sd_pid_notifyf(3)

1SD_NOTIFY(3)                       sd_notify                      SD_NOTIFY(3)
2
3
4

NAME

6       sd_notify, sd_notifyf, sd_pid_notify, sd_pid_notifyf,
7       sd_pid_notify_with_fds - Notify service manager about start-up
8       completion and other service status changes
9

SYNOPSIS

11       #include <systemd/sd-daemon.h>
12
13       int sd_notify(int unset_environment, const char *state);
14
15       int sd_notifyf(int unset_environment, const char *format, ...);
16
17       int sd_pid_notify(pid_t pid, int unset_environment, const char *state);
18
19       int sd_pid_notifyf(pid_t pid, int unset_environment,
20                          const char *format, ...);
21
22       int sd_pid_notify_with_fds(pid_t pid, int unset_environment,
23                                  const char *state, const int *fds,
24                                  unsigned n_fds);
25

DESCRIPTION

27       sd_notify() may be called by a service to notify the service manager
28       about state changes. It can be used to send arbitrary information,
29       encoded in an environment-block-like string. Most importantly, it can
30       be used for start-up completion notification.
31
32       If the unset_environment parameter is non-zero, sd_notify() will unset
33       the $NOTIFY_SOCKET environment variable before returning (regardless of
34       whether the function call itself succeeded or not). Further calls to
35       sd_notify() will then fail, but the variable is no longer inherited by
36       child processes.
37
38       The state parameter should contain a newline-separated list of variable
39       assignments, similar in style to an environment block. A trailing
40       newline is implied if none is specified. The string may contain any
41       kind of variable assignments, but the following shall be considered
42       well-known:
43
44       READY=1
45           Tells the service manager that service startup is finished, or the
46           service finished loading its configuration. This is only used by
47           systemd if the service definition file has Type=notify set. Since
48           there is little value in signaling non-readiness, the only value
49           services should send is "READY=1" (i.e.  "READY=0" is not defined).
50
51       RELOADING=1
52           Tells the service manager that the service is reloading its
53           configuration. This is useful to allow the service manager to track
54           the service's internal state, and present it to the user. Note that
55           a service that sends this notification must also send a "READY=1"
56           notification when it completed reloading its configuration. Reloads
57           are propagated in the same way as they are when initiated by the
58           user.
59
60       STOPPING=1
61           Tells the service manager that the service is beginning its
62           shutdown. This is useful to allow the service manager to track the
63           service's internal state, and present it to the user.
64
65       STATUS=...
66           Passes a single-line UTF-8 status string back to the service
67           manager that describes the service state. This is free-form and can
68           be used for various purposes: general state feedback, fsck-like
69           programs could pass completion percentages and failing programs
70           could pass a human-readable error message. Example:
71           "STATUS=Completed 66% of file system check..."
72
73       ERRNO=...
74           If a service fails, the errno-style error code, formatted as
75           string. Example: "ERRNO=2" for ENOENT.
76
77       BUSERROR=...
78           If a service fails, the D-Bus error-style error code. Example:
79           "BUSERROR=org.freedesktop.DBus.Error.TimedOut"
80
81       MAINPID=...
82           The main process ID (PID) of the service, in case the service
83           manager did not fork off the process itself. Example:
84           "MAINPID=4711"
85
86       WATCHDOG=1
87           Tells the service manager to update the watchdog timestamp. This is
88           the keep-alive ping that services need to issue in regular
89           intervals if WatchdogSec= is enabled for it. See systemd.service(5)
90           for information how to enable this functionality and
91           sd_watchdog_enabled(3) for the details of how the service can check
92           whether the watchdog is enabled.
93
94       WATCHDOG_USEC=...
95           Reset watchdog_usec value during runtime. Notice that this is not
96           available when using sd_event_set_watchdog() or
97           sd_watchdog_enabled(). Example : "WATCHDOG_USEC=20000000"
98
99       EXTEND_TIMEOUT_USEC=...
100           Tells the service manager to extend the startup, runtime or
101           shutdown service timeout corresponding the current state. The value
102           specified is a time in microseconds during which the service must
103           send a new message. A service timeout will occur if the message
104           isn't received, but only if the runtime of the current state is
105           beyond the original maximium times of TimeoutStartSec=,
106           RuntimeMaxSec=, and TimeoutStopSec=. See systemd.service(5) for
107           effects on the service timeouts.
108
109       FDSTORE=1
110           Stores additional file descriptors in the service manager. File
111           descriptors sent this way will be maintained per-service by the
112           service manager and will later be handed back using the usual file
113           descriptor passing logic at the next invocation of the service, see
114           sd_listen_fds(3). This is useful for implementing services that can
115           restart after an explicit request or a crash without losing state.
116           Any open sockets and other file descriptors which should not be
117           closed during the restart may be stored this way. Application state
118           can either be serialized to a file in /run, or better, stored in a
119           memfd_create(2) memory file descriptor. Note that the service
120           manager will accept messages for a service only if its
121           FileDescriptorStoreMax= setting is non-zero (defaults to zero, see
122           systemd.service(5)). If file descriptors sent are pollable (see
123           epoll_ctl(2)), then any EPOLLHUP or EPOLLERR event seen on them
124           will result in their automatic removal from the store. Multiple
125           arrays of file descriptors may be sent in separate messages, in
126           which case the arrays are combined. Note that the service manager
127           removes duplicate (pointing to the same object) file descriptors
128           before passing them to the service. Use sd_pid_notify_with_fds() to
129           send messages with "FDSTORE=1", see below.
130
131       FDSTOREREMOVE=1
132           Removes file descriptors from the file descriptor store. This field
133           needs to be combined with FDNAME= to specify the name of the file
134           descriptors to remove.
135
136       FDNAME=...
137           When used in combination with FDSTORE=1, specifies a name for the
138           submitted file descriptors. When used with FDSTOREREMOVE=1,
139           specifies the name for the file descriptors to remove. This name is
140           passed to the service during activation, and may be queried using
141           sd_listen_fds_with_names(3). File descriptors submitted without
142           this field set, will implicitly get the name "stored" assigned.
143           Note that, if multiple file descriptors are submitted at once, the
144           specified name will be assigned to all of them. In order to assign
145           different names to submitted file descriptors, submit them in
146           separate invocations of sd_pid_notify_with_fds(). The name may
147           consist of arbitrary ASCII characters except control characters or
148           ":". It may not be longer than 255 characters. If a submitted name
149           does not follow these restrictions, it is ignored.
150
151       It is recommended to prefix variable names that are not listed above
152       with X_ to avoid namespace clashes.
153
154       Note that systemd will accept status data sent from a service only if
155       the NotifyAccess= option is correctly set in the service definition
156       file. See systemd.service(5) for details.
157
158       Note that sd_notify() notifications may be attributed to units
159       correctly only if either the sending process is still around at the
160       time PID 1 processes the message, or if the sending process is
161       explicitly runtime-tracked by the service manager. The latter is the
162       case if the service manager originally forked off the process, i.e. on
163       all processes that match NotifyAccess=main or NotifyAccess=exec.
164       Conversely, if an auxiliary process of the unit sends an sd_notify()
165       message and immediately exits, the service manager might not be able to
166       properly attribute the message to the unit, and thus will ignore it,
167       even if NotifyAccess=all is set for it.
168
169       sd_notifyf() is similar to sd_notify() but takes a printf()-like format
170       string plus arguments.
171
172       sd_pid_notify() and sd_pid_notifyf() are similar to sd_notify() and
173       sd_notifyf() but take a process ID (PID) to use as originating PID for
174       the message as first argument. This is useful to send notification
175       messages on behalf of other processes, provided the appropriate
176       privileges are available. If the PID argument is specified as 0, the
177       process ID of the calling process is used, in which case the calls are
178       fully equivalent to sd_notify() and sd_notifyf().
179
180       sd_pid_notify_with_fds() is similar to sd_pid_notify() but takes an
181       additional array of file descriptors. These file descriptors are sent
182       along the notification message to the service manager. This is
183       particularly useful for sending "FDSTORE=1" messages, as described
184       above. The additional arguments are a pointer to the file descriptor
185       array plus the number of file descriptors in the array. If the number
186       of file descriptors is passed as 0, the call is fully equivalent to
187       sd_pid_notify(), i.e. no file descriptors are passed. Note that sending
188       file descriptors to the service manager on messages that do not expect
189       them (i.e. without "FDSTORE=1") they are immediately closed on
190       reception.
191

RETURN VALUE

193       On failure, these calls return a negative errno-style error code. If
194       $NOTIFY_SOCKET was not set and hence no status message could be sent, 0
195       is returned. If the status was sent, these functions return a positive
196       value. In order to support both service managers that implement this
197       scheme and those which do not, it is generally recommended to ignore
198       the return value of this call. Note that the return value simply
199       indicates whether the notification message was enqueued properly, it
200       does not reflect whether the message could be processed successfully.
201       Specifically, no error is returned when a file descriptor is attempted
202       to be stored using FDSTORE=1 but the service is not actually configured
203       to permit storing of file descriptors (see above).
204

NOTES

206       These APIs are implemented as a shared library, which can be compiled
207       and linked to with the libsystemd pkg-config(1) file.
208
209       These functions send a single datagram with the state string as payload
210       to the AF_UNIX socket referenced in the $NOTIFY_SOCKET environment
211       variable. If the first character of $NOTIFY_SOCKET is "@", the string
212       is understood as Linux abstract namespace socket. The datagram is
213       accompanied by the process credentials of the sending service, using
214       SCM_CREDENTIALS.
215

ENVIRONMENT

217       $NOTIFY_SOCKET
218           Set by the service manager for supervised processes for status and
219           start-up completion notification. This environment variable
220           specifies the socket sd_notify() talks to. See above for details.
221

EXAMPLES

223       Example 1. Start-up Notification
224
225       When a service finished starting up, it might issue the following call
226       to notify the service manager:
227
228           sd_notify(0, "READY=1");
229
230       Example 2. Extended Start-up Notification
231
232       A service could send the following after completing initialization:
233
234           sd_notifyf(0, "READY=1\n"
235                   "STATUS=Processing requests...\n"
236                   "MAINPID=%lu",
237                   (unsigned long) getpid());
238
239       Example 3. Error Cause Notification
240
241       A service could send the following shortly before exiting, on failure:
242
243           sd_notifyf(0, "STATUS=Failed to start up: %s\n"
244                   "ERRNO=%i",
245                   strerror(errno),
246                   errno);
247
248       Example 4. Store a File Descriptor in the Service Manager
249
250       To store an open file descriptor in the service manager, in order to
251       continue operation after a service restart without losing state, use
252       "FDSTORE=1":
253
254           sd_pid_notify_with_fds(0, 0, "FDSTORE=1\nFDNAME=foobar", &fd, 1);
255