1SYSTEMD(1) systemd SYSTEMD(1)
2
6 systemd, init - systemd system and service manager
7
9 /usr/lib/systemd/systemd [OPTIONS...]
10 init [OPTIONS...] {COMMAND}
12
14 systemd is a system and service manager for Linux operating systems.
15 When run as first process on boot (as PID 1), it acts as init system
16 that brings up and maintains userspace services. Separate instances are
17 started for logged-in users to start their services.
18 systemd is usually not invoked directly by the user, but is installed
20 as the /sbin/init symlink and started during early boot. The user
21 manager instances are started automatically through the
22 user@.service(5) service.
23 For compatibility with SysV, if the binary is called as init and is not
25 the first process on the machine (PID is not 1), it will execute
26 telinit and pass all command line arguments unmodified. That means init
27 and telinit are mostly equivalent when invoked from normal login
28 sessions. See telinit(8) for more information.
29 When run as a system instance, systemd interprets the configuration
31 file system.conf and the files in system.conf.d directories; when run
32 as a user instance, systemd interprets the configuration file user.conf
33 and the files in user.conf.d directories. See systemd-system.conf(5)
34 for more information.
35
37 systemd provides a dependency system between various entities called
38 "units" of 11 different types. Units encapsulate various objects that
39 are relevant for system boot-up and maintenance. The majority of units
40 are configured in unit configuration files, whose syntax and basic set
41 of options is described in systemd.unit(5), however some are created
42 automatically from other configuration files, dynamically from system
43 state or programmatically at runtime. Units may be "active" (meaning
44 started, bound, plugged in, ..., depending on the unit type, see
45 below), or "inactive" (meaning stopped, unbound, unplugged, ...), as
46 well as in the process of being activated or deactivated, i.e. between
47 the two states (these states are called "activating", "deactivating").
48 A special "failed" state is available as well, which is very similar to
49 "inactive" and is entered when the service failed in some way (process
50 returned error code on exit, or crashed, an operation timed out, or
51 after too many restarts). If this state is entered, the cause will be
52 logged, for later reference. Note that the various unit types may have
53 a number of additional substates, which are mapped to the five
54 generalized unit states described here.
55 The following unit types are available:
57 1. Service units, which start and control daemons and the processes
59 they consist of. For details, see systemd.service(5).
60 2. Socket units, which encapsulate local IPC or network sockets in the
62 system, useful for socket-based activation. For details about
63 socket units, see systemd.socket(5), for details on socket-based
64 activation and other forms of activation, see daemon(7).
65 3. Target units are useful to group units, or provide well-known
67 synchronization points during boot-up, see systemd.target(5).
68 4. Device units expose kernel devices in systemd and may be used to
70 implement device-based activation. For details, see
71 systemd.device(5).
72 5. Mount units control mount points in the file system, for details
74 see systemd.mount(5).
75 6. Automount units provide automount capabilities, for on-demand
77 mounting of file systems as well as parallelized boot-up. See
78 systemd.automount(5).
79 7. Timer units are useful for triggering activation of other units
81 based on timers. You may find details in systemd.timer(5).
82 8. Swap units are very similar to mount units and encapsulate memory
84 swap partitions or files of the operating system. They are
85 described in systemd.swap(5).
86 9. Path units may be used to activate other services when file system
88 objects change or are modified. See systemd.path(5).
89 10. Slice units may be used to group units which manage system
91 processes (such as service and scope units) in a hierarchical tree
92 for resource management purposes. See systemd.slice(5).
93 11. Scope units are similar to service units, but manage foreign
95 processes instead of starting them as well. See systemd.scope(5).
96 Units are named as their configuration files. Some units have special
98 semantics. A detailed list is available in systemd.special(7).
99 systemd knows various kinds of dependencies, including positive and
101 negative requirement dependencies (i.e. Requires= and Conflicts=) as
102 well as ordering dependencies (After= and Before=). NB: ordering and
103 requirement dependencies are orthogonal. If only a requirement
104 dependency exists between two units (e.g. foo.service requires
105 bar.service), but no ordering dependency (e.g. foo.service after
106 bar.service) and both are requested to start, they will be started in
107 parallel. It is a common pattern that both requirement and ordering
108 dependencies are placed between two units. Also note that the majority
109 of dependencies are implicitly created and maintained by systemd. In
110 most cases, it should be unnecessary to declare additional dependencies
111 manually, however it is possible to do this.
112 Application programs and units (via dependencies) may request state
114 changes of units. In systemd, these requests are encapsulated as 'jobs'
115 and maintained in a job queue. Jobs may succeed or can fail, their
116 execution is ordered based on the ordering dependencies of the units
117 they have been scheduled for.
118 On boot systemd activates the target unit default.target whose job is
120 to activate on-boot services and other on-boot units by pulling them in
121 via dependencies. Usually, the unit name is just an alias (symlink) for
122 either graphical.target (for fully-featured boots into the UI) or
123 multi-user.target (for limited console-only boots for use in embedded
124 or server environments, or similar; a subset of graphical.target).
125 However, it is at the discretion of the administrator to configure it
126 as an alias to any other target unit. See systemd.special(7) for
127 details about these target units.
128 On first boot, systemd will enable or disable units according to preset
130 policy. See systemd.preset(5) and "First Boot Semantics" in machine-
131 id(5).
132 systemd only keeps a minimal set of units loaded into memory.
134 Specifically, the only units that are kept loaded into memory are those
135 for which at least one of the following conditions is true:
136 1. It is in an active, activating, deactivating or failed state (i.e.
138 in any unit state except for "inactive")
139 2. It has a job queued for it
141 3. It is a dependency of at least one other unit that is loaded into
143 memory
144 4. It has some form of resource still allocated (e.g. a service unit
146 that is inactive but for which a process is still lingering that
147 ignored the request to be terminated)
148 5. It has been pinned into memory programmatically by a D-Bus call
150 systemd will automatically and implicitly load units from disk — if
152 they are not loaded yet — as soon as operations are requested for them.
153 Thus, in many respects, the fact whether a unit is loaded or not is
154 invisible to clients. Use systemctl list-units --all to comprehensively
155 list all units currently loaded. Any unit for which none of the
156 conditions above applies is promptly unloaded. Note that when a unit is
157 unloaded from memory its accounting data is flushed out too. However,
158 this data is generally not lost, as a journal log record is generated
159 declaring the consumed resources whenever a unit shuts down.
160 Processes systemd spawns are placed in individual Linux control groups
162 named after the unit which they belong to in the private systemd
163 hierarchy. (see Control Groups v2[1] for more information about control
164 groups, or short "cgroups"). systemd uses this to effectively keep
165 track of processes. Control group information is maintained in the
166 kernel, and is accessible via the file system hierarchy (beneath
167 /sys/fs/cgroup/), or in tools such as systemd-cgls(1) or ps(1) (ps xawf
168 -eo pid,user,cgroup,args is particularly useful to list all processes
169 and the systemd units they belong to.).
170 systemd is compatible with the SysV init system to a large degree: SysV
172 init scripts are supported and simply read as an alternative (though
173 limited) configuration file format. The SysV /dev/initctl interface is
174 provided, and compatibility implementations of the various SysV client
175 tools are available. In addition to that, various established Unix
176 functionality such as /etc/fstab or the utmp database are supported.
177 systemd has a minimal transaction system: if a unit is requested to
179 start up or shut down it will add it and all its dependencies to a
180 temporary transaction. Then, it will verify if the transaction is
181 consistent (i.e. whether the ordering of all units is cycle-free). If
182 it is not, systemd will try to fix it up, and removes non-essential
183 jobs from the transaction that might remove the loop. Also, systemd
184 tries to suppress non-essential jobs in the transaction that would stop
185 a running service. Finally it is checked whether the jobs of the
186 transaction contradict jobs that have already been queued, and
187 optionally the transaction is aborted then. If all worked out and the
188 transaction is consistent and minimized in its impact it is merged with
189 all already outstanding jobs and added to the run queue. Effectively
190 this means that before executing a requested operation, systemd will
191 verify that it makes sense, fixing it if possible, and only failing if
192 it really cannot work.
193 Note that transactions are generated independently of a unit's state at
195 runtime, hence, for example, if a start job is requested on an already
196 started unit, it will still generate a transaction and wake up any
197 inactive dependencies (and cause propagation of other jobs as per the
198 defined relationships). This is because the enqueued job is at the time
199 of execution compared to the target unit's state and is marked
200 successful and complete when both satisfy. However, this job also pulls
201 in other dependencies due to the defined relationships and thus leads
202 to, in our example, start jobs for any of those inactive units getting
203 queued as well.
204 systemd contains native implementations of various tasks that need to
206 be executed as part of the boot process. For example, it sets the
207 hostname or configures the loopback network device. It also sets up and
208 mounts various API file systems, such as /sys/ or /proc/.
209 For more information about the concepts and ideas behind systemd,
211 please refer to the Original Design Document[2].
212 Note that some but not all interfaces provided by systemd are covered
214 by the Interface Portability and Stability Promise[3].
215 Units may be generated dynamically at boot and system manager reload
217 time, for example based on other configuration files or parameters
218 passed on the kernel command line. For details, see
219 systemd.generator(7).
220 The D-Bus API of systemd is described in org.freedesktop.systemd1(5)
222 and org.freedesktop.LogControl1(5).
223 Systems which invoke systemd in a container or initrd environment
225 should implement the Container Interface[4] or initrd Interface[5]
226 specifications, respectively.
227
229 System unit directories
230 The systemd system manager reads unit configuration from various
231 directories. Packages that want to install unit files shall place
232 them in the directory returned by pkg-config systemd
233 --variable=systemdsystemunitdir. Other directories checked are
234 /usr/local/lib/systemd/system and /usr/lib/systemd/system. User
235 configuration always takes precedence. pkg-config systemd
236 --variable=systemdsystemconfdir returns the path of the system
237 configuration directory. Packages should alter the content of these
238 directories only with the enable and disable commands of the
239 systemctl(1) tool. Full list of directories is provided in
240 systemd.unit(5).
241 User unit directories
243 Similar rules apply for the user unit directories. However, here
244 the XDG Base Directory specification[6] is followed to find units.
245 Applications should place their unit files in the directory
246 returned by pkg-config systemd --variable=systemduserunitdir.
247 Global configuration is done in the directory reported by
248 pkg-config systemd --variable=systemduserconfdir. The enable and
249 disable commands of the systemctl(1) tool can handle both global
250 (i.e. for all users) and private (for one user) enabling/disabling
251 of units. Full list of directories is provided in systemd.unit(5).
252 SysV init scripts directory
254 The location of the SysV init script directory varies between
255 distributions. If systemd cannot find a native unit file for a
256 requested service, it will look for a SysV init script of the same
257 name (with the .service suffix removed).
258 SysV runlevel link farm directory
260 The location of the SysV runlevel link farm directory varies
261 between distributions. systemd will take the link farm into account
262 when figuring out whether a service shall be enabled. Note that a
263 service unit with a native unit configuration file cannot be
264 started by activating it in the SysV runlevel link farm.
265
267 SIGTERM
268 Upon receiving this signal the systemd system manager serializes
269 its state, reexecutes itself and deserializes the saved state
270 again. This is mostly equivalent to systemctl daemon-reexec.
271 systemd user managers will start the exit.target unit when this
273 signal is received. This is mostly equivalent to systemctl --user
274 start exit.target --job-mode=replace-irreversibly.
275 SIGINT
277 Upon receiving this signal the systemd system manager will start
278 the ctrl-alt-del.target unit. This is mostly equivalent to
279 systemctl start ctrl-alt-del.target
280 --job-mode=replace-irreversibly. If this signal is received more
281 than 7 times per 2s, an immediate reboot is triggered. Note that
282 pressing Ctrl+Alt+Del on the console will trigger this signal.
283 Hence, if a reboot is hanging, pressing Ctrl+Alt+Del more than 7
284 times in 2 seconds is a relatively safe way to trigger an immediate
285 reboot.
286 systemd user managers treat this signal the same way as SIGTERM.
288 SIGWINCH
290 When this signal is received the systemd system manager will start
291 the kbrequest.target unit. This is mostly equivalent to systemctl
292 start kbrequest.target.
293 This signal is ignored by systemd user managers.
295 SIGPWR
297 When this signal is received the systemd manager will start the
298 sigpwr.target unit. This is mostly equivalent to systemctl start
299 sigpwr.target.
300 SIGUSR1
302 When this signal is received the systemd manager will try to
303 reconnect to the D-Bus bus.
304 SIGUSR2
306 When this signal is received the systemd manager will log its
307 complete state in human-readable form. The data logged is the same
308 as printed by systemd-analyze dump.
309 SIGHUP
311 Reloads the complete daemon configuration. This is mostly
312 equivalent to systemctl daemon-reload.
313 SIGRTMIN+0
315 Enters default mode, starts the default.target unit. This is mostly
316 equivalent to systemctl isolate default.target.
317 SIGRTMIN+1
319 Enters rescue mode, starts the rescue.target unit. This is mostly
320 equivalent to systemctl isolate rescue.target.
321 SIGRTMIN+2
323 Enters emergency mode, starts the emergency.service unit. This is
324 mostly equivalent to systemctl isolate emergency.service.
325 SIGRTMIN+3
327 Halts the machine, starts the halt.target unit. This is mostly
328 equivalent to systemctl start halt.target
329 --job-mode=replace-irreversibly.
330 SIGRTMIN+4
332 Powers off the machine, starts the poweroff.target unit. This is
333 mostly equivalent to systemctl start poweroff.target
334 --job-mode=replace-irreversibly.
335 SIGRTMIN+5
337 Reboots the machine, starts the reboot.target unit. This is mostly
338 equivalent to systemctl start reboot.target
339 --job-mode=replace-irreversibly.
340 SIGRTMIN+6
342 Reboots the machine via kexec, starts the kexec.target unit. This
343 is mostly equivalent to systemctl start kexec.target
344 --job-mode=replace-irreversibly.
345 SIGRTMIN+7
347 Reboots userspace, starts the soft-reboot.target unit. This is
348 mostly equivalent to systemctl start soft-reboot.target
349 --job-mode=replace-irreversibly.
350 SIGRTMIN+13
352 Immediately halts the machine.
353 SIGRTMIN+14
355 Immediately powers off the machine.
356 SIGRTMIN+15
358 Immediately reboots the machine.
359 SIGRTMIN+16
361 Immediately reboots the machine with kexec.
362 SIGRTMIN+17
364 Immediately reboots the userspace.
365 SIGRTMIN+20
367 Enables display of status messages on the console, as controlled
368 via systemd.show_status=1 on the kernel command line.
369 SIGRTMIN+21
371 Disables display of status messages on the console, as controlled
372 via systemd.show_status=0 on the kernel command line.
373 SIGRTMIN+22
375 Sets the service manager's log level to "debug", in a fashion
376 equivalent to systemd.log_level=debug on the kernel command line.
377 SIGRTMIN+23
379 Restores the log level to its configured value. The configured
380 value is derived from – in order of priority – the value specified
381 with systemd.log-level= on the kernel command line, or the value
382 specified with LogLevel= in the configuration file, or the built-in
383 default of "info".
384 SIGRTMIN+24
386 Immediately exits the manager (only available for --user
387 instances).
388 SIGRTMIN+25
390 Upon receiving this signal the systemd manager will reexecute
391 itself. This is mostly equivalent to systemctl daemon-reexec except
392 that it will be done asynchronously.
393 The systemd system manager treats this signal the same way as
395 SIGTERM.
396 SIGRTMIN+26
398 Restores the log target to its configured value. The configured
399 value is derived from – in order of priority – the value specified
400 with systemd.log-target= on the kernel command line, or the value
401 specified with LogTarget= in the configuration file, or the
402 built-in default.
403 SIGRTMIN+27, SIGRTMIN+28
405 Sets the log target to "console" on SIGRTMIN+27 (or "kmsg" on
406 SIGRTMIN+28), in a fashion equivalent to systemd.log_target=console
407 (or systemd.log_target=kmsg on SIGRTMIN+28) on the kernel command
408 line.
409
411 The environment block for the system manager is initially set by the
412 kernel. (In particular, "key=value" assignments on the kernel command
413 line are turned into environment variables for PID 1). For the user
414 manager, the system manager sets the environment as described in the
415 "Environment Variables in Spawned Processes" section of
416 systemd.exec(5). The DefaultEnvironment= setting in the system manager
417 applies to all services including user@.service. Additional entries may
418 be configured (as for any other service) through the Environment= and
419 EnvironmentFile= settings for user@.service (see systemd.exec(5)).
420 Also, additional environment variables may be set through the
421 ManagerEnvironment= setting in systemd-system.conf(5) and systemd-
422 user.conf(5).
423 Some of the variables understood by systemd:
425 $SYSTEMD_LOG_LEVEL
427 The maximum log level of emitted messages (messages with a higher
428 log level, i.e. less important ones, will be suppressed). Either
429 one of (in order of decreasing importance) emerg, alert, crit, err,
430 warning, notice, info, debug, or an integer in the range 0...7. See
431 syslog(3) for more information.
432 This can be overridden with --log-level=.
434 $SYSTEMD_LOG_COLOR
436 A boolean. If true, messages written to the tty will be colored
437 according to priority.
438 This can be overridden with --log-color=.
440 $SYSTEMD_LOG_TIME
442 A boolean. If true, console log messages will be prefixed with a
443 timestamp.
444 This can be overridden with --log-time=.
446 $SYSTEMD_LOG_LOCATION
448 A boolean. If true, messages will be prefixed with a filename and
449 line number in the source code where the message originates.
450 This can be overridden with --log-location=.
452 $SYSTEMD_LOG_TID
454 A boolean. If true, messages will be prefixed with the current
455 numerical thread ID (TID).
456 $SYSTEMD_LOG_TARGET
458 The destination for log messages. One of console (log to the
459 attached tty), console-prefixed (log to the attached tty but with
460 prefixes encoding the log level and "facility", see syslog(3), kmsg
461 (log to the kernel circular log buffer), journal (log to the
462 journal), journal-or-kmsg (log to the journal if available, and to
463 kmsg otherwise), auto (determine the appropriate log target
464 automatically, the default), null (disable log output).
465 This can be overridden with --log-target=.
467 $SYSTEMD_LOG_RATELIMIT_KMSG
469 Whether to ratelimit kmsg or not. Takes a boolean. Defaults to
470 "true". If disabled, systemd will not ratelimit messages written to
471 kmsg.
472 $XDG_CONFIG_HOME, $XDG_CONFIG_DIRS, $XDG_DATA_HOME, $XDG_DATA_DIRS
474 The systemd user manager uses these variables in accordance to the
475 XDG Base Directory specification[6] to find its configuration.
476 $SYSTEMD_UNIT_PATH, $SYSTEMD_GENERATOR_PATH,
478 $SYSTEMD_ENVIRONMENT_GENERATOR_PATH
479 Controls where systemd looks for unit files and generators.
480 These variables may contain a list of paths, separated by colons
482 (":"). When set, if the list ends with an empty component ("...:"),
483 this list is prepended to the usual set of paths. Otherwise, the
484 specified list replaces the usual set of paths.
485 $SYSTEMD_PAGER
487 Pager to use when --no-pager is not given; overrides $PAGER. If
488 neither $SYSTEMD_PAGER nor $PAGER are set, a set of well-known
489 pager implementations are tried in turn, including less(1) and
490 more(1), until one is found. If no pager implementation is
491 discovered no pager is invoked. Setting this environment variable
492 to an empty string or the value "cat" is equivalent to passing
493 --no-pager.
494 Note: if $SYSTEMD_PAGERSECURE is not set, $SYSTEMD_PAGER (as well
496 as $PAGER) will be silently ignored.
497 $SYSTEMD_LESS
499 Override the options passed to less (by default "FRSXMK").
500 Users might want to change two options in particular:
502 K
504 This option instructs the pager to exit immediately when Ctrl+C
505 is pressed. To allow less to handle Ctrl+C itself to switch
506 back to the pager command prompt, unset this option.
507 If the value of $SYSTEMD_LESS does not include "K", and the
509 pager that is invoked is less, Ctrl+C will be ignored by the
510 executable, and needs to be handled by the pager.
511 X
513 This option instructs the pager to not send termcap
514 initialization and deinitialization strings to the terminal. It
515 is set by default to allow command output to remain visible in
516 the terminal even after the pager exits. Nevertheless, this
517 prevents some pager functionality from working, in particular
518 paged output cannot be scrolled with the mouse.
519 See less(1) for more discussion.
521 $SYSTEMD_LESSCHARSET
523 Override the charset passed to less (by default "utf-8", if the
524 invoking terminal is determined to be UTF-8 compatible).
525 $SYSTEMD_PAGERSECURE
527 Takes a boolean argument. When true, the "secure" mode of the pager
528 is enabled; if false, disabled. If $SYSTEMD_PAGERSECURE is not set
529 at all, secure mode is enabled if the effective UID is not the same
530 as the owner of the login session, see geteuid(2) and
531 sd_pid_get_owner_uid(3). In secure mode, LESSSECURE=1 will be set
532 when invoking the pager, and the pager shall disable commands that
533 open or create new files or start new subprocesses. When
534 $SYSTEMD_PAGERSECURE is not set at all, pagers which are not known
535 to implement secure mode will not be used. (Currently only less(1)
536 implements secure mode.)
537 Note: when commands are invoked with elevated privileges, for
539 example under sudo(8) or pkexec(1), care must be taken to ensure
540 that unintended interactive features are not enabled. "Secure" mode
541 for the pager may be enabled automatically as describe above.
542 Setting SYSTEMD_PAGERSECURE=0 or not removing it from the inherited
543 environment allows the user to invoke arbitrary commands. Note that
544 if the $SYSTEMD_PAGER or $PAGER variables are to be honoured,
545 $SYSTEMD_PAGERSECURE must be set too. It might be reasonable to
546 completely disable the pager using --no-pager instead.
547 $SYSTEMD_COLORS
549 Takes a boolean argument. When true, systemd and related utilities
550 will use colors in their output, otherwise the output will be
551 monochrome. Additionally, the variable can take one of the
552 following special values: "16", "256" to restrict the use of colors
553 to the base 16 or 256 ANSI colors, respectively. This can be
554 specified to override the automatic decision based on $TERM and
555 what the console is connected to.
556 $SYSTEMD_URLIFY
558 The value must be a boolean. Controls whether clickable links
559 should be generated in the output for terminal emulators supporting
560 this. This can be specified to override the decision that systemd
561 makes based on $TERM and other conditions.
562 $LISTEN_PID, $LISTEN_FDS, $LISTEN_FDNAMES
564 Set by systemd for supervised processes during socket-based
565 activation. See sd_listen_fds(3) for more information.
566 $NOTIFY_SOCKET
568 Set by systemd for supervised processes for status and start-up
569 completion notification. See sd_notify(3) for more information.
570 For further environment variables understood by systemd and its various
572 components, see Known Environment Variables[7].
573
575 When run as the system instance, systemd parses a number of options
576 listed below. They can be specified as kernel command line arguments
577 which are parsed from a number of sources depending on the environment
578 in which systemd is executed. If run inside a Linux container, these
579 options are parsed from the command line arguments passed to systemd
580 itself, next to any of the command line options listed in the Options
581 section above. If run outside of Linux containers, these arguments are
582 parsed from /proc/cmdline and from the "SystemdOptions" EFI variable
583 (on EFI systems) instead. Options from /proc/cmdline have higher
584 priority.
585 Note: use of "SystemdOptions" is deprecated.
587 The following variables are understood:
589 systemd.unit=, rd.systemd.unit=
591 Overrides the unit to activate on boot. Defaults to default.target.
592 This may be used to temporarily boot into a different boot unit,
593 for example rescue.target or emergency.service. See
594 systemd.special(7) for details about these units. The option
595 prefixed with "rd." is honored only in the initrd, while the one
596 that is not prefixed only in the main system.
597 systemd.dump_core
599 Takes a boolean argument or enables the option if specified without
600 an argument. If enabled, the systemd manager (PID 1) dumps core
601 when it crashes. Otherwise, no core dump is created. Defaults to
602 enabled.
603 systemd.crash_chvt
605 Takes a positive integer, or a boolean argument. Can be also
606 specified without an argument, with the same effect as a positive
607 boolean. If a positive integer (in the range 1–63) is specified,
608 the system manager (PID 1) will activate the specified virtual
609 terminal when it crashes. Defaults to disabled, meaning that no
610 such switch is attempted. If set to enabled, the virtual terminal
611 the kernel messages are written to is used instead.
612 systemd.crash_shell
614 Takes a boolean argument or enables the option if specified without
615 an argument. If enabled, the system manager (PID 1) spawns a shell
616 when it crashes, after a 10s delay. Otherwise, no shell is spawned.
617 Defaults to disabled, for security reasons, as the shell is not
618 protected by password authentication.
619 systemd.crash_reboot
621 Takes a boolean argument or enables the option if specified without
622 an argument. If enabled, the system manager (PID 1) will reboot the
623 machine automatically when it crashes, after a 10s delay.
624 Otherwise, the system will hang indefinitely. Defaults to disabled,
625 in order to avoid a reboot loop. If combined with
626 systemd.crash_shell, the system is rebooted after the shell exits.
627 systemd.confirm_spawn
629 Takes a boolean argument or a path to the virtual console where the
630 confirmation messages should be emitted. Can be also specified
631 without an argument, with the same effect as a positive boolean. If
632 enabled, the system manager (PID 1) asks for confirmation when
633 spawning processes using /dev/console. If a path or a console name
634 (such as "ttyS0") is provided, the virtual console pointed to by
635 this path or described by the give name will be used instead.
636 Defaults to disabled.
637 systemd.service_watchdogs=
639 Takes a boolean argument. If disabled, all service runtime
640 watchdogs (WatchdogSec=) and emergency actions (e.g. OnFailure= or
641 StartLimitAction=) are ignored by the system manager (PID 1); see
642 systemd.service(5). Defaults to enabled, i.e. watchdogs and failure
643 actions are processed normally. The hardware watchdog is not
644 affected by this option.
645 systemd.show_status
647 Takes a boolean argument or the constants error and auto. Can be
648 also specified without an argument, with the same effect as a
649 positive boolean. If enabled, the systemd manager (PID 1) shows
650 terse service status updates on the console during bootup. With
651 error, only messages about failures are shown, but boot is
652 otherwise quiet. auto behaves like false until there is a
653 significant delay in boot. Defaults to enabled, unless quiet is
654 passed as kernel command line option, in which case it defaults to
655 error. If specified overrides the system manager configuration file
656 option ShowStatus=, see systemd-system.conf(5).
657 systemd.status_unit_format=
659 Takes name, description or combined as the value. If name, the
660 system manager will use unit names in status messages. If combined,
661 the system manager will use unit names and description in status
662 messages. When specified, overrides the system manager
663 configuration file option StatusUnitFormat=, see systemd-
664 system.conf(5).
665 systemd.log_color, systemd.log_level=, systemd.log_location,
667 systemd.log_target=, systemd.log_time, systemd.log_tid,
668 systemd.log_ratelimit_kmsg
669 Controls log output, with the same effect as the
670 $SYSTEMD_LOG_COLOR, $SYSTEMD_LOG_LEVEL, $SYSTEMD_LOG_LOCATION,
671 $SYSTEMD_LOG_TARGET, $SYSTEMD_LOG_TIME, $SYSTEMD_LOG_TID and
672 $SYSTEMD_LOG_RATELIMIT_KMSG environment variables described above.
673 systemd.log_color, systemd.log_location, systemd.log_time,
674 systemd.log_tid and systemd.log_ratelimit_kmsg can be specified
675 without an argument, with the same effect as a positive boolean.
676 systemd.default_standard_output=, systemd.default_standard_error=
678 Controls default standard output and error output for services and
679 sockets. That is, controls the default for StandardOutput= and
680 StandardError= (see systemd.exec(5) for details). Takes one of
681 inherit, null, tty, journal, journal+console, kmsg, kmsg+console.
682 If the argument is omitted systemd.default-standard-output=
683 defaults to journal and systemd.default-standard-error= to inherit.
684 systemd.setenv=
686 Takes a string argument in the form VARIABLE=VALUE. May be used to
687 set default environment variables to add to forked child processes.
688 May be used more than once to set multiple variables.
689 systemd.machine_id=
691 Takes a 32 character hex value to be used for setting the
692 machine-id. Intended mostly for network booting where the same
693 machine-id is desired for every boot.
694 systemd.set_credential=, systemd.set_credential_binary=
696 Sets a system credential, which can then be propagated to system
697 services using the ImportCredential= or LoadCredential= setting,
698 see systemd.exec(5) for details. Takes a pair of credential name
699 and value, separated by a colon. The systemd.set_credential=
700 parameter expects the credential value in literal text form, the
701 systemd.set_credential_binary= parameter takes binary data encoded
702 in Base64. Note that the kernel command line is typically
703 accessible by unprivileged programs in /proc/cmdline. Thus, this
704 mechanism is not suitable for transferring sensitive data. Use it
705 only for data that is not sensitive (e.g. public keys/certificates,
706 rather than private keys), or in testing/debugging environments.
707 For further information see System and Service Credentials[8]
709 documentation.
710 systemd.import_credentials=
712 Takes a boolean argument. If false disables importing credentials
713 from the kernel command line, the DMI/SMBIOS OEM string table, the
714 qemu_fw_cfg subsystem or the EFI kernel stub.
715 quiet
717 Turn off status output at boot, much like systemd.show_status=no
718 would. Note that this option is also read by the kernel itself and
719 disables kernel log output. Passing this option hence turns off the
720 usual output from both the system manager and the kernel.
721 debug
723 Turn on debugging output. This is equivalent to
724 systemd.log_level=debug. Note that this option is also read by the
725 kernel itself and enables kernel debug output. Passing this option
726 hence turns on the debug output from both the system manager and
727 the kernel.
728 emergency, rd.emergency, -b
730 Boot into emergency mode. This is equivalent to
731 systemd.unit=emergency.target or rd.systemd.unit=emergency.target,
732 respectively, and provided for compatibility reasons and to be
733 easier to type.
734 rescue, rd.rescue, single, s, S, 1
736 Boot into rescue mode. This is equivalent to
737 systemd.unit=rescue.target or rd.systemd.unit=rescue.target,
738 respectively, and provided for compatibility reasons and to be
739 easier to type.
740 2, 3, 4, 5
742 Boot into the specified legacy SysV runlevel. These are equivalent
743 to systemd.unit=runlevel2.target, systemd.unit=runlevel3.target,
744 systemd.unit=runlevel4.target, and systemd.unit=runlevel5.target,
745 respectively, and provided for compatibility reasons and to be
746 easier to type.
747 locale.LANG=, locale.LANGUAGE=, locale.LC_CTYPE=, locale.LC_NUMERIC=,
749 locale.LC_TIME=, locale.LC_COLLATE=, locale.LC_MONETARY=,
750 locale.LC_MESSAGES=, locale.LC_PAPER=, locale.LC_NAME=,
751 locale.LC_ADDRESS=, locale.LC_TELEPHONE=, locale.LC_MEASUREMENT=,
752 locale.LC_IDENTIFICATION=
753 Set the system locale to use. This overrides the settings in
754 /etc/locale.conf. For more information, see locale.conf(5) and
755 locale(7).
756 For other kernel command line parameters understood by components of
758 the core OS, please refer to kernel-command-line(7).
759
761 During initialization the service manager will import credentials from
762 various sources into the system's set of credentials, which can then be
763 propagated into services and consumed by generators:
764 • When the service manager first initializes it will read system
766 credentials from SMBIOS Type 11 vendor strings
767 io.systemd.credential:name=value, and
768 io.systemd.credential.binary:name=value.
769 • At the same time it will import credentials from QEMU "fw_cfg".
771 (Note that the SMBIOS mechanism is generally preferred, because it
772 is faster and generic.)
773 • Credentials may be passed via the kernel command line, using the
775 systemd.set-credential= parameter, see above.
776 • Credentials may be passed from the UEFI environment via systemd-
778 stub(7).
779 • When the service manager is invoked during the initrd → host
781 transition it will import all files in /run/credentials/@initrd/ as
782 system credentials.
783 Invoke systemd-creds(1) as follows to see the list of credentials
785 passed into the system:
786 # systemd-creds --system list
788 For further information see System and Service Credentials[8]
790 documentation.
791 The service manager when run as PID 1 consumes the following system
793 credentials:
794 vmm.notify_socket
796 Contains a AF_VSOCK or AF_UNIX address where to send a READY=1
797 notification datagram when the system has finished booting. See
798 sd_notify(3) for more information. Note that in case the hypervisor
799 does not support SOCK_DGRAM over AF_VSOCK, SOCK_SEQPACKET will be
800 tried instead. The credential payload for AF_VSOCK should be in the
801 form "vsock:CID:PORT".
802 This feature is useful for hypervisors/VMMs or other processes on
804 the host to receive a notification via VSOCK when a virtual machine
805 has finished booting.
806 system.machine_id
808 Takes a 128bit hexadecimal ID to initialize /etc/machine-id from,
809 if the file is not set up yet. See machine-id(5) for details.
810
812 systemd is only very rarely invoked directly, since it is started early
813 and is already running by the time users may interact with it.
814 Normally, tools like systemctl(1) are used to give commands to the
815 manager. Since systemd is usually not invoked directly, the options
816 listed below are mostly useful for debugging and special purposes.
817 Introspection and debugging options
819 Those options are used for testing and introspection, and systemd may
820 be invoked with them at any time:
821 --dump-configuration-items
823 Dump understood unit configuration items. This outputs a terse but
824 complete list of configuration items understood in unit definition
825 files.
826 --dump-bus-properties
828 Dump exposed bus properties. This outputs a terse but complete list
829 of properties exposed on D-Bus.
830 --test
832 Determine the initial start-up transaction (i.e. the list of jobs
833 enqueued at start-up), dump it and exit — without actually
834 executing any of the determined jobs. This option is useful for
835 debugging only. Note that during regular service manager start-up
836 additional units not shown by this operation may be started,
837 because hardware, socket, bus or other kinds of activation might
838 add additional jobs as the transaction is executed. Use --system to
839 request the initial transaction of the system service manager (this
840 is also the implied default), combine with --user to request the
841 initial transaction of the per-user service manager instead.
842 --system, --user
844 When used in conjunction with --test, selects whether to calculate
845 the initial transaction for the system instance or for a per-user
846 instance. These options have no effect when invoked without --test,
847 as during regular (i.e. non---test) invocations the service manager
848 will automatically detect whether it shall operate in system or
849 per-user mode, by checking whether the PID it is run as is 1 or
850 not. Note that it is not supported booting and maintaining a system
851 with the service manager running in --system mode but with a PID
852 other than 1.
853 -h, --help
855 Print a short help text and exit.
856 --version
858 Print a short version string and exit.
859 Options that duplicate kernel command line settings
861 Those options correspond directly to options listed above in "Kernel
862 Command Line". Both forms may be used equivalently for the system
863 manager, but it is recommended to use the forms listed above in this
864 context, because they are properly namespaced. When an option is
865 specified both on the kernel command line and as a normal command line
866 argument, the latter has higher precedence.
867 When systemd is used as a user manager, the kernel command line is
869 ignored and only the options described below are understood.
870 Nevertheless, systemd is usually started in this mode through the
871 user@.service(5) service, which is shared between all users. It may be
872 more convenient to use configuration files to modify settings (see
873 systemd-user.conf(5)), or environment variables. See the "Environment"
874 section above for a discussion of how the environment block is set.
875 --unit=
877 Set default unit to activate on startup. If not specified, defaults
878 to default.target. See systemd.unit= above.
879 --dump-core
881 Enable core dumping on crash. This switch has no effect when
882 running as user instance. Same as systemd.dump_core= above.
883 --crash-vt=VT
885 Switch to a specific virtual console (VT) on crash. This switch has
886 no effect when running as user instance. Same as
887 systemd.crash_chvt= above (but not the different spelling!).
888 --crash-shell
890 Run a shell on crash. This switch has no effect when running as
891 user instance. See systemd.crash_shell= above.
892 --crash-reboot
894 Automatically reboot the system on crash. This switch has no effect
895 when running as user instance. See systemd.crash_reboot above.
896 --confirm-spawn
898 Ask for confirmation when spawning processes. This switch has no
899 effect when run as user instance. See systemd.confirm_spawn above.
900 --show-status
902 Show terse unit status information on the console during boot-up
903 and shutdown. See systemd.show_status above.
904 --log-color
906 Highlight important log messages. See systemd.log_color above.
907 --log-level=
909 Set log level. See systemd.log_level above.
910 --log-location
912 Include code location in log messages. See systemd.log_location
913 above.
914 --log-target=
916 Set log target. See systemd.log_target above.
917 --log-time=
919 Prefix console messages with timestamp. See systemd.log_time above.
920 --machine-id=
922 Override the machine-id set on the hard drive. See
923 systemd.machine_id= above.
924 --service-watchdogs
926 Globally enable/disable all service watchdog timeouts and emergency
927 actions. See systemd.service_watchdogs above.
928 --default-standard-output=, --default-standard-error=
930 Sets the default output or error output for all services and
931 sockets, respectively. See systemd.default_standard_output= and
932 systemd.default_standard_error= above.
933
935 /run/systemd/notify
936 Daemon status notification socket. This is an AF_UNIX datagram
937 socket and is used to implement the daemon notification logic as
938 implemented by sd_notify(3).
939 /run/systemd/private
941 Used internally as communication channel between systemctl(1) and
942 the systemd process. This is an AF_UNIX stream socket. This
943 interface is private to systemd and should not be used in external
944 projects.
945 /dev/initctl
947 Limited compatibility support for the SysV client interface, as
948 implemented by the systemd-initctl.service unit. This is a named
949 pipe in the file system. This interface is obsolete and should not
950 be used in new applications.
951
953 systemd 252
954 Kernel command-line arguments systemd.unified_cgroup_hierarchy and
955 systemd.legacy_systemd_cgroup_controller were deprecated. Please
956 switch to the unified cgroup hierarchy.
957
959 The systemd Homepage[9], systemd-system.conf(5), locale.conf(5),
960 systemctl(1), journalctl(1), systemd-notify(1), daemon(7), sd-
961 daemon(3), org.freedesktop.systemd1(5), systemd.unit(5),
962 systemd.special(7), pkg-config(1), kernel-command-line(7), bootup(7),
963 systemd.directives(7)
964
966 1. Control Groups v2
967 https://docs.kernel.org/admin-guide/cgroup-v2.html
968 2. Original Design Document
970 https://0pointer.de/blog/projects/systemd.html
971 3. Interface Portability and Stability Promise
973 https://systemd.io/PORTABILITY_AND_STABILITY/
974 4. Container Interface
976 https://systemd.io/CONTAINER_INTERFACE
977 5. initrd Interface
979 https://systemd.io/INITRD_INTERFACE/
980 6. XDG Base Directory specification
982 https://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html
983 7. Known Environment Variables
985 https://systemd.io/ENVIRONMENT
986 8. System and Service Credentials
988 https://systemd.io/CREDENTIALS
989 9. systemd Homepage
991 https://systemd.io/
992systemd 254 SYSTEMD(1)