1SYSTEMD-JOURNALD.SERVICE(8)systemd-journald.serviceSYSTEMD-JOURNALD.SERVICE(8)
2
6 systemd-journald.service, systemd-journald.socket, systemd-journald-
7 dev-log.socket, systemd-journald-audit.socket, systemd-
8 journald@.service, systemd-journald@.socket, systemd-journald-
9 varlink@.socket, systemd-journald - Journal service
10
12 systemd-journald.service
13 systemd-journald.socket
15 systemd-journald-dev-log.socket
17 systemd-journald-audit.socket
19 systemd-journald@.service
21 systemd-journald@.socket
23 systemd-journald-varlink@.socket
25 /usr/lib/systemd/systemd-journald
27
29 systemd-journald is a system service that collects and stores logging
30 data. It creates and maintains structured, indexed journals based on
31 logging information that is received from a variety of sources:
32 • Kernel log messages, via kmsg
34 • Simple system log messages, via the libc syslog(3) call
36 • Structured system log messages via the native Journal API, see
38 sd_journal_print(3) and Native Journal Protocol[1]
39 • Standard output and standard error of service units. For further
41 details see below.
42 • Audit records, originating from the kernel audit subsystem
44 The daemon will implicitly collect numerous metadata fields for each
46 log messages in a secure and unfakeable way. See systemd.journal-
47 fields(7) for more information about the collected metadata.
48 Log data collected by the journal is primarily text-based but can also
50 include binary data where necessary. Individual fields making up a log
51 record stored in the journal may be up to 2⁶⁴-1 bytes in size.
52 The journal service stores log data either persistently below
54 /var/log/journal or in a volatile way below /run/log/journal/ (in the
55 latter case it is lost at reboot). By default, log data is stored
56 persistently if /var/log/journal/ exists during boot, with an implicit
57 fallback to volatile storage otherwise. Use Storage= in
58 journald.conf(5) to configure where log data is placed, independently
59 of the existence of /var/log/journal/.
60 Note that journald will initially use volatile storage, until a call to
62 journalctl --flush (or sending SIGUSR1 to journald) will cause it to
63 switch to persistent logging (under the conditions mentioned above).
64 This is done automatically on boot via "systemd-journal-flush.service".
65 On systems where /var/log/journal/ does not exist yet but where
67 persistent logging is desired (and the default journald.conf is used),
68 it is sufficient to create the directory, and ensure it has the correct
69 access modes and ownership:
70 mkdir -p /var/log/journal
72 systemd-tmpfiles --create --prefix /var/log/journal
73 See journald.conf(5) for information about the configuration of this
75 service.
76
78 The systemd service manager invokes all service processes with standard
79 output and standard error connected to the journal by default. This
80 behaviour may be altered via the StandardOutput=/StandardError= unit
81 file settings, see systemd.exec(5) for details. The journal converts
82 the log byte stream received this way into individual log records,
83 splitting the stream at newline ("\n", ASCII 10) and NUL bytes.
84 If systemd-journald.service is stopped, the stream connections
86 associated with all services are terminated. Further writes to those
87 streams by the service will result in EPIPE errors. In order to react
88 gracefully in this case it is recommended that programs logging to
89 standard output/error ignore such errors. If the SIGPIPE UNIX signal
90 handler is not blocked or turned off, such write attempts will also
91 result in such process signals being generated, see signal(7). To
92 mitigate this issue, systemd service manager explicitly turns off the
93 SIGPIPE signal for all invoked processes by default (this may be
94 changed for each unit individually via the IgnoreSIGPIPE= option, see
95 systemd.exec(5) for details). After the standard output/standard error
96 streams have been terminated they may not be recovered until the
97 services they are associated with are restarted. Note that during
98 normal operation, systemd-journald.service stores copies of the file
99 descriptors for those streams in the service manager. If
100 systemd-journald.service is restarted using systemctl restart or
101 equivalent operation instead of a pair of separate systemctl stop and
102 systemctl start commands (or equivalent operations), these stream
103 connections are not terminated and survive the restart. It is thus safe
104 to restart systemd-journald.service, but stopping it is not
105 recommended.
106 Note that the log record metadata for records transferred via such
108 standard output/error streams reflect the metadata of the peer the
109 stream was originally created for. If the stream connection is passed
110 on to other processes (such as further child processes forked off the
111 main service process), the log records will not reflect their metadata,
112 but will continue to describe the original process. This is different
113 from the other logging transports listed above, which are inherently
114 record based and where the metadata is always associated with the
115 individual record.
116 In addition to the implicit standard output/error logging of services,
118 stream logging is also available via the systemd-cat(1) command line
119 tool.
120 Currently, the number of parallel log streams systemd-journald will
122 accept is limited to 4096. When this limit is reached further log
123 streams may be established but will receive EPIPE right from the
124 beginning.
125
127 Journal 'namespaces' are both a mechanism for logically isolating the
128 log stream of projects consisting of one or more services from the rest
129 of the system and a mechanism for improving performance. Multiple
130 journal namespaces may exist simultaneously, each defining its own,
131 independent log stream managed by its own instance of systemd-journald.
132 Namespaces are independent of each other, both in the data store and in
133 the IPC interface. By default only a single 'default' namespace exists,
134 managed by systemd-journald.service (and its associated socket units).
135 Additional namespaces are created by starting an instance of the
136 systemd-journald@.service service template. The instance name is the
137 namespace identifier, which is a short string used for referencing the
138 journal namespace. Service units may be assigned to a specific journal
139 namespace through the LogNamespace= unit file setting, see
140 systemd.exec(5) for details. The --namespace= switch of journalctl(1)
141 may be used to view the log stream of a specific namespace. If the
142 switch is not used the log stream of the default namespace is shown,
143 i.e. log data from other namespaces is not visible.
144 Services associated with a specific log namespace may log via syslog,
146 the native logging protocol of the journal and via stdout/stderr; the
147 logging from all three transports is associated with the namespace.
148 By default only the default namespace will collect kernel and audit log
150 messages.
151 The systemd-journald instance of the default namespace is configured
153 through /etc/systemd/journald.conf (see below), while the other
154 instances are configured through /etc/systemd/journald@NAMESPACE.conf.
155 The journal log data for the default namespace is placed in
156 /var/log/journal/MACHINE_ID (see below) while the data for the other
157 namespaces is located in /var/log/journal/MACHINE_ID.NAMESPACE.
158
160 SIGUSR1
161 Request that journal data from /run/ is flushed to /var/ in order
162 to make it persistent (if this is enabled). This must be used after
163 /var/ is mounted, as otherwise log data from /run/ is never flushed
164 to /var/ regardless of the configuration. Use the journalctl
165 --flush command to request flushing of the journal files, and wait
166 for the operation to complete. See journalctl(1) for details.
167 SIGUSR2
169 Request immediate rotation of the journal files. Use the journalctl
170 --rotate command to request journal file rotation, and wait for the
171 operation to complete.
172 SIGRTMIN+1
174 Request that all unwritten log data is written to disk. Use the
175 journalctl --sync command to trigger journal synchronization, and
176 wait for the operation to complete.
177
179 A few configuration parameters from journald.conf may be overridden on
180 the kernel command line:
181 systemd.journald.forward_to_syslog=, systemd.journald.forward_to_kmsg=,
183 systemd.journald.forward_to_console=, systemd.journald.forward_to_wall=
184 Enables/disables forwarding of collected log messages to syslog,
185 the kernel log buffer, the system console or wall.
186 See journald.conf(5) for information about these settings.
188 Note that these kernel command line options are only honoured by the
190 default namespace, see above.
191
193 Journal files are, by default, owned and readable by the
194 "systemd-journal" system group but are not writable. Adding a user to
195 this group thus enables them to read the journal files.
196 By default, each user, with a UID outside the range of system users,
198 dynamic service users, and the nobody user, will get their own set of
199 journal files in /var/log/journal/. See Users, Groups, UIDs and GIDs on
200 systemd systems[2] for more details about UID ranges. These journal
201 files will not be owned by the user, however, in order to avoid that
202 the user can write to them directly. Instead, file system ACLs are used
203 to ensure the user gets read access only.
204 Additional users and groups may be granted access to journal files via
206 file system access control lists (ACL). Distributions and
207 administrators may choose to grant read access to all members of the
208 "wheel" and "adm" system groups with a command such as the following:
209 # setfacl -Rnm g:wheel:rx,d:g:wheel:rx,g:adm:rx,d:g:adm:rx /var/log/journal/
211 Note that this command will update the ACLs both for existing journal
213 files and for future journal files created in the /var/log/journal/
214 directory.
215
217 /etc/systemd/journald.conf
218 Configure systemd-journald behavior. See journald.conf(5).
219 /run/log/journal/machine-id/*.journal,
221 /run/log/journal/machine-id/*.journal~,
222 /var/log/journal/machine-id/*.journal,
223 /var/log/journal/machine-id/*.journal~
224 systemd-journald writes entries to files in
225 /run/log/journal/machine-id/ or /var/log/journal/machine-id/ with
226 the ".journal" suffix. If the daemon is stopped uncleanly, or if
227 the files are found to be corrupted, they are renamed using the
228 ".journal~" suffix, and systemd-journald starts writing to a new
229 file. /run/ is used when /var/log/journal is not available, or
230 when Storage=volatile is set in the journald.conf(5) configuration
231 file.
232 When systemd-journald ceases writing to a journal file, it will be
234 renamed to "original-name@suffix.journal" (or
235 "original-name@suffix.journal~"). Such files are "archived" and
236 will not be written to any more.
237 In general, it is safe to read or copy any journal file (active or
239 archived). journalctl(1) and the functions in the sd-journal(3)
240 library should be able to read all entries that have been fully
241 written.
242 systemd-journald will automatically remove the oldest archived
244 journal files to limit disk use. See SystemMaxUse= and related
245 settings in journald.conf(5).
246 /dev/kmsg, /dev/log, /run/systemd/journal/dev-log,
248 /run/systemd/journal/socket, /run/systemd/journal/stdout
249 Sockets and other file node paths that systemd-journald will listen
250 on and are visible in the file system. In addition to these,
251 systemd-journald can listen for audit events using netlink(7),
252 depending on whether "systemd-journald-audit.socket" is enabled or
253 not.
254 If journal namespacing is used these paths are slightly altered to
256 include a namespace identifier, see above.
257
259 systemd(1), journalctl(1), journald.conf(5), systemd.journal-fields(7),
260 sd-journal(3), systemd-coredump(8), setfacl(1), sd_journal_print(3),
261 pydoc systemd.journal
262
264 1. Native Journal Protocol
265 https://systemd.io/JOURNAL_NATIVE_PROTOCOL
266 2. Users, Groups, UIDs and GIDs on systemd systems
268 https://systemd.io/UIDS-GIDS
269systemd 253 SYSTEMD-JOURNALD.SERVICE(8)