1SYSTEMD.SERVICE(5) systemd.service SYSTEMD.SERVICE(5)
2
6 systemd.service - Service unit configuration
7
9 service.service
10
12 A unit configuration file whose name ends in ".service" encodes
13 information about a process controlled and supervised by systemd.
14 This man page lists the configuration options specific to this unit
16 type. See systemd.unit(5) for the common options of all unit
17 configuration files. The common configuration items are configured in
18 the generic [Unit] and [Install] sections. The service specific
19 configuration options are configured in the [Service] section.
20 Additional options are listed in systemd.exec(5), which define the
22 execution environment the commands are executed in, and in
23 systemd.kill(5), which define the way the processes of the service are
24 terminated, and in systemd.resource-control(5), which configure
25 resource control settings for the processes of the service.
26 If a service is requested under a certain name but no unit
28 configuration file is found, systemd looks for a SysV init script by
29 the same name (with the .service suffix removed) and dynamically
30 creates a service unit from that script. This is useful for
31 compatibility with SysV. Note that this compatibility is quite
32 comprehensive but not 100%. For details about the incompatibilities,
33 see the Incompatibilities with SysV[1] document.
34 The systemd-run(1) command allows creating .service and .scope units
36 dynamically and transiently from the command line.
37
39 It is possible for systemd services to take a single argument via the
40 "service@argument.service" syntax. Such services are called
41 "instantiated" services, while the unit definition without the argument
42 parameter is called a "template". An example could be a dhcpcd@.service
43 service template which takes a network interface as a parameter to form
44 an instantiated service. Within the service file, this parameter or
45 "instance name" can be accessed with %-specifiers. See systemd.unit(5)
46 for details.
47
49 Implicit Dependencies
50 The following dependencies are implicitly added:
51 · Services with Type=dbus set automatically acquire dependencies of
53 type Requires= and After= on dbus.socket.
54 · Socket activated services are automatically ordered after their
56 activating .socket units via an automatic After= dependency.
57 Services also pull in all .socket units listed in Sockets= via
58 automatic Wants= and After= dependencies.
59 Additional implicit dependencies may be added as result of execution
61 and resource control parameters as documented in systemd.exec(5) and
62 systemd.resource-control(5).
63 Default Dependencies
65 The following dependencies are added unless DefaultDependencies=no is
66 set:
67 · Service units will have dependencies of type Requires= and After=
69 on sysinit.target, a dependency of type After= on basic.target as
70 well as dependencies of type Conflicts= and Before= on
71 shutdown.target. These ensure that normal service units pull in
72 basic system initialization, and are terminated cleanly prior to
73 system shutdown. Only services involved with early boot or late
74 system shutdown should disable this option.
75 · Instanced service units (i.e. service units with an "@" in their
77 name) are assigned by default a per-template slice unit (see
78 systemd.slice(5)), named after the template unit, containing all
79 instances of the specific template. This slice is normally stopped
80 at shutdown, together with all template instances. If that is not
81 desired, set DefaultDependencies=no in the template unit, and
82 either define your own per-template slice unit file that also sets
83 DefaultDependencies=no, or set Slice=system.slice (or another
84 suitable slice) in the template unit. Also see systemd.resource-
85 control(5).
86
88 Service files must include a [Service] section, which carries
89 information about the service and the process it supervises. A number
90 of options that may be used in this section are shared with other unit
91 types. These options are documented in systemd.exec(5), systemd.kill(5)
92 and systemd.resource-control(5). The options specific to the [Service]
93 section of service units are the following:
94 Type=
96 Configures the process start-up type for this service unit. One of
97 simple, exec, forking, oneshot, dbus, notify or idle:
98 · If set to simple (the default if ExecStart= is specified but
100 neither Type= nor BusName= are), the service manager will
101 consider the unit started immediately after the main service
102 process has been forked off. It is expected that the process
103 configured with ExecStart= is the main process of the service.
104 In this mode, if the process offers functionality to other
105 processes on the system, its communication channels should be
106 installed before the service is started up (e.g. sockets set up
107 by systemd, via socket activation), as the service manager will
108 immediately proceed starting follow-up units, right after
109 creating the main service process, and before executing the
110 service's binary. Note that this means systemctl start command
111 lines for simple services will report success even if the
112 service's binary cannot be invoked successfully (for example
113 because the selected User= doesn't exist, or the service binary
114 is missing).
115 · The exec type is similar to simple, but the service manager
117 will consider the unit started immediately after the main
118 service binary has been executed. The service manager will
119 delay starting of follow-up units until that point. (Or in
120 other words: simple proceeds with further jobs right after
121 fork() returns, while exec will not proceed before both fork()
122 and execve() in the service process succeeded.) Note that this
123 means systemctl start command lines for exec services will
124 report failure when the service's binary cannot be invoked
125 successfully (for example because the selected User= doesn't
126 exist, or the service binary is missing).
127 · If set to forking, it is expected that the process configured
129 with ExecStart= will call fork() as part of its start-up. The
130 parent process is expected to exit when start-up is complete
131 and all communication channels are set up. The child continues
132 to run as the main service process, and the service manager
133 will consider the unit started when the parent process exits.
134 This is the behavior of traditional UNIX services. If this
135 setting is used, it is recommended to also use the PIDFile=
136 option, so that systemd can reliably identify the main process
137 of the service. systemd will proceed with starting follow-up
138 units as soon as the parent process exits.
139 · Behavior of oneshot is similar to simple; however, the service
141 manager will consider the unit up after the main process exits.
142 It will then start follow-up units. RemainAfterExit= is
143 particularly useful for this type of service. Type=oneshot is
144 the implied default if neither Type= nor ExecStart= are
145 specified. Note that if this option is used without
146 RemainAfterExit= the service will never enter "active" unit
147 state, but directly transition from "activating" to
148 "deactivating" or "dead" since no process is configured that
149 shall run continuously. In particular this means that after a
150 service of this type ran (and which has RemainAfterExit= not
151 set) it will not show up as started afterwards, but as dead.
152 · Behavior of dbus is similar to simple; however, it is expected
154 that the service acquires a name on the D-Bus bus, as
155 configured by BusName=. systemd will proceed with starting
156 follow-up units after the D-Bus bus name has been acquired.
157 Service units with this option configured implicitly gain
158 dependencies on the dbus.socket unit. This type is the default
159 if BusName= is specified. A service unit of this type is
160 considered to be in the activating state until the specified
161 bus name is acquired. It is considered activated while the bus
162 name is taken. Once the bus name is released the service is
163 considered being no longer functional which has the effect that
164 the service manager attempts to terminate any remaining
165 processes belonging to the service. Services that drop their
166 bus name as part of their shutdown logic thus should be
167 prepared to receive a SIGTERM (or whichever signal is
168 configured in KillSignal=) as result.
169 · Behavior of notify is similar to exec; however, it is expected
171 that the service sends a notification message via sd_notify(3)
172 or an equivalent call when it has finished starting up. systemd
173 will proceed with starting follow-up units after this
174 notification message has been sent. If this option is used,
175 NotifyAccess= (see below) should be set to open access to the
176 notification socket provided by systemd. If NotifyAccess= is
177 missing or set to none, it will be forcibly set to main.
178 · Behavior of idle is very similar to simple; however, actual
180 execution of the service program is delayed until all active
181 jobs are dispatched. This may be used to avoid interleaving of
182 output of shell services with the status output on the console.
183 Note that this type is useful only to improve console output,
184 it is not useful as a general unit ordering tool, and the
185 effect of this service type is subject to a 5s timeout, after
186 which the service program is invoked anyway.
187 It is generally recommended to use Type=simple for long-running
189 services whenever possible, as it is the simplest and fastest
190 option. However, as this service type won't propagate service
191 start-up failures and doesn't allow ordering of other units against
192 completion of initialization of the service (which for example is
193 useful if clients need to connect to the service through some form
194 of IPC, and the IPC channel is only established by the service
195 itself — in contrast to doing this ahead of time through socket or
196 bus activation or similar), it might not be sufficient for many
197 cases. If so, notify or dbus (the latter only in case the service
198 provides a D-Bus interface) are the preferred options as they allow
199 service program code to precisely schedule when to consider the
200 service started up successfully and when to proceed with follow-up
201 units. The notify service type requires explicit support in the
202 service codebase (as sd_notify() or an equivalent API needs to be
203 invoked by the service at the appropriate time) — if it's not
204 supported, then forking is an alternative: it supports the
205 traditional UNIX service start-up protocol. Finally, exec might be
206 an option for cases where it is enough to ensure the service binary
207 is invoked, and where the service binary itself executes no or
208 little initialization on its own (and its initialization is
209 unlikely to fail). Note that using any type other than simple
210 possibly delays the boot process, as the service manager needs to
211 wait for service initialization to complete. It is hence
212 recommended not to needlessly use any types other than simple.
213 (Also note it is generally not recommended to use idle or oneshot
214 for long-running services.)
215 RemainAfterExit=
217 Takes a boolean value that specifies whether the service shall be
218 considered active even when all its processes exited. Defaults to
219 no.
220 GuessMainPID=
222 Takes a boolean value that specifies whether systemd should try to
223 guess the main PID of a service if it cannot be determined
224 reliably. This option is ignored unless Type=forking is set and
225 PIDFile= is unset because for the other types or with an explicitly
226 configured PID file, the main PID is always known. The guessing
227 algorithm might come to incorrect conclusions if a daemon consists
228 of more than one process. If the main PID cannot be determined,
229 failure detection and automatic restarting of a service will not
230 work reliably. Defaults to yes.
231 PIDFile=
233 Takes a path referring to the PID file of the service. Usage of
234 this option is recommended for services where Type= is set to
235 forking. The path specified typically points to a file below /run/.
236 If a relative path is specified it is hence prefixed with /run/.
237 The service manager will read the PID of the main process of the
238 service from this file after start-up of the service. The service
239 manager will not write to the file configured here, although it
240 will remove the file after the service has shut down if it still
241 exists. The PID file does not need to be owned by a privileged
242 user, but if it is owned by an unprivileged user additional safety
243 restrictions are enforced: the file may not be a symlink to a file
244 owned by a different user (neither directly nor indirectly), and
245 the PID file must refer to a process already belonging to the
246 service.
247 BusName=
249 Takes a D-Bus bus name that this service is reachable as. This
250 option is mandatory for services where Type= is set to dbus.
251 ExecStart=
253 Commands with their arguments that are executed when this service
254 is started. The value is split into zero or more command lines
255 according to the rules described below (see section "Command Lines"
256 below).
257 Unless Type= is oneshot, exactly one command must be given. When
259 Type=oneshot is used, zero or more commands may be specified.
260 Commands may be specified by providing multiple command lines in
261 the same directive, or alternatively, this directive may be
262 specified more than once with the same effect. If the empty string
263 is assigned to this option, the list of commands to start is reset,
264 prior assignments of this option will have no effect. If no
265 ExecStart= is specified, then the service must have
266 RemainAfterExit=yes and at least one ExecStop= line set. (Services
267 lacking both ExecStart= and ExecStop= are not valid.)
268 For each of the specified commands, the first argument must be
270 either an absolute path to an executable or a simple file name
271 without any slashes. Optionally, this filename may be prefixed with
272 a number of special characters:
273 Table 1. Special executable prefixes
275 ┌───────┬────────────────────────────┐
276 │Prefix │ Effect │
277 ├───────┼────────────────────────────┤
278 │"@" │ If the executable path is │
279 │ │ prefixed with "@", the │
280 │ │ second specified token │
281 │ │ will be passed as │
282 │ │ "argv[0]" to the executed │
283 │ │ process (instead of the │
284 │ │ actual filename), followed │
285 │ │ by the further arguments │
286 │ │ specified. │
287 ├───────┼────────────────────────────┤
288 │"-" │ If the executable path is │
289 │ │ prefixed with "-", an exit │
290 │ │ code of the command │
291 │ │ normally considered a │
292 │ │ failure (i.e. non-zero │
293 │ │ exit status or abnormal │
294 │ │ exit due to signal) is │
295 │ │ recorded, but has no │
296 │ │ further effect and is │
297 │ │ considered equivalent to │
298 │ │ success. │
299 ├───────┼────────────────────────────┤
300 │":" │ If the executable path is │
301 │ │ prefixed with ":", │
302 │ │ environment variable │
303 │ │ substitution (as described │
304 │ │ by the "Command Lines" │
305 │ │ section below) is not │
306 │ │ applied. │
307 ├───────┼────────────────────────────┤
308 │"+" │ If the executable path is │
309 │ │ prefixed with "+" then the │
310 │ │ process is executed with │
311 │ │ full privileges. In this │
312 │ │ mode privilege │
313 │ │ restrictions configured │
314 │ │ with User=, Group=, │
315 │ │ CapabilityBoundingSet= or │
316 │ │ the various file system │
317 │ │ namespacing options (such │
318 │ │ as PrivateDevices=, │
319 │ │ PrivateTmp=) are not │
320 │ │ applied to the invoked │
321 │ │ command line (but still │
322 │ │ affect any other │
323 │ │ ExecStart=, ExecStop=, ... │
324 │ │ lines). │
325 ├───────┼────────────────────────────┤
326 │"!" │ Similar to the "+" │
327 │ │ character discussed above │
328 │ │ this permits invoking │
329 │ │ command lines with │
330 │ │ elevated privileges. │
331 │ │ However, unlike "+" the │
332 │ │ "!" character exclusively │
333 │ │ alters the effect of │
334 │ │ User=, Group= and │
335 │ │ SupplementaryGroups=, i.e. │
336 │ │ only the stanzas that │
337 │ │ affect user and group │
338 │ │ credentials. Note that │
339 │ │ this setting may be │
340 │ │ combined with │
341 │ │ DynamicUser=, in which │
342 │ │ case a dynamic user/group │
343 │ │ pair is allocated before │
344 │ │ the command is invoked, │
345 │ │ but credential changing is │
346 │ │ left to the executed │
347 │ │ process itself. │
348 ├───────┼────────────────────────────┤
349 │"!!" │ This prefix is very │
350 │ │ similar to "!", however it │
351 │ │ only has an effect on │
352 │ │ systems lacking support │
353 │ │ for ambient process │
354 │ │ capabilities, i.e. without │
355 │ │ support for │
356 │ │ AmbientCapabilities=. It's │
357 │ │ intended to be used for │
358 │ │ unit files that take │
359 │ │ benefit of ambient │
360 │ │ capabilities to run │
361 │ │ processes with minimal │
362 │ │ privileges wherever │
363 │ │ possible while remaining │
364 │ │ compatible with systems │
365 │ │ that lack ambient │
366 │ │ capabilities support. Note │
367 │ │ that when "!!" is used, │
368 │ │ and a system lacking │
369 │ │ ambient capability support │
370 │ │ is detected any configured │
371 │ │ SystemCallFilter= and │
372 │ │ CapabilityBoundingSet= │
373 │ │ stanzas are implicitly │
374 │ │ modified, in order to │
375 │ │ permit spawned processes │
376 │ │ to drop credentials and │
377 │ │ capabilities themselves, │
378 │ │ even if this is configured │
379 │ │ to not be allowed. │
380 │ │ Moreover, if this prefix │
381 │ │ is used and a system │
382 │ │ lacking ambient capability │
383 │ │ support is detected │
384 │ │ AmbientCapabilities= will │
385 │ │ be skipped and not be │
386 │ │ applied. On systems │
387 │ │ supporting ambient │
388 │ │ capabilities, "!!" has no │
389 │ │ effect and is redundant. │
390 └───────┴────────────────────────────┘
391 "@", "-", ":", and one of "+"/"!"/"!!" may be used together and
392 they can appear in any order. However, only one of "+", "!", "!!"
393 may be used at a time. Note that these prefixes are also supported
394 for the other command line settings, i.e. ExecStartPre=,
395 ExecStartPost=, ExecReload=, ExecStop= and ExecStopPost=.
396 If more than one command is specified, the commands are invoked
398 sequentially in the order they appear in the unit file. If one of
399 the commands fails (and is not prefixed with "-"), other lines are
400 not executed, and the unit is considered failed.
401 Unless Type=forking is set, the process started via this command
403 line will be considered the main process of the daemon.
404 ExecStartPre=, ExecStartPost=
406 Additional commands that are executed before or after the command
407 in ExecStart=, respectively. Syntax is the same as for ExecStart=,
408 except that multiple command lines are allowed and the commands are
409 executed one after the other, serially.
410 If any of those commands (not prefixed with "-") fail, the rest are
412 not executed and the unit is considered failed.
413 ExecStart= commands are only run after all ExecStartPre= commands
415 that were not prefixed with a "-" exit successfully.
416 ExecStartPost= commands are only run after the commands specified
418 in ExecStart= have been invoked successfully, as determined by
419 Type= (i.e. the process has been started for Type=simple or
420 Type=idle, the last ExecStart= process exited successfully for
421 Type=oneshot, the initial process exited successfully for
422 Type=forking, "READY=1" is sent for Type=notify, or the BusName=
423 has been taken for Type=dbus).
424 Note that ExecStartPre= may not be used to start long-running
426 processes. All processes forked off by processes invoked via
427 ExecStartPre= will be killed before the next service process is
428 run.
429 Note that if any of the commands specified in ExecStartPre=,
431 ExecStart=, or ExecStartPost= fail (and are not prefixed with "-",
432 see above) or time out before the service is fully up, execution
433 continues with commands specified in ExecStopPost=, the commands in
434 ExecStop= are skipped.
435 Note that the execution of ExecStartPost= is taken into account for
437 the purpose of Before=/After= ordering constraints.
438 ExecCondition=
440 Optional commands that are executed before the command(s) in
441 ExecStartPre=. Syntax is the same as for ExecStart=, except that
442 multiple command lines are allowed and the commands are executed
443 one after the other, serially.
444 The behavior is like an ExecStartPre= and condition check hybrid:
446 when an ExecCondition= command exits with exit code 1 through 254
447 (inclusive), the remaining commands are skipped and the unit is not
448 marked as failed. However, if an ExecCondition= command exits with
449 255 or abnormally (e.g. timeout, killed by a signal, etc.), the
450 unit will be considered failed (and remaining commands will be
451 skipped). Exit code of 0 or those matching SuccessExitStatus= will
452 continue execution to the next command(s).
453 The same recommendations about not running long-running processes
455 in ExecStartPre= also applies to ExecCondition=. ExecCondition=
456 will also run the commands in ExecStopPost=, as part of stopping
457 the service, in the case of any non-zero or abnormal exits, like
458 the ones described above.
459 ExecReload=
461 Commands to execute to trigger a configuration reload in the
462 service. This argument takes multiple command lines, following the
463 same scheme as described for ExecStart= above. Use of this setting
464 is optional. Specifier and environment variable substitution is
465 supported here following the same scheme as for ExecStart=.
466 One additional, special environment variable is set: if known,
468 $MAINPID is set to the main process of the daemon, and may be used
469 for command lines like the following:
470 ExecReload=kill -HUP $MAINPID
472 Note however that reloading a daemon by sending a signal (as with
474 the example line above) is usually not a good choice, because this
475 is an asynchronous operation and hence not suitable to order
476 reloads of multiple services against each other. It is strongly
477 recommended to set ExecReload= to a command that not only triggers
478 a configuration reload of the daemon, but also synchronously waits
479 for it to complete. For example, dbus-broker(1) uses the following:
480 ExecReload=busctl call org.freedesktop.DBus \
482 /org/freedesktop/DBus org.freedesktop.DBus \
483 ReloadConfig
484 ExecStop=
486 Commands to execute to stop the service started via ExecStart=.
487 This argument takes multiple command lines, following the same
488 scheme as described for ExecStart= above. Use of this setting is
489 optional. After the commands configured in this option are run, it
490 is implied that the service is stopped, and any processes remaining
491 for it are terminated according to the KillMode= setting (see
492 systemd.kill(5)). If this option is not specified, the process is
493 terminated by sending the signal specified in KillSignal= or
494 RestartKillSignal= when service stop is requested. Specifier and
495 environment variable substitution is supported (including $MAINPID,
496 see above).
497 Note that it is usually not sufficient to specify a command for
499 this setting that only asks the service to terminate (for example,
500 by sending some form of termination signal to it), but does not
501 wait for it to do so. Since the remaining processes of the services
502 are killed according to KillMode= and KillSignal= or
503 RestartKillSignal= as described above immediately after the command
504 exited, this may not result in a clean stop. The specified command
505 should hence be a synchronous operation, not an asynchronous one.
506 Note that the commands specified in ExecStop= are only executed
508 when the service started successfully first. They are not invoked
509 if the service was never started at all, or in case its start-up
510 failed, for example because any of the commands specified in
511 ExecStart=, ExecStartPre= or ExecStartPost= failed (and weren't
512 prefixed with "-", see above) or timed out. Use ExecStopPost= to
513 invoke commands when a service failed to start up correctly and is
514 shut down again. Also note that the stop operation is always
515 performed if the service started successfully, even if the
516 processes in the service terminated on their own or were killed.
517 The stop commands must be prepared to deal with that case.
518 $MAINPID will be unset if systemd knows that the main process
519 exited by the time the stop commands are called.
520 Service restart requests are implemented as stop operations
522 followed by start operations. This means that ExecStop= and
523 ExecStopPost= are executed during a service restart operation.
524 It is recommended to use this setting for commands that communicate
526 with the service requesting clean termination. For post-mortem
527 clean-up steps use ExecStopPost= instead.
528 ExecStopPost=
530 Additional commands that are executed after the service is stopped.
531 This includes cases where the commands configured in ExecStop= were
532 used, where the service does not have any ExecStop= defined, or
533 where the service exited unexpectedly. This argument takes multiple
534 command lines, following the same scheme as described for
535 ExecStart=. Use of these settings is optional. Specifier and
536 environment variable substitution is supported. Note that – unlike
537 ExecStop= – commands specified with this setting are invoked when a
538 service failed to start up correctly and is shut down again.
539 It is recommended to use this setting for clean-up operations that
541 shall be executed even when the service failed to start up
542 correctly. Commands configured with this setting need to be able to
543 operate even if the service failed starting up half-way and left
544 incompletely initialized data around. As the service's processes
545 have been terminated already when the commands specified with this
546 setting are executed they should not attempt to communicate with
547 them.
548 Note that all commands that are configured with this setting are
550 invoked with the result code of the service, as well as the main
551 process' exit code and status, set in the $SERVICE_RESULT,
552 $EXIT_CODE and $EXIT_STATUS environment variables, see
553 systemd.exec(5) for details.
554 Note that the execution of ExecStopPost= is taken into account for
556 the purpose of Before=/After= ordering constraints.
557 RestartSec=
559 Configures the time to sleep before restarting a service (as
560 configured with Restart=). Takes a unit-less value in seconds, or a
561 time span value such as "5min 20s". Defaults to 100ms.
562 TimeoutStartSec=
564 Configures the time to wait for start-up. If a daemon service does
565 not signal start-up completion within the configured time, the
566 service will be considered failed and will be shut down again. The
567 precise action depends on the TimeoutStartFailureMode= option.
568 Takes a unit-less value in seconds, or a time span value such as
569 "5min 20s". Pass "infinity" to disable the timeout logic. Defaults
570 to DefaultTimeoutStartSec= from the manager configuration file,
571 except when Type=oneshot is used, in which case the timeout is
572 disabled by default (see systemd-system.conf(5)).
573 If a service of Type=notify sends "EXTEND_TIMEOUT_USEC=...", this
575 may cause the start time to be extended beyond TimeoutStartSec=.
576 The first receipt of this message must occur before
577 TimeoutStartSec= is exceeded, and once the start time has extended
578 beyond TimeoutStartSec=, the service manager will allow the service
579 to continue to start, provided the service repeats
580 "EXTEND_TIMEOUT_USEC=..." within the interval specified until the
581 service startup status is finished by "READY=1". (see
582 sd_notify(3)).
583 TimeoutStopSec=
585 This option serves two purposes. First, it configures the time to
586 wait for each ExecStop= command. If any of them times out,
587 subsequent ExecStop= commands are skipped and the service will be
588 terminated by SIGTERM. If no ExecStop= commands are specified, the
589 service gets the SIGTERM immediately. This default behavior can be
590 changed by the TimeoutStopFailureMode= option. Second, it
591 configures the time to wait for the service itself to stop. If it
592 doesn't terminate in the specified time, it will be forcibly
593 terminated by SIGKILL (see KillMode= in systemd.kill(5)). Takes a
594 unit-less value in seconds, or a time span value such as "5min
595 20s". Pass "infinity" to disable the timeout logic. Defaults to
596 DefaultTimeoutStopSec= from the manager configuration file (see
597 systemd-system.conf(5)).
598 If a service of Type=notify sends "EXTEND_TIMEOUT_USEC=...", this
600 may cause the stop time to be extended beyond TimeoutStopSec=. The
601 first receipt of this message must occur before TimeoutStopSec= is
602 exceeded, and once the stop time has extended beyond
603 TimeoutStopSec=, the service manager will allow the service to
604 continue to stop, provided the service repeats
605 "EXTEND_TIMEOUT_USEC=..." within the interval specified, or
606 terminates itself (see sd_notify(3)).
607 TimeoutAbortSec=
609 This option configures the time to wait for the service to
610 terminate when it was aborted due to a watchdog timeout (see
611 WatchdogSec=). If the service has a short TimeoutStopSec= this
612 option can be used to give the system more time to write a core
613 dump of the service. Upon expiration the service will be forcibly
614 terminated by SIGKILL (see KillMode= in systemd.kill(5)). The core
615 file will be truncated in this case. Use TimeoutAbortSec= to set a
616 sensible timeout for the core dumping per service that is large
617 enough to write all expected data while also being short enough to
618 handle the service failure in due time.
619 Takes a unit-less value in seconds, or a time span value such as
621 "5min 20s". Pass an empty value to skip the dedicated watchdog
622 abort timeout handling and fall back TimeoutStopSec=. Pass
623 "infinity" to disable the timeout logic. Defaults to
624 DefaultTimeoutAbortSec= from the manager configuration file (see
625 systemd-system.conf(5)).
626 If a service of Type=notify handles SIGABRT itself (instead of
628 relying on the kernel to write a core dump) it can send
629 "EXTEND_TIMEOUT_USEC=..." to extended the abort time beyond
630 TimeoutAbortSec=. The first receipt of this message must occur
631 before TimeoutAbortSec= is exceeded, and once the abort time has
632 extended beyond TimeoutAbortSec=, the service manager will allow
633 the service to continue to abort, provided the service repeats
634 "EXTEND_TIMEOUT_USEC=..." within the interval specified, or
635 terminates itself (see sd_notify(3)).
636 TimeoutSec=
638 A shorthand for configuring both TimeoutStartSec= and
639 TimeoutStopSec= to the specified value.
640 TimeoutStartFailureMode=, TimeoutStopFailureMode=
642 These options configure the action that is taken in case a daemon
643 service does not signal start-up within its configured
644 TimeoutStartSec=, respectively if it does not stop within
645 TimeoutStopSec=. Takes one of terminate, abort and kill. Both
646 options default to terminate.
647 If terminate is set the service will be gracefully terminated by
649 sending the signal specified in KillSignal= (defaults to SIGTERM,
650 see systemd.kill(5)). If the service does not terminate the
651 FinalKillSignal= is sent after TimeoutStopSec=. If abort is set,
652 WatchdogSignal= is sent instead and TimeoutAbortSec= applies before
653 sending FinalKillSignal=. This setting may be used to analyze
654 services that fail to start-up or shut-down intermittently. By
655 using kill the service is immediately terminated by sending
656 FinalKillSignal= without any further timeout. This setting can be
657 used to expedite the shutdown of failing services.
658 RuntimeMaxSec=
660 Configures a maximum time for the service to run. If this is used
661 and the service has been active for longer than the specified time
662 it is terminated and put into a failure state. Note that this
663 setting does not have any effect on Type=oneshot services, as they
664 terminate immediately after activation completed. Pass "infinity"
665 (the default) to configure no runtime limit.
666 If a service of Type=notify sends "EXTEND_TIMEOUT_USEC=...", this
668 may cause the runtime to be extended beyond RuntimeMaxSec=. The
669 first receipt of this message must occur before RuntimeMaxSec= is
670 exceeded, and once the runtime has extended beyond RuntimeMaxSec=,
671 the service manager will allow the service to continue to run,
672 provided the service repeats "EXTEND_TIMEOUT_USEC=..." within the
673 interval specified until the service shutdown is achieved by
674 "STOPPING=1" (or termination). (see sd_notify(3)).
675 WatchdogSec=
677 Configures the watchdog timeout for a service. The watchdog is
678 activated when the start-up is completed. The service must call
679 sd_notify(3) regularly with "WATCHDOG=1" (i.e. the "keep-alive
680 ping"). If the time between two such calls is larger than the
681 configured time, then the service is placed in a failed state and
682 it will be terminated with SIGABRT (or the signal specified by
683 WatchdogSignal=). By setting Restart= to on-failure, on-watchdog,
684 on-abnormal or always, the service will be automatically restarted.
685 The time configured here will be passed to the executed service
686 process in the WATCHDOG_USEC= environment variable. This allows
687 daemons to automatically enable the keep-alive pinging logic if
688 watchdog support is enabled for the service. If this option is
689 used, NotifyAccess= (see below) should be set to open access to the
690 notification socket provided by systemd. If NotifyAccess= is not
691 set, it will be implicitly set to main. Defaults to 0, which
692 disables this feature. The service can check whether the service
693 manager expects watchdog keep-alive notifications. See
694 sd_watchdog_enabled(3) for details. sd_event_set_watchdog(3) may
695 be used to enable automatic watchdog notification support.
696 Restart=
698 Configures whether the service shall be restarted when the service
699 process exits, is killed, or a timeout is reached. The service
700 process may be the main service process, but it may also be one of
701 the processes specified with ExecStartPre=, ExecStartPost=,
702 ExecStop=, ExecStopPost=, or ExecReload=. When the death of the
703 process is a result of systemd operation (e.g. service stop or
704 restart), the service will not be restarted. Timeouts include
705 missing the watchdog "keep-alive ping" deadline and a service
706 start, reload, and stop operation timeouts.
707 Takes one of no, on-success, on-failure, on-abnormal, on-watchdog,
709 on-abort, or always. If set to no (the default), the service will
710 not be restarted. If set to on-success, it will be restarted only
711 when the service process exits cleanly. In this context, a clean
712 exit means any of the following:
713 · exit code of 0;
715 · for types other than Type=oneshot, one of the signals SIGHUP,
717 SIGINT, SIGTERM, or SIGPIPE;
718 · exit statuses and signals specified in SuccessExitStatus=.
720 If set to on-failure, the service will be restarted when the
722 process exits with a non-zero exit code, is terminated by a signal
723 (including on core dump, but excluding the aforementioned four
724 signals), when an operation (such as service reload) times out, and
725 when the configured watchdog timeout is triggered. If set to
726 on-abnormal, the service will be restarted when the process is
727 terminated by a signal (including on core dump, excluding the
728 aforementioned four signals), when an operation times out, or when
729 the watchdog timeout is triggered. If set to on-abort, the service
730 will be restarted only if the service process exits due to an
731 uncaught signal not specified as a clean exit status. If set to
732 on-watchdog, the service will be restarted only if the watchdog
733 timeout for the service expires. If set to always, the service will
734 be restarted regardless of whether it exited cleanly or not, got
735 terminated abnormally by a signal, or hit a timeout.
736 Table 2. Exit causes and the effect of the Restart= settings on
738 them
739 ┌──────────────┬────┬────────┬────────────┬────────────┬─────────────┬──────────┬─────────────┐
740 │Restart │ no │ always │ on-success │ on-failure │ on-abnormal │ on-abort │ on-watchdog │
741 │settings/Exit │ │ │ │ │ │ │ │
742 │causes │ │ │ │ │ │ │ │
743 ├──────────────┼────┼────────┼────────────┼────────────┼─────────────┼──────────┼─────────────┤
744 │Clean exit │ │ X │ X │ │ │ │ │
745 │code or │ │ │ │ │ │ │ │
746 │signal │ │ │ │ │ │ │ │
747 ├──────────────┼────┼────────┼────────────┼────────────┼─────────────┼──────────┼─────────────┤
748 │Unclean exit │ │ X │ │ X │ │ │ │
749 │code │ │ │ │ │ │ │ │
750 ├──────────────┼────┼────────┼────────────┼────────────┼─────────────┼──────────┼─────────────┤
751 │Unclean │ │ X │ │ X │ X │ X │ │
752 │signal │ │ │ │ │ │ │ │
753 ├──────────────┼────┼────────┼────────────┼────────────┼─────────────┼──────────┼─────────────┤
754 │Timeout │ │ X │ │ X │ X │ │ │
755 ├──────────────┼────┼────────┼────────────┼────────────┼─────────────┼──────────┼─────────────┤
756 │Watchdog │ │ X │ │ X │ X │ │ X │
757 └──────────────┴────┴────────┴────────────┴────────────┴─────────────┴──────────┴─────────────┘
758 As exceptions to the setting above, the service will not be
759 restarted if the exit code or signal is specified in
760 RestartPreventExitStatus= (see below) or the service is stopped
761 with systemctl stop or an equivalent operation. Also, the services
762 will always be restarted if the exit code or signal is specified in
763 RestartForceExitStatus= (see below).
764 Note that service restart is subject to unit start rate limiting
766 configured with StartLimitIntervalSec= and StartLimitBurst=, see
767 systemd.unit(5) for details. A restarted service enters the failed
768 state only after the start limits are reached.
769 Setting this to on-failure is the recommended choice for
771 long-running services, in order to increase reliability by
772 attempting automatic recovery from errors. For services that shall
773 be able to terminate on their own choice (and avoid immediate
774 restarting), on-abnormal is an alternative choice.
775 SuccessExitStatus=
777 Takes a list of exit status definitions that, when returned by the
778 main service process, will be considered successful termination, in
779 addition to the normal successful exit status 0 and, except for
780 Type=oneshot, the signals SIGHUP, SIGINT, SIGTERM, and SIGPIPE.
781 Exit status definitions can be numeric termination statuses,
782 termination status names, or termination signal names, separated by
783 spaces. See the Process Exit Codes section in systemd.exec(5) for a
784 list of termination status names (for this setting only the part
785 without the "EXIT_" or "EX_" prefix should be used). See signal(7)
786 for a list of signal names.
787 Note that this setting does not change the mapping between numeric
789 exit statuses and their names, i.e. regardless how this setting is
790 used 0 will still be mapped to "SUCCESS" (and thus typically shown
791 as "0/SUCCESS" in tool outputs) and 1 to "FAILURE" (and thus
792 typically shown as "1/FAILURE"), and so on. It only controls what
793 happens as effect of these exit statuses, and how it propagates to
794 the state of the service as a whole.
795 This option may appear more than once, in which case the list of
797 successful exit statuses is merged. If the empty string is assigned
798 to this option, the list is reset, all prior assignments of this
799 option will have no effect.
800 Example 1. A service with the SuccessExitStatus= setting
802 SuccessExitStatus=TEMPFAIL 250 SIGKILL
804 Exit status 75 (TEMPFAIL), 250, and the termination signal SIGKILL
806 are considered clean service terminations.
807 Note: systemd-analyze exit-status may be used to list exit statuses
809 and translate between numerical status values and names.
810 RestartPreventExitStatus=
812 Takes a list of exit status definitions that, when returned by the
813 main service process, will prevent automatic service restarts,
814 regardless of the restart setting configured with Restart=. Exit
815 status definitions can either be numeric exit codes or termination
816 signal names, and are separated by spaces. Defaults to the empty
817 list, so that, by default, no exit status is excluded from the
818 configured restart logic. For example:
819 RestartPreventExitStatus=1 6 SIGABRT
821 ensures that exit codes 1 and 6 and the termination signal SIGABRT
823 will not result in automatic service restarting. This option may
824 appear more than once, in which case the list of restart-preventing
825 statuses is merged. If the empty string is assigned to this option,
826 the list is reset and all prior assignments of this option will
827 have no effect.
828 Note that this setting has no effect on processes configured via
830 ExecStartPre=, ExecStartPost=, ExecStop=, ExecStopPost= or
831 ExecReload=, but only on the main service process, i.e. either the
832 one invoked by ExecStart= or (depending on Type=, PIDFile=, ...)
833 the otherwise configured main process.
834 RestartForceExitStatus=
836 Takes a list of exit status definitions that, when returned by the
837 main service process, will force automatic service restarts,
838 regardless of the restart setting configured with Restart=. The
839 argument format is similar to RestartPreventExitStatus=.
840 RootDirectoryStartOnly=
842 Takes a boolean argument. If true, the root directory, as
843 configured with the RootDirectory= option (see systemd.exec(5) for
844 more information), is only applied to the process started with
845 ExecStart=, and not to the various other ExecStartPre=,
846 ExecStartPost=, ExecReload=, ExecStop=, and ExecStopPost= commands.
847 If false, the setting is applied to all configured commands the
848 same way. Defaults to false.
849 NonBlocking=
851 Set the O_NONBLOCK flag for all file descriptors passed via
852 socket-based activation. If true, all file descriptors >= 3 (i.e.
853 all except stdin, stdout, stderr), excluding those passed in via
854 the file descriptor storage logic (see FileDescriptorStoreMax= for
855 details), will have the O_NONBLOCK flag set and hence are in
856 non-blocking mode. This option is only useful in conjunction with a
857 socket unit, as described in systemd.socket(5) and has no effect on
858 file descriptors which were previously saved in the file-descriptor
859 store for example. Defaults to false.
860 NotifyAccess=
862 Controls access to the service status notification socket, as
863 accessible via the sd_notify(3) call. Takes one of none (the
864 default), main, exec or all. If none, no daemon status updates are
865 accepted from the service processes, all status update messages are
866 ignored. If main, only service updates sent from the main process
867 of the service are accepted. If exec, only service updates sent
868 from any of the main or control processes originating from one of
869 the Exec*= commands are accepted. If all, all services updates from
870 all members of the service's control group are accepted. This
871 option should be set to open access to the notification socket when
872 using Type=notify or WatchdogSec= (see above). If those options are
873 used but NotifyAccess= is not configured, it will be implicitly set
874 to main.
875 Note that sd_notify() notifications may be attributed to units
877 correctly only if either the sending process is still around at the
878 time PID 1 processes the message, or if the sending process is
879 explicitly runtime-tracked by the service manager. The latter is
880 the case if the service manager originally forked off the process,
881 i.e. on all processes that match main or exec. Conversely, if an
882 auxiliary process of the unit sends an sd_notify() message and
883 immediately exits, the service manager might not be able to
884 properly attribute the message to the unit, and thus will ignore
885 it, even if NotifyAccess=all is set for it.
886 Hence, to eliminate all race conditions involving lookup of the
888 client's unit and attribution of notifications to units correctly,
889 sd_notify_barrier() may be used. This call acts as a
890 synchronization point and ensures all notifications sent before
891 this call have been picked up by the service manager when it
892 returns successfully. Use of sd_notify_barrier() is needed for
893 clients which are not invoked by the service manager, otherwise
894 this synchronization mechanism is unnecessary for attribution of
895 notifications to the unit.
896 Sockets=
898 Specifies the name of the socket units this service shall inherit
899 socket file descriptors from when the service is started. Normally,
900 it should not be necessary to use this setting, as all socket file
901 descriptors whose unit shares the same name as the service (subject
902 to the different unit name suffix of course) are passed to the
903 spawned process.
904 Note that the same socket file descriptors may be passed to
906 multiple processes simultaneously. Also note that a different
907 service may be activated on incoming socket traffic than the one
908 which is ultimately configured to inherit the socket file
909 descriptors. Or, in other words: the Service= setting of .socket
910 units does not have to match the inverse of the Sockets= setting of
911 the .service it refers to.
912 This option may appear more than once, in which case the list of
914 socket units is merged. Note that once set, clearing the list of
915 sockets again (for example, by assigning the empty string to this
916 option) is not supported.
917 FileDescriptorStoreMax=
919 Configure how many file descriptors may be stored in the service
920 manager for the service using sd_pid_notify_with_fds(3)'s
921 "FDSTORE=1" messages. This is useful for implementing services that
922 can restart after an explicit request or a crash without losing
923 state. Any open sockets and other file descriptors which should not
924 be closed during the restart may be stored this way. Application
925 state can either be serialized to a file in /run, or better, stored
926 in a memfd_create(2) memory file descriptor. Defaults to 0, i.e. no
927 file descriptors may be stored in the service manager. All file
928 descriptors passed to the service manager from a specific service
929 are passed back to the service's main process on the next service
930 restart (see sd_listen_fds(3) for details about the precise
931 protocol used and the order in which the file descriptors are
932 passed). Any file descriptors passed to the service manager are
933 automatically closed when POLLHUP or POLLERR is seen on them, or
934 when the service is fully stopped and no job is queued or being
935 executed for it. If this option is used, NotifyAccess= (see above)
936 should be set to open access to the notification socket provided by
937 systemd. If NotifyAccess= is not set, it will be implicitly set to
938 main.
939 USBFunctionDescriptors=
941 Configure the location of a file containing USB FunctionFS[2]
942 descriptors, for implementation of USB gadget functions. This is
943 used only in conjunction with a socket unit with ListenUSBFunction=
944 configured. The contents of this file are written to the ep0 file
945 after it is opened.
946 USBFunctionStrings=
948 Configure the location of a file containing USB FunctionFS strings.
949 Behavior is similar to USBFunctionDescriptors= above.
950 OOMPolicy=
952 Configure the Out-Of-Memory (OOM) killer policy. On Linux, when
953 memory becomes scarce the kernel might decide to kill a running
954 process in order to free up memory and reduce memory pressure. This
955 setting takes one of continue, stop or kill. If set to continue and
956 a process of the service is killed by the kernel's OOM killer this
957 is logged but the service continues running. If set to stop the
958 event is logged but the service is terminated cleanly by the
959 service manager. If set to kill and one of the service's processes
960 is killed by the OOM killer the kernel is instructed to kill all
961 remaining processes of the service, too. Defaults to the setting
962 DefaultOOMPolicy= in systemd-system.conf(5) is set to, except for
963 services where Delegate= is turned on, where it defaults to
964 continue.
965 Use the OOMScoreAdjust= setting to configure whether processes of
967 the unit shall be considered preferred or less preferred candidates
968 for process termination by the Linux OOM killer logic. See
969 systemd.exec(5) for details.
970 Check systemd.exec(5) and systemd.kill(5) for more settings.
972
974 This section describes command line parsing and variable and specifier
975 substitutions for ExecStart=, ExecStartPre=, ExecStartPost=,
976 ExecReload=, ExecStop=, and ExecStopPost= options.
977 Multiple command lines may be concatenated in a single directive by
979 separating them with semicolons (these semicolons must be passed as
980 separate words). Lone semicolons may be escaped as "\;".
981 Each command line is split on whitespace, with the first item being the
983 command to execute, and the subsequent items being the arguments.
984 Double quotes ("...") and single quotes ('...') may be used to wrap a
985 whole item (the opening quote may appear only at the beginning or after
986 whitespace that is not quoted, and the closing quote must be followed
987 by whitespace or the end of line), in which case everything until the
988 next matching quote becomes part of the same argument. Quotes
989 themselves are removed. C-style escapes are also supported. The table
990 below contains the list of known escape patterns. Only escape patterns
991 which match the syntax in the table are allowed; other patterns may be
992 added in the future and unknown patterns will result in a warning. In
993 particular, any backslashes should be doubled. Finally, a trailing
994 backslash ("\") may be used to merge lines.
995 This syntax is inspired by shell syntax, but only the meta-characters
997 and expansions described in the following paragraphs are understood,
998 and the expansion of variables is different. Specifically, redirection
999 using "<", "<<", ">", and ">>", pipes using "|", running programs in
1000 the background using "&", and other elements of shell syntax are not
1001 supported.
1002 The command to execute may contain spaces, but control characters are
1004 not allowed.
1005 The command line accepts "%" specifiers as described in
1007 systemd.unit(5).
1008 Basic environment variable substitution is supported. Use "${FOO}" as
1010 part of a word, or as a word of its own, on the command line, in which
1011 case it will be erased and replaced by the exact value of the
1012 environment variable (if any) including all whitespace it contains,
1013 always resulting in exactly a single argument. Use "$FOO" as a separate
1014 word on the command line, in which case it will be replaced by the
1015 value of the environment variable split at whitespace, resulting in
1016 zero or more arguments. For this type of expansion, quotes are
1017 respected when splitting into words, and afterwards removed.
1018 If the command is not a full (absolute) path, it will be resolved to a
1020 full path using a fixed search path determinted at compilation time.
1021 Searched directories include /usr/local/bin/, /usr/bin/, /bin/ on
1022 systems using split /usr/bin/ and /bin/ directories, and their sbin/
1023 counterparts on systems using split bin/ and sbin/. It is thus safe to
1024 use just the executable name in case of executables located in any of
1025 the "standard" directories, and an absolute path must be used in other
1026 cases. Using an absolute path is recommended to avoid ambiguity. Hint:
1027 this search path may be queried using systemd-path
1028 search-binaries-default.
1029 Example:
1031 Environment="ONE=one" 'TWO=two two'
1033 ExecStart=echo $ONE $TWO ${TWO}
1034 This will execute /bin/echo with four arguments: "one", "two", "two",
1036 and "two two".
1037 Example:
1039 Environment=ONE='one' "TWO='two two' too" THREE=
1041 ExecStart=/bin/echo ${ONE} ${TWO} ${THREE}
1042 ExecStart=/bin/echo $ONE $TWO $THREE
1043 This results in /bin/echo being called twice, the first time with
1045 arguments "'one'", "'two two' too", "", and the second time with
1046 arguments "one", "two two", "too".
1047 To pass a literal dollar sign, use "$$". Variables whose value is not
1049 known at expansion time are treated as empty strings. Note that the
1050 first argument (i.e. the program to execute) may not be a variable.
1051 Variables to be used in this fashion may be defined through
1053 Environment= and EnvironmentFile=. In addition, variables listed in the
1054 section "Environment variables in spawned processes" in
1055 systemd.exec(5), which are considered "static configuration", may be
1056 used (this includes e.g. $USER, but not $TERM).
1057 Note that shell command lines are not directly supported. If shell
1059 command lines are to be used, they need to be passed explicitly to a
1060 shell implementation of some kind. Example:
1061 ExecStart=sh -c 'dmesg | tac'
1063 Example:
1065 ExecStart=echo one ; echo "two two"
1067 This will execute echo two times, each time with one argument: "one"
1069 and "two two", respectively. Because two commands are specified,
1070 Type=oneshot must be used.
1071 Example:
1073 ExecStart=echo / >/dev/null & \; \
1075 ls
1076 This will execute echo with five arguments: "/", ">/dev/null", "&",
1078 ";", and "ls".
1079 Table 3. C escapes supported in command lines and environment variables
1081 ┌────────┬─────────────────────────┐
1082 │Literal │ Actual value │
1083 ├────────┼─────────────────────────┤
1084 │"\a" │ bell │
1085 ├────────┼─────────────────────────┤
1086 │"\b" │ backspace │
1087 ├────────┼─────────────────────────┤
1088 │"\f" │ form feed │
1089 ├────────┼─────────────────────────┤
1090 │"\n" │ newline │
1091 ├────────┼─────────────────────────┤
1092 │"\r" │ carriage return │
1093 ├────────┼─────────────────────────┤
1094 │"\t" │ tab │
1095 ├────────┼─────────────────────────┤
1096 │"\v" │ vertical tab │
1097 ├────────┼─────────────────────────┤
1098 │"\\" │ backslash │
1099 ├────────┼─────────────────────────┤
1100 │"\"" │ double quotation mark │
1101 ├────────┼─────────────────────────┤
1102 │"\'" │ single quotation mark │
1103 ├────────┼─────────────────────────┤
1104 │"\s" │ space │
1105 ├────────┼─────────────────────────┤
1106 │"\xxx" │ character number xx in │
1107 │ │ hexadecimal encoding │
1108 ├────────┼─────────────────────────┤
1109 │"\nnn" │ character number nnn in │
1110 │ │ octal encoding │
1111 └────────┴─────────────────────────┘
1112
1114 Example 2. Simple service
1115 The following unit file creates a service that will execute
1117 /usr/sbin/foo-daemon. Since no Type= is specified, the default
1118 Type=simple will be assumed. systemd will assume the unit to be started
1119 immediately after the program has begun executing.
1120 [Unit]
1122 Description=Foo
1123 [Service]
1125 ExecStart=/usr/sbin/foo-daemon
1126 [Install]
1128 WantedBy=multi-user.target
1129 Note that systemd assumes here that the process started by systemd will
1131 continue running until the service terminates. If the program
1132 daemonizes itself (i.e. forks), please use Type=forking instead.
1133 Since no ExecStop= was specified, systemd will send SIGTERM to all
1135 processes started from this service, and after a timeout also SIGKILL.
1136 This behavior can be modified, see systemd.kill(5) for details.
1137 Note that this unit type does not include any type of notification when
1139 a service has completed initialization. For this, you should use other
1140 unit types, such as Type=notify if the service understands systemd's
1141 notification protocol, Type=forking if the service can background
1142 itself or Type=dbus if the unit acquires a DBus name once
1143 initialization is complete. See below.
1144 Example 3. Oneshot service
1146 Sometimes, units should just execute an action without keeping active
1148 processes, such as a filesystem check or a cleanup action on boot. For
1149 this, Type=oneshot exists. Units of this type will wait until the
1150 process specified terminates and then fall back to being inactive. The
1151 following unit will perform a cleanup action:
1152 [Unit]
1154 Description=Cleanup old Foo data
1155 [Service]
1157 Type=oneshot
1158 ExecStart=/usr/sbin/foo-cleanup
1159 [Install]
1161 WantedBy=multi-user.target
1162 Note that systemd will consider the unit to be in the state "starting"
1164 until the program has terminated, so ordered dependencies will wait for
1165 the program to finish before starting themselves. The unit will revert
1166 to the "inactive" state after the execution is done, never reaching the
1167 "active" state. That means another request to start the unit will
1168 perform the action again.
1169 Type=oneshot are the only service units that may have more than one
1171 ExecStart= specified. For units with multiple commands (Type=oneshot),
1172 all commands will be run again.
1173 For Type=oneshot, Restart=always and Restart=on-success are not
1175 allowed.
1176 Example 4. Stoppable oneshot service
1178 Similarly to the oneshot services, there are sometimes units that need
1180 to execute a program to set up something and then execute another to
1181 shut it down, but no process remains active while they are considered
1182 "started". Network configuration can sometimes fall into this category.
1183 Another use case is if a oneshot service shall not be executed each
1184 time when they are pulled in as a dependency, but only the first time.
1185 For this, systemd knows the setting RemainAfterExit=yes, which causes
1187 systemd to consider the unit to be active if the start action exited
1188 successfully. This directive can be used with all types, but is most
1189 useful with Type=oneshot and Type=simple. With Type=oneshot, systemd
1190 waits until the start action has completed before it considers the unit
1191 to be active, so dependencies start only after the start action has
1192 succeeded. With Type=simple, dependencies will start immediately after
1193 the start action has been dispatched. The following unit provides an
1194 example for a simple static firewall.
1195 [Unit]
1197 Description=Simple firewall
1198 [Service]
1200 Type=oneshot
1201 RemainAfterExit=yes
1202 ExecStart=/usr/local/sbin/simple-firewall-start
1203 ExecStop=/usr/local/sbin/simple-firewall-stop
1204 [Install]
1206 WantedBy=multi-user.target
1207 Since the unit is considered to be running after the start action has
1209 exited, invoking systemctl start on that unit again will cause no
1210 action to be taken.
1211 Example 5. Traditional forking services
1213 Many traditional daemons/services background (i.e. fork, daemonize)
1215 themselves when starting. Set Type=forking in the service's unit file
1216 to support this mode of operation. systemd will consider the service to
1217 be in the process of initialization while the original program is still
1218 running. Once it exits successfully and at least a process remains (and
1219 RemainAfterExit=no), the service is considered started.
1220 Often, a traditional daemon only consists of one process. Therefore, if
1222 only one process is left after the original process terminates, systemd
1223 will consider that process the main process of the service. In that
1224 case, the $MAINPID variable will be available in ExecReload=,
1225 ExecStop=, etc.
1226 In case more than one process remains, systemd will be unable to
1228 determine the main process, so it will not assume there is one. In that
1229 case, $MAINPID will not expand to anything. However, if the process
1230 decides to write a traditional PID file, systemd will be able to read
1231 the main PID from there. Please set PIDFile= accordingly. Note that the
1232 daemon should write that file before finishing with its initialization.
1233 Otherwise, systemd might try to read the file before it exists.
1234 The following example shows a simple daemon that forks and just starts
1236 one process in the background:
1237 [Unit]
1239 Description=Some simple daemon
1240 [Service]
1242 Type=forking
1243 ExecStart=/usr/sbin/my-simple-daemon -d
1244 [Install]
1246 WantedBy=multi-user.target
1247 Please see systemd.kill(5) for details on how you can influence the way
1249 systemd terminates the service.
1250 Example 6. DBus services
1252 For services that acquire a name on the DBus system bus, use Type=dbus
1254 and set BusName= accordingly. The service should not fork (daemonize).
1255 systemd will consider the service to be initialized once the name has
1256 been acquired on the system bus. The following example shows a typical
1257 DBus service:
1258 [Unit]
1260 Description=Simple DBus service
1261 [Service]
1263 Type=dbus
1264 BusName=org.example.simple-dbus-service
1265 ExecStart=/usr/sbin/simple-dbus-service
1266 [Install]
1268 WantedBy=multi-user.target
1269 For bus-activatable services, do not include a [Install] section in the
1271 systemd service file, but use the SystemdService= option in the
1272 corresponding DBus service file, for example
1273 (/usr/share/dbus-1/system-services/org.example.simple-dbus-service.service):
1274 [D-BUS Service]
1276 Name=org.example.simple-dbus-service
1277 Exec=/usr/sbin/simple-dbus-service
1278 User=root
1279 SystemdService=simple-dbus-service.service
1280 Please see systemd.kill(5) for details on how you can influence the way
1282 systemd terminates the service.
1283 Example 7. Services that notify systemd about their initialization
1285 Type=simple services are really easy to write, but have the major
1287 disadvantage of systemd not being able to tell when initialization of
1288 the given service is complete. For this reason, systemd supports a
1289 simple notification protocol that allows daemons to make systemd aware
1290 that they are done initializing. Use Type=notify for this. A typical
1291 service file for such a daemon would look like this:
1292 [Unit]
1294 Description=Simple notifying service
1295 [Service]
1297 Type=notify
1298 ExecStart=/usr/sbin/simple-notifying-service
1299 [Install]
1301 WantedBy=multi-user.target
1302 Note that the daemon has to support systemd's notification protocol,
1304 else systemd will think the service has not started yet and kill it
1305 after a timeout. For an example of how to update daemons to support
1306 this protocol transparently, take a look at sd_notify(3). systemd will
1307 consider the unit to be in the 'starting' state until a readiness
1308 notification has arrived.
1309 Please see systemd.kill(5) for details on how you can influence the way
1311 systemd terminates the service.
1312
1314 systemd(1), systemctl(1), systemd-system.conf(5), systemd.unit(5),
1315 systemd.exec(5), systemd.resource-control(5), systemd.kill(5),
1316 systemd.directives(7), systemd-run(1)
1317
1319 1. Incompatibilities with SysV
1320 https://www.freedesktop.org/wiki/Software/systemd/Incompatibilities
1321 2. USB FunctionFS
1323 https://www.kernel.org/doc/Documentation/usb/functionfs.txt
1324systemd 246 SYSTEMD.SERVICE(5)