to create a central repository where prebuilt packages are stored. In order to use this repository, it is necessary to make Portage aware of it by having the PORTAGE_BINHOST variable point to it. For instance, if the prebuilt packages are on ftp://buildhost/gentoo:

FILE /etc/portage/make.confAdd PORTAGE_BINHOST location
PORTAGE_BINHOST="ftp://buildhost/gentoo"

To install a prebuilt package, add the --getbinpkg option to the emerge command alongside of the --usepkg option. The former tells emerge to download the prebuilt package from the previously defined server while the latter asks emerge to try to install the prebuilt package first before fetching the sources and compiling it.

For instance, to install gnumeric with prebuilt packages:

root #emerge --usepkg --getbinpkg gnumeric

More information about emerge's prebuilt package options can be found in the emerge man page:

user $man emerge

Distributing prebuilt packages to others

If prebuilt packages are to be distributed to others, then make sure that this is permitted. Check the distribution terms of the upstream package for this. For example, for a package released under the GNU GPL, sources must be made available along with the binaries.

Ebuilds may define a bindist restriction in their RESTRICT variable if built binaries are not distributable. Sometimes this restriction is conditional on one or more USE flags.

By default, Portage will not mask any packages because of restrictions. This can be changed globally by setting the ACCEPT_RESTRICT variable in /etc/portage/make.conf. For example, to mask packages that have a bindist restriction, add the following line to make.conf:

FILE /etc/portage/make.confOnly accept binary distributable packages
ACCEPT_RESTRICT="* -bindist"

It is also possible to override the ACCEPT_RESTRICT variable by passing the --accept-restrict option to the emerge command. For example, --accept-restrict=-bindist will temporarily mask packages with a bindist restriction.

Also consider setting the ACCEPT_LICENSE variable when distributing packages. See the Licenses section for this.

Important
It is entirely the responsibility of each user to comply with packages' license terms and with laws of each user's country. The metadata variables defined by ebuilds (RESTRICT or LICENSE) can provide guidance when distribution of binaries is not permitted, however output from Portage or questions answered by the Gentoo developers are not legal statements and should not be relied upon as such. Be cautious to abide by the law of your physical location.

Fetching files

Userfetch

Portage is normally run as the root user. Setting FEATURES="userfetch" will allow Portage to drop root privileges while fetching package sources and run with user/group permissions of portage:portage. This is a small security improvement.

If userfetch is set in FEATURES be sure to change the owner of all the files beneath /var/db/repos/gentoo using the chown command with root privileges:

root #chown --recursive --verbose portage:portage /var/db/repos/gentoo

Verify distfiles

To re-verify the integrity and (potentially) re-download previously removed/corrupted distfiles for all currently installed packages, run:

root #emerge --ask --fetchonly --emptytree @world





Runlevels

Booting the system

When the system is booted, lots of text floats by. When paying close attention, one will notice this text is (usually) the same every time the system is rebooted. The sequence of all these actions is called the boot sequence and is (more or less) statically defined.

First, the boot loader will load the kernel image that is defined in the boot loader configuration. Then, the boot loader instructs the CPU to execute kernel. When the kernel is loaded and run, it initializes all kernel-specific structures and tasks and starts the init process.

This process then makes sure that all filesystems (defined in /etc/fstab) are mounted and ready to be used. Then it executes several scripts located in /etc/init.d/, which will start the services needed in order to have a successfully booted system.

Finally, when all scripts are executed, init activates the terminals (in most cases just the virtual consoles which are hidden beneath Alt+F1, Alt+F2, etc.) attaching a special process called agetty to it. This process will then make sure users are able to log on through these terminals by running login.

Initscripts

Now init doesn't just execute the scripts in /etc/init.d/ randomly. Even more, it doesn't run all scripts in /etc/init.d/, only the scripts it is told to execute. It decides which scripts to execute by looking into /etc/runlevels/.

First, init runs all scripts from /etc/init.d/ that have symbolic links inside /etc/runlevels/boot/. Usually, it will start the scripts in alphabetical order, but some scripts have dependency information in them, telling the system that another script must be run before they can be started.

When all /etc/runlevels/boot/ referenced scripts are executed, init continues with running the scripts that have a symbolic link to them in /etc/runlevels/default/. Again, it will use the alphabetical order to decide what script to run first, unless a script has dependency information in it, in which case the order is changed to provide a valid start-up sequence. The latter is also the reason why commands used during the installation of Gentoo Linux used default, as in rc-update add sshd default.

How init works

Of course init doesn't decide all that by itself. It needs a configuration file that specifies what actions need to be taken. This configuration file is /etc/inittab.

Remember the boot sequence that was just described - init's first action is to mount all file systems. This is defined in the following line from /etc/inittab:

FILE /etc/inittabInitialization command
si::sysinit:/sbin/openrc sysinit

This line tells init that it must run /sbin/openrc sysinit to initialize the system. The /sbin/openrc script takes care of the initialization, so one might say that init doesn't do much - it delegates the task of initializing the system to another process.

Second, init executed all scripts that had symbolic links in /etc/runlevels/boot/. This is defined in the following line:

FILE /etc/inittabBoot command invocation
rc::bootwait:/sbin/openrc boot

Again the OpenRC script performs the necessary tasks. Note that the option given to OpenRC (boot) is the same as the sub-directory of /etc/runlevels/ that is used.

Now init checks its configuration file to see what runlevel it should run. To decide this, it reads the following line from /etc/inittab:

FILE /etc/inittabDefault runlevel selection
id:3:initdefault:

In this case (which the majority of Gentoo users will use), the runlevel id is 3. Using this information, init checks what it must run to start runlevel 3:

FILE /etc/inittabRunlevel definitions
l0:0:wait:/sbin/openrc shutdown
l1:S1:wait:/sbin/openrc single
l2:2:wait:/sbin/openrc nonetwork
l3:3:wait:/sbin/openrc default
l4:4:wait:/sbin/openrc default
l5:5:wait:/sbin/openrc default
l6:6:wait:/sbin/openrc reboot

The line that defines level 3, again, uses the openrc script to start the services (now with argument default). Again note that the argument of openrc is the same as the subdirectory from /etc/runlevels/.

When OpenRC has finished, init decides what virtual consoles it should activate and what commands need to be run at each console:

FILE /etc/inittabTerminal definitions
c1:12345:respawn:/sbin/agetty 38400 tty1 linux
c2:12345:respawn:/sbin/agetty 38400 tty2 linux
c3:12345:respawn:/sbin/agetty 38400 tty3 linux
c4:12345:respawn:/sbin/agetty 38400 tty4 linux
c5:12345:respawn:/sbin/agetty 38400 tty5 linux
c6:12345:respawn:/sbin/agetty 38400 tty6 linux

Available runlevels

In a previous section, we saw that init uses a numbering scheme to decide what runlevel it should activate. A runlevel is a state in which the system is running and contains a collection of scripts (runlevel scripts or initscripts) that must be executed when entering or leaving a runlevel.

In Gentoo, there are seven runlevels defined: three internal runlevels, and four user-defined runlevels. The internal runlevels are called sysinit, shutdown and reboot and do exactly what their names imply: initialize the system, powering off the system, and rebooting the system.

The user-defined runlevels are those with an accompanying /etc/runlevels/ subdirectory: boot, default, nonetwork and single. The boot runlevel starts all system-necessary services which all other runlevels use. The remaining three runlevels differ in what services they start: default is used for day-to-day operations, nonetwork is used in case no network connectivity is required, and single is used when the system needs to be fixed.

Working with initscripts

The scripts that the openrc process starts are called init scripts. Each script in /etc/init.d/ can be executed with the arguments start, stop, restart, zap, status, ineed, iuse, iwant, needsme, usesme, or wantsme.

To start, stop, or restart a service (and all depending services), the start, stop, and restart arguments should be used:

root #rc-service postfix start
Note
Only the services that need the given service are stopped or restarted. The other depending services (those that use the service but don't need it) are left untouched.

To stop a service, but not the services that depend on it, use the --nodeps option together with the stop argument:

root #rc-service --nodeps postfix stop

To see what status a service has (started, stopped, ...) use the status argument:

root #rc-service postfix status

If the status information shows that the service is running, but in reality it is not, then reset the status information to "stopped" with the zap argument:

root #rc-service postfix zap

To also ask what dependencies the service has, use iwant, iuse or ineed. With ineed it is possible to see the services that are really necessary for the correct functioning of the service. iwant or iuse, on the other hand, shows the services that can be used by the service, but are not necessary for the correct functioning.

root #rc-service postfix ineed

Similarly, it is possible to ask what services require the service (needsme) or can use it (usesme or wantsme):

root #rc-service postfix needsme

Updating runlevels

rc-update

Gentoo's init system uses a dependency-tree to decide what service needs to be started first. As this is a tedious task that we wouldn't want our users to have to do manually, we have created tools that ease the administration of the runlevels and init scripts.

With rc-update it is possible to add and remove init scripts to a runlevel. The rc-update tool will then automatically ask the depscan.sh script to rebuild the dependency tree.

Adding and removing services

In earlier instructions, init scripts have already been added to the "default" runlevel. What "default" means has been explained earlier in this document. Next to the runlevel, the rc-update script requires a second argument that defines the action: add, del, or show.

To add or remove an init script, just give rc-update the add or del argument, followed by the init script and the runlevel. For instance:

root #rc-update del postfix default

The rc-update -v show command will show all the available init scripts and list at which runlevels they will execute:

root #rc-update -v show

It is also possible to run rc-update show (without -v) to just view enabled init scripts and their runlevels.

Configuring services

Why additional configuration is needed

Init scripts can be quite complex. It is therefore not really desirable to have the users edit the init script directly, as it would make it more error-prone. It is however important to be able to configure such a service. For instance, users might want to give more options to the service itself.

A second reason to have this configuration outside the init script is to be able to update the init scripts without the fear that the user's configuration changes will be undone.

conf.d directory

Gentoo provides an easy way to configure such a service: every init script that can be configured has a file in /etc/conf.d/. For instance, the apache2 initscript (called /etc/init.d/apache2) has a configuration file called /etc/conf.d/apache2, which can contain the options to give to the Apache 2 server when it is started:

FILE /etc/conf.d/apache2Example options for apache2 init script
APACHE2_OPTS="-D PHP5"

Such a configuration file contains only variables (just like /etc/portage/make.conf does), making it very easy to configure services. It also allows us to provide more information about the variables (as comments).

Writing initscripts

Another useful resource is OpenRC's service script guide.

Is it necessary?

No, writing an init script is usually not necessary as Gentoo provides ready-to-use init scripts for all provided services. However, some users might have installed a service without using Portage, in which case they will most likely have to create an init script.

Do not use the init script provided by the service if it isn't explicitly written for Gentoo: Gentoo's init scripts are not compatible with the init scripts used by other distributions! That is, unless the other distribution is using OpenRC!

Layout

The basic layout of an init script is shown below.

CODE Example initscript layout (traditional)
#!/sbin/openrc-run
  
depend() {
#  (Dependency information)
}
  
start() {
#  (Commands necessary to start the service)
}
  
stop() {
#  (Commands necessary to stop the service)
}
CODE Example initscript layout (updated)
#!/sbin/openrc-run
command=/usr/bin/foo
command_args="${foo_args} --bar"
pidfile=/var/run/foo.pid
name="FooBar Daemon"
 
description="FooBar is a daemon that drinks"
extra_started_commands="drink"
description_drink="Opens mouth and reflexively swallows"
 
depend() {
#  (Dependency information)
}
 
start_pre() {
#  (Commands necessary to prepare to start the service)
    # Ensure that our dirs are correct
    checkpath --directory --owner foo:foo --mode 0775 \
        /var/run/foo /var/cache/foo
}
  
stop_post() {
#  (Commands necessary to clean up after the service)
    # Clean any spills
    rm -rf /var/cache/foo/*
}
 
drink() {
    ebegin "Starting to drink"
    ${command} --drink beer
    eend $? "Failed to drink any beer :("
}

Every init script requires the start() function or command variable to be defined. All other sections are optional.

Dependencies

There are three dependency-alike settings that can be defined which influence the start-up or sequencing of init scripts: want, use and need. Next to these, there are also two order-influencing methods called before and after. These last two are no dependencies per se - they do not make the original init script fail if the selected one isn't scheduled to start (or fails to start).

  • The use settings informs the init system that this script uses functionality offered by the selected script, but does not directly depend on it. A good example would be use logger or use dns. If those services are available, they will be put in good use, but if the system does not have a logger or DNS server the services will still work. If the services exist, then they are started before the script that uses them.
  • The want setting is similar to use with one exception. use only considers services which were added to an init level. want will try to start any available service even if not added to an init level.
  • The need setting is a hard dependency. It means that the script that is needing another script will not start before the other script is launched successfully. Also, if that other script is restarted, then this one will be restarted as well.
  • When using before, then the given script is launched before the selected one if the selected one is part of the init level. So an init script xdm that defines before alsasound will start before the alsasound script, but only if alsasound is scheduled to start as well in the same init level. If alsasound is not scheduled to start too, then this particular setting has no effect and xdm will be started when the init system deems it most appropriate.
  • Similarly, after informs the init system that the given script should be launched after the selected one if the selected one is part of the init level. If not, then the setting has no effect and the script will be launched by the init system when it deems it most appropriate.

It should be clear from the above that need is the only "true" dependency setting as it affects if the script will be started or not. All the others are merely pointers towards the init system to clarify in which order scripts can be (or should be) launched.

Now, look at many of Gentoo's available init scripts and notice that some have dependencies on things that are no init scripts. These "things" we call virtuals.

A virtual dependency is a dependency that a service provides, but that is not provided solely by that service. An init script can depend on a system logger, but there are many system loggers available (metalogd, syslog-ng, sysklogd, ...). As the script cannot need every single one of them (no sensible system has all these system loggers installed and running) we made sure that all these services provide a virtual dependency.

For instance, take a look at the postfix dependency information:

FILE /etc/init.d/postfixDependency information of the postfix service
depend() {
  need net
  use logger dns
  provide mta
}

As can be seen, the postfix service:

  • Requires the (virtual) net dependency (which is provided by, for instance, /etc/init.d/net.eth0).
  • Uses the (virtual) logger dependency (which is provided by, for instance, /etc/init.d/syslog-ng).
  • Uses the (virtual) dns dependency (which is provided by, for instance, /etc/init.d/named).
  • Provides the (virtual) mta dependency (which is common for all mail servers).

Controlling the order

As described in the previous section, it is possible to tell the init system what order it should use for starting (or stopping) scripts. This ordering is handled both through the dependency settings use and need, but also through the order settings before and after. As we have described these earlier already, let's take a look at the portmap service as an example of such init script.

FILE /etc/init.d/portmapDependency information of the portmap service
depend() {
  need net
  before inetd
  before xinetd
}

It is possible to use the "*" glob to catch all services in the same runlevel, although this isn't advisable.

CODE Using the * glob
depend() {
  before *
}

If the service must write to local disks, it should need localmount. If it places anything in /var/run/ such as a PID file, then it should start after bootmisc:

CODE Dependency setting with needing localmount and after bootmisc
depend() {
  need localmount
  after bootmisc
}

Standard functions

Next to the depend() functionality, it is also necessary to define the start() function. This one contains all the commands necessary to initialize the service. It is advisable to use the ebegin and eend functions to inform the user about what is happening:

CODE Example start() function
start() {
  if [ "${RC_CMD}" = "restart" ];
  then
    # Do something in case a restart requires more than stop, start
  fi
  
  ebegin "Starting my_service"
  start-stop-daemon --start --exec /path/to/my_service \
    --pidfile /path/to/my_pidfile
  eend $?
}

Both --exec and --pidfile should be used in start and stop functions. If the service does not create a pidfile, then use --make-pidfile if possible, though it is recommended to test this to be sure. Otherwise, don't use pidfiles. It is also possible to add --quiet to the start-stop-daemon options, but this is not recommended unless the service is extremely verbose. Using --quiet may hinder debugging if the service fails to start.

Another notable setting used in the above example is to check the contents of the RC_CMD variable. Unlike the previous init script system, the newer OpenRC system does not support script-specific restart functionality. Instead, the script needs to check the contents of the RC_CMD variable to see if a function (be it start() or stop()) is called as part of a restart or not.

Note
Make sure that --exec actually calls a service and not just a shell script that launches services and exits - that's what the init script is supposed to do.

For more examples of the start() function, please read the source code of the available init scripts in the /etc/init.d/ directory.

Another function that can (but does not have to) be defined is stop(). The init system is intelligent enough to fill in this function by itself if start-stop-daemon is used.

CODE Example stop() function
stop() {
  ebegin "Stopping my_service"
  start-stop-daemon --stop --exec /path/to/my_service \
    --pidfile /path/to/my_pidfile
  eend $?
}

If the service runs some other script (for example, Bash, Python, or Perl), and this script later changes names (for example, foo.py to foo), then it is necessary to add --name to start-stop-daemon. This must specify the name that the script will be changed to. In this example, a service starts foo.py, which changes names to foo:

CODE Example definition for a service that starts the foo script
start() {
  ebegin "Starting my_script"
  start-stop-daemon --start --exec /path/to/my_script \
    --pidfile /path/to/my_pidfile --name foo
  eend $?
}

start-stop-daemon has an excellent man page available if more information is needed:

user $man start-stop-daemon

Gentoo's init script syntax is based on the POSIX Shell so people are free to use sh-compatible constructs inside their init scripts. Keep other constructs, like bash-specific ones, out of the init scripts to ensure that the scripts remain functional regardless of the change Gentoo might do on its init system.

Adding custom options

If the initscript needs to support more options than the ones we have already encountered, then add the option to one of the following variables, and create a function with the same name as the option. For instance, to support an option called restartdelay:

  • extra_commands - Command is available with the service in any state
  • extra_started_commands - Command is available when the service is started
  • extra_stopped_commands - Command is available when the service is stopped


CODE Example definition of restartdelay method
extra_started_commands="restartdelay"
  
restartdelay() {
  stop
  sleep 3    # Wait 3 seconds before starting again
  start
}
Important
The restart() function cannot be overridden in OpenRC!

Service configuration variables

In order to support configuration files in /etc/conf.d/, no specifics need to be implemented: when the init script is executed, the following files are automatically sourced (i.e. the variables are available to use):

  • /etc/conf.d/YOUR_INIT_SCRIPT
  • /etc/conf.d/basic
  • /etc/rc.conf

Also, if the init script provides a virtual dependency (such as net), the file associated with that dependency (such as /etc/conf.d/net) will be sourced too.

Changing runlevel behavior

Who might benefit

Many laptop users know the situation: at home they need to start net.eth0, but they don't want to start net.eth0 while on the road (as there is no network available). With Gentoo the runlevel behavior can be altered at will.

For instance, a second "default" runlevel can be created which can be booted that has other init scripts assigned to it. At boot time, the user can then select what default runlevel to use.

Using softlevel

First of all, create the runlevel directory for the second "default" runlevel. As an example we create the offline runlevel:

root #mkdir /etc/runlevels/offline

Add the necessary init scripts to the newly created runlevel. For instance, to have an exact copy of the current default runlevel but without net.eth0:

root #cd /etc/runlevels/default
root #for service in *; do rc-update add $service offline; done
root #rc-update del net.eth0 offline
root #rc-update show offline
(Partial sample Output)
               acpid | offline
          domainname | offline
               local | offline
            net.eth0 |

Even though net.eth0 has been removed from the offline runlevel, udev might want to attempt to start any devices it detects and launch the appropriate services, a functionality that is called hotplugging. By default, Gentoo does not enable hotplugging.

To enable hotplugging, but only for a selected set of scripts, use the rc_hotplug variable in /etc/rc.conf:

FILE /etc/rc.confEnable hotplugging of the WLAN interface
rc_hotplug="net.wlan !net.*"
Note
For more information on device initiated services, please see the comments inside /etc/rc.conf.

Edit the bootloader configuration and add a new entry for the offline runlevel. In that entry, add softlevel=offline as a boot parameter.

Using bootlevel

Using bootlevel is completely analogous to softlevel. The only difference here is that a second "boot" runlevel is defined instead of a second "default" runlevel.





Environment variables

Introduction

An environment variable is a named object that contains information used by one or more applications. By using environment variables one can easily change a configuration setting for one or more applications.

Important examples

The following table lists a number of variables used by a Linux system and describes their use. Example values are presented after the table.

Variable Description
PATH This variable contains a colon-separated list of directories in which the system looks for executable files. If a name is entered of an executable (such as ls, rc-update, or emerge) but this executable is not located in a listed directory, then the system will not execute it (unless the full path is entered as the command, such as /bin/ls).
ROOTPATH This variable has the same function as PATH, but this one only lists the directories that should be checked when the root-user enters a command.
LDPATH This variable contains a colon-separated list of directories in which the dynamical linker searches through to find a library.
MANPATH This variable contains a colon-separated list of directories in which the man command searches for the man pages.
INFODIR This variable contains a colon-separated list of directories in which the info command searches for the info pages.
PAGER This variable contains the path to the program used to list the contents of files through (such as less or more).
EDITOR This variable contains the path to the program used to change the contents of files with (such as nano or vi).
KDEDIRS This variable contains a colon-separated list of directories which contain KDE-specific material.
CONFIG_PROTECT This variable contains a space-delimited list of directories which should be protected by Portage during package updates.
CONFIG_PROTECT_MASK This variable contains a space-delimited list of directories which should not be protected by Portage during package updates.

Below is an example definition of all these variables:

CODE Example settings for the mentioned variables
PATH="/bin:/usr/bin:/usr/local/bin:/opt/bin:/usr/games/bin"
ROOTPATH="/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin"
LDPATH="/lib:/usr/lib:/usr/local/lib:/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.3"
MANPATH="/usr/share/man:/usr/local/share/man"
INFODIR="/usr/share/info:/usr/local/share/info"
PAGER="/usr/bin/less"
EDITOR="/usr/bin/vim"
KDEDIRS="/usr"
# Directories that are protected during package updates.
# Note the use of the \ (backslashes) on the end of the following lines which interprets to a single space-delimited line.
CONFIG_PROTECT="/usr/X11R6/lib/X11/xkb /opt/tomcat/conf \
                /usr/kde/3.1/share/config /usr/share/texmf/tex/generic/config/ \
                /usr/share/texmf/tex/platex/config/ /usr/share/config"
# Directories that are _not_ protected during package updates.
CONFIG_PROTECT_MASK="/etc/gconf"

Defining variables globally

The env.d directory

To centralize the definitions of these variables, Gentoo introduced the /etc/env.d/ directory. Inside this directory a number of files are available, such as 50baselayout, gcc/config-x86_64-pc-linux-gnu, etc. which contain the variables needed by the application mentioned in their name.

For instance, when gcc is installed, a file called gcc/config-x86_64-pc-linux-gnu was created by the ebuild which contains the definitions of the following variables:

FILE /etc/env.d/gcc/config-x86_64-pc-linux-gnuDefault gcc enabled environment variables for GCC 13
GCC_PATH="/usr/x86_64-pc-linux-gnu/gcc-bin/13"
LDPATH="/usr/lib/gcc/x86_64-pc-linux-gnu/13:/usr/lib/gcc/x86_64-pc-linux-gnu/13/32"
MANPATH="/usr/share/gcc-data/x86_64-pc-linux-gnu/13/man"
INFOPATH="/usr/share/gcc-data/x86_64-pc-linux-gnu/13/info"
STDCXX_INCDIR="g++-v13"
CTARGET="x86_64-pc-linux-gnu"
GCC_SPECS=""
MULTIOSDIRS="../lib64:../lib"

Other distributions might tell the system administrator to change or add such environment variable definitions in /etc/profile or other locations. Gentoo on the other hand makes it easy for the sysadmins (and for Portage) to maintain and manage the environment variables without having to pay attention to the numerous files that can contain environment variables.

For instance, when gcc is updated, the associated file(s) under /etc/env.d/gcc are updated too without requesting any administrative interaction.

There is still the occasionally where a system administrator is asked to set a certain environment variable system-wide. As an example, take the http_proxy variable. Instead of editing a file under the /etc/profile directory, create a file named /etc/env.d/99local and enter the definition in it:

FILE /etc/env.d/99localSetting a global environment variable
http_proxy="proxy.server.com:8080"

By using the same file for all customized environment variables, system administrators have a quick overview on the variables they have defined themselves.

env-update

Several files within the /etc/env.d directory add definitions to the PATH variable. This is not a mistake: when the env-update command is executed, it will append the several definitions before it atomically updates each environment variable, thereby making it easy for packages (or system administrators) to add their own environment variable settings without interfering with the already existing values.

The env-update script will append the values in the alphabetical order of the /etc/env.d/ files. The file names must begin with two decimal digits.

CODE Update order used by env-update
09sandbox    50baselayout     51dconf
     +------------+----------------+-----------+
CONFIG_PROTECT_MASK="/etc/sandbox.d /etc/gentoo-release /etc/dconf ..."

The concatenation of variables does not always happen, only with the following variables: ADA_INCLUDE_PATH, ADA_OBJECTS_PATH, CLASSPATH, KDEDIRS, PATH, LDPATH, MANPATH, INFODIR, INFOPATH, ROOTPATH, CONFIG_PROTECT, CONFIG_PROTECT_MASK, PRELINK_PATH, PRELINK_PATH_MASK, PKG_CONFIG_PATH, and PYTHONPATH. For all other variables the latest defined value (in alphabetical order of the files in /etc/env.d/) is used.

It is possible to add more variables into this list of concatenate-variables by adding the variable name to either COLON_SEPARATED or SPACE_SEPARATED variables (also inside an /etc/env.d/ file).

When executing env-update, the script will create all environment variables and place them in /etc/profile.env (which is used by /etc/profile). It will also extract the information from the LDPATH variable and use that to create /etc/ld.so.conf. After this, it will run ldconfig to recreate the /etc/ld.so.cache file used by the dynamical linker.

To notice the effect of env-update immediately after running it, execute the following command to update the environment. Users who have installed Gentoo themselves will probably remember this from the installation instructions:

root #env-update && source /etc/profile
Note
The above command only updates the variables in the current terminal, new consoles, and their children. Thus, if the user is working in X11, he needs to either type source /etc/profile in every new terminal opened or restart X so that all new terminals source the new variables. If a login manager is used, it is necessary to become root and restart the /etc/init.d/xdm service.
Important
It is not possible to use shell variables when defining other variables. This means things like FOO="$BAR" (where $BAR is another variable) are forbidden.

Defining variables locally

User specific

It might not be necessary to define an environment variable globally. For instance, one might want to add /home/my_user/bin and the current working directory (the directory the user is in) to the PATH variable but do not want all other users on the system to have that in their PATH too. To define an environment variable locally, use ~/.bashrc or ~/.bash_profile:

FILE ~/.bashrcExtending PATH for local usage
# A colon followed by no directory is treated as the current working directory
PATH="${PATH}:/home/my_user/bin:"

After logout/login, the PATH variable will be updated.

Session specific

Sometimes even stricter definitions are requested. For instance, a user might want to be able to use binaries from a temporary directory created without using the path to the binaries themselves or editing ~/.bashrc for the short time necessary.

In this case, just define the PATH variable in the current session by using the export command. As long as the user does not log out, the PATH variable will be using the temporary settings.

root #export PATH="${PATH}:/home/my_user/tmp/usr/bin"