s6-rc

From Gentoo Wiki
Jump to:navigation Jump to:search

s6-rc is a service manager for s6-based systems, i.e. a suite of programs that can start and stop services, both long-running daemons and one-time initialization scripts, in the proper order according to a dependency tree. It can be used as an init system component, with a role similar to that of OpenRC for sysvinit + OpenRC systems. A high level overview of s6-rc is available here. The package's documentation is provided in HTML format, and can be read on a text user interface using for example www-client/links. However, a man page port of the documentation is available: app-doc/s6-rc-man-pages.

Installation

USE flags

USE flags for sys-apps/s6-rc Service manager for the s6 supervision suite

Data provided by the Gentoo Package Database · Last update: 2024-10-24 09:38 More information about USE flags

Emerge

root #emerge --ask sys-apps/s6-rc
Important
s6-rc is currently on the testing branch. Users with systems on the stable branch will need to add the package to /etc/portage/package.accept_keywords (if using Portage) to be able to install it, and also add dev-libs/skalibs, dev-lang/execline and sys-apps/s6. While it is generally not advised to mix packages of stable and testing branches, the skarnet.org software stack only depends on the libc, so in this case it should be safe.

Configuration

Files

  • /run/s6-rc - Default live state directory.
  • /etc/s6-rc/compiled - Default pathname of the compiled service database used by s6-rc-init to initialize s6-rc.

Usage

s6-rc manages services. A service is usually thought of as being provided by a server program that runs as a long-lived process. In this case, the service is up while the program is running and ready to provide the service, and down otherwise. s6-rc calls this kind of service a longrun, and arranges to run the corresponding program as a supervised process using s6 tools.

A service can also be thought of as being provided after some sequence of actions is performed, without involving a long-lived process. For example, making a filesystem available by mounting it, or making a network interface ready for traffic by bringing it up and configuring it with a static address, could also be considered services. In this case the service is up after the corresponding actions are successfully performed. s6-rc calls this kind of service a oneshot, and arranges to perform the corresponding actions in a controlled, reproducible environment. This environment is also set up by an s6 supervision tree even though oneshots do not involve supervised processes. For a oneshot, a transition to down state might mean just doing nothing and 'forgetting' that the corresponding actions were performed, so a subsequent transition to up would simply redo them. Or it might mean performing some sort of inverse procedure, like e.g. unmounting a filesystem or bringing down a network interface.

Services can also depend on other services. This generally means that a service can only transition to and remain in up state only as long as all of the services it depends on are also in up state. The dependencies of a service might in turn have dependencies themselves, creating dependency chains. s6-rc supports dependency specifications and arranges to perform service state transitions accordingly.

Compiled service databases

Performing requested state transitions for a set of services submitted to s6-rc (called the selection) requires it to compute a dependency graph, and then use it to compute, from the full set of managed services, the complete subset of those that need state transitions (called the closed selection), so that the requested operation can be performed while also honoring all dependency specifications. For this purpose, s6-rc stores its managed longruns and oneshots in a compiled service database. This database is a directory that contains a set of files and subdirectories with formats optimized for efficient computation of dependency graphs and closed selections. A compiled database is created with the s6-rc-compile program. It cannot be modified except by using the s6-rc-bundle program (see live updates to the service database), and can be placed in a read-only filesystem.

s6-rc-compile accepts the (absoulte or relative to its working directory) pathname of the database to create, and a nonempty set of directory pathnames. These directories must contain service definitions in a format called s6-rc-compile's source format. A service definition is a subdirectory, or symbolic link to directory, that must contain a regular file named type. This file specifies one of the s6-rc's supported service types: longrun, oneshot and bundle (see service bundles). The contents of a service definition directory depend on the service's type. The name of the directory defines the service name that is used by all s6-rc programs to refer to the service. Service names cannot be duplicated and cannot contain a slash ('/') or a newline; they can contain spaces and tabs, but using anything other than alphanumerical characters, underscores ('_') and dashes ('-') is discouraged.

Currently, s6-rc-compile also unconditionally adds two s6-rc support services to each compiled service database: s6rc-oneshot-runner, a longrun that is used to implement oneshots (see oneshot definitions), and s6rc-fdholder, a longrun that is used to implement pipelines (see longrun pipelining). Service names that start with s6rc- or s6-rc- are reserved, and s6-rc-compile fails with an error if it finds one in the submitted service definitions. s6-rc-compile also ignores, in any of the supplied directories, files that are not directories or symbolic links to directory, and files with names that start with a dot ('.'). For the full description of s6-rc-compile, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

The examples/source directory in the package's /usr/share/doc subdirectory contains an example set of service definitions in s6-rc-compile's source format for a few programs, taken from a working Linux system running Busybox and skarnet.org packages (without private information and without the supporting files elsewhere, like e.g. configuration files in /etc).

Longrun definitions

A longrun definition in s6-rc-compile's source format is a variation of an s6 service directory. It must contain a file named run, and can contain files named finish, notification-fd, nosetsid, timeout-finish, and timeout-kill, with the same meaning and required content as for s6. As of version 0.5.0.0, a longrun definition can also contain max-death-tally and down-signal files, which s6 added support for in versions 2.7.1.0 and 2.7.2.0, respectively. On the other hand, down files are ignored because longruns are managed using s6-rc tools that take care of these files themselves. And log subdirectories or symbolic links are ignored too, because their functionality is replaced by pipelines (see longrun pipelining).

Important
Support for max-death-tally and down-signal files in longrun definitions was actually introduced in s6-rc version 0.4.1.0, but a bug prevented it from working. The bug was fixed in version 0.5.0.0

The type file of a longrun definition must contain the word longrun followed by a newline. A longrun definition can also contain optional, regular files named dependencies (see service dependencies), timeout-up and timeout-down (see managing services). If the service is a member of a pipeline, the longrun definition must also contain a regular file named producer-for or consumer-for, and can contain an optional, regular file named pipeline-name. And finally, a longrun definition can also contain subdirectories named data and env, for use by the run and finish files. Their content is ignored by s6-rc-compile. env is customarily an environment directory used by an s6-envdir invocation in run or finish, and data is customarily used for other support files. For example, to hold rules directories or rules files for tools like s6-fdholderd and s6-ipcserver-access, or to hold a check file for s6-notifyoncheck.

Longrun definitions are 'compiled' by s6-rc-compile to actual s6 service directories. The program automatically sets appropriate permissions for files with meaning defined by s6 (e.g. 0755 for run and finish, 0644 for notification-fd, etc.), so s6-rc service definitions do not need to have them set correctly. It only matters that s6-rc-compile's effective user is allowed to read them. The compiled s6 service directories are copied to the live state directory and linked from the scan directory of a running s6-svscan process by the s6-rc-init program (see initializing s6-rc). The data and env subdirectories are copied verbatim to the s6 service directory, and down files are created or removed in it as needed by s6-rc programs. For the full description of the longrun definition source format, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory.

Example definition in source format for a longrun named test-longrun:

user $ls -l test-longrun
total 16
-rw-r--r-- 1 user user 130 Mar 20 12:00 finish
-rw-r--r-- 1 user user   2 Mar 20 12:00 notification-fd
-rw-r--r-- 1 user user  94 Mar 20 12:00 run
-rw-r--r-- 1 user user   8 Mar 20 12:00 type
FILE test-longrun/type
longrun
FILE test-longrun/run
#!/bin/execlineb -P
foreground { echo test-daemon starting }
fdmove -c 2 1
test-daemon --s6=3
FILE test-longrun/notification-fd
3
FILE test-longrun/finish
#!/bin/execlineb -S0
ifelse { test $1 = 256 } {
   echo test-daemon killed by signal $2
}
   echo test-daemon exited with code $1

This service's transition to up state spawns a long-lived process that executes a hypothetical test-daemon program. It is assumed that it supports the s6 readiness notification protocol when passed an --s6 option. The notification channel's file descriptor is 3, so there is a notification-fd file specifying that. The type file indicates that this service is a longrun, and the rest of the files are standard s6 service directory files. Messages are printed to run's and finish's standard output, so that service events can be tracked.

Oneshot definitions

A oneshot definition in s6-rc-compile's source format is a directory that must contain a regular file named up, and can optionally contain a regular file named down. This files encode the actions that must be performed to transition the service to up and down state, respectively. When s6-rc is asked to start or stop the service, it behaves as if it executed an s6-sudo -e supervision-socket up or s6-sudo -e supervision-socket down command, respectively, where supervision-socket would be the UNIX domain socket of an s6-sudod process that runs with the same environment that is set up for longruns, and has been invoked as s6-sudod execlineb -P. execlineb is the script parser and launcher from the execline package (dev-lang/execline). This means that the up and down files can have execline syntax (quoted strings, {}-blocks, backslash sequences, #-coments, etc.), and that it is possible to have them contain full execline scripts. However, because a program name that can be found by PATH search, or a program's absolute pathname, followed by program arguments, is a valid execline script, it is possible to use shell scripts for the oneshot's actions by simply invoking them in the up and down files.

An absent down file is equivalent to an empty down file. Because invoking execlineb with an empty file does nothing but making it exit with a 0 code, stopping the corresponding service performs no actions other than updating s6-rc's notion of the service state. The type file of a oneshot definition must contain the word oneshot followed by a newline. A oneshot definition can also contain regular files named dependencies (see service dependencies), timeout-up and timeout-down (see managing services). For the full description of the oneshot definition source format, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory.

Oneshot definitions are 'compiled' by s6-rc-compile to an array of strings, suitable as the 'argv' argument of a POSIX execve() call. s6-rc-compile actually performs execlineb-style parsing itself using the execline package's library, libexecline. At the time of a service state transition, what s6-rc actually does is perform an s6-sudo invocation with the -e option and the UNIX domain socket of an internal support service named s6rc-oneshot-runner. s6rc-oneshot-runner spawns a long-lived s6-sudod process; when a connection is made by an s6-sudoc client to this process' socket, an s6-rc internal program, s6-rc-oneshot-run, is executed with the arguments supplied by the client. s6-rc-oneshot-run accepts an action name ('up' or 'down') and the numerical encoding of a oneshot, that it uses to retrieve the corresponding compiled array of strings from the service database, and makes the execve() call that actually performs the oneshot's actions. This ensures that the s6-sudoc - s6-sudod mechanism is only used to run s6-rc-oneshot-run, so that only actions explicitly encoded in the database are executed. Because the s6rc-oneshot-runner service is an s6-rc longrun, the corresponding s6-sudod process is supervised. Therefore, it runs with the environment set up by the s6 supervision tree used for longruns, which is inherited by its s6-rc-oneshot-run child process. This ensures that oneshot's actions are always performed in this known environment, and using s6-sudo's -e option ensures that it is not accidentally modified by the s6-sudoc client's environment. s6-rc-compile automatically makes s6rc-oneshot-runner a direct dependency of every oneshot. For further information about s6-rc-oneshot-run, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example definition in source format for a oneshot named test-oneshot:

user $ls -l test-oneshot
total 12
-rw-r--r-- 1 user user 118 Mar 20 12:00 down
-rw-r--r-- 1 user user   8 Mar 20 12:00 type
-rw-r--r-- 1 user user 119 Mar 20 12:00 up
FILE test-oneshot/type
oneshot
FILE test-oneshot/up
foreground { echo "Calling /home/user/s6-rc-scripts/real-script.sh..." }
/home/user/s6-rc-scripts/real-script.sh start
FILE test-oneshot/down
foreground { echo "Calling /home/user/s6-rc-scripts/real-script.sh..." }
/home/user/s6-rc-scripts/real-script.sh stop

This service's transition to up state executes a hypothetical real-script.sh shell script, located in /home/user/s6-rc-scripts, with a start argument. Its transition to down state executes the same script, but with a stop argument instead. This shows how to use shell scripts, or any external executable file, with s6-rc. Messages are printed before the script's invocation so that service events can be tracked. The type file indicates that this service is a oneshot. The oneshot's actions are actually encoded in real-script.sh:

FILE /home/user/s6-rc-scripts/real-script.sh
#!/bin/sh
echo real-script.sh invoked with arguments $@
case "$1" in
   start) echo Performing start actions ;;
   stop) echo Performing stop actions ;;
   *) echo Unrecognized argument ;;
esac

Service bundles

A service bundle is just a named group of services, created for administrative purposes. Because a bundle is also created with a service definition directory supplied to s6-rc-compile, and bundle names can be used in most places that expect a service name, the term service is loosely used to also refer to bundles, and the term atomic service is used to refer to either longruns or oneshots, but not bundles, when the distinction matters. In particular, bundles can be used to implement runlevel-like functionality if so desired, combined with the s6-rc -pu change command (see service dependencies). For examples of this, see live updates to the service database.

A service bundle definition in s6-rc-compile's source format must contain a regular file named contents. The type file must contain the word bundle followed by a newline. The contents file must contain a list of service names, one per line; the bundle will cointain all the corresponding services. Whitespace at the beginning of a line is ignored, but trailing whitespace is not. Lines that start with a hash sign ('#') are treated as comments and ignored. Names that refer to other bundles can be used in contents; s6-rc-compile interprets them as referring to the set of atomic services contained in the named bundle. For the full description of the bundle definition source format, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory.

Example definition in source format for a service bundle named test-bundle:

user $ls -l test-bundle
total 8
-rw-r--r-- 1 user user 26 Mar 20 12:00 contents
-rw-r--r-- 1 user user  7 Mar 20 12:00 type
FILE test-bundle/type
bundle
FILE test-bundle/contents
test-longrun
test-oneshot

This bundle is comprised by atomic services test-longrun and test-oneshot. The type file indicates that this service is not an atomic service, but a bundle.

Extracting information from a compiled database

Because the format of an s6-rc compiled service database is not human-readable, the package provides the s6-rc-db program for analyzing a compiled database, extracting information from it, and printing it in a user-friendly format. By default, the database read by s6-rc-db is the one currently associated with live state directory /run/s6-rc (see initializing s6-rc), unless it is passed an -l option followed by the (absolute or relative to the working directory) pathname of a different live state directory, or it is passed a -c option followed by the pathname of the database.

s6-rc-db accepts a subcommand that tells it what to print. Some of these subcommands are:

  • help: prints a help message.
  • list: lists service names that satisfy some condition specified by the subcommand's argument:
    • all (i.e. s6-rc-db list all): lists the names of all atomic services and service bundles contained in the database.
    • services: lists the names of all atomic services contained in the database, i.e. the list does no include service bundle names.
    • bundles: lists the names of all service bundles contained in the database.
    • longruns: lists the names of all longruns contained in the database. This includes the s6-rc supporting services automatically added by s6-rc-compile.
    • oneshots: lists the names of all oneshots contained in the database.
  • type: displays the type of the service ('oneshot', 'longrun' or 'bundle') corresponding to the name supplied as the subcommand's argument.
  • contents: displays the list of atomic services contained in the service bundle corresponding to the name supplied as the subcommand's argument.
  • atomics: displays the fully resolved list of atomic services represented by a list of service names supplied as the subcommand's arguments. Atomic service names in the list will be displayed in the output as-is, and service bundle names will be replaced in the output by the names of its contained atomic services. Each service name will be listed only once, with no repetitions. This computation of an atomic service list is also performed by the s6-rc program, see managing services.
  • script: displays the actions executed during state transitions of the oneshot corresponding to the name supplied as the subcommand's argument. If a -u option is passed to s6-rc-db (before the subcommand name), the output is the string sequence produced by s6-rc-compile's parsing of the up file from the oneshot's definition in s6-rc-compile's source format. If a -d option is passed to s6-rc-db, the output is the string sequence produced by s6-rc-compile's parsing of the down file from the oneshot's definition. If there is neither a -u option nor a -d option, s6-rc-db behaves as if the -u option had been specified. Each string in the sequence is terminated by a null character, so it is usually necessary to pipe s6-rc-db's output to something like xargs -0.

Other s6-rc-db subcommands can be used to display dependency information (see service dependencies), state transition timeouts (see managing services) and pipeline information (see longrun pipelining). For the full description of s6-rc-db, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example directories srv-src1 and srv-src2 containing service definition subdirectories in s6-rc-compile's source format:

user $ls -l srv-src*
srv-src1:
total 12
drwxr-xr-x 2 user user 4096 Mar 20 12:00 test-bundle
drwxr-xr-x 2 user user 4096 Mar 20 12:00 test-longrun
drwxr-xr-x 2 user user 4096 Mar 20 12:00 test-oneshot

srv-src2:
total 12
drwxr-xr-x 2 user user 4096 Mar 20 12:00 all-services
drwxr-xr-x 2 user user 4096 Mar 20 12:00 another-longrun
drwxr-xr-x 2 user user 4096 Mar 20 12:00 another-oneshot

Directory srv-src1 contains definitions for the example services described in previous sections. Directory srv-src2 contains another set of service definitions. all-services is a bundle:

FILE srv-src2/all-services/type
bundle
FILE srv-src2/all-services/contents
test-bundle
another-longrun
another-oneshot

Note that the contents file contains names of both atomic services (another-longrun and another-oneshot) and bundles (test-bundle).

An s6-rc compiled service database test-database1 can be created from srv-src1 and srv-src2 by invoking s6-rc-compile with their pathnames as arguments:

user $s6-rc-compile -v 3 -u $(id -u) test-database1 srv-src1 srv-src2
s6-rc-compile: info: parsing srv-src1/test-bundle
s6-rc-compile: info: test-bundle has type bundle
s6-rc-compile: info: parsing srv-src1/test-longrun
s6-rc-compile: info: test-longrun has type longrun
s6-rc-compile: info: parsing srv-src1/test-oneshot
s6-rc-compile: info: test-oneshot has type oneshot
s6-rc-compile: info: parsing srv-src2/all-services
s6-rc-compile: info: all-services has type bundle
s6-rc-compile: info: parsing srv-src2/another-longrun
s6-rc-compile: info: another-longrun has type longrun
s6-rc-compile: info: parsing srv-src2/another-oneshot
s6-rc-compile: info: another-oneshot has type oneshot
s6-rc-compile: info: making bundles for pipelines
s6-rc-compile: info: resolving bundle names
s6-rc-compile: info: converting bundle array
s6-rc-compile: info: resolving service names
s6-rc-compile: info: converting service dependency array
s6-rc-compile: info: checking database correctness
s6-rc-compile: info: writing compiled information to test-database1
s6-rc-compile: info: writing test-database1/n
s6-rc-compile: info: writing test-database1/resolve.cdb
s6-rc-compile: info: writing test-database1/db
s6-rc-compile: info: writing test-database1/servicedirs

The -u option is needed so that the s6-rc program can be used to manage oneshots by the nonprivileged user that created the database. The -v 3 option makes the output of s6-rc-compile more verbose. The resulting database is a directory:

user $ls -l test-database1
total 16
-rw-r--r-- 1 user user  646 Mar 20 12:05 db
-rw-r--r-- 1 user user    0 Mar 20 12:05 lock
-rw-r--r-- 1 user user   24 Mar 20 12:05 n
-rw-r--r-- 1 user user 2397 Mar 20 12:05 resolve.cdb
drwxr-xr-x 6 user user 4096 Mar 20 12:05 servicedirs

The servicedirs subdirectory contains the s6 service directories generated for all longruns, including those of automatically included s6-rc supporting services:

user $ls -l test-database1/servicedirs
total 16
drwxr-xr-x 2 user user 4096 Mar 20 12:05 another-longrun
drwxr-xr-x 3 user user 4096 Mar 20 12:05 s6rc-fdholder
drwxr-xr-x 3 user user 4096 Mar 20 12:05 s6rc-oneshot-runner
drwxr-xr-x 2 user user 4096 Mar 20 12:05 test-longrun
user $ls -l test-database1/servicedirs/test-longrun
total 12
-rwxr-xr-x 1 user user 130 Mar 20 12:05 finish
-rw-r--r-- 1 user user   2 Mar 20 12:05 notification-fd
-rwxr-xr-x 1 user user  94 Mar 20 12:05 run

This shows that the run and finish files are executable, as expected by s6-supervise, even if the corresponding ones in the s6-rc service definition were not. Listing all service names in the resulting compiled database:

user $s6-rc-db -c test-database1 list all | sort
all-services
another-longrun
another-oneshot
s6rc-fdholder
s6rc-oneshot-runner
test-bundle
test-longrun
test-oneshot

This shows that the created database contains the union of the services defined in subdirectories of srv-src1, and the services defined in subdirectories of srv-src2. Listing service names by type:

user $s6-rc-db -c test-database1 list longruns | sort
another-longrun
s6rc-fdholder
s6rc-oneshot-runner
test-longrun
user $s6-rc-db -c test-database1 list oneshots | sort
another-oneshot
test-oneshot
user $s6-rc-db -c test-database1 list bundles | sort
all-services
test-bundle

Displaying the type of different services:

user $s6-rc-db -c test-database1 type all-services
bundle
user $s6-rc-db -c test-database1 type test-longrun
longrun
user $s6-rc-db -c test-database1 type another-oneshot
oneshot

Displaying the contents of service bundles:

user $s6-rc-db -c test-database1 contents test-bundle | sort
test-longrun
test-oneshot
user $s6-rc-db -c test-database1 contents all-services | sort
another-longrun
another-oneshot
test-longrun
test-oneshot

This shows that, because the contents file in the definition of bundle all-services named the bundle test-bundle, all atomic services contained in the latter became members of the former. Displaying the string sequence compiled from the up and down files of a oneshot:

user $s6-rc-db -uc test-database1 script test-oneshot | xargs -0 printf '"%s"\n'
"foreground"
" echo"
" Calling /home/user/s6-rc-scripts/real-script.sh..."
""
"/home/user/s6-rc-scripts/real-script.sh"
"start"
user $s6-rc-db -dc test-database1 script test-oneshot | xargs -0 printf '"%s"\n'
"foreground"
" echo"
" Calling /home/user/s6-rc-scripts/real-script.sh..."
""
"/home/user/s6-rc-scripts/real-script.sh"
"stop"

Notice the result of execlineb-style processing: the quoted strings that were arguments of the echo utility become single elements in the compiled sequence of strings even if they contain whitespace, and {}-blocks become a sequence of arguments that start with a space, terminated by an empty string, a format recognized by all execline programs, including foreground.

Resolving some atomic service sets:

user $s6-rc-db -c test-database1 atomics another-oneshot test-bundle | sort
another-oneshot
test-longrun
test-oneshot

This shows that because another-oneshot is an atomic service, it is included as-is in the output, and because test-bundle is a bundle, it is replaced in the output by the set of atomic services it contains.

user $s6-rc-db -c test-database1 atomics another-longrun all-services | sort
another-longrun
another-oneshot
test-longrun
test-oneshot

This shows that bundle all-services is replaced by the set of atomic services it contains, and because that set includes another-longrun, this service is listed only once.

Initializing s6-rc

s6-rc must be initialized with a compiled service database and an s6 supervision tree, that is used to supervise longruns and to provide the environment used for performing oneshots' actions. The program that performs this initialization is s6-rc-init.

s6-rc-init accepts the absolute pathname of a scan directory, that must have a corresponding s6-svscan process already running. The service database is assumed to be /etc/s6-rc/compiled. A different (absolute) pathname can be specified with a -c option. s6-rc-init creates a live state directory that associates the s6-rc service database with the s6 supervision tree, and keeps track of service states. The live state directory must be in a read-write filesystem, and defines an s6-rc 'instance'. The live state directory is /run/s6-rc by default; a different (absolute) pathname can be specified with an -l option. /run/s6-rc is actually a symbolic link to a directory created by s6-rc-init in /run; this is necessary for the operation of the s6-rc-update program (see live updates to the service database). Similarly, if the -l option is used, s6-rc-init it will create both a symbolic link with that pathname, and a subdirectory in the directory containing the symlink.

A compiled database that becomes associated with a live state directory by using s6-rc-init, s6-rc-update or s6-rc-format-upgrade is said to be live. A database that is not live can be freely moved around in the filesystem, but once it becomes live, it must not move anymore and must not be deleted. A database stops being live if it is deassociated from the live state directory by s6-rc-update or s6-rc-format-upgrade, or if the corresponding s6 supervision torn down (e.g. by s6-svscanctl -t).

The s6-rc live state directory also contains a copy of the compiled s6 service directories of all longruns in the service database. s6-rc-init creates a symbolic link to each of them in the specified scan directory, so that the corresponding program can be run as a supervised process, and performs the equivalent an s6-svscanctl -a command to make s6-svscan perform a scan. The initial state of all atomic services after the invocation of s6-rc-init is down. In the case of longruns, s6-rc-init enforces this by creating down files in each of their service directories, so that their s6-supervise processes do not execute the run file. s6-rc-init also accepts a -p option followed by a prefix, that prepends the specified prefix to the name of every longrun when creating the symlinks in the scan directory. For example, if a longrun's name is name, s6-rc-init -p instance-1: will name the symlink instance-1:name. This allows the sharing of a single s6 supervision tree among different 'instances' of s6-rc, i.e. among s6-rc-managed services associated with different live state directories.

Support services s6rc-oneshot-runner and s6rc-fdholfer use UNIX domain sockets, and s6 access control tools to decide whether to grant or refuse service to clients that connect to these sockets. By default, they only grant access to root, but this can be modified by passing -u and -g options to s6-rc-compile, followed by a comma-separated list of numerical user IDs and group IDs, respectively. These options are used to construct rules directories: user IDs specified with -u are used for the uid subdirectory, and group IDs specified with -g are used for the gid subdirectory. If any of these options are used, root's UID or the GID of a group it is a member of must be explicitly specified to grant it access. For further information about s6-rc-compile, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example s6 scan directory containing a supervised logger with a FIFO, that logs supervision tree messages in a similar way to that of the catch-all logger of an s6 and s6-rc-based init system:

user $ls -ld * .s6-svscan
drwxr-xr-x 2 user user 4096 Mar 30 12:00 logger
drwxr-xr-x 2 user user 4096 Mar 30 12:00 .s6-svscan
user $ls -l logger
total 4
prw------- 1 user user   0 Mar 30 12:00 fifo
-rwxr-xr-x 1 user user 144 Mar 30 12:00 run
FILE logger/run
#!/bin/execlineb -P
redirfd -w 1 /dev/null
redirfd -rnb 0 fifo
foreground {
   fdmove -c 1 2
   echo Starting logger...
}
s6-log t ../../logdir

Starting the supervision tree, assuming that the working directory is the scan directory:

user $redirfd -wnb 1 logger/fifo s6-svscan &
[1] 2460
Starting logger...
user $ps xf -o pid,ppid,pgrp,euser,args
 PID  PPID  PGRP EUSER    COMMAND
...
2130  2123  2130 user     -bash
2460  2130  2460 user      \_ s6-svscan
2461  2460  2460 user          \_ s6-supervise logger
2462  2461  2462 user              \_ s6-log t ../../logdir
...

Initializing s6-rc with this supervision tree and the example database test-database1 from section "extracting information from a compiled database", assuming that it is located in /home/user:

user $s6-rc-init -c /home/user/test-database1 -l /home/user/live "$(pwd)"
user $ls -l
total 12
lrwxrwxrwx 1 user user   43 Mar 30 12:10 another-longrun -> /home/user/live/servicedirs/another-longrun
drwxr-xr-x 4 user user 4096 Mar 30 12:05 logger
lrwxrwxrwx 1 user user   41 Mar 30 12:10 s6rc-fdholder -> /home/user/live/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   47 Mar 30 12:10 s6rc-oneshot-runner -> /home/user/live/servicedirs/s6rc-oneshot-runner
lrwxrwxrwx 1 user user   40 Mar 30 12:10 test-longrun -> /home/user/live/servicedirs/test-longrun

This creates the live state directory in /home/user, and, as shown, symbolic links in the scan directory.

user $ls -l ../live
lrwxrwxrwx 1 user user 22 Mar 30 12:10 ../live -> live:s6-rc-init:peEEK3

This shows that /home/user/live is actually a symbolic link to subdirectory live:s6-rc-init:peEEK3 of /home/user:

user $ls -l ../live:s6-rc-init:peEEK3
total 8
lrwxrwxrwx 1 user user   25 Mar 30 12:10 compiled -> /home/user/test-database1
-rw-r--r-- 1 user user    0 Mar 30 12:10 prefix
lrwxrwxrwx 1 user user   15 Mar 30 12:10 scandir -> /home/user/scan
drwxr-xr-x 6 user user 4096 Mar 30 12:10 servicedirs
-rw-r--r-- 1 user user    6 Mar 30 12:10 state

Updated supervision tree:

user $ps xf -o pid,ppid,pgrp,euser,args
 PID  PPID  PGRP EUSER    COMMAND
...
2130  2123  2130 user     -bash
2460  2130  2460 user      \_ s6-svscan
2461  2460  2460 user          \_ s6-supervise logger
2462  2461  2462 user          |   \_ s6-log t ../../logdir
2471  2460  2460 user          \_ s6-supervise test-longrun
2472  2460  2460 user          \_ s6-supervise s6rc-oneshot-runner
2473  2460  2460 user          \_ s6-supervise s6rc-fdholder
2474  2460  2460 user          \_ s6-supervise another-longrun
...

This shows that none of the longruns is in up state: only their supervisor is running. Service directory of test-longrun:

user $ls -l ../live/servicedirs/test-longrun
total 20
-rw-r--r-- 1 user user    0 Mar 30 12:10 down
drwx-ws--T 2 user user 4096 Mar 30 12:10 event
-rwxr-xr-x 1 user user  130 Mar 30 12:10 finish
-rw-r--r-- 1 user user    2 Mar 30 12:10 notification-fd
-rwxr-xr-x 1 user user   94 Mar 30 12:10 run
drwx------ 2 user user 4096 Mar 30 12:10 supervise

This shows that s6-rc-init created a down file. Subdirectories event and supervise are created by the corresponding s6-supervise process (process 2471 in this example). From s6's point of view, the service is down:

user $s6-svstat test-longrun
down (exitcode 0) 337 seconds, ready 337 seconds

Another example using prefixes:

user $mv srv-src2/all-services srv-src2/.all-services
user $s6-rc-compile -u $(id -u) db-instance-1 srv-src1
user $s6-rc-compile -u $(id -u) db-instance-2 srv-src2

Service definitions in directories srv-src1 and srv-src2 are now compiled to separate databases, db-instance-1 and db-instance-2. The definition directory of bundle all-services was renamed to start with a dot so that s6-rc-compile ingores it. After changing to scan directory /home/user/scan and starting a supervision tree again with command redirfd -wnb 1 logger/fifo s6-svscan &:

user $s6-rc-init -c /home/user/db-instance-1 -p instance-1: -l /home/user/live1 /home/user/scan
user $s6-rc-init -c /home/user/db-instance-2 -p instance-2: -l /home/user/live2 /home/user/scan
user $ls -l *
total 24
lrwxrwxrwx 1 user user   42 Apr 29 12:10 instance-1:s6rc-fdholder -> /home/user/live1/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   48 Apr 29 12:10 instance-1:s6rc-oneshot-runner -> /home/user/live1/servicedirs/s6rc-oneshot-runner
lrwxrwxrwx 1 user user   41 Apr 29 12:10 instance-1:test-longrun -> /home/user/live1/servicedirs/test-longrun
lrwxrwxrwx 1 user user   44 Apr 29 12:11 instance-2:another-longrun -> /home/user/live2/servicedirs/another-longrun
lrwxrwxrwx 1 user user   42 Apr 29 12:11 instance-2:s6rc-fdholder -> /home/user/live2/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   48 Apr 29 12:11 instance-2:s6rc-oneshot-runner -> /home/user/live2/servicedirs/s6rc-oneshot-runner
drwxr-xr-x 4 user user 4096 Apr 29 12:05 logger

This shows that the symbolic links to the s6 service directories have different prefixes depending on which live state directory the corresponding logrun is associated with. Live state directory /home/user/live1 is associated with compiled service database db-instance-1, and live state directory /home/user/live2 is associated with compiled service database db-instance-2. Each of these can be thought as a different s6-rc 'instance', even if they share the same s6 supervision tree and scan directory. The corresponding services are independent:

user $ls -l ../live*
lrwxrwxrwx 1 user user   23 Apr 29 12:10 ../live1 -> live1:s6-rc-init:ISOd2e
lrwxrwxrwx 1 user user   23 Apr 29 12:11 ../live2 -> live2:s6-rc-init:AtGbln

'../live1:s6-rc-init:ISOd2e':
total 12
lrwxrwxrwx 1 user user   24 Apr 29 12:10 compiled -> /home/user/db-instance-1
-rw-r--r-- 1 user user   11 Apr 29 12:10 prefix
lrwxrwxrwx 1 user user   15 Apr 29 12:10 scandir -> /home/user/scan
drwxr-xr-x 5 user user 4096 Apr 29 12:10 servicedirs
-rw-r--r-- 1 user user    4 Apr 29 12:10 state

'../live2:s6-rc-init:AtGbln':
total 12
lrwxrwxrwx 1 user user   24 Apr 29 12:11 compiled -> /home/user/db-instance-2
-rw-r--r-- 1 user user   11 Apr 29 12:11 prefix
lrwxrwxrwx 1 user user   15 Apr 29 12:11 scandir -> /home/user/scan
drwxr-xr-x 5 user user 4096 Apr 29 12:11 servicedirs
-rw-r--r-- 1 user user    4 Apr 29 12:11 state
user $s6-rc-db -l ../live1 list all | sort
s6rc-fdholder
s6rc-oneshot-runner
test-bundle
test-longrun
test-oneshot
user $s6-rc-db -l ../live2 list all | sort
another-longrun
another-oneshot
s6rc-fdholder
s6rc-oneshot-runner

Managing services

The s6-rc package provides one program, also named s6-rc, to perform most of the service management tasks after initialization has been done. s6-rc accepts a subcommand that specifies what to do. The subcommands are:

  • help: prints a help message.
  • list: displays the fully resolved list of atomic services represented by a list of service names supplied as the subcommand's arguments, just like s6-rc-db atomics (see extracting information from a compiled database). This list of atomic services is called the selection.
  • change: computes the selection in the same way as the list subcommand, and performs state transitions for all atomic services in the resulting list. If a -u option ('up') is passed to s6-rc (before the subcommand name), the requested action is to start all services in the selection, i.e. transition them to up state. If a -d option ('down') is passed to s6-rc, the requested action is to stop all services in the selection, i.e. transition them to down state. If there is neither a -u nor a -d option, s6-rc behaves as if the -u option had been supplied. If an -n option ('dry run') followed by an integral value is passed to s6-rc, state transitions are simulated instead of actually performed (see later).
  • diff: checks the consistency between s6-rc's view of the state of all longruns (as recorded in the live state directory), and s6's view (as recorded in the state file of the service's supervise control directory). This subcommand should normally report nothing, but directly using some s6 programs with s6-rc-managed longruns can cause state discrepancies. For example, if a longrun is in down state, and an s6-svc -u command with its symlink in the scan directory is used instead of an s6-rc -u change command with its service name, not only would that run the underlying program (without also starting services that were specified as dependencies), but it would also make s6 view the process' state as up (as reported by s6-svstat), while s6-rc would still view the corresponding service as down. The diff subcommand prints a line for each longrun it finds that has a state inconsistency, prepended by a plus sign ('+') if s6 considers it to be up, and a minus plus sign ('-') if s6 considers it to be down.
  • listall: displays the computed closed selection, see service dependencies).

Unlike s6-rc-db, which operates exclusively on a compiled database, s6-rc operates on the live state directory. Therefore, it has knowledge of not only the associated database, but also of service states. The live state directory is /run/s6-rc by default, but a different one can be used by passing an -l option to s6-rc with the directory's pathname (absoulte or relative to s6-rc's working directory). s6-rc-db can be used both with live and non-live databases, but in the latter case, the database's pathname has to be specified with a -c option.

Because s6-rc is aware of service states, it accepts an -a ('active') option for both its list and change subcommands, that adds the current set of atomic services in up state to the initial selection derived from the subcomannd's arguments. s6-rc also accepts the -d option for its list subcommand, but the meaning in this case is to complement the initial selection derived from the subcomannd's arguments (plus the active services if the -a option is also present), i.e. the final selection is the set of all atomic services in the live state directory's associated database that are not in the initial selection. For completeness, s6-rc also accepts an explicit -u option for its list subcommand, that does nothing. For both list and change, the supplied list of service names can be empty, meaning an (initial) empty selection. s6-rc list is mostly useful in conjunction with the -a or -d options.

s6-rc also has a notion of transition success and transition failure. For a longrun, if its underlying program supports the s6 readiness notification protocol, transition to up state is considered successful if the supervised process is up and ready before the expiry of a timeout. Otherwise, transition to up is considered successful if the supervised process is up before the expiry of a timeout, i.e. when its supervisor has spawned a run child. The s6-rc -u change command uses an s6-svc -uwU or s6-svc -uwu command, respectively, to start the supervised process, so it will wait until it receives a notification of the relevant event from s6-supervise. Transition to down state is considered successful if the supervised process is really down before the expiry of a timeout, i.e. if it is down and the finish file, if present, was executed and has exited (or got killed by s6-supervise because of a timeout-finish file). The s6-rc -d change command uses an s6-svc -dwD command to stop the supervised process, so it will wait until it receives the really down notification from s6-supervise. If a transition to up fails, the s6-rc -u change command then uses an s6-svc -d command to bring the supervised process to a known down state. For a oneshot, transition to up or down state is considered successful if the equivalent of executing the up or down file from its service definition, respectively, that is performed by the s6-rc-oneshot-run program, exits with code 0 before the expiry of a timeout (see oneshot definitions). The s6-rc change command waits for s6-rc-oneshot-run to finish. For the full description of the s6-rc program, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Supporting service s6rc-oneshot-runner uses a rules directory to decide whether to grant or refuse service to clients that connect to its UNIX domain socket. The s6-rc program executes a sudo -e command with s6rc-oneshot-runner's socket to start or stop a oneshot, so s6-rc's effective user must be allowed by this service's rules directory. Setup of s6rc-oneshot-runner's rules directory can be done using s6-rc-compile's -u and -g options, see initializing s6-rc.

The timeout for a state transition can be configured by including regular files named timeout-up and timeout-down in the definition of an atomic service submitted to s6-rc-compile, that must contain an integral time value in milliseconds. The former specifies the timeout for the transition to up, and the latter, for the transition to down. If any of these files is absent or contains the value 0, the corresponding timeout is infinite, i.e. the s6-rc change command will wait forever for the notification of the relevant event from s6-supervise in the case of a longrun, and will wait forever for s6-rc-oneshot-run to exit, in the case of a oneshot. For further information about configuration of state transition timeouts, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory. The timeout subcommand of the s6-rc-db program prints the state transition timeouts. The up timeout is printed if a -u option is passed to s6-rc-db, and the down timeout is printed if a -d option is passed to s6-rc-db. If there is neither a -u option nor a -d option, s6-rc-db behaves as if the -u option had been specified. For further information, please consult s6-rc-db's HTML documentation in the package's /usr/share/doc subdirectory.

s6-rc change calls the internal program s6-rc-dryrun instead of s6-svc or s6-sudo when the -n option is specified, which makes s6-rc print to standard output what the command would have done without the -n option. The option must be followed by integral time value in milliseconds; s6-rc-dryrun will sleep for the specified time before exiting, to simulate an s6-svc or s6-sudo invocation that does not complete immediately. For further information about s6-rc-dryrun, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example usage of s6-rc to print a list of all atomic services in the compiled database that is live, assuming the setup from the first example of section "initializing s6-rc":

user $s6-rc -dl ../live list | sort
another-longrun
another-oneshot
s6rc-fdholder
s6rc-oneshot-runner
test-longrun
test-oneshot

This works because no service names are provided, so the initial selection is empty, and then the -d option complements the selection. The complement of the empty set is 'the universe', i.e. the set of all atomic services in the compiled database. So this is the same as using an s6-rc-db -l ../live list services | sort command. Simulating a transition to up for service test-longrun:

user $s6-rc -n 10 -ul ../live change test-longrun
s6-rc-dryrun: s6-svc -uwU -T 0 -- ../live/servicedirs/test-longrun

The service is a longrun, so this shows that it would be started with an s6-svc -u command. Option -wU makes the command wait for an up and ready notification from s6-supervise. Actually starting the service:

user $time s6-rc -ul ../live change test-longrun
real	0m10.010s
user	0m0.002s
sys	0m0.000s
user $cat ../logdir/current | s6-tai64nlocal
2018-03-30 12:19:25.012770836 test-daemon starting
user $s6-svstat test-longrun
up (pid 2448) 38 seconds, ready 28 seconds

This shows that s6-rc -u change waited the approximately 10 seconds that were necessary for test-daemon become ready and notify s6-supervise.

user $ls -l ../live/servicedirs/test-longrun
total 20
drwx-ws--T 2 user user 4096 Mar 30 12:19 event
-rwxr-xr-x 1 user user  130 Mar 30 12:10 finish
-rw-r--r-- 1 user user    2 Mar 30 12:10 notification-fd
-rwxr-xr-x 1 user user   94 Mar 30 12:10 run
drwx------ 2 user user 4096 Mar 30 12:19 supervise

This shows that the down file created by s6-rc-init was removed from test-longrun's compiled s6 service directory by the s6-rc -u change command. Simulating a transition to up for service test-oneshot:

user $s6-rc -n 10 -ul ../live change test-oneshot
s6-rc-dryrun: s6-svc -uwU -T 0 -- ../live/servicedirs/s6rc-oneshot-runner
s6-rc-dryrun: s6-sudo -el0 -t 30000 -T 0 -- ../live/servicedirs/s6rc-oneshot-runner/s up 3  # test-oneshot

The service is a oneshot, so this shows that it would be started with an s6-sudo -e command that connects to s6rc-oneshot-runner's socket (/home/user/live/servicedirs/s6rc-oneshot-runner/s), and passes arguments 'up' and '3'. This in turn triggers execution of an s6-rc-oneshot-run up 3 command. The output of s6-rc-dryrun is intended to be human readable, so it shows that numeric identifier 3 corresponds indeed to service test-oneshot in the compiled service database. Because s6rc-oneshot-runner is a dependency of test-oneshot and was in down state, s6-rc would start it first (see service dependencies), and because s6rc-oneshot-runner is a longrun, it would use an s6-svc -u command to do that. Actually starting the service:

user $s6-rc -ul ../live change test-oneshot
Calling /home/user/s6-rc-scripts/real-script.sh...
real-script.sh invoked with arguments start
Performing start actions

Because the s6-sudoc - s6-sudod mechanism is used for oneshots, and s6-sudoc transmits the file descriptors of its standard input, output and error to s6-sudod via fd-passing, messages from the oneshot's executed programs are printed on s6-rc's standard output, which is the interactive shell's controlling terminal in this case. Displaying the list of atomic services in up state:

user $s6-rc -al ../live list | sort
s6rc-oneshot-runner
test-longrun
test-oneshot

This works because no service names are provided, so the initial selection is empty, and then the -a option adds the set of atomic services in up state to it. It is also shown that s6rc-oneshot-runner is also in up state.

Service bundle names can be used with s6-rc as a substitute of their contained atomic services:

user $s6-rc-db -l ../live contents test-bundle | sort
test-longrun
test-oneshot
user $s6-rc -n 10 -dl ../live change test-bundle
s6-rc-dryrun: s6-sudo -el0 -t 30000 -T 0 -- ../live/servicedirs/s6rc-oneshot-runner/s down 3  # test-oneshot
s6-rc-dryrun: s6-svc -dwD -T 0 -- ../live/servicedirs/test-longrun

This shows that stopping bundle test-bundle is the same as individually stopping test-longrun and test-oneshot. test-longrun would be stopped with an s6-svc -d command; option -wD makes the command wait for a really down notification from s6-supervise. test-oneshot would be stopped with an s6-sudo -e command that connects to s6rc-oneshot-runner's socket, and passes arguments 'down' and '3'. This in turn triggers execution of an s6-rc-oneshot-run down 3 command.

user $s6-rc -dl ../live change test-bundle
Calling /home/user/s6-rc-scripts/real-script.sh...
real-script.sh invoked with arguments stop
Performing stop actions
user $s6-svstat test-longrun
down (exitcode 0) 17 seconds, ready 17 seconds
user $cat ../logdir/current | s6-tai64nlocal
2018-03-30 12:19:25.012770836 test-daemon starting
2018-03-30 12:25:50.314271979 test-daemon exited with code 0

This shows that s6-rc -d change stopped all atomic services contained in the bundle.

user $ls -l ../live/servicedirs/test-longrun
total 20
-rw-r--r-- 1 user user    0 Mar 30 12:25 down
drwx-ws--T 2 user user 4096 Mar 30 12:25 event
-rwxr-xr-x 1 user user  130 Mar 30 12:10 finish
-rw-r--r-- 1 user user    2 Mar 30 12:10 notification-fd
-rwxr-xr-x 1 user user   94 Mar 30 12:10 run
drwx------ 2 user user 4096 Mar 30 12:25 supervise

This shows that s6-rc -d change created a down file in this longrun's service directory.

user $s6-rc -ul ../live list all-services | sort
another-longrun
another-oneshot
test-longrun
test-oneshot

Here the resulting selection of the s6-rc list command is the contents of bundle all-service, so that's what is shown. This is the same as using an s6-rc-db -l ../live atomics all-services | sort command. Starting all atomic services in this selection:

user $s6-rc -ul ../live change all-services
Calling /home/user/s6-rc-scripts/real-script.sh...
real-script.sh invoked with arguments start
Performing start actions
user $s6-rc -al ../live list | sort
another-longrun
another-oneshot
s6rc-oneshot-runner
test-longrun
test-oneshot

Updated supervision tree:

user $ps xf -o pid,ppid,pgrp,euser,args
 PID  PPID  PGRP EUSER    COMMAND
...
2130  2123  2130 user     -bash
2460  2130  2460 user      \_ s6-svscan
2461  2460  2460 user          \_ s6-supervise logger
2462  2461  2462 user          |   \_ s6-log t ../../logdir
2471  2460  2460 user          \_ s6-supervise test-longrun
2496  2471  2496 user          |   \_ test-daemon --s6=3
2472  2460  2460 user          \_ s6-supervise s6rc-oneshot-runner
2495  2472  2495 user          |   \_ s6-ipcserverd -1 -- s6-ipcserver-access ... -- s6-sudod ... -- /usr/libexec/s6-rc-oneshot-run ...
2473  2460  2460 user          \_ s6-supervise s6rc-fdholder
2474  2460  2460 user          \_ s6-supervise another-longrun
2494  2474  2494 user              \_ test-daemon
...

This shows that longrun another-longrun just spawns another test-daemon process, with readiness notification turned off, and that s6rc-oneshot-runner launches an s6-ipcserverd process that, in turn, spawns an s6-sudod child for each client connection to its corresponding socket. s6-sudod executes program s6-rc-oneshot-run with the arguments supplied by the client.

user $s6-svc -d another-longrun
user $s6-svstat another-longrun
down (exitcode 0) 21 seconds, normally up, ready 21 seconds

Service another-longrun is now actually in down state, because the underlying program is no longer running, but s6-rc doesn't know about it. This type of inconsistencies can be reported by an s6-rc diff command:

user $s6-rc -l ../live diff
-another-longrun

The minus sign ('-') confirms that s6 considers it down. A subsequent s6-rc -d change command that that contains another-longrun in its closed selection (see service dependencies) will bring the service state in sync again:

user $s6-rc -dl ../live change another-longrun
user $s6-rc -l ../live diff
user $s6-svstat another-longrun
down (exitcode 0) 48 seconds, ready 48 seconds

The s6-rc -d change command uses an s6-svc -dwD command, which is a no operation and returns immediatly because the supervised process is already down, but the service state is now also down in s6-rc's view, so the the output of s6-rc diff is empty.

Stopping all services in the live compiled database:

user $s6-rc -dal ../live change
Calling /home/user/s6-rc-scripts/real-script.sh...
real-script.sh invoked with arguments stop
Performing stop actions
user $cat ../logdir/current | s6-tai64nlocal
2018-03-30 12:19:25.012770836 test-daemon starting
2018-03-30 12:25:50.314271979 test-daemon exited with code 0
2018-03-30 12:41:08.974825979 test-daemon starting
2018-03-30 12:43:25.258589540 test-daemon exited with code 0

This shows all of test-longrun's state transitions, as reported by messages printed by its run and finish scripts and logged by the supervised s6-log process. Requesting a list of atomic services in up state using s6-rc -al ../live list should now return an empty result.

To demonstrate usage of s6-rc diff once again, service test-longrun can transition to up with an s6-svc -u command without s6-rc noticing:

user $s6-svc -u test-longrun
user $s6-svstat test-longrun
up (pid 2847) 6 seconds, normally down
user $s6-rc -l ../live diff
+test-longrun

The plus sign ('+') confirms that s6 considers it up. Consistency can be restored by using an s6-rc -ul ../live change test-longrun command to align s6-rc's view of service state with s6's, or an s6-svc -d test-longrun command to align s6's view of service state with s6-rc's.

Service dependencies

s6-rc supports the specification of dependencies between services. Service A can be defined to directly depend on service B, with the following meaning:

  • If s6-rc it is asked to start A, then B is automatically started first, and if that transition was successful, the requested operation is performed.
  • If s6-rc it is asked to stop B, then A is automatically stopped first, and if that transition was successful, the requested operation is performed.

Therefore, transition failures (see managing services) might result in not performing the requested operation, leaving the service in the state it was before.

In turn, a service's dependencies can have dependencies themselves, forming dependency chains: if service A directly depends on service B, and B directly depends on service C, then service A is said to depend (with no qualification) on both B and C. The aforementioned procedure for state transitions is performed recursively when there are dependency chains, so asking s6-rc to start A will result in a cascade of transitions to up to honor dependencies: C first, B second, A last. Similarly, asking s6-rc to stop C will result in transitions to down for A, B and C, in that order. This means that dependencies introduce, for the purpose of state transitions, a partial ordering of services. The s6-rc change command parallelizes the transitions as much as it can. Independent services are processed in parallel, but ordering constraints introduced by dependencies may result in serial processing for one or more subsets of services.

An atomic service definition in s6-rc-compile's source format (i.e. not bundle definitions) can specify direct dependencies by including a regular file named dependencies, that must contain a list of service names, one per line. Whitespace at the beginning of a line is ignored, but trailing whitespace is not. Lines that start with a hash sign ('#') are treated as comments and ignored. dependencies files can contain service bundle names; s6-rc considers the service being defined to directly depend on every atomic service contained in the specified bundle. For further information about configuration of service dependencies, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory.

Because dependencies may cause state transitions for services that are not in the computed selection of an s6-rc change command, the s6-rc program must also compute a forward dependency graph for transitions to up, or a reverse dependency graph for transitions to down, and then it must compute the closed selection, which is the set of all atomic services in the database that need a state transition. An atomic service is in the command's closed selection if:

  • it is in the selection, or
  • a service in the selection depends on it and the requested operation was to start services, or
  • it depends on a service in the selection and the requested operation was to stop services.

The format of a compiled service database allows s6-rc to perform these computations efficiently. Dependency correctness, i.e. the absence of dependency cycles in the submitted set of service definitions, is verified by s6-rc-compile at the time of database creation.

The s6-rc-db program (see extracting information from a compiled database) has two subcommands that display dependency information:

  • dependencies: displays the direct dependencies of the service corresponding to the name supplied as the subcommand's argument. If the specified service is a bundle, it displays the union of the direct dependencies of all the atomic services contained in the bundle.
  • all-dependencies: displays the fully resolved list of atomic services represented by the list of service names supplied as the subcommand's arguments, in the same way as the atomics subcommand, plus all services in either the forward or the reverse dependency graph. The forward dependency graph is used if a -u option is passed to s6-rc-db, and the reverse dependency graph is used if a -d option is passed to s6-rc-db. If there is neither a -u option nor a -d option, s6-rc-db behaves as if the -u option had been specified.

Finally, the s6-rc program also has a listall subcommand that displays the closed selection computed from the service names supplied as arguments. Since this is done with a depencency graph, the result is the same as that of s6-rc-db all-dependencies, except that the latter can also be used with a compiled database that is not live. The forward dependency graph is used if the -u option is passed to s6-rc, and the reverse dependency graph is used if the -d option is passed to s6-rc. If a -p option ('prune') is passed to s6-rc combined with the change subcommand, if a transition to up was requested, services in the computed closed selection are started and every other service not in down state is stopped. If a transition to down was requested, services in the computed closed selection are stopped and every other service not in up state is started. In both cases, all transitions to down are performed before all transitions to up.

For further information about the s6-rc and s6-rc-db programs, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example set of service definitions with dependencies:

user $ls -l srv-src3/*
srv-src3/service-A:
total 24
drwxr-xr-x 2 user user 4096 Apr 11 12:00 env
-rw-r--r-- 1 user user   44 Apr 11 12:00 finish
-rw-r--r-- 1 user user    2 Apr 11 12:00 notification-fd
-rw-r--r-- 1 user user  161 Apr 11 12:00 run
-rw-r--r-- 1 user user    6 Apr 11 12:00 timeout-up
-rw-r--r-- 1 user user    8 Apr 11 12:00 type

srv-src3/service-B:
total 20
-rw-r--r-- 1 user user 10 Apr 11 12:00 dependencies
-rw-r--r-- 1 user user 44 Apr 11 12:00 finish
-rw-r--r-- 1 user user  2 Apr 11 12:00 notification-fd
-rw-r--r-- 1 user user 92 Apr 11 12:00 run
-rw-r--r-- 1 user user  8 Apr 11 12:00 type

srv-src3/service-C:
total 12
-rw-r--r-- 1 user user  50 Apr 11 12:00 down
-rw-r--r-- 1 user user   8 Apr 11 12:00 type
-rw-r--r-- 1 user user 150 Apr 11 12:00 up

srv-src3/service-D:
total 16
-rw-r--r-- 1 user user 20 Apr 11 12:00 dependencies
-rw-r--r-- 1 user user 44 Apr 11 12:00 finish
-rw-r--r-- 1 user user 85 Apr 11 12:00 run
-rw-r--r-- 1 user user  8 Apr 11 12:00 type

srv-src3/service-E:
total 20
drwxr-xr-x 2 user user 4096 Apr 11 12:00 env
-rw-r--r-- 1 user user  177 Apr 11 12:00 finish
-rw-r--r-- 1 user user    2 Apr 11 12:00 notification-fd
-rw-r--r-- 1 user user  158 Apr 11 12:00 run
-rw-r--r-- 1 user user    8 Apr 11 12:00 type

srv-src3/service-F:
total 16
-rw-r--r-- 1 user user 10 Apr 11 12:00 dependencies
-rw-r--r-- 1 user user 76 Apr 11 12:00 down
-rw-r--r-- 1 user user  8 Apr 11 12:00 type
-rw-r--r-- 1 user user 50 Apr 11 12:00 up

srv-src3/service-G:
total 12
-rw-r--r-- 1 user user 44 Apr 11 12:00 finish
-rw-r--r-- 1 user user 85 Apr 11 12:00 run
-rw-r--r-- 1 user user  8 Apr 11 12:00 type

Dependency specifications for these services are set up according to the following diagram: 

A <-- B <--+ D
           |
      C <--+

      E <--  F

             G

Legend X <-- Y means "Y directly depends on X". This diagram shows that service D depends on services A, B and C. Service A is a longrun with no dependencies:

FILE srv-src3/service-A/type
longrun
FILE srv-src3/service-A/run
#!/bin/execlineb -P
foreground { echo Service A starting }
s6-envdir env
foreground {
   importas -D 1 T LONGRUN_TIME
   sleep $T
}
fdmove -c 2 1
test-daemon --s6=3
FILE srv-src3/service-A/notification-fd
3
FILE srv-src3/service-A/timeout-up
15000
FILE srv-src3/service-A/finish
#!/bin/execlineb -P
echo Service A stopping

The underlying program is a hypothetical test-daemon program that supports the s6 readiness notification protocol when passed an --s6 option. The notification channel's file descriptor is 3, so there is a notification-fd file specifying that. The time taken by this service to be up and ready is the value of environment variable LONGRUN_TIME (1 second if unset), plus the actual time test-daemon takes to be up and ready. A timeout value of 15 seconds (15000 milliseconds) is set for this service's transition to up. The environment is set by the s6 supervision tree and modified by the contents of environment subdirectory env, which will be copied verbatim to service A's compiled s6 service directory. Both the run and finish scripts print messages to standard output, so that service events can be tracked.

Service B also runs test-daemon, and has service A as a direct dependency:

FILE srv-src3/service-B/type
longrun
FILE srv-src3/service-B/dependencies
service-A
FILE srv-src3/service-B/run
#!/bin/execlineb -P
foreground { echo Service B starting }
fdmove -c 2 1
test-daemon --s6=3
FILE srv-src3/service-B/notification-fd
3
FILE srv-src3/service-B/finish
#!/bin/execlineb -P
echo Service B stopping

Service C is a oneshot with no dependencies that prints messages to standard output, so that service events can be tracked, but does nothing else:

FILE srv-src3/service-C/type
oneshot
FILE srv-src3/service-C/up
foreground {
   pipeline -d { echo Service C starting }
   s6-tai64n
}
s6-envdir -I /home/user/s6-rc-env/service-C
importas -D 0 E ONESHOT_STATUS
exit $E
FILE srv-src3/service-C/down
pipeline -d { echo Service C stopping }
s6-tai64n

The up and down files contain full execline scripts in this case. The former can have different exit codes depending on environment variable ONESHOT_STATUS (0 if unset). The environment is set by the s6 supervision tree, inherited by s6-rc-oneshot-run, and modified by the contents of environment directory /home/user/s6-rc-env/service-C. Because oneshots don't have a corresponding s6 service directory, this envdir must be elsewhere, and specified via an absolute pathname. s6-tai64n invocations are used to prepend a timestamp to printed messages.

Service D also runs test-daemon with readiness notification turned off, and has two direct dependencies:

FILE srv-src3/service-D/type
longrun
FILE srv-src3/service-D/dependencies
service-B
service-C

This means it has both a longrun (service B) and a oneshot (service C) as dependencies.

FILE srv-src3/service-D/run
#!/bin/execlineb -P
foreground { echo Service D starting }
fdmove -c 2 1
test-daemon
FILE srv-src3/service-D/finish
#!/bin/execlineb -P
echo Service D stopping

Service E is a longrun with no dependencies:

FILE srv-src3/service-E/type
longrun
FILE srv-src3/service-E/run
#!/bin/execlineb -P
foreground { echo Service E starting }
s6-envdir env
importas -D no Y LONGRUN_FAIL
if { test $Y != yes }
fdmove -c 2 1
test-daemon --s6=3
FILE srv-src3/service-E/notification-fd
3

This run execline script exits 1 if environment variable LONGRUN_FAIL has the value yes, and runs test-daemon otherwise. The environment is set by the s6 supervision tree and modified by the contents of environment subdirectory env.

FILE srv-src3/service-E/finish
#!/bin/execlineb -P
s6-envdir env
importas -D no Y LONGRUN_FAIL
ifte {
   echo Service E stopping
} {
   foreground { echo Service E failed permanently }
   exit 125
}
test $Y != yes

This finish execline script exits 125 if environment variable LONGRUN_FAIL has the value yes, signalling permanent failure to s6-supervise.

Service F is a oneshot that only prints messages to standard output, and that has a longrun as its direct dependency:

FILE srv-src3/service-F/type
oneshot
FILE srv-src3/service-F/dependencies
service-E
FILE srv-src3/service-F/up
pipeline -d { echo Service F starting }
s6-tai64n
FILE srv-src3/service-F/down
foreground {
   pipeline -d { echo Service F stopping }
   s6-tai64n
}
sleep 10

And finally, service G is a longrun with no dependencies and no relation with any other service, that also runs test-daemon with readiness notification turned off:

FILE srv-src3/service-G/type
longrun
FILE srv-src3/service-G/run
#!/bin/execlineb -P
foreground { echo Service G starting }
fdmove -c 2 1
test-daemon
FILE srv-src3/service-G/finish
#!/bin/execlineb -P
echo Service G stopping

Creating an s6-rc compiled database with this set of services and using s6-rc-db to show direct dependencies:

user $s6-rc-compile -u $(id -u) test-database2 srv-src3
user $for i in A B C D E F G
> do echo service-$i:
> s6-rc-db -c test-database2 dependencies service-$i
> echo
> done
service-A:

service-B:
service-A

service-C:
s6rc-oneshot-runner

service-D:
service-C
service-B

service-E:

service-F:
service-E
s6rc-oneshot-runner

service-G:

Note that for every oneshot the implicit dependency on service s6rc-oneshot-runner is also displayed (see oneshot definitions). Initializing s6-rc, assuming that there is a supervision tree with /home/user/scan as the scan directory and with a logger like in the example from section "initializing s6-rc", that database test-database2 is in /home/user, and that the working directory is the scan directory:

user $s6-rc-init -c /home/user/test-database2 -l /home/user/live /home/user/scan
user $s6-rc -ul ../live list service-{D,F,G} | sort
service-D
service-F
service-G
user $s6-rc -ul ../live listall service-{D,F,G} | sort
s6rc-oneshot-runner
service-A
service-B
service-C
service-D
service-E
service-F
service-G

Here s6-rc's selection is the same as the list of services provided as arguments, because they are all atomic services, and s6-rc -u listall shows the resulting closed selection using the forward dependency graph: services D, F and G are included because they are in the selection, services A, B and C are included because they are dependencies of service D, service E is included because it is a dependency of service F, and s6rc-oneshot-runner is included because services C and F are oneshots and depend on it. An s6-rc-db -ul ../live all-dependencies service-{D,F,G} | sort command would have produced the same output. Simulating the transition of services to up state:

user $s6-rc -n 10 -ul ../live change service-{D,F,G}
s6-rc-dryrun: s6-svc -uwu -T 0 -- ../live/servicedirs/service-G
s6-rc-dryrun: s6-svc -uwU -T 0 -- ../live/servicedirs/service-E
s6-rc-dryrun: s6-svc -uwU -T 15000 -- ../live/servicedirs/service-A
s6-rc-dryrun: s6-svc -uwU -T 0 -- ../live/servicedirs/s6rc-oneshot-runner
s6-rc-dryrun: s6-svc -uwU -T 0 -- ../live/servicedirs/service-B
s6-rc-dryrun: s6-sudo -el0 -t 30000 -T 0 -- ../live/servicedirs/s6rc-oneshot-runner/s up 8  # service-F
s6-rc-dryrun: s6-sudo -el0 -t 30000 -T 0 -- ../live/servicedirs/s6rc-oneshot-runner/s up 7  # service-C
s6-rc-dryrun: s6-svc -uwu -T 0 -- ../live/servicedirs/service-D

Services D and G are longruns that don't use readiness notificacion, so s6-svc's -wu option (wait for the up notification) is used instead of -wU (wait for the up and ready notification). And because service A's definition specified an up timeout of 15 seconds, the corresponding s6-svc invocation uses a -T 15000 option instead of -T 0 ("infinite").

Actually starting the services:

user $time s6-rc -ul ../live change service-{D,F,G} | s6-tai64nlocal
2018-04-12 11:39:59.886267000 Service C starting
2018-04-12 11:40:09.801486747 Service F starting

real	0m21.194s
user	0m0.015s
sys	0m0.010s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-12 11:39:59.683503297 Service G starting
2018-04-12 11:39:59.685378175 Service E starting
2018-04-12 11:39:59.686330763 Service A starting
2018-04-12 11:40:10.784500674 Service B starting
2018-04-12 11:40:20.793244854 Service D starting

This shows how s6-rc change tried to parallelize state transitions as much as it could: services A, C, E and G were started immediately and in parallel, but services B, D and F were not. Service B's transition happens approximately 11 seconds after service A's, because it is a dependency and that is the time it took service A to become up and ready, and service D's transition happens approximately 10 seconds after service B's for the same reason. It didn't happen later than that because its other dependencies, services A and C, had already successfully transitioned to up. Service F's transition happens approximately 10 seconds after service E's, because it is a dependency and that is the time it took service E to become up and ready. Service D's, F's, and G's transitions are almost instant, because service F is a oneshot, and services D and G do not wait for readiness, so the whole operation took about 21 seconds, and finished after service D's transition to up. Because all state transitions were successful, all services in the selection are now in up state. Stopping all services:

user $time s6-rc -dal ../live change | s6-tai64nlocal
2018-04-12 11:42:37.012906156 Service F stopping
2018-04-12 11:42:37.022828479 Service C stopping

real	0m10.019s
user	0m0.005s
sys	0m0.006s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-12 11:39:59.683503297 Service G starting
2018-04-12 11:39:59.685378175 Service E starting
2018-04-12 11:39:59.686330763 Service A starting
2018-04-12 11:40:10.784500674 Service B starting
2018-04-12 11:40:20.793244854 Service D starting
2018-04-12 11:42:37.014092710 Service G stopping
2018-04-12 11:42:37.017188595 Service D stopping
2018-04-12 11:42:37.022690818 Service B stopping
2018-04-12 11:42:37.029667303 Service A stopping
2018-04-12 11:42:47.023984194 Service E stopping

This shows that most state transitions were almost instant except service F's, because of the sleep 10 command in its down file. Service E's transition happens after service F's transition completed because it is a dependency, so the whole operation took about 10 seconds and finished after service E's transition to down.

Forcing a rotation of the logging directory with an s6-svc -a logger command so that the current file gets emptied, and redoing the steps with one modification:

user $echo 10 >service-A/env/LONGRUN_TIME
user $time s6-rc -ul ../live change service-{D,F,G} | s6-tai64nlocal
2018-04-12 11:47:02.955608987 Service C starting
2018-04-12 11:47:12.969170358 Service F starting
s6-svlisten1: fatal: timed out
s6-rc: warning: unable to start service service-A: command exited 99

real	0m15.025s
user	0m0.003s
sys	0m0.005s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-12 11:47:02.943057415 Service G starting
2018-04-12 11:47:02.948378126 Service E starting
2018-04-12 11:47:02.948776158 Service A starting
2018-04-12 11:47:17.962572120 Service A stopping

This shows that the attempt to start service A failed: an extra delay of 10 seconds caused by the setting of environment variable LONGRUN_TIME caused the transition to take longer than the maximum 15 seconds configured in its service definiton. And because of that failure, services that depend on service A were not started -in this case, services B and D-. The whole operation took about 15 seconds and finished after service A's transition timed out. The final "Service A stopping" message in the log is caused by the extra s6-svc -d command that s6-rc -u change executes for longruns that fail to start.

user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-C
service-E
service-F
service-G

This shows that after the s6-rc -u change command, services A, B and D are not in up state. Stopping all services with an s6-rc -dal ../live change, forcing another rotation of the logging directory, and redoing the steps with more modifications:

user $rm service-A/env/LONGRUN_TIME
user $echo 7 >/home/user/s6-rc-env/service-C/ONESHOT_STATUS
user $time s6-rc -ul ../live change service-{D,F,G} | s6-tai64nlocal
2018-04-12 11:53:24.643007310 Service C starting
s6-rc: warning: unable to start service service-C: command exited 7
2018-04-12 11:53:34.642459632 Service F starting

real	0m21.031s
user	0m0.004s
sys	0m0.005s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-12 11:53:24.632173839 Service E starting
2018-04-12 11:53:24.637136736 Service G starting
2018-04-12 11:53:24.637411837 Service A starting
2018-04-12 11:53:35.650247605 Service B starting

This shows that the attempt to start service C failed: the equivalent of executing the up file of its service definition that program s6-rc-oneshot-run performs exited with a nonzero code (7), caused by the setting of environment variable ONESHOT_STATUS. And because service D depends on service C, the former was not started. The whole operation took about 21 seconds and finished after service B's transition to up.

user $s6-rc -al ../live list | sort
s6rc-oneshot-runner
service-A
service-B
service-E
service-F
service-G

This shows that after the s6-rc -u change command, services C and D are not in up state. Stopping all services, forcing another rotation of the logging directory, and redoing the steps with more modifications:

user $rm /home/user/s6-rc-env/service-C/ONESHOT_STATUS
user $echo yes >service-E/env/LONGRUN_FAIL
user $time s6-rc -ul ../live change service-{D,F,G} | s6-tai64nlocal
s6-svlisten1: fatal: ../live/servicedirs/service-E failed permanently: the finish script exited 125
s6-rc: warning: unable to start service service-E: command exited 1
2018-04-12 12:08:45.991490743 Service C starting

real	0m21.238s
user	0m0.006s
sys	0m0.011s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-12 12:08:45.689700612 Service G starting
2018-04-12 12:08:45.720197503 Service E starting
2018-04-12 12:08:45.740891903 Service A starting
2018-04-12 12:08:45.880561107 Service E failed permanently
2018-04-12 12:08:56.846140697 Service B starting
2018-04-12 12:09:06.853497873 Service D starting

This shows that the attempt to start service E failed: the setting of environment variable LONGRUN_FAIL to yes caused the run script in its s6 service directory to exit immediately, and the finish script to exit with code 125, which signals permanent failure to s6. And because service F depends on service E, the former was not started. The whole operation took about 21 seconds and finished after service D's transition to up.

user $s6-rc -al ../live list | sort
s6rc-oneshot-runner
service-A
service-B
service-C
service-D
service-G

This shows that after the s6-rc -u change command, services E and F are not in up state.

Live updates to the service database

The s6-rc package provides a program, s6-rc-bundle, that allows adding or removing service bundles to or from a compiled database, without having to recompile it from source using s6-rc-compile. It is also notable in that it can be used even if the database is live, i.e. already associated with an s6 supervision tree and an s6-rc live state directory. s6-rc-bundle accepts a subcommand that tells it what to do:

  • add: adds a service bundle to the database. It accepts a list of service names: the first one is the name of the bundle, and the others are the contents of the bundle. The effect is the same as that of a bundle definition in s6-rc-compile's source format that contains those service names in its contents file.
  • delete: removes a service bundle from the database. It accepts a single argument specifying the name of the bundle.
  • multiple: allows removal and addition of several service bundles in one go. It expects its arguments to be in the format execline's execlineb program generates when parsing the block syntax, so the forward compatible way to use it is in an execline script or execlineb -c command: the invocation can be written using a the syntax s6-rc-bundle multiple { d1 d2 ... } n1 { c11 c12 ... } n2 { c21 c22 ... } ..., where d1, d2, ... are service names specifying bundles to delete, n1, n2, ... are service names specifying bundles to add, c11, c12, ... are service names specifying the contents of bundle n1, c21, c22, ... are service names specifying the contents of bundle n2, and so on. The list d1 d2 ... can be empty, resulting in no deletions. Everything after the first {}-block can be omitted, resulting in no additions.

s6-rc-bundle operates on the database associated with live state directory /run/s6-rc, unless it is passed an -l option followed by the (absolute or relative to the working directory) pathname of a different live state directory, or it is passed a -c option followed by the pathname of a database. For more information about s6-rc-bundle, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Using s6-rc-bundle is the only way to modify a compiled database. Any other modification has to be done on the service definitions in source format that were used to create the database, and then s6-rc-compile must be used to make a new database. Because a database that is live cannot be deleted, the s6-rc package provides a tool that allows its replacement: the s6-rc-update program, an online service database switcher. If the replaced database contained a service that is not present in the replacing database, the affected service will be deleted: its name will no longer be recognized as valid by tools like the s6-rc and s6-rc-db programs. Therefore, the replacing database must contain all wanted services, both modified and unmodified ones, and generally means that service definitions in s6-rc-compile's source format must be kept around.

s6-rc-update accepts the replacing database's absolute pathname as an argument, and performs the database switch for live state directory /run/s6-rc, unless it is passed an -l option followed by the absolute pathname of a different live state directory. It also ensures consistency of service states: it analyses both the replaced and replacing database's contents to determine which services are present in both (based on name), and preserves their corresponding state, except for services that change types (longrun to oneshot and vice-versa, atomic service to bundle and vice-versa) or direct dependencies. In those cases, if the service is in up state at program startup, s6-rc-update restarts it: it stops it before the database replacement, and then starts it after the database replacement, so if there are new services defined as direct dependencies, s6-rc-update will start them too after the database replacement. Deleted services are stopped before the replacement.

s6-rc-update also accepts an -f option followed by the pathname of a file, the conversion file, that can currently be used for two purposes: service renaming and forced restarts. The file must contain a sequence of lines of text, each of which is independently lexed into words as if by the execlineb program, so it can contain #-comments, quoted strings, etc. A line of the form old-service -> new-service renames service old-service to new-service. Therefore, all appearances of name old-service in service definitions of the replaced database, and all appearances of name new-service in service definitions of the replacing database, are understood to refer to the same service for the purpose of computing needed state transitions. Otherwise, they are considered to name unrelated services. A line of the form service-name restart makes s6-rc-update forcibly restart service service-name, and a line of the form old-service -> new-service restart makes s6-rc-update perform both a rename and a forced restart.

Note
If a renamed service is a longrun that does not end up getting restarted by s6-rc-update, and the ps utility is used after the service database switch, its s6-supervise parent will appear in the output with the old service name as agrument, because that is how it was invoked by s6-svscan. This is purely cosmetic and has no impact on the service; nevertheless, if the 'correct' ouput is desired, the file supplied with the -f option must also specify a forced restart of the longrun.

s6-rc-update performs the computed state transitions using an s6-rc -d change command before the database switch, and an s6-rc -u change command after the database switch. It also accepts an -n ('dry run') option that makes it simulate the state transitions without switching the database, printing the s6-rc invocations to standard output with the s6-rc-dryrun internal program. For the full description of s6-rc-compile, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Important
The absolute pathname specified to s6-rc-update with an -l option must be that of the symbolic link to directory created by s6-rc-init, not that of the actual directory.

Finally, s6-rc provides the s6-rc-format-upgrade program, that can be used to upgrade the format of a compiled service database that is live and was created with an earlier, incompatible version of s6-rc-compile. It is intended to be used after upgrading the s6-rc package to a possibly backwards-incompatible version, without disrupting the live state directory and associated s6 supervision tree (e.g. if s6-rc is used as an init system component). It accepts the absolute pathname of a compiled database, and operates on live state directory /run/s6-rc, unless it is passed an -l option followed by the (absolute or relative to the working directory) pathname of a different live state directory. For more information about s6-rc-format-upgrade, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Important
The replacing database submitted to s6-rc-format-upgrade must have been created with the s6-rc-compile program from the upgraded package, and the exact same service definitions in source format that were originally used to compile the replaced database. So, for example, if database A was created from service definitions S1, S2, ..., SN with s6-rc-compile from version 0.3.0.1, and an upgrade is made to version 0.4.0.0, then, after the ugrade, a new database B must be created from S1, S2, ..., SN with s6-rc-compile, and then s6-rc-format-upgrade must be invoked with database B's pathname (and that of the live state directory A is currently associated with, using the -l option, if it is not /run/s6-rc). Please carefully review the details of the upgrade procedure in s6-rc-format-upgrade's HTML documentation (in the package's /usr/share/doc subdirectory).
Warning
As of s6-rc version 0.4.0.0, s6-rc-update cannot be safely used with a live state directory currently associated with a database that was created with the s6-rc-compile program from an earlier, backwards-incompatible version of the package.

s6-rc-format-upgrade currently handles upgrades from version 0.3.0.1 to version 0.4.0.0, and performs a database switch like s6-rc-update does, but without changing any service states (i.e. without executing any s6-rc change commands). However, to upgrade to version 0.4.0.0, or any later version compatible with the 0.4.x.x series, without disruption of services in a database that is live, an intermediate upgrade to version 0.3.0.1 is needed[1] using the following procedure:

  1. Upgrade to s6-rc version 0.3.0.1.
  2. Recompile all live service databases using s6-rc-compile from their corresponding service definitions in source format.
  3. Perform all live database switches using s6-rc-update (s6-rc-format-upgrade was introduced in version 0.4.0.0).
  4. Upgrade to s6-rc version 0.4.0.0 or any later, compatible version.
  5. Recompile all live service databases again using s6-rc-compile.
  6. Perform all database format upgrades using s6-rc-format-upgrade.
Note
s6-rc packages have four number verions of the form A.B.C.D. For the purpose of deciding whether s6-rc-format-upgrade should be used, two versions are considered incompatible if they differ in either the first (A) or second (B) number.

Example usage of s6-rc-update, assuming the setup from section "service dependencies", and assuming that the working directory is the scan directory:

user $ls -l ../live
lrwxrwxrwx 1 user user 22 Apr 16 12:14 ../live -> live:s6-rc-init:y0jVfr
user $ls -l
total 8
drwxr-xr-x 4 user user 4096 Apr 16 12:00 logger
lrwxrwxrwx 1 user user   41 Apr 16 12:14 s6rc-fdholder -> /home/user/live/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   47 Apr 16 12:14 s6rc-oneshot-runner -> /home/user/live/servicedirs/s6rc-oneshot-runner
lrwxrwxrwx 1 user user   37 Apr 16 12:14 service-A -> /home/user/live/servicedirs/service-A
lrwxrwxrwx 1 user user   37 Apr 16 12:14 service-B -> /home/user/live/servicedirs/service-B
lrwxrwxrwx 1 user user   37 Apr 16 12:14 service-D -> /home/user/live/servicedirs/service-D
lrwxrwxrwx 1 user user   37 Apr 16 12:14 service-E -> /home/user/live/servicedirs/service-E
lrwxrwxrwx 1 user user   37 Apr 16 12:14 service-G -> /home/user/live/servicedirs/service-G

List of services in up state:

user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-A
service-B
service-C
service-D
service-E
service-F
service-G

The associated service database is going to be replaced by one created from this set of service definitions in s6-rc-compile's source format:

user $ls -l srv-src4/*
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src4/service-C-renamed -> ../srv-src3/service-C
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src4/service-F -> ../srv-src3/service-F
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src4/service-G -> ../srv-src3/service-G

srv-src4/service-B:
total 0
lrwxrwxrwx 1 user user 31 Apr 14 12:00 finish -> ../../srv-src3/service-B/finish
lrwxrwxrwx 1 user user 40 Apr 14 12:00 notification-fd -> ../../srv-src3/service-B/notification-fd
lrwxrwxrwx 1 user user 28 Apr 14 12:00 run -> ../../srv-src3/service-B/run
lrwxrwxrwx 1 user user 29 Apr 14 12:00 type -> ../../srv-src3/service-B/type

srv-src4/service-D:
total 4
-rw-r--r-- 1 user user 28 Apr 14 12:00 dependencies
lrwxrwxrwx 1 user user 31 Apr 14 12:00 finish -> ../../srv-src3/service-D/finish
lrwxrwxrwx 1 user user 28 Apr 14 12:00 run -> ../../srv-src3/service-D/run
lrwxrwxrwx 1 user user 29 Apr 14 12:00 type -> ../../srv-src3/service-D/type

srv-src4/service-E:
total 12
-rw-r--r-- 1 user user 50 Apr 14 12:00 down
-rw-r--r-- 1 user user  8 Apr 14 12:00 type
-rw-r--r-- 1 user user 50 Apr 14 12:00 up

This shows that directories containing service definitions are allowed to have symbolic links to files and directories placed elsewhere. The new database is mostly the same as database test-database2, except that it contains no service A (so it will be removed after the database switch), renames service C to service-C-renamed, and turns service E into a oneshot. So all references to service A must be removed, and all references to service C must use the new name:

FILE srv-src4/service-D/dependencies
service-B
service-C-renamed

Service A is gone from service-D/dependencies, and so is file service-B/dependencies. Service E needs a new definition:

FILE srv-src4/service-E/type
oneshot
FILE srv-src4/service-E/up
pipeline -d { echo Service E starting }
s6-tai64n
FILE srv-src4/service-E/down
pipeline -d { echo Service E stopping }
s6-tai64n

To actually rename service C, a conversion file must be supplied:

FILE conversion-file1
service-C -> service-C-renamed

Compiling the new database using an s6-rc-compile -u $(id -u) test-database3 srv-src4 command, assuming it will be placed in /home/user along with the conversion file, and simulating the effect of invoking s6-rc-update:

user $s6-rc-update -n -l /home/user/live -f ../conversion-file1 /home/user/test-database3
s6-rc-dryrun: s6-rc -v 1 -t 0 -l /home/user/live -d -- change service-E service-A
s6-rc-dryrun: s6-rc -v 1 -t 0 -l /home/user/live -u -- change service-F service-E service-D service-B

This shows that s6-rc-update would use an s6-rc -d change command to stop some services, switch the database, and then use an s6-rc -u change command to start some other services, which implies that (possibly updated) dependency specifications would be honored. Actually performing the database switch, after forcing a rotation of the logging directory with an s6-svc -a logger command so that the current file gets emptied:

user $time s6-rc-update -l /home/user/live -f ../conversion-file1 /home/user/test-database3 | s6-tai64nlocal
2018-04-16 12:17:38.591813680 Service F stopping
2018-04-16 12:17:48.616989175 Service E starting
2018-04-16 12:17:48.623626677 Service F starting

real	0m20.040s
user	0m0.006s
sys	0m0.012s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-16 12:17:38.591475904 Service D stopping
2018-04-16 12:17:38.596463367 Service B stopping
2018-04-16 12:17:38.603899826 Service A stopping
2018-04-16 12:17:48.601233571 Service E stopping
2018-04-16 12:17:48.617483162 Service B starting
2018-04-16 12:17:58.624084755 Service D starting

This shows that because service E changes type, the original longrun was stopped, and the new oneshot was started. Longruns' messages appear in the supervision tree's logging directory, but because of s6-sudoc's fd-passing, oneshots' messages are printed to s6-rc-update's standard output. Therefore, the "Service E stopping" and "Service E starting" messages appear in different places. This also shows that because service A was removed, it was stopped first, and because services B and D depended on it and were up before the database switch, both were restarted. Note that transitions to down have to be serial in the order D, B, A, and transitions to up have to be serial in the order B, D. Finally, because service F continues to depend on service E, and the latter had to restart, F was restarted too, again using serial transitions. Service F was stopped first, and service E was stopped about 10 seconds later, because that was the duration of service F's transition. Then, after the database switch, their transitions to up were almost instant. The whole operation took about 20 seconds: 10 to stop service F plus 10 to start service B.

The symbolic link to the live state directory was updated:

user $ls -l ../live
lrwxrwxrwx 1 user user 24 Apr 16 12:17 ../live -> live:s6-rc-update:SGBH3e

The output of ls previously showed live:s6-rc-init:y0jVfr. The s6 scan directory has also updated all symbolic links:

user $ls -l
total 8
drwxr-xr-x 4 user user 4096 Apr 16 12:00 logger
lrwxrwxrwx 1 user user   41 Apr 16 12:17 s6rc-fdholder -> /home/user/live/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   47 Apr 16 12:17 s6rc-oneshot-runner -> /home/user/live/servicedirs/s6rc-oneshot-runner
lrwxrwxrwx 1 user user   37 Apr 16 12:17 service-B -> /home/user/live/servicedirs/service-B
lrwxrwxrwx 1 user user   37 Apr 16 12:17 service-D -> /home/user/live/servicedirs/service-D
lrwxrwxrwx 1 user user   37 Apr 16 12:17 service-G -> /home/user/live/servicedirs/service-G

There is neither a service-A symlink, because the corresponding service no longer exists, nor a service-E symlink, because the corresponding service is no longer a longrun.

user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-B
service-C-renamed
service-D
service-E
service-F
service-G

This shows that, except for service A, the same set of services are in up state after the database switch, and that the renaming of service C has been performed indeed. Note that the service C has not been restarted; there were no "Service C starting" or "Service C stopping" messages printed to s6-rc-update's standard output. Also note that, because service G is both in test-database2 and test-database3 and not in any dependency chain that involves modified or restarted services, it is left alone by s6-rc-update and remains in up state during the database switch.

One more example with another set of service definitions:

user $ls -l srv-src5/*
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src5/service-B-renamed -> ../srv-src4/service-B
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src5/service-C-renamed -> ../srv-src3/service-C
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src5/service-E -> ../srv-src4/service-E
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src5/service-F -> ../srv-src3/service-F
lrwxrwxrwx 1 user user   21 Apr 14 12:00 srv-src5/service-G -> ../srv-src3/service-G

srv-src5/service-D:
total 4
-rw-r--r-- 1 user user 46 Apr 14 12:00 dependencies
lrwxrwxrwx 1 user user 31 Apr 14 12:00 finish -> ../../srv-src3/service-D/finish
lrwxrwxrwx 1 user user 28 Apr 14 12:00 run -> ../../srv-src3/service-D/run
lrwxrwxrwx 1 user user 29 Apr 14 12:00 type -> ../../srv-src3/service-D/type

srv-src5/service-H:
total 12
-rw-r--r-- 1 user user 44 Apr 14 12:00 finish
-rw-r--r-- 1 user user 85 Apr 14 12:00 run
-rw-r--r-- 1 user user  8 Apr 14 12:00 type

Service B is is going to be renamed, and a new service H is going to be created as a dependency of service D:

FILE srv-src5/service-D/dependencies
service-B-renamed
service-C-renamed
service-H

Service H is a longrun with no dependencies that runs program test-daemon with readiness notification turned off:

FILE srv-src5/service-H/type
longrun
FILE srv-src5/service-H/run
#!/bin/execlineb -P
foreground { echo Service H starting }
fdmove -c 2 1
test-daemon
FILE srv-src5/service-H/finish
#!/bin/execlineb -P
echo Service H stopping

Conversion file for renaming service B that also forces a restart of service G (which is otherwise unmodified):

FILE conversion-file2
service-B -> service-B-renamed
service-G restart

Compiling a new database with these service definitions using a s6-rc-compile -u $(id -u) test-database4 srv-src5 command, assuming it will also be placed in /home/user along with the conversion file, forcing another rotation of the supervision tree's logging directory, and then performing the database switch:

user $time s6-rc-update -l /home/user/live -f ../conversion-file2 /home/user/test-database4
real	0m0.127s
user	0m0.007s
sys	0m0.007s
user $cat ../logdir/current | s6-tai64nlocal
2018-04-16 12:23:07.375107786 Service G stopping
2018-04-16 12:23:07.376383751 Service D stopping
2018-04-16 12:23:07.478780248 Service H starting
2018-04-16 12:23:07.479606503 Service G starting
2018-04-16 12:23:07.496890435 Service D starting

This shows that service G was restarted indeed because of the conversion file, but also service D, because it was in up state before the database switch and its dependencies changed after the switch: service H was added as a direct dependency. On the other hand, service B was only renamed, so it didn't restart. The whole operation took place almost instantly, because not waiting for readiness notification makes service D's, G's and H's transitions almost instant.

user $ls -l ../live
lrwxrwxrwx 1 user user 24 Apr 16 12:23 ../live -> live:s6-rc-update:5RCX6s

This shows that the live state directory symlink was modified once again; the output of ls previously showed live:s6-rc-update:SGBH3e

user $ls -l
total 12
drwxr-xr-x 4 user user 4096 Apr 16 12:00 logger
lrwxrwxrwx 1 user user   41 Apr 16 12:23 s6rc-fdholder -> /home/user/live/servicedirs/s6rc-fdholder
lrwxrwxrwx 1 user user   47 Apr 16 12:23 s6rc-oneshot-runner -> /home/user/live/servicedirs/s6rc-oneshot-runner
lrwxrwxrwx 1 user user   45 Apr 16 12:23 service-B-renamed -> /home/user/live/servicedirs/service-B-renamed
lrwxrwxrwx 1 user user   37 Apr 16 12:23 service-D -> /home/user/live/servicedirs/service-D
lrwxrwxrwx 1 user user   37 Apr 16 12:23 service-G -> /home/user/live/servicedirs/service-G
lrwxrwxrwx 1 user user   37 Apr 16 12:23 service-H -> /home/user/live/servicedirs/service-H

This shows that the symlink for service B in the s6 scan directory was renamed, and that a new symlink was created for service H.

user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-B-renamed
service-C-renamed
service-D
service-E
service-F
service-G
service-H

This shows again that the same set of services are in up state after the database switch, and that the renaming of service B has been performed indeed. Adding two runlevel-like bundles to the compiled service database while it is live:

user $s6-rc-db -l ../live list bundles
user $execlineb -Pc 's6-rc-bundle -l ../live multiple { } runlevel-X { service-D } runlevel-Y { service-F service-G }'
user $s6-rc-db -l ../live list bundles | sort
runlevel-X
runlevel-Y

This shows that both bundles were simultaneously added. It is equivalent to using an s6-rc-bundle -l ../live add runlevel-X service-D command followed by an s6-rc-bundle -l ../live add runlevel-Y service-F service-G command. Forcing another rotation of the supervision tree's logging directory, and then using an s6-rc -pu change command to emulate a runlevel change:

user $s6-rc -pul ../live change runlevel-X | s6-tai64nlocal
2018-04-16 12:26:08.604157313 Service F stopping
2018-04-16 12:26:18.632999649 Service E stopping
user $cat ../logdir/current | s6-tai64nlocal
2018-04-16 12:26:08.603648681 Service G stopping
user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-B-renamed
service-C-renamed
service-D
service-H

This shows that only atomic services that are members of bundle runlevel-X, and their dependencies, are now up. Deleting the bundle from the compiled service database:

user $s6-rc-bundle -l ../live delete runlevel-X
user $s6-rc-db -l ../live list bundles
runlevel-Y

This shows that only bundle runlevel-Y remains. It is equivalent to using an execlineb -Pc 's6-rc-bundle -l ../live multiple { runlevel-X }' command. Performing another 'runlevel change':

user $s6-rc -pul ../live change runlevel-Y | s6-tai64nlocal
2018-04-16 12:29:39.558335051 Service C stopping
2018-04-16 12:29:39.569330382 Service E starting
2018-04-16 12:29:39.576582824 Service F starting
user $cat ../logdir/current | s6-tai64nlocal
2018-04-16 12:26:08.603648681 Service G stopping
2018-04-16 12:29:39.552344150 Service D stopping
2018-04-16 12:29:39.559029637 Service B stopping
2018-04-16 12:29:39.561047890 Service H stopping
2018-04-16 12:29:39.569014150 Service G starting
user $s6-rc -ual ../live list | sort
s6rc-oneshot-runner
service-E
service-F
service-G

This shows that atomic services that were members of (former) bundle runlevel-X were stopped, and only atomic services that are members of bundle runlevel-Y, and their dependencies, are now up.

Longrun pipelining

The s6 supervision tools support the daemontools mechanism that allows connecting the standard output of a supervised process to the standard input of another one with a pipe, normally for logging purposes. This mechanism is used when a subdirectory or symbolic link to directory named log is present in a service directory. s6-rc generalizes this idea with pipelines: it allows arbirary length chains of longruns that have their standard output connected to the standard input of the next one in the chain with a pipe. Moreover, as of version 0.4.0.0, s6-rc also allows tying the standard input of a longrun with the standard output of two or more other longruns, so that multiple streams of messages from different supervised processes can be collapsed into a single stream. This means that s6-rc pipelines are really funnels, but the name 'pipeline' is retained for compatibility with earlier versions that did not support this feature.

The s6-rc-compile program creates pipelines when the definition directories of their member services contain regular files named producer-for and consumer-for. A producer-for file must contain a single service name; the standard output of the service being defined will then be connected to the standard input of the service named in the file as a result. s6-rc requires that the definition directory of the latter also contain a consumer-for file that names the former, and s6-rc-compile implicitly makes the latter a direct dependency of the former, as if by the presence a dependencies file. This ensures that if the pipeline mechanism is used for logging, loggers will be up before starting the 'main' services. A consumer-for file must contain a list of service names. If more that one service is named in a consumer-for file, the funnel-style collapsing of message streams will take place. Pipeline consistency, e.g. absence of collisions or cycles, is verified by s6-rc-compile at service database creation time, as well as the existence of a consumer-for file in the definition of each service named in a producer-for file, and a producer-for file in the definition of each service named in a consumer-for file. To create a logging chain with s6-rc services, definitions of 'main' services must contain a producer-for file naming their loggers, and the loggers must be created just like any other s6-rc longrun, with definition directories that contain a consumer-for file naming (only) their main services.

The definition directory of a service with a consumer-for file and no producer-for file, named the corresponding pipeline's last consumer, can also contain an optional, regular file named pipeline-name, that must contain a single service name; s6-rc-compile will then create a service bundle with this name that contains all the services that are members of the pipeline. A pipeline-name in the definition directory of a service that is not a pipeline's last consumer is ignored. For further information about pipelines, please consult s6-rc-compile's HTML documentation in the package's /usr/share/doc subdirectory.

s6-rc implements pipelines using an internal support service named s6rc-fdholder. This service spawns a long-lived process that executes the s6-fdholderd program, i.e. the fd-holder daemon from s6, and runs an internal s6-rc program, s6-rc-fdholder-filler, chained from s6-ipcclient. s6-rc-fdholder-filler is a short-lived UCSPI IPC client, that reads a file containing service names, creates a pipe for each of them using the POSIX pipe() call, connects to the socket of an s6-fdholderd process, and transfers the file descriptors associated with the reading and writing end of all pipes to the server with a set dump operation, i.e. as if by s6-fdholder-setdumpc. The file with service names is created by s6-rc-compile in s6rc-fdholder's s6 service directory. s6-rc-compile also automatically makes s6rc-fdholder a direct dependency of every longrun that is a member of a pipeline, and creates a wrapper run execline script in their compiled s6 service directory. The wrapper script retrieves the file descriptors that correspond to the reading and / or writing ends of the pipes the service needs with s6-fdholder-retrieve invocations, and then executes the actual run file supplied in its s6-rc definition (currently installed as run.user in the compiled s6 service directory). The same happens with the finish file, if there is one. For further information about s6-rc-fdholder-filler, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

s6rc-fdholder uses a rules directory to decide whether to grant or refuse service to clients that connect to its UNIX domain socket. The effective user of all processes in an s6 supervision tree is initially that of s6-svscan. The s6-supervise process of every longrun that is a member of a pipeline will spawn a child process that executes the wrapper run execline script with the s6-fdholder-retrieve invocations, so s6-svscan's effective user must be allowed by s6rc-fdholder's rules directory. Setup of this rules directory can be done using s6-rc-compile's -u and -g options, see initializing s6-rc. If s6-svscan's effective user is root, the effective user of the s6-fdholderd process can be adjusted by passing an -h option to s6-rc-compile followed by an account's username; s6rc-fdholder's run script will then include an s6-envuidgid invocation to drop privileges accordingly. For further information about s6-rc-compile, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Finally, the s6-rc-db program (see extracting information from a compiled database) has a pipeline subcommand that accepts a service name argument. If the corresponding service is a member of a pipeline, it displays information about it, otherwise, it just prints the supplied name. For further information about s6-rc-db, please consult the HTML documentation in the package's /usr/share/doc subdirectory.

Example set of service definitions using pipelines:

user $ls -l srv-src6/*
srv-src6/final-logger:
total 16
-rw-r--r-- 1 user user  18 Apr 18 12:00 consumer-for
-rw-r--r-- 1 user user  20 Apr 18 12:00 pipeline-name
-rw-r--r-- 1 user user 130 Apr 18 12:00 run
-rw-r--r-- 1 user user   8 Apr 18 12:00 type

srv-src6/reader-A:
total 16
-rw-r--r-- 1 user user  10 Apr 18 12:00 consumer-for
-rw-r--r-- 1 user user  13 Apr 18 12:00 producer-for
-rw-r--r-- 1 user user 115 Apr 18 12:00 run
-rw-r--r-- 1 user user   8 Apr 18 12:00 type

srv-src6/reader-B:
total 16
-rw-r--r-- 1 user user  10 Apr 18 12:00 consumer-for
-rw-r--r-- 1 user user  13 Apr 18 12:00 producer-for
-rw-r--r-- 1 user user 115 Apr 18 12:00 run
-rw-r--r-- 1 user user   8 Apr 18 12:00 type

srv-src6/service-A:
total 12
-rw-r--r-- 1 user user  9 Apr 18 12:00 producer-for
-rw-r--r-- 1 user user 32 Apr 18 12:00 run
-rw-r--r-- 1 user user  8 Apr 18 12:00 type

srv-src6/service-B:
total 12
-rw-r--r-- 1 user user  9 Apr 18 12:00 producer-for
-rw-r--r-- 1 user user 32 Apr 18 12:00 run
-rw-r--r-- 1 user user  8 Apr 18 12:00 type

Service A spawns a process that executes a hypothetical test-daemon program. It is assumed that this program prints a message of the form "Message #N", where N is an incrementing number, each time it receives a SIGHUP signal.

FILE srv-src6/service-A/type
longrun
FILE srv-src6/service-A/run
#!/bin/execlineb -P
test-daemon
FILE srv-src6/service-A/producer-for
reader-A

Reader A spawns a process that reads lines of text from its standard input and prints them to its standard output, prepended with the string "This comes from service A:":

FILE srv-src6/reader-A/type
longrun
FILE srv-src6/reader-A/run
#!/usr/bin/python3
import sys
for line in sys.stdin: print ('This comes from service A:', line, end='', flush=True)
FILE srv-src6/reader-A/consumer-for
service-A

Service A's standard ouput will be connected to reader A's standard input with a pipe. Service B and reader B constitute a similar pair of services:

FILE srv-src6/service-B/type
longrun
FILE srv-src6/service-B/run
#!/bin/execlineb -P
test-daemon
FILE srv-src6/service-B/producer-for
reader-B
FILE srv-src6/reader-B/type
longrun
FILE srv-src6/reader-B/run
#!/usr/bin/python3
import sys
for line in sys.stdin: print ('This comes from service B:', line, end='', flush=True)
FILE srv-src6/reader-B/consumer-for
service-B

And finally, service final-reader spawns an s6-log process. Messages read from its standard input are logged in logging directory /home/user/logdir:

FILE srv-src6/final-logger/type
longrun
FILE srv-src6/final-logger/run
#!/bin/execlineb -P
redirfd -w 1 /dev/null
foreground {
   fdmove -c 1 2
   echo Starting final-logger
}
s6-log /home/user/logdir

The s6-rc pipeline ties reader A's and B's standard output with this logger's standard input:

FILE srv-src6/reader-A/producer-for
final-logger
FILE srv-src6/reader-B/producer-for
final-logger
FILE srv-src6/final-logger/consumer-for
reader-A
reader-B

final-logger is the pipeline's last consumer, so it has a pipeline-name file that specifies a name for the pipeline:

FILE srv-src6/final-logger/pipeline-name
service-AB-pipeline

A diagram that represents this pipeline: 

service A --> reader A -->+
                          |
service B --> reader B -->+
                          |
                          +--> final-logger

After creating a database with these service definitions using a s6-rc-compile -u $(id -u) test-database5 srv-src6 command:

user $s6-rc-db -c test-database5 list all | sort
final-logger
reader-A
reader-B
s6rc-fdholder
s6rc-oneshot-runner
service-A
service-AB-pipeline
service-B
user $s6-rc-db -c test-database5 type service-AB-pipeline
bundle

This shows that the database contains an actual service named service-AB-pipeline, and that it is a bundle. s6-rc-compile must be invoked with an -u option so that the nonprivileged effective user of each longrun's process is allowed to retrieve the file descriptors associated with this pipeline from the fd-holder.

user $s6-rc-db -c test-database5 contents service-AB-pipeline | sort
final-logger
reader-A
reader-B
service-A
service-B

This shows that the bundle contains all services associated this pipeline.

user $for i in {service,reader}-{A,B} final-logger
> do echo $i:
> s6-rc-db -c test-database5 dependencies $i
> echo
> done
service-A:
reader-A
s6rc-fdholder

service-B:
reader-B
s6rc-fdholder

reader-A:
final-logger
s6rc-fdholder

reader-B:
final-logger
s6rc-fdholder

final-logger:
s6rc-fdholder

This shows that s6-rc-compile made s6rc-fdholder a direct dependency of all members of the pipeline and made each service associated with the writing end of a pipe a dependency of the service associated with the corresponding reading end.

Starting all the services after launching an s6 supervision tree and initializing s6-rc, assuming that the working directory is the scan directory and that ../live is the relative pathname of the live state directory:

user $s6-rc -ul ../live change service-AB-pipeline
user $ps xf -o pid,ppid,pgrp,euser,args
 PID  PPID  PGRP EUSER    COMMAND
...
2128  2121  2128 user     -bash
2473  2128  2473 user      \_ s6-svscan
2477  2473  2473 user          \_ s6-supervise reader-A
2506  2477  2506 user          |   \_ /usr/lib/python-exec/python3.6/python3 ./run.user
2478  2473  2473 user          \_ s6-supervise service-B
2511  2478  2511 user          |   \_ test-daemon
2479  2473  2473 user          \_ s6-supervise s6rc-oneshot-runner
2480  2473  2473 user          \_ s6-supervise service-A
2515  2480  2515 user          |   \_ test-daemon
2481  2473  2473 user          \_ s6-supervise s6rc-fdholder
2489  2481  2489 user          |   \_ s6-fdholderd -1 -i data/rules
2482  2473  2473 user          \_ s6-supervise reader-B
2505  2482  2505 user          |   \_ /usr/lib/python-exec/python3.6/python3 ./run.user
2483  2473  2473 user          \_ s6-supervise final-logger
2498  2483  2498 user              \_ s6-log /home/user/logdir
...

This shows that all services in the pipeline are now in up state, including s6rc-fdholder, which is a supervised s6-fdholderd process. Sending a few SIGHUP signals to the test-daemon processes that correspond to services A and B:

user $for i in A A B B A B B A
>do
>s6-svc -h service-$i
>done
user $cat /home/user/logdir/current
This comes from service A: Message #1
This comes from service A: Message #2
This comes from service B: Message #1
This comes from service B: Message #2
This comes from service A: Message #3
This comes from service B: Message #3
This comes from service B: Message #4
This comes from service A: Message #4

This shows that all messages passed through the pipeline and got logged in the logging directory. Using s6-rc-db to display information about the pipeline:

user $s6-rc-db -l ../live pipeline reader-B
service-B | reader-B
reader-B | final-logger
service-A | reader-A
reader-A | final-logger

Reader B is a member of pipeline service-AB-pipeline, so that is what is printed.

s6-rc as an init system component

s6-rc can be used as the service manager of an init system based on s6. Such an init system runs program s6-svscan from the s6 package as process 1 for most of the machine's uptime, and separate programs (usually execline scripts), the stage1 init and the stage3 init, during the boot and shutdown sequences, respectively. Therefore, process 1 is the root of an s6 supervision tree for most of the machine's uptime, and has an associated scan directory, usually in a read-write tmpfs. s6-rc uses this supervision tree and a compiled service database, the boot-time service database, that must be available during the machine's boot sequence, usually in the rootfs.

Because s6-rc must be initialized using the s6-rc-init program (see initializing s6-rc) when the s6-svscan process is already running, this takes place in the stage2 init, a program (usually an execline or shell script) that is launched by the stage1 init and blocks until the init system's catch-all logger is started. The s6-rc live state database is usually created in the same read-write tmpfs that is used for the init system's scan directory. The tmpfs is customarily mounted on /run, and the live state directory is left to be /run/s6-rc, s6-rc's default, so that no -l options have to be specified each time to s6-rc tools. After initializing s6-rc, atomic services (oneshots and longruns) that are needed to complete the machine's initialization, and longruns that are wanted in up state at the end of the boot sequence, must be started. This is customarily done by defining a service bundle that contains all those services, and starting it in the stage2 init with an s6-rc -u change command naming that bundle.

The boot-time service database becomes live during the machine's boot sequence, so any service modification must be generally performed by creating a new database using s6-rc-compile, and then replacing the currently live one with this database using the s6-rc-update program (see live updates to the service database). If during the machine's uptime the s6-rc package is upgraded to a backwards-incompatible version, the current live database must be recompiled after the package upgrade using the same service definitions in s6-rc-compile's source format, and the s6-rc-format-upgrade program must be used to switch databases.

Here is a summary of possible ways of performing modifications to s6-rc-managed services after boot:

  • Bundle modifications can be performed directly on the live database using s6-rc-bundle, without requiring a database switch, and these changes are 'sticky' (i.e. are preserved across machine reboots).
  • Longrun file modifications could be performed directly in the copy of their compiled s6 service directory (including modifications of the contents of their env and data subdirectories, if present) using the corresponding symlink in the init system's scan directory, but this is a non-atomic operation that can have issues if the corresponding supervised process terminates and gets restarted by its supervisor in the middle of the directory update [2]. Also, changes done this way are not 'sticky', because the service directory would be restored to the contents of the boot-time service database on the next reboot, so this is generally discouraged.
  • Supporting files of oneshots (e.g. external scripts invoked from an up or down file) must be generally referred to by absolute pathname, can be modified in their corresponding locations, and these changes are 'sticky'.
  • Any other modification must generally be done using s6-rc-update with a new compiled service database; changes are 'sticky' this way and safe.

Finally, before s6-svscan replaces itself with the stage3 init during the shutdown sequence, all s6-rc-managed services must be stopped, usually with an s6-rc -da change command. This takes place in an an execline or shell script, the shutdown script, that is invoked from an s6-svscan diverted signal handler.

Removal

Unmerge

root #emerge --ask --depclean --verbose sys-apps/s6-rc

All live state directories, compiled service databases and service definitions in s6-rc-compile's source format must be manually deleted if no longer wanted after removing the package. And obviously, if s6-rc is being used as an init system component, an alternative init system must be installed in parallel, and the machine rebooted to use it (possibly by reconfiguring the bootloader), before the package is removed, or otherwise the machine will become unbootable.

See also

  • OpenRC — a dependency-based init system for Unix-like systems that maintains compatibility with the system-provided init system
  • Systemd — a modern SysV-style init and rc replacement for Linux systems.

References

  1. ↑ Laurent Bercot, On upgrading from s6-rc version 0.3.0.1 to version 0.4.0.0, May 20th, 2018.
  2. ↑ Laurent Bercot, On updating a live service directory, October 29th, 2017.
Retrieved from "https://wiki.gentoo.org/index.php?title=S6-rc&oldid=1284017"