1DAEMON(7) daemon DAEMON(7)
2
6 daemon - Writing and packaging system daemons
7
9 A daemon is a service process that runs in the background and
10 supervises the system or provides functionality to other processes.
11 Traditionally, daemons are implemented following a scheme originating
12 in SysV Unix. Modern daemons should follow a simpler yet more powerful
13 scheme (here called "new-style" daemons), as implemented by systemd(1).
14 This manual page covers both schemes, and in particular includes
15 recommendations for daemons that shall be included in the systemd init
16 system.
17 SysV Daemons
19 When a traditional SysV daemon starts, it should execute the following
20 steps as part of the initialization. Note that these steps are
21 unnecessary for new-style daemons (see below), and should only be
22 implemented if compatibility with SysV is essential.
23 1. Close all open file descriptors except standard input, output, and
25 error (i.e. the first three file descriptors 0, 1, 2). This ensures
26 that no accidentally passed file descriptor stays around in the
27 daemon process. On Linux, this is best implemented by iterating
28 through /proc/self/fd, with a fallback of iterating from file
29 descriptor 3 to the value returned by getrlimit() for
30 RLIMIT_NOFILE.
31 2. Reset all signal handlers to their default. This is best done by
33 iterating through the available signals up to the limit of _NSIG
34 and resetting them to SIG_DFL.
35 3. Reset the signal mask using sigprocmask().
37 4. Sanitize the environment block, removing or resetting environment
39 variables that might negatively impact daemon runtime.
40 5. Call fork(), to create a background process.
42 6. In the child, call setsid() to detach from any terminal and create
44 an independent session.
45 7. In the child, call fork() again, to ensure that the daemon can
47 never re-acquire a terminal again.
48 8. Call exit() in the first child, so that only the second child (the
50 actual daemon process) stays around. This ensures that the daemon
51 process is re-parented to init/PID 1, as all daemons should be.
52 9. In the daemon process, connect /dev/null to standard input, output,
54 and error.
55 10. In the daemon process, reset the umask to 0, so that the file modes
57 passed to open(), mkdir() and suchlike directly control the access
58 mode of the created files and directories.
59 11. In the daemon process, change the current directory to the root
61 directory (/), in order to avoid that the daemon involuntarily
62 blocks mount points from being unmounted.
63 12. In the daemon process, write the daemon PID (as returned by
65 getpid()) to a PID file, for example /run/foobar.pid (for a
66 hypothetical daemon "foobar") to ensure that the daemon cannot be
67 started more than once. This must be implemented in race-free
68 fashion so that the PID file is only updated when it is verified at
69 the same time that the PID previously stored in the PID file no
70 longer exists or belongs to a foreign process.
71 13. In the daemon process, drop privileges, if possible and applicable.
73 14. From the daemon process, notify the original process started that
75 initialization is complete. This can be implemented via an unnamed
76 pipe or similar communication channel that is created before the
77 first fork() and hence available in both the original and the
78 daemon process.
79 15. Call exit() in the original process. The process that invoked the
81 daemon must be able to rely on that this exit() happens after
82 initialization is complete and all external communication channels
83 are established and accessible.
84 The BSD daemon() function should not be used, as it implements only a
86 subset of these steps.
87 A daemon that needs to provide compatibility with SysV systems should
89 implement the scheme pointed out above. However, it is recommended to
90 make this behavior optional and configurable via a command line
91 argument to ease debugging as well as to simplify integration into
92 systems using systemd.
93 New-Style Daemons
95 Modern services for Linux should be implemented as new-style daemons.
96 This makes it easier to supervise and control them at runtime and
97 simplifies their implementation.
98 For developing a new-style daemon, none of the initialization steps
100 recommended for SysV daemons need to be implemented. New-style init
101 systems such as systemd make all of them redundant. Moreover, since
102 some of these steps interfere with process monitoring, file descriptor
103 passing and other functionality of the init system, it is recommended
104 not to execute them when run as new-style service.
105 Note that new-style init systems guarantee execution of daemon
107 processes in a clean process context: it is guaranteed that the
108 environment block is sanitized, that the signal handlers and mask is
109 reset and that no left-over file descriptors are passed. Daemons will
110 be executed in their own session, with standard input connected to
111 /dev/null and standard output/error connected to the systemd-
112 journald.service(8) logging service, unless otherwise configured. The
113 umask is reset.
114 It is recommended for new-style daemons to implement the following:
116 1. If SIGTERM is received, shut down the daemon and exit cleanly.
118 2. If SIGHUP is received, reload the configuration files, if this
120 applies.
121 3. Provide a correct exit code from the main daemon process, as this
123 is used by the init system to detect service errors and problems.
124 It is recommended to follow the exit code scheme as defined in the
125 LSB recommendations for SysV init scripts[1].
126 4. If possible and applicable, expose the daemon's control interface
128 via the D-Bus IPC system and grab a bus name as last step of
129 initialization.
130 5. For integration in systemd, provide a .service unit file that
132 carries information about starting, stopping and otherwise
133 maintaining the daemon. See systemd.service(5) for details.
134 6. As much as possible, rely on the init system's functionality to
136 limit the access of the daemon to files, services and other
137 resources, i.e. in the case of systemd, rely on systemd's resource
138 limit control instead of implementing your own, rely on systemd's
139 privilege dropping code instead of implementing it in the daemon,
140 and similar. See systemd.exec(5) for the available controls.
141 7. If D-Bus is used, make your daemon bus-activatable by supplying a
143 D-Bus service activation configuration file. This has multiple
144 advantages: your daemon may be started lazily on-demand; it may be
145 started in parallel to other daemons requiring it — which maximizes
146 parallelization and boot-up speed; your daemon can be restarted on
147 failure without losing any bus requests, as the bus queues requests
148 for activatable services. See below for details.
149 8. If your daemon provides services to other local processes or remote
151 clients via a socket, it should be made socket-activatable
152 following the scheme pointed out below. Like D-Bus activation, this
153 enables on-demand starting of services as well as it allows
154 improved parallelization of service start-up. Also, for state-less
155 protocols (such as syslog, DNS), a daemon implementing socket-based
156 activation can be restarted without losing a single request. See
157 below for details.
158 9. If applicable, a daemon should notify the init system about startup
160 completion or status updates via the sd_notify(3) interface.
161 10. Instead of using the syslog() call to log directly to the system
163 syslog service, a new-style daemon may choose to simply log to
164 standard error via fprintf(), which is then forwarded to syslog by
165 the init system. If log levels are necessary, these can be encoded
166 by prefixing individual log lines with strings like "<4>" (for log
167 level 4 "WARNING" in the syslog priority scheme), following a
168 similar style as the Linux kernel's printk() level system. For
169 details, see sd-daemon(3) and systemd.exec(5).
170 These recommendations are similar but not identical to the Apple MacOS
172 X Daemon Requirements[2].
173
175 New-style init systems provide multiple additional mechanisms to
176 activate services, as detailed below. It is common that services are
177 configured to be activated via more than one mechanism at the same
178 time. An example for systemd: bluetoothd.service might get activated
179 either when Bluetooth hardware is plugged in, or when an application
180 accesses its programming interfaces via D-Bus. Or, a print server
181 daemon might get activated when traffic arrives at an IPP port, or when
182 a printer is plugged in, or when a file is queued in the printer spool
183 directory. Even for services that are intended to be started on system
184 bootup unconditionally, it is a good idea to implement some of the
185 various activation schemes outlined below, in order to maximize
186 parallelization. If a daemon implements a D-Bus service or listening
187 socket, implementing the full bus and socket activation scheme allows
188 starting of the daemon with its clients in parallel (which speeds up
189 boot-up), since all its communication channels are established already,
190 and no request is lost because client requests will be queued by the
191 bus system (in case of D-Bus) or the kernel (in case of sockets) until
192 the activation is completed.
193 Activation on Boot
195 Old-style daemons are usually activated exclusively on boot (and
196 manually by the administrator) via SysV init scripts, as detailed in
197 the LSB Linux Standard Base Core Specification[1]. This method of
198 activation is supported ubiquitously on Linux init systems, both
199 old-style and new-style systems. Among other issues, SysV init scripts
200 have the disadvantage of involving shell scripts in the boot process.
201 New-style init systems generally employ updated versions of activation,
202 both during boot-up and during runtime and using more minimal service
203 description files.
204 In systemd, if the developer or administrator wants to make sure that a
206 service or other unit is activated automatically on boot, it is
207 recommended to place a symlink to the unit file in the .wants/
208 directory of either multi-user.target or graphical.target, which are
209 normally used as boot targets at system startup. See systemd.unit(5)
210 for details about the .wants/ directories, and systemd.special(7) for
211 details about the two boot targets.
212 Socket-Based Activation
214 In order to maximize the possible parallelization and robustness and
215 simplify configuration and development, it is recommended for all
216 new-style daemons that communicate via listening sockets to employ
217 socket-based activation. In a socket-based activation scheme, the
218 creation and binding of the listening socket as primary communication
219 channel of daemons to local (and sometimes remote) clients is moved out
220 of the daemon code and into the init system. Based on per-daemon
221 configuration, the init system installs the sockets and then hands them
222 off to the spawned process as soon as the respective daemon is to be
223 started. Optionally, activation of the service can be delayed until the
224 first inbound traffic arrives at the socket to implement on-demand
225 activation of daemons. However, the primary advantage of this scheme is
226 that all providers and all consumers of the sockets can be started in
227 parallel as soon as all sockets are established. In addition to that,
228 daemons can be restarted with losing only a minimal number of client
229 transactions, or even any client request at all (the latter is
230 particularly true for state-less protocols, such as DNS or syslog),
231 because the socket stays bound and accessible during the restart, and
232 all requests are queued while the daemon cannot process them.
233 New-style daemons which support socket activation must be able to
235 receive their sockets from the init system instead of creating and
236 binding them themselves. For details about the programming interfaces
237 for this scheme provided by systemd, see sd_listen_fds(3) and sd-
238 daemon(3). For details about porting existing daemons to socket-based
239 activation, see below. With minimal effort, it is possible to implement
240 socket-based activation in addition to traditional internal socket
241 creation in the same codebase in order to support both new-style and
242 old-style init systems from the same daemon binary.
243 systemd implements socket-based activation via .socket units, which are
245 described in systemd.socket(5). When configuring socket units for
246 socket-based activation, it is essential that all listening sockets are
247 pulled in by the special target unit sockets.target. It is recommended
248 to place a WantedBy=sockets.target directive in the "[Install]" section
249 to automatically add such a dependency on installation of a socket
250 unit. Unless DefaultDependencies=no is set, the necessary ordering
251 dependencies are implicitly created for all socket units. For more
252 information about sockets.target, see systemd.special(7). It is not
253 necessary or recommended to place any additional dependencies on socket
254 units (for example from multi-user.target or suchlike) when one is
255 installed in sockets.target.
256 Bus-Based Activation
258 When the D-Bus IPC system is used for communication with clients,
259 new-style daemons should employ bus activation so that they are
260 automatically activated when a client application accesses their IPC
261 interfaces. This is configured in D-Bus service files (not to be
262 confused with systemd service unit files!). To ensure that D-Bus uses
263 systemd to start-up and maintain the daemon, use the SystemdService=
264 directive in these service files to configure the matching systemd
265 service for a D-Bus service. e.g.: For a D-Bus service whose D-Bus
266 activation file is named org.freedesktop.RealtimeKit.service, make sure
267 to set SystemdService=rtkit-daemon.service in that file to bind it to
268 the systemd service rtkit-daemon.service. This is needed to make sure
269 that the daemon is started in a race-free fashion when activated via
270 multiple mechanisms simultaneously.
271 Device-Based Activation
273 Often, daemons that manage a particular type of hardware should be
274 activated only when the hardware of the respective kind is plugged in
275 or otherwise becomes available. In a new-style init system, it is
276 possible to bind activation to hardware plug/unplug events. In systemd,
277 kernel devices appearing in the sysfs/udev device tree can be exposed
278 as units if they are tagged with the string "systemd". Like any other
279 kind of unit, they may then pull in other units when activated (i.e.
280 plugged in) and thus implement device-based activation. systemd
281 dependencies may be encoded in the udev database via the SYSTEMD_WANTS=
282 property. See systemd.device(5) for details. Often, it is nicer to pull
283 in services from devices only indirectly via dedicated targets.
284 Example: Instead of pulling in bluetoothd.service from all the various
285 bluetooth dongles and other hardware available, pull in
286 bluetooth.target from them and bluetoothd.service from that target.
287 This provides for nicer abstraction and gives administrators the option
288 to enable bluetoothd.service via controlling a bluetooth.target.wants/
289 symlink uniformly with a command like enable of systemctl(1) instead of
290 manipulating the udev ruleset.
291 Path-Based Activation
293 Often, runtime of daemons processing spool files or directories (such
294 as a printing system) can be delayed until these file system objects
295 change state, or become non-empty. New-style init systems provide a way
296 to bind service activation to file system changes. systemd implements
297 this scheme via path-based activation configured in .path units, as
298 outlined in systemd.path(5).
299 Timer-Based Activation
301 Some daemons that implement clean-up jobs that are intended to be
302 executed in regular intervals benefit from timer-based activation. In
303 systemd, this is implemented via .timer units, as described in
304 systemd.timer(5).
305 Other Forms of Activation
307 Other forms of activation have been suggested and implemented in some
308 systems. However, there are often simpler or better alternatives, or
309 they can be put together of combinations of the schemes above. Example:
310 Sometimes, it appears useful to start daemons or .socket units when a
311 specific IP address is configured on a network interface, because
312 network sockets shall be bound to the address. However, an alternative
313 to implement this is by utilizing the Linux IP_FREEBIND socket option,
314 as accessible via FreeBind=yes in systemd socket files (see
315 systemd.socket(5) for details). This option, when enabled, allows
316 sockets to be bound to a non-local, not configured IP address, and
317 hence allows bindings to a particular IP address before it actually
318 becomes available, making such an explicit dependency to the configured
319 address redundant. Another often suggested trigger for service
320 activation is low system load. However, here too, a more convincing
321 approach might be to make proper use of features of the operating
322 system, in particular, the CPU or I/O scheduler of Linux. Instead of
323 scheduling jobs from userspace based on monitoring the OS scheduler, it
324 is advisable to leave the scheduling of processes to the OS scheduler
325 itself. systemd provides fine-grained access to the CPU and I/O
326 schedulers. If a process executed by the init system shall not
327 negatively impact the amount of CPU or I/O bandwidth available to other
328 processes, it should be configured with CPUSchedulingPolicy=idle and/or
329 IOSchedulingClass=idle. Optionally, this may be combined with
330 timer-based activation to schedule background jobs during runtime and
331 with minimal impact on the system, and remove it from the boot phase
332 itself.
333
335 Writing systemd Unit Files
336 When writing systemd unit files, it is recommended to consider the
337 following suggestions:
338 1. If possible, do not use the Type=forking setting in service files.
340 But if you do, make sure to set the PID file path using PIDFile=.
341 See systemd.service(5) for details.
342 2. If your daemon registers a D-Bus name on the bus, make sure to use
344 Type=dbus in the service file if possible.
345 3. Make sure to set a good human-readable description string with
347 Description=.
348 4. Do not disable DefaultDependencies=, unless you really know what
350 you do and your unit is involved in early boot or late system
351 shutdown.
352 5. Normally, little if any dependencies should need to be defined
354 explicitly. However, if you do configure explicit dependencies,
355 only refer to unit names listed on systemd.special(7) or names
356 introduced by your own package to keep the unit file operating
357 system-independent.
358 6. Make sure to include an "[Install]" section including installation
360 information for the unit file. See systemd.unit(5) for details. To
361 activate your service on boot, make sure to add a
362 WantedBy=multi-user.target or WantedBy=graphical.target directive.
363 To activate your socket on boot, make sure to add
364 WantedBy=sockets.target. Usually, you also want to make sure that
365 when your service is installed, your socket is installed too, hence
366 add Also=foo.socket in your service file foo.service, for a
367 hypothetical program foo.
368 Installing systemd Service Files
370 At the build installation time (e.g. make install during package
371 build), packages are recommended to install their systemd unit files in
372 the directory returned by pkg-config systemd
373 --variable=systemdsystemunitdir (for system services) or pkg-config
374 systemd --variable=systemduserunitdir (for user services). This will
375 make the services available in the system on explicit request but not
376 activate them automatically during boot. Optionally, during package
377 installation (e.g. rpm -i by the administrator), symlinks should be
378 created in the systemd configuration directories via the enable command
379 of the systemctl(1) tool to activate them automatically on boot.
380 Packages using autoconf(1) are recommended to use a configure script
382 excerpt like the following to determine the unit installation path
383 during source configuration:
384 PKG_PROG_PKG_CONFIG
386 AC_ARG_WITH([systemdsystemunitdir],
387 [AS_HELP_STRING([--with-systemdsystemunitdir=DIR], [Directory for systemd service files])],,
388 [with_systemdsystemunitdir=auto])
389 AS_IF([test "x$with_systemdsystemunitdir" = "xyes" -o "x$with_systemdsystemunitdir" = "xauto"], [
390 def_systemdsystemunitdir=$($PKG_CONFIG --variable=systemdsystemunitdir systemd)
391 AS_IF([test "x$def_systemdsystemunitdir" = "x"],
393 [AS_IF([test "x$with_systemdsystemunitdir" = "xyes"],
394 [AC_MSG_ERROR([systemd support requested but pkg-config unable to query systemd package])])
395 with_systemdsystemunitdir=no],
396 [with_systemdsystemunitdir="$def_systemdsystemunitdir"])])
397 AS_IF([test "x$with_systemdsystemunitdir" != "xno"],
398 [AC_SUBST([systemdsystemunitdir], [$with_systemdsystemunitdir])])
399 AM_CONDITIONAL([HAVE_SYSTEMD], [test "x$with_systemdsystemunitdir" != "xno"])
400 This snippet allows automatic installation of the unit files on systemd
402 machines, and optionally allows their installation even on machines
403 lacking systemd. (Modification of this snippet for the user unit
404 directory is left as an exercise for the reader.)
405 Additionally, to ensure that make distcheck continues to work, it is
407 recommended to add the following to the top-level Makefile.am file in
408 automake(1)-based projects:
409 DISTCHECK_CONFIGURE_FLAGS = \
411 --with-systemdsystemunitdir=$$dc_install_base/$(systemdsystemunitdir)
412 Finally, unit files should be installed in the system with an automake
414 excerpt like the following:
415 if HAVE_SYSTEMD
417 systemdsystemunit_DATA = \
418 foobar.socket \
419 foobar.service
420 endif
421 In the rpm(8) .spec file, use snippets like the following to
423 enable/disable the service during installation/deinstallation. This
424 makes use of the RPM macros shipped along systemd. Consult the
425 packaging guidelines of your distribution for details and the
426 equivalent for other package managers.
427 At the top of the file:
429 BuildRequires: systemd
431 %{?systemd_requires}
432 And as scriptlets, further down:
434 %post
436 %systemd_post foobar.service foobar.socket
437 %preun
439 %systemd_preun foobar.service foobar.socket
440 %postun
442 %systemd_postun
443 If the service shall be restarted during upgrades, replace the
445 "%postun" scriptlet above with the following:
446 %postun
448 %systemd_postun_with_restart foobar.service
449 Note that "%systemd_post" and "%systemd_preun" expect the names of all
451 units that are installed/removed as arguments, separated by spaces.
452 "%systemd_postun" expects no arguments. "%systemd_postun_with_restart"
453 expects the units to restart as arguments.
454 To facilitate upgrades from a package version that shipped only SysV
456 init scripts to a package version that ships both a SysV init script
457 and a native systemd service file, use a fragment like the following:
458 %triggerun -- foobar < 0.47.11-1
460 if /sbin/chkconfig --level 5 foobar ; then
461 /bin/systemctl --no-reload enable foobar.service foobar.socket >/dev/null 2>&1 || :
462 fi
463 Where 0.47.11-1 is the first package version that includes the native
465 unit file. This fragment will ensure that the first time the unit file
466 is installed, it will be enabled if and only if the SysV init script is
467 enabled, thus making sure that the enable status is not changed. Note
468 that chkconfig is a command specific to Fedora which can be used to
469 check whether a SysV init script is enabled. Other operating systems
470 will have to use different commands here.
471
473 Since new-style init systems such as systemd are compatible with
474 traditional SysV init systems, it is not strictly necessary to port
475 existing daemons to the new style. However, doing so offers additional
476 functionality to the daemons as well as simplifying integration into
477 new-style init systems.
478 To port an existing SysV compatible daemon, the following steps are
480 recommended:
481 1. If not already implemented, add an optional command line switch to
483 the daemon to disable daemonization. This is useful not only for
484 using the daemon in new-style init systems, but also to ease
485 debugging.
486 2. If the daemon offers interfaces to other software running on the
488 local system via local AF_UNIX sockets, consider implementing
489 socket-based activation (see above). Usually, a minimal patch is
490 sufficient to implement this: Extend the socket creation in the
491 daemon code so that sd_listen_fds(3) is checked for already passed
492 sockets first. If sockets are passed (i.e. when sd_listen_fds()
493 returns a positive value), skip the socket creation step and use
494 the passed sockets. Secondly, ensure that the file system socket
495 nodes for local AF_UNIX sockets used in the socket-based activation
496 are not removed when the daemon shuts down, if sockets have been
497 passed. Third, if the daemon normally closes all remaining open
498 file descriptors as part of its initialization, the sockets passed
499 from the init system must be spared. Since new-style init systems
500 guarantee that no left-over file descriptors are passed to executed
501 processes, it might be a good choice to simply skip the closing of
502 all remaining open file descriptors if sockets are passed.
503 3. Write and install a systemd unit file for the service (and the
505 sockets if socket-based activation is used, as well as a path unit
506 file, if the daemon processes a spool directory), see above for
507 details.
508 4. If the daemon exposes interfaces via D-Bus, write and install a
510 D-Bus activation file for the service, see above for details.
511
513 It is recommended to follow the general guidelines for placing package
514 files, as discussed in file-hierarchy(7).
515
517 systemd(1), sd-daemon(3), sd_listen_fds(3), sd_notify(3), daemon(3),
518 systemd.service(5), file-hierarchy(7)
519
521 1. LSB recommendations for SysV init scripts
522 http://refspecs.linuxbase.org/LSB_3.1.1/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
523 2. Apple MacOS X Daemon Requirements
525 https://developer.apple.com/library/mac/documentation/MacOSX/Conceptual/BPSystemStartup/Chapters/CreatingLaunchdJobs.html
526systemd 239 DAEMON(7)