1SD_NOTIFY(3) sd_notify SD_NOTIFY(3)
2
6 sd_notify, sd_notifyf, sd_pid_notify, sd_pid_notifyf,
7 sd_pid_notify_with_fds, sd_notify_barrier - Notify service manager
8 about start-up completion and other service status changes
9
11 #include <systemd/sd-daemon.h>
12 int sd_notify(int unset_environment, const char *state);
14 int sd_notifyf(int unset_environment, const char *format, ...);
16 int sd_pid_notify(pid_t pid, int unset_environment, const char *state);
18 int sd_pid_notifyf(pid_t pid, int unset_environment,
20 const char *format, ...);
21 int sd_pid_notify_with_fds(pid_t pid, int unset_environment,
23 const char *state, const int *fds,
24 unsigned n_fds);
25 int sd_notify_barrier(int unset_environment, uint64_t timeout);
27
29 sd_notify() may be called by a service to notify the service manager
30 about state changes. It can be used to send arbitrary information,
31 encoded in an environment-block-like string. Most importantly, it can
32 be used for start-up completion notification.
33 If the unset_environment parameter is non-zero, sd_notify() will unset
35 the $NOTIFY_SOCKET environment variable before returning (regardless of
36 whether the function call itself succeeded or not). Further calls to
37 sd_notify() will then fail, but the variable is no longer inherited by
38 child processes.
39 The state parameter should contain a newline-separated list of variable
41 assignments, similar in style to an environment block. A trailing
42 newline is implied if none is specified. The string may contain any
43 kind of variable assignments, but the following shall be considered
44 well-known:
45 READY=1
47 Tells the service manager that service startup is finished, or the
48 service finished loading its configuration. This is only used by
49 systemd if the service definition file has Type=notify set. Since
50 there is little value in signaling non-readiness, the only value
51 services should send is "READY=1" (i.e. "READY=0" is not defined).
52 RELOADING=1
54 Tells the service manager that the service is reloading its
55 configuration. This is useful to allow the service manager to track
56 the service's internal state, and present it to the user. Note that
57 a service that sends this notification must also send a "READY=1"
58 notification when it completed reloading its configuration. Reloads
59 are propagated in the same way as they are when initiated by the
60 user.
61 STOPPING=1
63 Tells the service manager that the service is beginning its
64 shutdown. This is useful to allow the service manager to track the
65 service's internal state, and present it to the user.
66 STATUS=...
68 Passes a single-line UTF-8 status string back to the service
69 manager that describes the service state. This is free-form and can
70 be used for various purposes: general state feedback, fsck-like
71 programs could pass completion percentages and failing programs
72 could pass a human-readable error message. Example:
73 "STATUS=Completed 66% of file system check..."
74 ERRNO=...
76 If a service fails, the errno-style error code, formatted as
77 string. Example: "ERRNO=2" for ENOENT.
78 BUSERROR=...
80 If a service fails, the D-Bus error-style error code. Example:
81 "BUSERROR=org.freedesktop.DBus.Error.TimedOut"
82 MAINPID=...
84 The main process ID (PID) of the service, in case the service
85 manager did not fork off the process itself. Example:
86 "MAINPID=4711"
87 WATCHDOG=1
89 Tells the service manager to update the watchdog timestamp. This is
90 the keep-alive ping that services need to issue in regular
91 intervals if WatchdogSec= is enabled for it. See systemd.service(5)
92 for information how to enable this functionality and
93 sd_watchdog_enabled(3) for the details of how the service can check
94 whether the watchdog is enabled.
95 WATCHDOG=trigger
97 Tells the service manager that the service detected an internal
98 error that should be handled by the configured watchdog options.
99 This will trigger the same behaviour as if WatchdogSec= is enabled
100 and the service did not send "WATCHDOG=1" in time. Note that
101 WatchdogSec= does not need to be enabled for "WATCHDOG=trigger" to
102 trigger the watchdog action. See systemd.service(5) for information
103 about the watchdog behavior.
104 WATCHDOG_USEC=...
106 Reset watchdog_usec value during runtime. Notice that this is not
107 available when using sd_event_set_watchdog() or
108 sd_watchdog_enabled(). Example : "WATCHDOG_USEC=20000000"
109 EXTEND_TIMEOUT_USEC=...
111 Tells the service manager to extend the startup, runtime or
112 shutdown service timeout corresponding the current state. The value
113 specified is a time in microseconds during which the service must
114 send a new message. A service timeout will occur if the message
115 isn't received, but only if the runtime of the current state is
116 beyond the original maximum times of TimeoutStartSec=,
117 RuntimeMaxSec=, and TimeoutStopSec=. See systemd.service(5) for
118 effects on the service timeouts.
119 FDSTORE=1
121 Stores additional file descriptors in the service manager. File
122 descriptors sent this way will be maintained per-service by the
123 service manager and will later be handed back using the usual file
124 descriptor passing logic at the next invocation of the service, see
125 sd_listen_fds(3). This is useful for implementing services that can
126 restart after an explicit request or a crash without losing state.
127 Any open sockets and other file descriptors which should not be
128 closed during the restart may be stored this way. Application state
129 can either be serialized to a file in /run/, or better, stored in a
130 memfd_create(2) memory file descriptor. Note that the service
131 manager will accept messages for a service only if its
132 FileDescriptorStoreMax= setting is non-zero (defaults to zero, see
133 systemd.service(5)). If FDPOLL=0 is not set and the file
134 descriptors sent are pollable (see epoll_ctl(2)), then any EPOLLHUP
135 or EPOLLERR event seen on them will result in their automatic
136 removal from the store. Multiple arrays of file descriptors may be
137 sent in separate messages, in which case the arrays are combined.
138 Note that the service manager removes duplicate (pointing to the
139 same object) file descriptors before passing them to the service.
140 Use sd_pid_notify_with_fds() to send messages with "FDSTORE=1", see
141 below.
142 FDSTOREREMOVE=1
144 Removes file descriptors from the file descriptor store. This field
145 needs to be combined with FDNAME= to specify the name of the file
146 descriptors to remove.
147 FDNAME=...
149 When used in combination with FDSTORE=1, specifies a name for the
150 submitted file descriptors. When used with FDSTOREREMOVE=1,
151 specifies the name for the file descriptors to remove. This name is
152 passed to the service during activation, and may be queried using
153 sd_listen_fds_with_names(3). File descriptors submitted without
154 this field set, will implicitly get the name "stored" assigned.
155 Note that, if multiple file descriptors are submitted at once, the
156 specified name will be assigned to all of them. In order to assign
157 different names to submitted file descriptors, submit them in
158 separate invocations of sd_pid_notify_with_fds(). The name may
159 consist of arbitrary ASCII characters except control characters or
160 ":". It may not be longer than 255 characters. If a submitted name
161 does not follow these restrictions, it is ignored.
162 FDPOLL=0
164 When used in combination with FDSTORE=1, disables polling of the
165 stored file descriptors regardless of whether or not they are
166 pollable. As this option disables automatic cleanup of the stored
167 file descriptors on EPOLLERR and EPOLLHUP, care must be taken to
168 ensure proper manual cleanup. Use of this option is not generally
169 recommended except for when automatic cleanup has unwanted behavior
170 such as prematurely discarding file descriptors from the store.
171 BARRIER=1
173 Tells the service manager that the client is explicitly requesting
174 synchronization by means of closing the file descriptor sent with
175 this command. The service manager guarantees that the processing of
176 a
177 BARRIER=1 command will only happen after all previous notification
178 messages sent before this command have been processed. Hence, this
179 command accompanied with a single file descriptor can be used to
180 synchronize against reception of all previous status messages. Note
181 that this command cannot be mixed with other notifications, and has
182 to be sent in a separate message to the service manager, otherwise
183 all assignments will be ignored. Note that sending 0 or more than 1
184 file descriptor with this command is a violation of the protocol.
185 It is recommended to prefix variable names that are not listed above
187 with X_ to avoid namespace clashes.
188 Note that systemd will accept status data sent from a service only if
190 the NotifyAccess= option is correctly set in the service definition
191 file. See systemd.service(5) for details.
192 Note that sd_notify() notifications may be attributed to units
194 correctly only if either the sending process is still around at the
195 time PID 1 processes the message, or if the sending process is
196 explicitly runtime-tracked by the service manager. The latter is the
197 case if the service manager originally forked off the process, i.e. on
198 all processes that match NotifyAccess=main or NotifyAccess=exec.
199 Conversely, if an auxiliary process of the unit sends an sd_notify()
200 message and immediately exits, the service manager might not be able to
201 properly attribute the message to the unit, and thus will ignore it,
202 even if NotifyAccess=all is set for it.
203 Hence, to eliminate all race conditions involving lookup of the
205 client's unit and attribution of notifications to units correctly,
206 sd_notify_barrier() may be used. This call acts as a synchronization
207 point and ensures all notifications sent before this call have been
208 picked up by the service manager when it returns successfully. Use of
209 sd_notify_barrier() is needed for clients which are not invoked by the
210 service manager, otherwise this synchronization mechanism is
211 unnecessary for attribution of notifications to the unit.
212 sd_notifyf() is similar to sd_notify() but takes a printf()-like format
214 string plus arguments.
215 sd_pid_notify() and sd_pid_notifyf() are similar to sd_notify() and
217 sd_notifyf() but take a process ID (PID) to use as originating PID for
218 the message as first argument. This is useful to send notification
219 messages on behalf of other processes, provided the appropriate
220 privileges are available. If the PID argument is specified as 0, the
221 process ID of the calling process is used, in which case the calls are
222 fully equivalent to sd_notify() and sd_notifyf().
223 sd_pid_notify_with_fds() is similar to sd_pid_notify() but takes an
225 additional array of file descriptors. These file descriptors are sent
226 along the notification message to the service manager. This is
227 particularly useful for sending "FDSTORE=1" messages, as described
228 above. The additional arguments are a pointer to the file descriptor
229 array plus the number of file descriptors in the array. If the number
230 of file descriptors is passed as 0, the call is fully equivalent to
231 sd_pid_notify(), i.e. no file descriptors are passed. Note that sending
232 file descriptors to the service manager on messages that do not expect
233 them (i.e. without "FDSTORE=1") they are immediately closed on
234 reception.
235 sd_notify_barrier() allows the caller to synchronize against reception
237 of previously sent notification messages and uses the "BARRIER=1"
238 command. It takes a relative timeout value in microseconds which is
239 passed to ppoll(2). A value of UINT64_MAX is interpreted as infinite
240 timeout.
241
243 On failure, these calls return a negative errno-style error code. If
244 $NOTIFY_SOCKET was not set and hence no status message could be sent, 0
245 is returned. If the status was sent, these functions return a positive
246 value. In order to support both service managers that implement this
247 scheme and those which do not, it is generally recommended to ignore
248 the return value of this call. Note that the return value simply
249 indicates whether the notification message was enqueued properly, it
250 does not reflect whether the message could be processed successfully.
251 Specifically, no error is returned when a file descriptor is attempted
252 to be stored using FDSTORE=1 but the service is not actually configured
253 to permit storing of file descriptors (see above).
254
256 These APIs are implemented as a shared library, which can be compiled
257 and linked to with the libsystemd pkg-config(1) file.
258 These functions send a single datagram with the state string as payload
260 to the AF_UNIX socket referenced in the $NOTIFY_SOCKET environment
261 variable. If the first character of $NOTIFY_SOCKET is "@", the string
262 is understood as Linux abstract namespace socket. The datagram is
263 accompanied by the process credentials of the sending service, using
264 SCM_CREDENTIALS.
265
267 $NOTIFY_SOCKET
268 Set by the service manager for supervised processes for status and
269 start-up completion notification. This environment variable
270 specifies the socket sd_notify() talks to. See above for details.
271
273 Example 1. Start-up Notification
274 When a service finished starting up, it might issue the following call
276 to notify the service manager:
277 sd_notify(0, "READY=1");
279 Example 2. Extended Start-up Notification
281 A service could send the following after completing initialization:
283 sd_notifyf(0, "READY=1\n"
285 "STATUS=Processing requests...\n"
286 "MAINPID=%lu",
287 (unsigned long) getpid());
288 Example 3. Error Cause Notification
290 A service could send the following shortly before exiting, on failure:
292 sd_notifyf(0, "STATUS=Failed to start up: %s\n"
294 "ERRNO=%i",
295 strerror(errno),
296 errno);
297 Example 4. Store a File Descriptor in the Service Manager
299 To store an open file descriptor in the service manager, in order to
301 continue operation after a service restart without losing state, use
302 "FDSTORE=1":
303 sd_pid_notify_with_fds(0, 0, "FDSTORE=1\nFDNAME=foobar", &fd, 1);
305 Example 5. Eliminating race conditions
307 When the client sending the notifications is not spawned by the service
309 manager, it may exit too quickly and the service manager may fail to
310 attribute them correctly to the unit. To prevent such races, use
311 sd_notify_barrier() to synchronize against reception of all
312 notifications sent before this call is made.
313 sd_notify(0, "READY=1");
315 /* set timeout to 5 seconds */
316 sd_notify_barrier(0, 5 * 1000000);
317
320 systemd(1), sd-daemon(3), sd_listen_fds(3),
321 sd_listen_fds_with_names(3), sd_watchdog_enabled(3), daemon(7),
322 systemd.service(5)
323systemd 249 SD_NOTIFY(3)