]> git.ipfire.org Git - thirdparty/systemd.git/blob - man/systemd.service.xml
treewide: Correct typos and spell plural of bus consistent
[thirdparty/systemd.git] / man / systemd.service.xml
1 <?xml version='1.0'?> <!--*-nxml-*-->
2 <!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
3 "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
4
5 <!--
6 This file is part of systemd.
7
8 Copyright 2010 Lennart Poettering
9
10 systemd is free software; you can redistribute it and/or modify it
11 under the terms of the GNU Lesser General Public License as published by
12 the Free Software Foundation; either version 2.1 of the License, or
13 (at your option) any later version.
14
15 systemd is distributed in the hope that it will be useful, but
16 WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 Lesser General Public License for more details.
19
20 You should have received a copy of the GNU Lesser General Public License
21 along with systemd; If not, see <http://www.gnu.org/licenses/>.
22 -->
23
24 <refentry id="systemd.service">
25 <refentryinfo>
26 <title>systemd.service</title>
27 <productname>systemd</productname>
28
29 <authorgroup>
30 <author>
31 <contrib>Developer</contrib>
32 <firstname>Lennart</firstname>
33 <surname>Poettering</surname>
34 <email>lennart@poettering.net</email>
35 </author>
36 </authorgroup>
37 </refentryinfo>
38
39 <refmeta>
40 <refentrytitle>systemd.service</refentrytitle>
41 <manvolnum>5</manvolnum>
42 </refmeta>
43
44 <refnamediv>
45 <refname>systemd.service</refname>
46 <refpurpose>Service unit configuration</refpurpose>
47 </refnamediv>
48
49 <refsynopsisdiv>
50 <para><filename><replaceable>service</replaceable>.service</filename></para>
51 </refsynopsisdiv>
52
53 <refsect1>
54 <title>Description</title>
55
56 <para>A unit configuration file whose name ends in
57 <filename>.service</filename> encodes information about a process
58 controlled and supervised by systemd.</para>
59
60 <para>This man page lists the configuration options specific to
61 this unit type. See
62 <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>
63 for the common options of all unit configuration files. The common
64 configuration items are configured in the generic
65 <literal>[Unit]</literal> and <literal>[Install]</literal>
66 sections. The service specific configuration options are
67 configured in the <literal>[Service]</literal> section.</para>
68
69 <para>Additional options are listed in
70 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
71 which define the execution environment the commands are executed
72 in, and in
73 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
74 which define the way the processes of the service are terminated,
75 and in
76 <citerefentry><refentrytitle>systemd.resource-control</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
77 which configure resource control settings for the processes of the
78 service.</para>
79
80 <para>Unless <varname>DefaultDependencies=</varname> is set to
81 <option>false</option>, service units will implicitly have
82 dependencies of type <varname>Requires=</varname> and
83 <varname>After=</varname> on <filename>basic.target</filename> as
84 well as dependencies of type <varname>Conflicts=</varname> and
85 <varname>Before=</varname> on
86 <filename>shutdown.target</filename>. These ensure that normal
87 service units pull in basic system initialization, and are
88 terminated cleanly prior to system shutdown. Only services
89 involved with early boot or late system shutdown should disable
90 this option.</para>
91
92 <para>If a service is requested under a certain name but no unit
93 configuration file is found, systemd looks for a SysV init script
94 by the same name (with the <filename>.service</filename> suffix
95 removed) and dynamically creates a service unit from that script.
96 This is useful for compatibility with SysV. Note that this
97 compatibility is quite comprehensive but not 100%. For details
98 about the incompatibilities, see the <ulink
99 url="http://www.freedesktop.org/wiki/Software/systemd/Incompatibilities">Incompatibilities
100 with SysV</ulink> document.
101 </para>
102 </refsect1>
103
104 <refsect1>
105 <title>Options</title>
106
107 <para>Service files must include a <literal>[Service]</literal>
108 section, which carries information about the service and the
109 process it supervises. A number of options that may be used in
110 this section are shared with other unit types. These options are
111 documented in
112 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
113 and
114 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
115 The options specific to the <literal>[Service]</literal> section
116 of service units are the following:</para>
117
118 <variablelist class='unit-directives'>
119 <varlistentry>
120 <term><varname>Type=</varname></term>
121
122 <listitem><para>Configures the process start-up type for this
123 service unit. One of
124 <option>simple</option>,
125 <option>forking</option>,
126 <option>oneshot</option>,
127 <option>dbus</option>,
128 <option>notify</option> or
129 <option>idle</option>.</para>
130
131 <para>If set to <option>simple</option> (the default if
132 neither <varname>Type=</varname> nor
133 <varname>BusName=</varname>, but <varname>ExecStart=</varname>
134 are specified), it is expected that the process configured
135 with <varname>ExecStart=</varname> is the main process of the
136 service. In this mode, if the process offers functionality to
137 other processes on the system, its communication channels
138 should be installed before the daemon is started up (e.g.
139 sockets set up by systemd, via socket activation), as systemd
140 will immediately proceed starting follow-up units.</para>
141
142 <para>If set to <option>forking</option>, it is expected that
143 the process configured with <varname>ExecStart=</varname> will
144 call <function>fork()</function> as part of its start-up. The
145 parent process is expected to exit when start-up is complete
146 and all communication channels are set up. The child continues
147 to run as the main daemon process. This is the behavior of
148 traditional UNIX daemons. If this setting is used, it is
149 recommended to also use the <varname>PIDFile=</varname>
150 option, so that systemd can identify the main process of the
151 daemon. systemd will proceed with starting follow-up units as
152 soon as the parent process exits.</para>
153
154 <para>Behavior of <option>oneshot</option> is similar to
155 <option>simple</option>; however, it is expected that the
156 process has to exit before systemd starts follow-up units.
157 <varname>RemainAfterExit=</varname> is particularly useful for
158 this type of service. This is the implied default if neither
159 <varname>Type=</varname> or <varname>ExecStart=</varname> are
160 specified.</para>
161
162 <para>Behavior of <option>dbus</option> is similar to
163 <option>simple</option>; however, it is expected that the
164 daemon acquires a name on the D-Bus bus, as configured by
165 <varname>BusName=</varname>. systemd will proceed with
166 starting follow-up units after the D-Bus bus name has been
167 acquired. Service units with this option configured implicitly
168 gain dependencies on the <filename>dbus.socket</filename>
169 unit. This type is the default if <varname>BusName=</varname>
170 is specified.</para>
171
172 <para>Behavior of <option>notify</option> is similar to
173 <option>simple</option>; however, it is expected that the
174 daemon sends a notification message via
175 <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>
176 or an equivalent call when it has finished starting up.
177 systemd will proceed with starting follow-up units after this
178 notification message has been sent. If this option is used,
179 <varname>NotifyAccess=</varname> (see below) should be set to
180 open access to the notification socket provided by systemd. If
181 <varname>NotifyAccess=</varname> is not set, it will be
182 implicitly set to <option>main</option>. Note that currently
183 <varname>Type=</varname><option>notify</option> will not work
184 if used in combination with
185 <varname>PrivateNetwork=</varname><option>yes</option>.</para>
186
187 <para>Behavior of <option>idle</option> is very similar to
188 <option>simple</option>; however, actual execution of the
189 service binary is delayed until all jobs are dispatched. This
190 may be used to avoid interleaving of output of shell services
191 with the status output on the console.</para>
192 </listitem>
193 </varlistentry>
194
195 <varlistentry>
196 <term><varname>RemainAfterExit=</varname></term>
197
198 <listitem><para>Takes a boolean value that specifies whether
199 the service shall be considered active even when all its
200 processes exited. Defaults to <option>no</option>.</para>
201 </listitem>
202 </varlistentry>
203
204 <varlistentry>
205 <term><varname>GuessMainPID=</varname></term>
206
207 <listitem><para>Takes a boolean value that specifies whether
208 systemd should try to guess the main PID of a service if it
209 cannot be determined reliably. This option is ignored unless
210 <option>Type=forking</option> is set and
211 <option>PIDFile=</option> is unset because for the other types
212 or with an explicitly configured PID file, the main PID is
213 always known. The guessing algorithm might come to incorrect
214 conclusions if a daemon consists of more than one process. If
215 the main PID cannot be determined, failure detection and
216 automatic restarting of a service will not work reliably.
217 Defaults to <option>yes</option>.</para>
218 </listitem>
219 </varlistentry>
220
221 <varlistentry>
222 <term><varname>PIDFile=</varname></term>
223
224 <listitem><para>Takes an absolute file name pointing to the
225 PID file of this daemon. Use of this option is recommended for
226 services where <varname>Type=</varname> is set to
227 <option>forking</option>. systemd will read the PID of the
228 main process of the daemon after start-up of the service.
229 systemd will not write to the file configured here.</para>
230 </listitem>
231 </varlistentry>
232
233 <varlistentry>
234 <term><varname>BusName=</varname></term>
235
236 <listitem><para>Takes a D-Bus bus name that this service is
237 reachable as. This option is mandatory for services where
238 <varname>Type=</varname> is set to
239 <option>dbus</option>.</para>
240 </listitem>
241 </varlistentry>
242
243 <varlistentry>
244 <term><varname>BusPolicy=</varname></term>
245
246 <listitem><para>If specified, a custom
247 <ulink url="https://code.google.com/p/d-bus/">kdbus</ulink>
248 endpoint will be created and installed as the default bus node
249 for the service. Such a custom endpoint can hold an own set of
250 policy rules that are enforced on top of the bus-wide ones.
251 The custom endpoint is named after the service it was created
252 for, and its node will be bind-mounted over the default bus
253 node location, so the service can only access the bus through
254 its own endpoint. Note that custom bus endpoints default to a
255 'deny all' policy. Hence, if at least one
256 <varname>BusPolicy=</varname> directive is given, you have to
257 make sure to add explicit rules for everything the service
258 should be able to do.</para>
259 <para>The value of this directive is comprised
260 of two parts; the bus name, and a verb to
261 specify to granted access, which is one of
262 <option>see</option>,
263 <option>talk</option>, or
264 <option>own</option>.
265 <option>talk</option> implies
266 <option>see</option>, and <option>own</option>
267 implies both <option>talk</option> and
268 <option>see</option>.
269 If multiple access levels are specified for the
270 same bus name, the most powerful one takes
271 effect.
272 </para>
273 <para>Examples:</para>
274 <programlisting>BusPolicy=org.freedesktop.systemd1 talk</programlisting>
275 <programlisting>BusPolicy=org.foo.bar see</programlisting>
276 <para>This option is only available on kdbus enabled systems.</para>
277 </listitem>
278 </varlistentry>
279
280 <varlistentry>
281 <term><varname>ExecStart=</varname></term>
282 <listitem><para>Commands with their arguments that are
283 executed when this service is started. The value is split into
284 zero or more command lines is according to the rules described
285 below (see section "Command Lines" below).
286 </para>
287
288 <para>When <varname>Type</varname> is not
289 <option>oneshot</option>, only one command may and must be
290 given. When <varname>Type=oneshot</varname> is used, zero or
291 more commands may be specified. This can be specified by
292 providing multiple command lines in the same directive, or
293 alternatively, this directive may be specified more than once
294 with the same effect. If the empty string is assigned to this
295 option, the list of commands to start is reset, prior
296 assignments of this option will have no effect. If no
297 <varname>ExecStart=</varname> is specified, then the service
298 must have <varname>RemainAfterExit=yes</varname> set.</para>
299
300 <para>For each of the specified commands, the first argument
301 must be an absolute path to an executable. Optionally, if this
302 file name is prefixed with <literal>@</literal>, the second
303 token will be passed as <literal>argv[0]</literal> to the
304 executed process, followed by the further arguments specified.
305 If the absolute filename is prefixed with
306 <literal>-</literal>, an exit code of the command normally
307 considered a failure (i.e. non-zero exit status or abnormal
308 exit due to signal) is ignored and considered success. If both
309 <literal>-</literal> and <literal>@</literal> are used, they
310 can appear in either order.</para>
311
312 <para>If more than one command is specified, the commands are
313 invoked sequentially in the order they appear in the unit
314 file. If one of the commands fails (and is not prefixed with
315 <literal>-</literal>), other lines are not executed, and the
316 unit is considered failed.</para>
317
318 <para>Unless <varname>Type=forking</varname> is set, the
319 process started via this command line will be considered the
320 main process of the daemon.</para>
321 </listitem>
322 </varlistentry>
323
324 <varlistentry>
325 <term><varname>ExecStartPre=</varname></term>
326 <term><varname>ExecStartPost=</varname></term>
327 <listitem><para>Additional commands that are executed before
328 or after the command in <varname>ExecStart=</varname>,
329 respectively. Syntax is the same as for
330 <varname>ExecStart=</varname>, except that multiple command
331 lines are allowed and the commands are executed one after the
332 other, serially.</para>
333
334 <para>If any of those commands (not prefixed with
335 <literal>-</literal>) fail, the rest are not executed and the
336 unit is considered failed.</para>
337
338 <para>Note that <varname>ExecStartPre=</varname> may not be
339 used to start long-running processes. All processes forked
340 off by processes invoked via <varname>ExecStartPre=</varname> will
341 be killed before the next service process is run.</para>
342 </listitem>
343 </varlistentry>
344
345 <varlistentry>
346 <term><varname>ExecReload=</varname></term>
347 <listitem><para>Commands to execute to trigger a configuration
348 reload in the service. This argument takes multiple command
349 lines, following the same scheme as described for
350 <varname>ExecStart=</varname> above. Use of this setting is
351 optional. Specifier and environment variable substitution is
352 supported here following the same scheme as for
353 <varname>ExecStart=</varname>.</para>
354
355 <para>One additional, special environment variable is set: if
356 known, <varname>$MAINPID</varname> is set to the main process
357 of the daemon, and may be used for command lines like the
358 following:</para>
359
360 <programlisting>/bin/kill -HUP $MAINPID</programlisting>
361
362 <para>Note however that reloading a daemon by sending a signal
363 (as with the example line above) is usually not a good choice,
364 because this is an asynchronous operation and hence not
365 suitable to order reloads of multiple services against each
366 other. It is strongly recommended to set
367 <varname>ExecReload=</varname> to a command that not only
368 triggers a configuration reload of the daemon, but also
369 synchronously waits for it to complete.</para>
370 </listitem>
371 </varlistentry>
372
373 <varlistentry>
374 <term><varname>ExecStop=</varname></term>
375 <listitem><para>Commands to execute to stop the service
376 started via <varname>ExecStart=</varname>. This argument takes
377 multiple command lines, following the same scheme as described
378 for <varname>ExecStart=</varname> above. Use of this setting
379 is optional. After the commands configured in this option are
380 run, all processes remaining for a service are terminated
381 according to the <varname>KillMode=</varname> setting (see
382 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>).
383 If this option is not specified, the process is terminated
384 immediately when service stop is requested. Specifier and
385 environment variable substitution is supported (including
386 <varname>$MAINPID</varname>, see above).</para></listitem>
387 </varlistentry>
388
389 <varlistentry>
390 <term><varname>ExecStopPost=</varname></term>
391 <listitem><para>Additional commands that are executed after
392 the service was stopped. This includes cases where the
393 commands configured in <varname>ExecStop=</varname> were used,
394 where the service does not have any
395 <varname>ExecStop=</varname> defined, or where the service
396 exited unexpectedly. This argument takes multiple command
397 lines, following the same scheme as described for
398 <varname>ExecStart</varname>. Use of these settings is
399 optional. Specifier and environment variable substitution is
400 supported.</para></listitem>
401 </varlistentry>
402
403 <varlistentry>
404 <term><varname>RestartSec=</varname></term>
405 <listitem><para>Configures the time to sleep before restarting
406 a service (as configured with <varname>Restart=</varname>).
407 Takes a unit-less value in seconds, or a time span value such
408 as "5min 20s". Defaults to 100ms.</para></listitem>
409 </varlistentry>
410
411 <varlistentry>
412 <term><varname>TimeoutStartSec=</varname></term>
413 <listitem><para>Configures the time to wait for start-up. If a
414 daemon service does not signal start-up completion within the
415 configured time, the service will be considered failed and
416 will be shut down again. Takes a unit-less value in seconds,
417 or a time span value such as "5min 20s". Pass
418 <literal>0</literal> to disable the timeout logic. Defaults to
419 <varname>DefaultTimeoutStartSec=</varname> from the manager
420 configuration file, except when
421 <varname>Type=oneshot</varname> is used, in which case the
422 timeout is disabled by default (see
423 <citerefentry><refentrytitle>systemd-system.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>).
424 </para></listitem>
425 </varlistentry>
426
427 <varlistentry>
428 <term><varname>TimeoutStopSec=</varname></term>
429 <listitem><para>Configures the time to wait for stop. If a
430 service is asked to stop, but does not terminate in the
431 specified time, it will be terminated forcibly via
432 <constant>SIGTERM</constant>, and after another timeout of
433 equal duration with <constant>SIGKILL</constant> (see
434 <varname>KillMode=</varname> in
435 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>).
436 Takes a unit-less value in seconds, or a time span value such
437 as "5min 20s". Pass <literal>0</literal> to disable the
438 timeout logic. Defaults to
439 <varname>DefaultTimeoutStopSec=</varname> from the manager
440 configuration file (see
441 <citerefentry><refentrytitle>systemd-system.conf</refentrytitle><manvolnum>5</manvolnum></citerefentry>).
442 </para></listitem>
443 </varlistentry>
444
445 <varlistentry>
446 <term><varname>TimeoutSec=</varname></term>
447 <listitem><para>A shorthand for configuring both
448 <varname>TimeoutStartSec=</varname> and
449 <varname>TimeoutStopSec=</varname> to the specified value.
450 </para></listitem>
451 </varlistentry>
452
453 <varlistentry>
454 <term><varname>WatchdogSec=</varname></term>
455 <listitem><para>Configures the watchdog timeout for a service.
456 The watchdog is activated when the start-up is completed. The
457 service must call
458 <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>
459 regularly with <literal>WATCHDOG=1</literal> (i.e. the
460 "keep-alive ping"). If the time between two such calls is
461 larger than the configured time, then the service is placed in
462 a failed state and it will be terminated with
463 <varname>SIGABRT</varname>. By setting
464 <varname>Restart=</varname> to <option>on-failure</option> or
465 <option>always</option>, the service will be automatically
466 restarted. The time configured here will be passed to the
467 executed service process in the
468 <varname>WATCHDOG_USEC=</varname> environment variable. This
469 allows daemons to automatically enable the keep-alive pinging
470 logic if watchdog support is enabled for the service. If this
471 option is used, <varname>NotifyAccess=</varname> (see below)
472 should be set to open access to the notification socket
473 provided by systemd. If <varname>NotifyAccess=</varname> is
474 not set, it will be implicitly set to <option>main</option>.
475 Defaults to 0, which disables this feature.</para></listitem>
476 </varlistentry>
477
478 <varlistentry>
479 <term><varname>Restart=</varname></term>
480 <listitem><para>Configures whether the service shall be
481 restarted when the service process exits, is killed, or a
482 timeout is reached. The service process may be the main
483 service process, but it may also be one of the processes
484 specified with <varname>ExecStartPre=</varname>,
485 <varname>ExecStartPost=</varname>,
486 <varname>ExecStop=</varname>,
487 <varname>ExecStopPost=</varname>, or
488 <varname>ExecReload=</varname>. When the death of the process
489 is a result of systemd operation (e.g. service stop or
490 restart), the service will not be restarted. Timeouts include
491 missing the watchdog "keep-alive ping" deadline and a service
492 start, reload, and stop operation timeouts.</para>
493
494 <para>Takes one of
495 <option>no</option>,
496 <option>on-success</option>,
497 <option>on-failure</option>,
498 <option>on-abnormal</option>,
499 <option>on-watchdog</option>,
500 <option>on-abort</option>, or
501 <option>always</option>.
502 If set to <option>no</option> (the default), the service will
503 not be restarted. If set to <option>on-success</option>, it
504 will be restarted only when the service process exits cleanly.
505 In this context, a clean exit means an exit code of 0, or one
506 of the signals
507 <constant>SIGHUP</constant>,
508 <constant>SIGINT</constant>,
509 <constant>SIGTERM</constant> or
510 <constant>SIGPIPE</constant>, and
511 additionally, exit statuses and signals specified in
512 <varname>SuccessExitStatus=</varname>. If set to
513 <option>on-failure</option>, the service will be restarted
514 when the process exits with a non-zero exit code, is
515 terminated by a signal (including on core dump, but excluding
516 the aforementioned four signals), when an operation (such as
517 service reload) times out, and when the configured watchdog
518 timeout is triggered. If set to <option>on-abnormal</option>,
519 the service will be restarted when the process is terminated
520 by a signal (including on core dump, excluding the
521 aforementioned four signals), when an operation times out, or
522 when the watchdog timeout is triggered. If set to
523 <option>on-abort</option>, the service will be restarted only
524 if the service process exits due to an uncaught signal not
525 specified as a clean exit status. If set to
526 <option>on-watchdog</option>, the service will be restarted
527 only if the watchdog timeout for the service expires. If set
528 to <option>always</option>, the service will be restarted
529 regardless of whether it exited cleanly or not, got terminated
530 abnormally by a signal, or hit a timeout.</para>
531
532 <table>
533 <title>Exit causes and the effect of the <varname>Restart=</varname> settings on them</title>
534
535 <tgroup cols='2'>
536 <colspec colname='path' />
537 <colspec colname='expl' />
538 <thead>
539 <row>
540 <entry>Restart settings/Exit causes</entry>
541 <entry><option>no</option></entry>
542 <entry><option>always</option></entry>
543 <entry><option>on-success</option></entry>
544 <entry><option>on-failure</option></entry>
545 <entry><option>on-abnormal</option></entry>
546 <entry><option>on-abort</option></entry>
547 <entry><option>on-watchdog</option></entry>
548 </row>
549 </thead>
550 <tbody>
551 <row>
552 <entry>Clean exit code or signal</entry>
553 <entry/>
554 <entry>X</entry>
555 <entry>X</entry>
556 <entry/>
557 <entry/>
558 <entry/>
559 <entry/>
560 </row>
561 <row>
562 <entry>Unclean exit code</entry>
563 <entry/>
564 <entry>X</entry>
565 <entry/>
566 <entry>X</entry>
567 <entry/>
568 <entry/>
569 <entry/>
570 </row>
571 <row>
572 <entry>Unclean signal</entry>
573 <entry/>
574 <entry>X</entry>
575 <entry/>
576 <entry>X</entry>
577 <entry>X</entry>
578 <entry>X</entry>
579 <entry/>
580 </row>
581 <row>
582 <entry>Timeout</entry>
583 <entry/>
584 <entry>X</entry>
585 <entry/>
586 <entry>X</entry>
587 <entry>X</entry>
588 <entry/>
589 <entry/>
590 </row>
591 <row>
592 <entry>Watchdog</entry>
593 <entry/>
594 <entry>X</entry>
595 <entry/>
596 <entry>X</entry>
597 <entry>X</entry>
598 <entry/>
599 <entry>X</entry>
600 </row>
601 </tbody>
602 </tgroup>
603 </table>
604
605 <para>As exceptions to the setting above the service will not
606 be restarted if the exit code or signal is specified in
607 <varname>RestartPreventExitStatus=</varname> (see below).
608 Also, the services will always be restarted if the exit code
609 or signal is specified in
610 <varname>RestartForceExitStatus=</varname> (see below).</para>
611
612 <para>Setting this to <option>on-failure</option> is the
613 recommended choice for long-running services, in order to
614 increase reliability by attempting automatic recovery from
615 errors. For services that shall be able to terminate on their
616 own choice (and avoid immediate restarting),
617 <option>on-abnormal</option> is an alternative choice.</para>
618 </listitem>
619 </varlistentry>
620
621 <varlistentry>
622 <term><varname>SuccessExitStatus=</varname></term>
623 <listitem><para>Takes a list of exit status definitions that
624 when returned by the main service process will be considered
625 successful termination, in addition to the normal successful
626 exit code 0 and the signals <constant>SIGHUP</constant>,
627 <constant>SIGINT</constant>, <constant>SIGTERM</constant>, and
628 <constant>SIGPIPE</constant>. Exit status definitions can
629 either be numeric exit codes or termination signal names,
630 separated by spaces. For example:
631 <programlisting>SuccessExitStatus=1 2 8
632 SIGKILL</programlisting> ensures that exit codes 1, 2, 8 and
633 the termination signal <constant>SIGKILL</constant> are
634 considered clean service terminations.
635 </para>
636
637 <para>Note that if a process has a signal handler installed
638 and exits by calling
639 <citerefentry><refentrytitle>_exit</refentrytitle><manvolnum>2</manvolnum></citerefentry>
640 in response to a signal, the information about the signal is
641 lost. Programs should instead perform cleanup and kill
642 themselves with the same signal instead. See
643 <ulink url="http://www.cons.org/cracauer/sigint.html">Proper
644 handling of SIGINT/SIGQUIT — How to be a proper
645 program</ulink>.</para>
646
647 <para>This option may appear more than once, in which case the
648 list of successful exit statuses is merged. If the empty
649 string is assigned to this option, the list is reset, all
650 prior assignments of this option will have no
651 effect.</para></listitem>
652 </varlistentry>
653
654 <varlistentry>
655 <term><varname>RestartPreventExitStatus=</varname></term>
656 <listitem><para>Takes a list of exit status definitions that
657 when returned by the main service process will prevent
658 automatic service restarts, regardless of the restart setting
659 configured with <varname>Restart=</varname>. Exit status
660 definitions can either be numeric exit codes or termination
661 signal names, and are separated by spaces. Defaults to the
662 empty list, so that, by default, no exit status is excluded
663 from the configured restart logic. For example:
664 <programlisting>RestartPreventExitStatus=1 6
665 SIGABRT</programlisting> ensures that exit codes 1 and 6 and
666 the termination signal <constant>SIGABRT</constant> will not
667 result in automatic service restarting. This option may appear
668 more than once, in which case the list of restart-preventing
669 statuses is merged. If the empty string is assigned to this
670 option, the list is reset and all prior assignments of this
671 option will have no effect.</para></listitem>
672 </varlistentry>
673
674 <varlistentry>
675 <term><varname>RestartForceExitStatus=</varname></term>
676 <listitem><para>Takes a list of exit status definitions that
677 when returned by the main service process will force automatic
678 service restarts, regardless of the restart setting configured
679 with <varname>Restart=</varname>. The argument format is
680 similar to
681 <varname>RestartPreventExitStatus=</varname>.</para></listitem>
682 </varlistentry>
683
684 <varlistentry>
685 <term><varname>PermissionsStartOnly=</varname></term>
686 <listitem><para>Takes a boolean argument. If true, the
687 permission-related execution options, as configured with
688 <varname>User=</varname> and similar options (see
689 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
690 for more information), are only applied to the process started
691 with
692 <varname>ExecStart=</varname>, and not to the various other
693 <varname>ExecStartPre=</varname>,
694 <varname>ExecStartPost=</varname>,
695 <varname>ExecReload=</varname>,
696 <varname>ExecStop=</varname>, and
697 <varname>ExecStopPost=</varname>
698 commands. If false, the setting is applied to all configured
699 commands the same way. Defaults to false.</para></listitem>
700 </varlistentry>
701
702 <varlistentry>
703 <term><varname>RootDirectoryStartOnly=</varname></term>
704 <listitem><para>Takes a boolean argument. If true, the root
705 directory, as configured with the
706 <varname>RootDirectory=</varname> option (see
707 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
708 for more information), is only applied to the process started
709 with <varname>ExecStart=</varname>, and not to the various
710 other <varname>ExecStartPre=</varname>,
711 <varname>ExecStartPost=</varname>,
712 <varname>ExecReload=</varname>, <varname>ExecStop=</varname>,
713 and <varname>ExecStopPost=</varname> commands. If false, the
714 setting is applied to all configured commands the same way.
715 Defaults to false.</para></listitem>
716 </varlistentry>
717
718 <varlistentry>
719 <term><varname>NonBlocking=</varname></term>
720 <listitem><para>Set the <constant>O_NONBLOCK</constant> flag
721 for all file descriptors passed via socket-based activation.
722 If true, all file descriptors >= 3 (i.e. all except stdin,
723 stdout, and stderr) will have the
724 <constant>O_NONBLOCK</constant> flag set and hence are in
725 non-blocking mode. This option is only useful in conjunction
726 with a socket unit, as described in
727 <citerefentry><refentrytitle>systemd.socket</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
728 Defaults to false.</para></listitem>
729 </varlistentry>
730
731 <varlistentry>
732 <term><varname>NotifyAccess=</varname></term>
733 <listitem><para>Controls access to the service status
734 notification socket, as accessible via the
735 <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>
736 call. Takes one of <option>none</option> (the default),
737 <option>main</option> or <option>all</option>. If
738 <option>none</option>, no daemon status updates are accepted
739 from the service processes, all status update messages are
740 ignored. If <option>main</option>, only service updates sent
741 from the main process of the service are accepted. If
742 <option>all</option>, all services updates from all members of
743 the service's control group are accepted. This option should
744 be set to open access to the notification socket when using
745 <varname>Type=notify</varname> or
746 <varname>WatchdogSec=</varname> (see above). If those options
747 are used but <varname>NotifyAccess=</varname> is not
748 configured, it will be implicitly set to
749 <option>main</option>.</para></listitem>
750 </varlistentry>
751
752 <varlistentry>
753 <term><varname>Sockets=</varname></term>
754 <listitem><para>Specifies the name of the socket units this
755 service shall inherit socket file descriptors from when the
756 service is started. Normally it should not be necessary to use
757 this setting as all socket file descriptors whose unit shares
758 the same name as the service (subject to the different unit
759 name suffix of course) are passed to the spawned
760 process.</para>
761
762 <para>Note that the same socket file descriptors may be passed
763 to multiple processes simultaneously. Also note that a
764 different service may be activated on incoming socket traffic
765 than the one which is ultimately configured to inherit the
766 socket file descriptors. Or in other words: the
767 <varname>Service=</varname> setting of
768 <filename>.socket</filename> units does not have to match the
769 inverse of the <varname>Sockets=</varname> setting of the
770 <filename>.service</filename> it refers to.</para>
771
772 <para>This option may appear more than once, in which case the
773 list of socket units is merged. If the empty string is
774 assigned to this option, the list of sockets is reset, and all
775 prior uses of this setting will have no
776 effect.</para></listitem>
777 </varlistentry>
778
779 <varlistentry>
780 <term><varname>StartLimitInterval=</varname></term>
781 <term><varname>StartLimitBurst=</varname></term>
782
783 <listitem><para>Configure service start rate limiting. By
784 default, services which are started more than 5 times within
785 10 seconds are not permitted to start any more times until the
786 10 second interval ends. With these two options, this rate
787 limiting may be modified. Use
788 <varname>StartLimitInterval=</varname> to configure the
789 checking interval (defaults to
790 <varname>DefaultStartLimitInterval=</varname> in manager
791 configuration file, set to 0 to disable any kind of rate
792 limiting). Use <varname>StartLimitBurst=</varname> to
793 configure how many starts per interval are allowed (defaults
794 to <varname>DefaultStartLimitBurst=</varname> in manager
795 configuration file). These configuration options are
796 particularly useful in conjunction with
797 <varname>Restart=</varname>; however, they apply to all kinds
798 of starts (including manual), not just those triggered by the
799 <varname>Restart=</varname> logic. Note that units which are
800 configured for <varname>Restart=</varname> and which reach the
801 start limit are not attempted to be restarted anymore;
802 however, they may still be restarted manually at a later
803 point, from which point on, the restart logic is again
804 activated. Note that <command>systemctl reset-failed</command>
805 will cause the restart rate counter for a service to be
806 flushed, which is useful if the administrator wants to
807 manually start a service and the start limit interferes with
808 that.</para></listitem>
809 </varlistentry>
810
811 <varlistentry>
812 <term><varname>StartLimitAction=</varname></term>
813
814 <listitem><para>Configure the action to take if the rate limit
815 configured with <varname>StartLimitInterval=</varname> and
816 <varname>StartLimitBurst=</varname> is hit. Takes one of
817 <option>none</option>,
818 <option>reboot</option>,
819 <option>reboot-force</option>,
820 <option>reboot-immediate</option>,
821 <option>poweroff</option>,
822 <option>poweroff-force</option> or
823 <option>poweroff-immediate</option>. If
824 <option>none</option> is set, hitting the rate limit will
825 trigger no action besides that the start will not be
826 permitted. <option>reboot</option> causes a reboot following
827 the normal shutdown procedure (i.e. equivalent to
828 <command>systemctl reboot</command>).
829 <option>reboot-force</option> causes a forced reboot which
830 will terminate all processes forcibly but should cause no
831 dirty file systems on reboot (i.e. equivalent to
832 <command>systemctl reboot -f</command>) and
833 <option>reboot-immediate</option> causes immediate execution
834 of the
835 <citerefentry><refentrytitle>reboot</refentrytitle><manvolnum>2</manvolnum></citerefentry>
836 system call, which might result in data loss. Similar,
837 <option>poweroff</option>, <option>poweroff-force</option>,
838 <option>poweroff-immediate</option> have the effect of
839 powering down the system with similar semantics. Defaults to
840 <option>none</option>.</para></listitem>
841 </varlistentry>
842
843 <varlistentry>
844 <term><varname>FailureAction=</varname></term>
845 <listitem><para>Configure the action to take when the service
846 enters a failed state. Takes the same values as
847 <varname>StartLimitAction=</varname> and executes the same
848 actions. Defaults to <option>none</option>. </para></listitem>
849 </varlistentry>
850
851 <varlistentry>
852 <term><varname>RebootArgument=</varname></term>
853 <listitem><para>Configure the optional argument for the
854 <citerefentry><refentrytitle>reboot</refentrytitle><manvolnum>2</manvolnum></citerefentry>
855 system call if <varname>StartLimitAction=</varname> or
856 <varname>FailureAction=</varname> is a reboot action. This
857 works just like the optional argument to <command>systemctl
858 reboot</command> command.</para></listitem>
859 </varlistentry>
860
861 <varlistentry>
862 <term><varname>FileDescriptorStoreMax=</varname></term>
863 <listitem><para>Configure how many file descriptors may be
864 stored in the service manager for the service using
865 <citerefentry><refentrytitle>sd_pid_notify_with_fds</refentrytitle><manvolnum>3</manvolnum></citerefentry>'s
866 <literal>FDSTORE=1</literal> messages. This is useful for
867 implementing service restart schemes where the state is
868 serialized to <filename>/run</filename> and the file
869 descriptors passed to the service manager, to allow restarts
870 without losing state. Defaults to 0, i.e. no file descriptors
871 may be stored in the service manager by default. All file
872 descriptors passed to the service manager from a specific
873 service are passed back to the service's main process on the
874 next service restart. Any file descriptors passed to the
875 service manager are automatically closed when POLLHUP or
876 POLLERR is seen on them, or when the service is fully stopped
877 and no job queued or being executed for it.</para></listitem>
878 </varlistentry>
879
880 </variablelist>
881
882 <para>Check
883 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
884 and
885 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
886 for more settings.</para>
887
888 </refsect1>
889
890 <refsect1>
891 <title>Command lines</title>
892
893 <para>This section describes command line parsing and
894 variable and specifier substitutions for
895 <varname>ExecStart=</varname>,
896 <varname>ExecStartPre=</varname>,
897 <varname>ExecStartPost=</varname>,
898 <varname>ExecReload=</varname>,
899 <varname>ExecStop=</varname>, and
900 <varname>ExecStopPost=</varname> options.</para>
901
902 <para>Multiple command lines may be concatenated in a single
903 directive by separating them with semicolons (these semicolons
904 must be passed as separate words). Lone semicolons may be escaped
905 as <literal>\;</literal>.</para>
906
907 <para>Each command line is split on whitespace, with the first
908 item being the command to execute, and the subsequent items being
909 the arguments. Double quotes ("...") and single quotes ('...') may
910 be used, in which case everything until the next matching quote
911 becomes part of the same argument. C-style escapes are also
912 supported, see table below. Quotes themselves are removed after
913 parsing and escape sequences substituted. In addition, a trailing
914 backslash (<literal>\</literal>) may be used to merge lines.
915 </para>
916
917 <para>This syntax is intended to be very similar to shell syntax,
918 but only the meta-characters and expansions described in the
919 following paragraphs are understood. Specifically, redirection
920 using
921 <literal>&lt;</literal>,
922 <literal>&lt;&lt;</literal>,
923 <literal>&gt;</literal>, and
924 <literal>&gt;&gt;</literal>, pipes using
925 <literal>|</literal>, running programs in the background using
926 <literal>&amp;</literal>, and <emphasis>other elements of shell
927 syntax are not supported</emphasis>.</para>
928
929 <para>The command to execute must an absolute path name. It may
930 contain spaces, but control characters are not allowed.</para>
931
932 <para>The command line accepts <literal>%</literal> specifiers as
933 described in
934 <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>.
935 Note that the first argument of the command line (i.e. the program
936 to execute) may not include specifiers.</para>
937
938 <para>Basic environment variable substitution is supported. Use
939 <literal>${FOO}</literal> as part of a word, or as a word of its
940 own, on the command line, in which case it will be replaced by the
941 value of the environment variable including all whitespace it
942 contains, resulting in a single argument. Use
943 <literal>$FOO</literal> as a separate word on the command line, in
944 which case it will be replaced by the value of the environment
945 variable split at whitespace resulting in zero or more arguments.
946 For this type of expansion, quotes and respected when splitting
947 into words, and afterwards removed.</para>
948
949 <para>Example:</para>
950
951 <programlisting>Environment="ONE=one" 'TWO=two two'
952 ExecStart=/bin/echo $ONE $TWO ${TWO}</programlisting>
953
954 <para>This will execute <command>/bin/echo</command> with four
955 arguments: <literal>one</literal>, <literal>two</literal>,
956 <literal>two</literal>, and <literal>two two</literal>.</para>
957
958 <para>Example:</para>
959 <programlisting>Environment=ONE='one' "TWO='two two' too" THREE=
960 ExecStart=/bin/echo ${ONE} ${TWO} ${THREE}
961 ExecStart=/bin/echo $ONE $TWO $THREE</programlisting>
962 <para>This results in <filename>echo</filename> being
963 called twice, the first time with arguments
964 <literal>'one'</literal>,
965 <literal>'two two' too</literal>, <literal></literal>,
966 and the second time with arguments
967 <literal>one</literal>, <literal>two two</literal>,
968 <literal>too</literal>.
969 </para>
970
971 <para>To pass a literal dollar sign, use <literal>$$</literal>.
972 Variables whose value is not known at expansion time are treated
973 as empty strings. Note that the first argument (i.e. the program
974 to execute) may not be a variable.</para>
975
976 <para>Variables to be used in this fashion may be defined through
977 <varname>Environment=</varname> and
978 <varname>EnvironmentFile=</varname>. In addition, variables listed
979 in the section "Environment variables in spawned processes" in
980 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
981 which are considered "static configuration", may be used (this
982 includes e.g. <varname>$USER</varname>, but not
983 <varname>$TERM</varname>).</para>
984
985 <para>Note that shell command lines are not directly supported. If
986 shell command lines are to be used, they need to be passed
987 explicitly to a shell implementation of some kind. Example:</para>
988 <programlisting>ExecStart=/bin/sh -c 'dmesg | tac'</programlisting>
989
990 <para>Example:</para>
991
992 <programlisting>ExecStart=/bin/echo one ; /bin/echo "two two"</programlisting>
993
994 <para>This will execute <command>/bin/echo</command> two times,
995 each time with one argument: <literal>one</literal> and
996 <literal>two two</literal>, respectively. Because two commands are
997 specified, <varname>Type=oneshot</varname> must be used.</para>
998
999 <para>Example:</para>
1000
1001 <programlisting>ExecStart=/bin/echo / &gt;/dev/null &amp; \; \
1002 /bin/ls</programlisting>
1003
1004 <para>This will execute <command>/bin/echo</command>
1005 with five arguments: <literal>/</literal>,
1006 <literal>&gt;/dev/null</literal>,
1007 <literal>&amp;</literal>, <literal>;</literal>, and
1008 <literal>/bin/ls</literal>.</para>
1009
1010 <table>
1011 <title>C escapes supported in command lines and environment variables</title>
1012 <tgroup cols='2'>
1013 <colspec colname='escape' />
1014 <colspec colname='meaning' />
1015 <thead>
1016 <row>
1017 <entry>Literal</entry>
1018 <entry>Actual value</entry>
1019 </row>
1020 </thead>
1021 <tbody>
1022 <row>
1023 <entry><literal>\a</literal></entry>
1024 <entry>bell</entry>
1025 </row>
1026 <row>
1027 <entry><literal>\b</literal></entry>
1028 <entry>backspace</entry>
1029 </row>
1030 <row>
1031 <entry><literal>\f</literal></entry>
1032 <entry>form feed</entry>
1033 </row>
1034 <row>
1035 <entry><literal>\n</literal></entry>
1036 <entry>newline</entry>
1037 </row>
1038 <row>
1039 <entry><literal>\r</literal></entry>
1040 <entry>carriage return</entry>
1041 </row>
1042 <row>
1043 <entry><literal>\t</literal></entry>
1044 <entry>tab</entry>
1045 </row>
1046 <row>
1047 <entry><literal>\v</literal></entry>
1048 <entry>vertical tab</entry>
1049 </row>
1050 <row>
1051 <entry><literal>\\</literal></entry>
1052 <entry>backslash</entry>
1053 </row>
1054 <row>
1055 <entry><literal>\"</literal></entry>
1056 <entry>double quotation mark</entry>
1057 </row>
1058 <row>
1059 <entry><literal>\'</literal></entry>
1060 <entry>single quotation mark</entry>
1061 </row>
1062 <row>
1063 <entry><literal>\s</literal></entry>
1064 <entry>space</entry>
1065 </row>
1066 <row>
1067 <entry><literal>\x<replaceable>xx</replaceable></literal></entry>
1068 <entry>character number <replaceable>xx</replaceable> in hexadecimal encoding</entry>
1069 </row>
1070 <row>
1071 <entry><literal>\<replaceable>nnn</replaceable></literal></entry>
1072 <entry>character number <replaceable>nnn</replaceable> in octal encoding</entry>
1073 </row>
1074 </tbody>
1075 </tgroup>
1076 </table>
1077 </refsect1>
1078
1079 <refsect1>
1080 <title>Examples</title>
1081
1082 <example>
1083 <title>Simple service</title>
1084
1085 <para>The following unit file creates a service that will
1086 execute <filename>/usr/sbin/foo-daemon</filename>. Since no
1087 <varname>Type=</varname> is specified, the default
1088 <varname>Type=</varname><option>simple</option> will be assumed.
1089 systemd will assume the unit to be started immediately after the
1090 program has begun executing.</para>
1091
1092 <programlisting>[Unit]
1093 Description=Foo
1094
1095 [Service]
1096 ExecStart=/usr/sbin/foo-daemon
1097
1098 [Install]
1099 WantedBy=multi-user.target</programlisting>
1100
1101 <para>Note that systemd assumes here that the process started by
1102 systemd will continue running until the service terminates. If
1103 the program daemonizes itself (i.e. forks), please use
1104 <varname>Type=</varname><option>forking</option> instead.</para>
1105
1106 <para>Since no <varname>ExecStop=</varname> was specified,
1107 systemd will send SIGTERM to all processes started from this
1108 service, and after a timeout also SIGKILL. This behavior can be
1109 modified, see
1110 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
1111 for details.</para>
1112
1113 <para>Note that this unit type does not include any type of
1114 notification when a service has completed initialization. For
1115 this, you should use other unit types, such as
1116 <varname>Type=</varname><option>notify</option> if the service
1117 understands systemd's notification protocol,
1118 <varname>Type=</varname><option>forking</option> if the service
1119 can background itself or
1120 <varname>Type=</varname><option>dbus</option> if the unit
1121 acquires a DBus name once initialization is complete. See
1122 below.</para>
1123 </example>
1124
1125 <example>
1126 <title>Oneshot service</title>
1127
1128 <para>Sometimes units should just execute an action without
1129 keeping active processes, such as a filesystem check or a
1130 cleanup action on boot. For this,
1131 <varname>Type=</varname><option>oneshot</option> exists. Units
1132 of this type will wait until the process specified terminates
1133 and then fall back to being inactive. The following unit will
1134 perform a cleanup action:</para>
1135
1136 <programlisting>[Unit]
1137 Description=Cleanup old Foo data
1138
1139 [Service]
1140 Type=oneshot
1141 ExecStart=/usr/sbin/foo-cleanup
1142
1143 [Install]
1144 WantedBy=multi-user.target</programlisting>
1145
1146 <para>Note that systemd will consider the unit to be in the
1147 state 'starting' until the program has terminated, so ordered
1148 dependencies will wait for the program to finish before starting
1149 themselves. The unit will revert to the 'inactive' state after
1150 the execution is done, never reaching the 'active' state. That
1151 means another request to start the unit will perform the action
1152 again.</para>
1153
1154 <para><varname>Type=</varname><option>oneshot</option> are the
1155 only service units that may have more than one
1156 <varname>ExecStart=</varname> specified. They will be executed
1157 in order until either they are all successful or one of them
1158 fails.</para>
1159 </example>
1160
1161 <example>
1162 <title>Stoppable oneshot service</title>
1163
1164 <para>Similarly to the oneshot services, there are sometimes
1165 units that need to execute a program to set up something and
1166 then execute another to shut it down, but no process remains
1167 active while they are considered 'started'. Network
1168 configuration can sometimes fall into this category. Another use
1169 case is if a oneshot service shall not be executed a each time
1170 when they are pulled in as a dependency, but only the first
1171 time.</para>
1172
1173 <para>For this, systemd knows the setting
1174 <varname>RemainAfterExit=</varname><option>yes</option>, which
1175 causes systemd to consider the unit to be active if the start
1176 action exited successfully. This directive can be used with all
1177 types, but is most useful with
1178 <varname>Type=</varname><option>oneshot</option> and
1179 <varname>Type=</varname><option>simple</option>. With
1180 <varname>Type=</varname><option>oneshot</option> systemd waits
1181 until the start action has completed before it considers the
1182 unit to be active, so dependencies start only after the start
1183 action has succeeded. With
1184 <varname>Type=</varname><option>simple</option> dependencies
1185 will start immediately after the start action has been
1186 dispatched. The following unit provides an example for a simple
1187 static firewall.</para>
1188
1189 <programlisting>[Unit]
1190 Description=Simple firewall
1191
1192 [Service]
1193 Type=oneshot
1194 RemainAfterExit=yes
1195 ExecStart=/usr/local/sbin/simple-firewall-start
1196 ExecStop=/usr/local/sbin/simple-firewall-stop
1197
1198 [Install]
1199 WantedBy=multi-user.target</programlisting>
1200
1201 <para>Since the unit is considered to be running after the start
1202 action has exited, invoking <command>systemctl start</command>
1203 on that unit again will cause no action to be taken.</para>
1204 </example>
1205
1206 <example>
1207 <title>Traditional forking services</title>
1208
1209 <para>Many traditional daemons/services background (i.e. fork,
1210 daemonize) themselves when starting. Set
1211 <varname>Type=</varname><option>forking</option> in the
1212 service's unit file to support this mode of operation. systemd
1213 will consider the service to be in the process of initialization
1214 while the original program is still running. Once it exits
1215 successfully and at least a process remains (and
1216 <varname>RemainAfterExit=</varname><option>no</option>), the
1217 service is considered started.</para>
1218
1219 <para>Often a traditional daemon only consists of one process.
1220 Therefore, if only one process is left after the original
1221 process terminates, systemd will consider that process the main
1222 process of the service. In that case, the
1223 <varname>$MAINPID</varname> variable will be available in
1224 <varname>ExecReload=</varname>, <varname>ExecStop=</varname>,
1225 etc.</para>
1226
1227 <para>In case more than one process remains, systemd will be
1228 unable to determine the main process, so it will not assume
1229 there is one. In that case, <varname>$MAINPID</varname> will not
1230 expand to anything. However, if the process decides to write a
1231 traditional PID file, systemd will be able to read the main PID
1232 from there. Please set <varname>PIDFile=</varname> accordingly.
1233 Note that the daemon should write that file before finishing
1234 with its initialization, otherwise systemd might try to read the
1235 file before it exists.</para>
1236
1237 <para>The following example shows a simple daemon that forks and
1238 just starts one process in the background:</para>
1239
1240 <programlisting>[Unit]
1241 Description=Some simple daemon
1242
1243 [Service]
1244 Type=forking
1245 ExecStart=/usr/sbin/my-simple-daemon -d
1246
1247 [Install]
1248 WantedBy=multi-user.target</programlisting>
1249
1250 <para>Please see
1251 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
1252 for details on how you can influence the way systemd terminates
1253 the service.</para>
1254 </example>
1255
1256 <example>
1257 <title>DBus services</title>
1258
1259 <para>For services that acquire a name on the DBus system bus,
1260 use <varname>Type=</varname><option>dbus</option> and set
1261 <varname>BusName=</varname> accordingly. The service should not
1262 fork (daemonize). systemd will consider the service to be
1263 initialized once the name has been acquired on the system bus.
1264 The following example shows a typical DBus service:</para>
1265
1266 <programlisting>[Unit]
1267 Description=Simple DBus service
1268
1269 [Service]
1270 Type=dbus
1271 BusName=org.example.simple-dbus-service
1272 ExecStart=/usr/sbin/simple-dbus-service
1273
1274 [Install]
1275 WantedBy=multi-user.target</programlisting>
1276
1277 <para>For <emphasis>bus-activatable</emphasis> services, don't
1278 include a <literal>[Install]</literal> section in the systemd
1279 service file, but use the <varname>SystemdService=</varname>
1280 option in the corresponding DBus service file, for example
1281 (<filename>/usr/share/dbus-1/system-services/org.example.simple-dbus-service.service</filename>):</para>
1282
1283 <programlisting>[D-BUS Service]
1284 Name=org.example.simple-dbus-service
1285 Exec=/usr/sbin/simple-dbus-service
1286 User=root
1287 SystemdService=simple-dbus-service.service</programlisting>
1288
1289 <para>Please see
1290 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
1291 for details on how you can influence the way systemd terminates
1292 the service.</para>
1293 </example>
1294
1295 <example>
1296 <title>Services that notify systemd about their initialization</title>
1297
1298 <para><varname>Type=</varname><option>simple</option> services
1299 are really easy to write, but have the major disadvantage of
1300 systemd not being able to tell when initialization of the given
1301 service is complete. For this reason, systemd supports a simple
1302 notification protocol that allows daemons to make systemd aware
1303 that they are done initializing. Use
1304 <varname>Type=</varname><option>notify</option> for this. A
1305 typical service file for such a daemon would look like
1306 this:</para>
1307
1308 <programlisting>[Unit]
1309 Description=Simple notifying service
1310
1311 [Service]
1312 Type=notify
1313 ExecStart=/usr/sbin/simple-notifying-service
1314
1315 [Install]
1316 WantedBy=multi-user.target</programlisting>
1317
1318 <para>Note that the daemon has to support systemd's notification
1319 protocol, else systemd will think the service hasn't started yet
1320 and kill it after a timeout. For an example of how to update
1321 daemons to support this protocol transparently, take a look at
1322 <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>.
1323 systemd will consider the unit to be in the 'starting' state
1324 until a readiness notification has arrived.</para>
1325
1326 <para>Please see
1327 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
1328 for details on how you can influence the way systemd terminates
1329 the service.</para>
1330 </example>
1331 </refsect1>
1332
1333 <refsect1>
1334 <title>See Also</title>
1335 <para>
1336 <citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
1337 <citerefentry><refentrytitle>systemctl</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
1338 <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
1339 <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
1340 <citerefentry><refentrytitle>systemd.resource-control</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
1341 <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
1342 <citerefentry><refentrytitle>systemd.directives</refentrytitle><manvolnum>7</manvolnum></citerefentry>
1343 </para>
1344 </refsect1>
1345
1346 </refentry>