2 <!DOCTYPE refentry PUBLIC
"-//OASIS//DTD DocBook XML V4.5//EN"
3 "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
4 <!-- SPDX-License-Identifier: LGPL-2.1-or-later -->
6 <refentry id=
"sd_notify"
7 xmlns:
xi=
"http://www.w3.org/2001/XInclude">
10 <title>sd_notify
</title>
11 <productname>systemd
</productname>
15 <refentrytitle>sd_notify
</refentrytitle>
16 <manvolnum>3</manvolnum>
20 <refname>sd_notify
</refname>
21 <refname>sd_notifyf
</refname>
22 <refname>sd_pid_notify
</refname>
23 <refname>sd_pid_notifyf
</refname>
24 <refname>sd_pid_notify_with_fds
</refname>
25 <refname>sd_pid_notifyf_with_fds
</refname>
26 <refname>sd_notify_barrier
</refname>
27 <refname>sd_pid_notify_barrier
</refname>
28 <refpurpose>Notify service manager about start-up completion and other service status changes
</refpurpose>
33 <funcsynopsisinfo>#include
<systemd/sd-daemon.h
></funcsynopsisinfo>
36 <funcdef>int
<function>sd_notify
</function></funcdef>
37 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
38 <paramdef>const char *
<parameter>state
</parameter></paramdef>
42 <funcdef>int
<function>sd_notifyf
</function></funcdef>
43 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
44 <paramdef>const char *
<parameter>format
</parameter></paramdef>
45 <paramdef>…
</paramdef>
49 <funcdef>int
<function>sd_pid_notify
</function></funcdef>
50 <paramdef>pid_t
<parameter>pid
</parameter></paramdef>
51 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
52 <paramdef>const char *
<parameter>state
</parameter></paramdef>
56 <funcdef>int
<function>sd_pid_notifyf
</function></funcdef>
57 <paramdef>pid_t
<parameter>pid
</parameter></paramdef>
58 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
59 <paramdef>const char *
<parameter>format
</parameter></paramdef>
60 <paramdef>…
</paramdef>
64 <funcdef>int
<function>sd_pid_notify_with_fds
</function></funcdef>
65 <paramdef>pid_t
<parameter>pid
</parameter></paramdef>
66 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
67 <paramdef>const char *
<parameter>state
</parameter></paramdef>
68 <paramdef>const int *
<parameter>fds
</parameter></paramdef>
69 <paramdef>unsigned
<parameter>n_fds
</parameter></paramdef>
73 <funcdef>int
<function>sd_pid_notifyf_with_fds
</function></funcdef>
74 <paramdef>pid_t
<parameter>pid
</parameter></paramdef>
75 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
76 <paramdef>const int *
<parameter>fds
</parameter></paramdef>
77 <paramdef>size_t
<parameter>n_fds
</parameter></paramdef>
78 <paramdef>const char *
<parameter>format
</parameter></paramdef>
79 <paramdef>…
</paramdef>
83 <funcdef>int
<function>sd_notify_barrier
</function></funcdef>
84 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
85 <paramdef>uint64_t
<parameter>timeout
</parameter></paramdef>
89 <funcdef>int
<function>sd_pid_notify_barrier
</function></funcdef>
90 <paramdef>pid_t
<parameter>pid
</parameter></paramdef>
91 <paramdef>int
<parameter>unset_environment
</parameter></paramdef>
92 <paramdef>uint64_t
<parameter>timeout
</parameter></paramdef>
98 <title>Description
</title>
100 <para><function>sd_notify()
</function> may be called by a service to notify the service manager about
101 state changes. It can be used to send arbitrary information, encoded in an environment-block-like
102 string. Most importantly, it can be used for start-up completion notification.
</para>
104 <para>If the
<parameter>unset_environment
</parameter> parameter is non-zero,
105 <function>sd_notify()
</function> will unset the
<varname>$NOTIFY_SOCKET
</varname> environment variable
106 before returning (regardless of whether the function call itself succeeded or not). Further calls to
107 <function>sd_notify()
</function> will then fail, but the variable is no longer inherited by child
110 <para>The
<parameter>state
</parameter> parameter should contain a newline-separated list of variable
111 assignments, similar in style to an environment block. A trailing newline is implied if none is
112 specified. The string may contain any kind of variable assignments, but the following shall be considered
119 <listitem><para>Tells the service manager that service startup is finished, or the service finished
120 re-loading its configuration. This is only used by systemd if the service definition file has
121 <varname>Type=notify
</varname> or
<varname>Type=notify-reload
</varname> set. Since there is little
122 value in signaling non-readiness, the only value services should send is
<literal>READY=
1</literal>
123 (i.e.
<literal>READY=
0</literal> is not defined).
</para></listitem>
127 <term>RELOADING=
1</term>
129 <listitem><para>Tells the service manager that the service is beginning to reload its
130 configuration. This is useful to allow the service manager to track the service's internal state, and
131 present it to the user. Note that a service that sends this notification must also send a
132 <literal>READY=
1</literal> notification when it completed reloading its configuration. Reloads the
133 service manager is notified about with this mechanisms are propagated in the same way as they are
134 when originally initiated through the service manager. This message is particularly relevant for
135 <varname>Type=notify-reload
</varname> services, to inform the service manager that the request to
136 reload the service has been received and is now being processed.
</para></listitem>
140 <term>MONOTONIC_USEC=…
</term>
142 <listitem><para>A field carrying the monotonic timestamp (as per
143 <constant>CLOCK_MONOTONIC
</constant>) formatted in decimal in μs, when the notification message was
144 generated by the client. This is typically used in combination with
<literal>RELOADING=
1</literal>,
145 to allow the service manager to properly synchronize reload cycles. See
146 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
147 for details, specifically
<literal>Type=notify-reload
</literal>.
</para></listitem>
151 <term>STOPPING=
1</term>
153 <listitem><para>Tells the service manager that the service is beginning its shutdown. This is useful
154 to allow the service manager to track the service's internal state, and present it to the
155 user.
</para></listitem>
159 <term>STATUS=…
</term>
161 <listitem><para>Passes a single-line UTF-
8 status string back to the service manager that describes
162 the service state. This is free-form and can be used for various purposes: general state feedback,
163 fsck-like programs could pass completion percentages and failing programs could pass a human-readable
164 error message. Example:
<literal>STATUS=Completed
66% of file system
165 check…
</literal></para></listitem>
169 <term>NOTIFYACCESS=…
</term>
171 <listitem><para>Reset the access to the service status notification socket during runtime, overriding
172 <varname>NotifyAccess=
</varname> setting in the service unit file. See
173 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
174 for details, specifically
<literal>NotifyAccess=
</literal> for a list of accepted
175 values.
</para></listitem>
181 <listitem><para>If a service fails, the errno-style error code, formatted as string. Example:
182 <literal>ERRNO=
2</literal> for ENOENT.
</para></listitem>
186 <term>BUSERROR=…
</term>
188 <listitem><para>If a service fails, the D-Bus error-style error code. Example:
189 <literal>BUSERROR=org.freedesktop.DBus.Error.TimedOut
</literal></para></listitem>
193 <term>EXIT_STATUS=…
</term>
195 <listitem><para>If a service exits, the return value of its
<function>main()
</function> function.
200 <term>MAINPID=…
</term>
202 <listitem><para>The main process ID (PID) of the service, in case the service manager did not fork
203 off the process itself. Example:
<literal>MAINPID=
4711</literal></para></listitem>
207 <term>WATCHDOG=
1</term>
209 <listitem><para>Tells the service manager to update the watchdog timestamp. This is the keep-alive
210 ping that services need to issue in regular intervals if
<varname>WatchdogSec=
</varname> is enabled
212 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
213 for information how to enable this functionality and
214 <citerefentry><refentrytitle>sd_watchdog_enabled
</refentrytitle><manvolnum>3</manvolnum></citerefentry>
215 for the details of how the service can check whether the watchdog is enabled.
</para></listitem>
219 <term>WATCHDOG=trigger
</term>
221 <listitem><para>Tells the service manager that the service detected an internal error that should be
222 handled by the configured watchdog options. This will trigger the same behaviour as if
223 <varname>WatchdogSec=
</varname> is enabled and the service did not send
<literal>WATCHDOG=
1</literal>
224 in time. Note that
<varname>WatchdogSec=
</varname> does not need to be enabled for
225 <literal>WATCHDOG=trigger
</literal> to trigger the watchdog action. See
226 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
227 for information about the watchdog behavior.
</para></listitem>
231 <term>WATCHDOG_USEC=…
</term>
233 <listitem><para>Reset
<varname>watchdog_usec
</varname> value during runtime. Notice that this is not
234 available when using
<function>sd_event_set_watchdog()
</function> or
235 <function>sd_watchdog_enabled()
</function>. Example :
236 <literal>WATCHDOG_USEC=
20000000</literal></para></listitem>
240 <term>EXTEND_TIMEOUT_USEC=…
</term>
242 <listitem><para>Tells the service manager to extend the startup, runtime or shutdown service timeout
243 corresponding the current state. The value specified is a time in microseconds during which the
244 service must send a new message. A service timeout will occur if the message isn't received, but only
245 if the runtime of the current state is beyond the original maximum times of
246 <varname>TimeoutStartSec=
</varname>,
<varname>RuntimeMaxSec=
</varname>, and
247 <varname>TimeoutStopSec=
</varname>. See
248 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>
249 for effects on the service timeouts.
</para></listitem>
253 <term>FDSTORE=
1</term>
255 <listitem><para>Stores additional file descriptors in the service manager. File descriptors sent this
256 way will be maintained per-service by the service manager and will later be handed back using the
257 usual file descriptor passing logic at the next invocation of the service (e.g. when it is
259 <citerefentry><refentrytitle>sd_listen_fds
</refentrytitle><manvolnum>3</manvolnum></citerefentry>.
260 This is useful for implementing services that can restart after an explicit request or a crash
261 without losing state. Any open sockets and other file descriptors which should not be closed during
262 the restart may be stored this way. Application state can either be serialized to a file in
263 <filename>/run/
</filename>, or better, stored in a
264 <citerefentry><refentrytitle>memfd_create
</refentrytitle><manvolnum>2</manvolnum></citerefentry>
265 memory file descriptor. Note that the service manager will accept messages for a service only if its
266 <varname>FileDescriptorStoreMax=
</varname> setting is non-zero (defaults to zero, see
267 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>). If
268 <varname>FDPOLL=
0</varname> is not set and the file descriptors sent are pollable (see
269 <citerefentry><refentrytitle>epoll_ctl
</refentrytitle><manvolnum>2</manvolnum></citerefentry>), then
270 any
<constant>EPOLLHUP
</constant> or
<constant>EPOLLERR
</constant> event seen on them will result in
271 their automatic removal from the store. Multiple arrays of file descriptors may be sent in separate
272 messages, in which case the arrays are combined. Note that the service manager removes duplicate
273 (pointing to the same object) file descriptors before passing them to the service. When a service is
274 stopped, its file descriptor store is discarded and all file descriptors in it are closed. Use
275 <function>sd_pid_notify_with_fds()
</function> to send messages with
<literal>FDSTORE=
1</literal>, see
276 below. The service manager will set the
<varname>$FDSTORE
</varname> environment variable for services
277 that have the file descriptor store enabled.
</para></listitem>
281 <term>FDSTOREREMOVE=
1</term>
283 <listitem><para>Removes file descriptors from the file descriptor store. This field needs to be
284 combined with
<varname>FDNAME=
</varname> to specify the name of the file descriptors to
285 remove.
</para></listitem>
289 <term>FDNAME=…
</term>
291 <listitem><para>When used in combination with
<varname>FDSTORE=
1</varname>, specifies a name for the
292 submitted file descriptors. When used with
<varname>FDSTOREREMOVE=
1</varname>, specifies the name for
293 the file descriptors to remove. This name is passed to the service during activation, and may be
295 <citerefentry><refentrytitle>sd_listen_fds_with_names
</refentrytitle><manvolnum>3</manvolnum></citerefentry>. File
296 descriptors submitted without this field set, will implicitly get the name
<literal>stored
</literal>
297 assigned. Note that, if multiple file descriptors are submitted at once, the specified name will be
298 assigned to all of them. In order to assign different names to submitted file descriptors, submit
299 them in separate invocations of
<function>sd_pid_notify_with_fds()
</function>. The name may consist
300 of arbitrary ASCII characters except control characters or
<literal>:
</literal>. It may not be longer
301 than
255 characters. If a submitted name does not follow these restrictions, it is
302 ignored.
</para></listitem>
306 <term>FDPOLL=
0</term>
308 <listitem><para>When used in combination with
<varname>FDSTORE=
1</varname>, disables polling of the
309 stored file descriptors regardless of whether or not they are pollable. As this option disables
310 automatic cleanup of the stored file descriptors on EPOLLERR and EPOLLHUP, care must be taken to
311 ensure proper manual cleanup. Use of this option is not generally recommended except for when
312 automatic cleanup has unwanted behavior such as prematurely discarding file descriptors from the
313 store.
</para></listitem>
317 <term>BARRIER=
1</term>
319 <listitem><para>Tells the service manager that the client is explicitly requesting synchronization by
320 means of closing the file descriptor sent with this command. The service manager guarantees that the
321 processing of a
<varname>BARRIER=
1</varname> command will only happen after all previous notification
322 messages sent before this command have been processed. Hence, this command accompanied with a single
323 file descriptor can be used to synchronize against reception of all previous status messages. Note
324 that this command cannot be mixed with other notifications, and has to be sent in a separate message
325 to the service manager, otherwise all assignments will be ignored. Note that sending
0 or more than
1
326 file descriptor with this command is a violation of the protocol.
</para></listitem>
330 <para>It is recommended to prefix variable names that are not listed above with
<varname>X_
</varname> to
331 avoid namespace clashes.
</para>
333 <para>Note that systemd will accept status data sent from a service only if the
334 <varname>NotifyAccess=
</varname> option is correctly set in the service definition file. See
335 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry> for
338 <para>Note that
<function>sd_notify()
</function> notifications may be attributed to units correctly only
339 if either the sending process is still around at the time PID
1 processes the message, or if the sending
340 process is explicitly runtime-tracked by the service manager. The latter is the case if the service
341 manager originally forked off the process, i.e. on all processes that match
342 <varname>NotifyAccess=
</varname><option>main
</option> or
343 <varname>NotifyAccess=
</varname><option>exec
</option>. Conversely, if an auxiliary process of the unit
344 sends an
<function>sd_notify()
</function> message and immediately exits, the service manager might not be
345 able to properly attribute the message to the unit, and thus will ignore it, even if
346 <varname>NotifyAccess=
</varname><option>all
</option> is set for it.
</para>
348 <para>Hence, to eliminate all race conditions involving lookup of the client's unit and attribution of
349 notifications to units correctly,
<function>sd_notify_barrier()
</function> may be used. This call acts as
350 a synchronization point and ensures all notifications sent before this call have been picked up by the
351 service manager when it returns successfully. Use of
<function>sd_notify_barrier()
</function> is needed
352 for clients which are not invoked by the service manager, otherwise this synchronization mechanism is
353 unnecessary for attribution of notifications to the unit.
</para>
355 <para><function>sd_notifyf()
</function> is similar to
<function>sd_notify()
</function> but takes a
356 <function>printf()
</function>-like format string plus arguments.
</para>
358 <para><function>sd_pid_notify()
</function> and
<function>sd_pid_notifyf()
</function> are similar to
359 <function>sd_notify()
</function> and
<function>sd_notifyf()
</function> but take a process ID (PID) to use
360 as originating PID for the message as first argument. This is useful to send notification messages on
361 behalf of other processes, provided the appropriate privileges are available. If the PID argument is
362 specified as
0, the process ID of the calling process is used, in which case the calls are fully
363 equivalent to
<function>sd_notify()
</function> and
<function>sd_notifyf()
</function>.
</para>
365 <para><function>sd_pid_notify_with_fds()
</function> is similar to
<function>sd_pid_notify()
</function>
366 but takes an additional array of file descriptors. These file descriptors are sent along the notification
367 message to the service manager. This is particularly useful for sending
<literal>FDSTORE=
1</literal>
368 messages, as described above. The additional arguments are a pointer to the file descriptor array plus
369 the number of file descriptors in the array. If the number of file descriptors is passed as
0, the call
370 is fully equivalent to
<function>sd_pid_notify()
</function>, i.e. no file descriptors are passed. Note
371 that file descriptors sent to the service manager on a message without
<literal>FDSTORE=
1</literal> are
372 immediately closed on reception.
</para>
374 <para><function>sd_pid_notifyf_with_fds()
</function> is a combination of
375 <function>sd_pid_notify_with_fds()
</function> and
<function>sd_notifyf()
</function>, i.e. it accepts both
376 a PID and a set of file descriptors as input, and processes a format string to generate the state
379 <para><function>sd_notify_barrier()
</function> allows the caller to synchronize against reception of
380 previously sent notification messages and uses the
<varname>BARRIER=
1</varname> command. It takes a
381 relative
<varname>timeout
</varname> value in microseconds which is passed to
382 <citerefentry><refentrytitle>ppoll
</refentrytitle><manvolnum>2</manvolnum>
383 </citerefentry>. A value of UINT64_MAX is interpreted as infinite timeout.
</para>
385 <para><function>sd_pid_notify_barrier()
</function> is just like
<function>sd_notify_barrier()
</function>,
386 but allows specifying the originating PID for the notification message.
</para>
390 <title>Return Value
</title>
392 <para>On failure, these calls return a negative errno-style error code. If
393 <varname>$NOTIFY_SOCKET
</varname> was not set and hence no status message could be sent,
0 is
394 returned. If the status was sent, these functions return a positive value. In order to support both
395 service managers that implement this scheme and those which do not, it is generally recommended to ignore
396 the return value of this call. Note that the return value simply indicates whether the notification
397 message was enqueued properly, it does not reflect whether the message could be processed
398 successfully. Specifically, no error is returned when a file descriptor is attempted to be stored using
399 <varname>FDSTORE=
1</varname> but the service is not actually configured to permit storing of file
400 descriptors (see above).
</para>
406 <xi:include href=
"libsystemd-pkgconfig.xml" xpointer=
"pkgconfig-text"/>
407 <xi:include href=
"threads-aware.xml" xpointer=
"getenv"/>
409 <para>These functions send a single datagram with the state string as payload to the socket referenced in
410 the
<varname>$NOTIFY_SOCKET
</varname> environment variable. If the first character of
411 <varname>$NOTIFY_SOCKET
</varname> is
<literal>/
</literal> or
<literal>@
</literal>, the string is
412 understood as an
<constant>AF_UNIX
</constant> or Linux abstract namespace socket (respectively), and in
413 both cases the datagram is accompanied by the process credentials of the sending service, using
414 SCM_CREDENTIALS. If the string starts with
<literal>vsock:
</literal> then the string is understood as an
415 <constant>AF_VSOCK
</constant> address, which is useful for hypervisors/VMMs or other processes on the
416 host to receive a notification when a virtual machine has finished booting. Note that in case the
417 hypervisor does not support
<constant>SOCK_DGRAM
</constant> over
<constant>AF_VSOCK
</constant>,
418 <constant>SOCK_SEQPACKET
</constant> will be used instead. The address should be in the form:
419 <literal>vsock:CID:PORT
</literal>. Note that unlike other uses of vsock, the CID is mandatory and cannot
420 be
<literal>VMADDR_CID_ANY
</literal>. Note that PID1 will send the VSOCK packets from a privileged port
421 (i.e.: lower than
1024), as an attempt to address concerns that unprivileged processes in the guest might
422 try to send malicious notifications to the host, driving it to make destructive decisions based on
427 <title>Environment
</title>
429 <variablelist class='environment-variables'
>
431 <term><varname>$NOTIFY_SOCKET
</varname></term>
433 <listitem><para>Set by the service manager for supervised processes for status and start-up
434 completion notification. This environment variable specifies the socket
435 <function>sd_notify()
</function> talks to. See above for details.
</para></listitem>
441 <title>Examples
</title>
444 <title>Start-up Notification
</title>
446 <para>When a service finished starting up, it might issue the following call to notify the service
449 <programlisting>sd_notify(
0,
"READY=1");
</programlisting>
453 <title>Extended Start-up Notification
</title>
455 <para>A service could send the following after completing initialization:
</para>
458 sd_notifyf(
0,
"READY=1\n"
459 "STATUS=Processing requests…\n"
461 (unsigned long) getpid());
</programlisting>
465 <title>Error Cause Notification
</title>
467 <para>A service could send the following shortly before exiting, on failure:
</para>
470 sd_notifyf(
0,
"STATUS=Failed to start up: %s\n"
472 strerror_r(errnum, (char[
1024]){},
1024),
473 errnum);
</programlisting>
477 <title>Store a File Descriptor in the Service Manager
</title>
479 <para>To store an open file descriptor in the service manager, in order to continue operation after a
480 service restart without losing state, use
<literal>FDSTORE=
1</literal>:
</para>
482 <programlisting>sd_pid_notify_with_fds(
0,
0,
"FDSTORE=1\nFDNAME=foobar",
&fd,
1);
</programlisting>
486 <title>Eliminating race conditions
</title>
488 <para>When the client sending the notifications is not spawned by the service manager, it may exit too
489 quickly and the service manager may fail to attribute them correctly to the unit. To prevent such
490 races, use
<function>sd_notify_barrier()
</function> to synchronize against reception of all
491 notifications sent before this call is made.
</para>
494 sd_notify(
0,
"READY=1");
495 /* set timeout to
5 seconds */
496 sd_notify_barrier(
0,
5 *
1000000);
502 <title>See Also
</title>
504 <citerefentry><refentrytitle>systemd
</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
505 <citerefentry><refentrytitle>sd-daemon
</refentrytitle><manvolnum>3</manvolnum></citerefentry>,
506 <citerefentry><refentrytitle>sd_listen_fds
</refentrytitle><manvolnum>3</manvolnum></citerefentry>,
507 <citerefentry><refentrytitle>sd_listen_fds_with_names
</refentrytitle><manvolnum>3</manvolnum></citerefentry>,
508 <citerefentry><refentrytitle>sd_watchdog_enabled
</refentrytitle><manvolnum>3</manvolnum></citerefentry>,
509 <citerefentry><refentrytitle>daemon
</refentrytitle><manvolnum>7</manvolnum></citerefentry>,
510 <citerefentry><refentrytitle>systemd.service
</refentrytitle><manvolnum>5</manvolnum></citerefentry>