1 systemd System and Service Manager
5 * Two new unit types have been introduced:
7 Scope units are very similar to service units, however, are
8 created out of pre-existing processes -- instead of PID 1
9 forking off the processes. By using scope units it is
10 possible for system services and applications to group their
11 own child processes (worker processes) in a powerful way
12 which then maybe used to organize them, or kill them
13 together, or apply resource limits on them.
15 Slice units may be used to partition system resources in an
16 hierarchial fashion and then assign other units to them. By
17 default there are now three slices: system.slice (for all
18 system services), user.slice (for all user sessions),
19 machine.slice (for VMs and containers).
21 Slices and scopes have been introduced primarily in
22 context of the work to move cgroup handling to a
23 single-writer scheme, where only PID 1
24 creates/removes/manages cgroups.
26 * There's a new concept of "transient" units. In contrast to
27 normal units these units are created via an API at runtime,
28 not from configuration from disk. More specifically this
29 means it is now possible to run arbitrary programs as
30 independent services, with all execution parameters passed
31 in via bus APIs rather than read from disk. Transient units
32 make systemd substantially more dynamic then it ever was,
33 and useful as a general batch manager.
35 * logind has been updated to make use of scope and slice units
36 for managing user sessions. As a user logs in he will get
37 his own private slice unit, to which all sessions are added
38 as scope units. We also added support for automatically
39 adding an instance of user@.service for the user into the
40 slice. Effectively logind will no longer create cgroup
41 hierarchies on its own now, it will defer entirely to PID 1
42 for this by means of scope, service and slice units. Since
43 user sessions this way become entities managed by PID 1
44 the output of "systemctl" is now a lot more comprehensive.
46 * A new mini-daemon "systemd-machined" has been added which
47 may be used by virtualization managers to register local
48 VMs/containers. nspawn has been updated accordingly, and
49 libvirt will be updated shortly. machined will collect a bit
50 of meta information about the VMs/containers, and assign
51 them their own scope unit (see above). The collected
52 meta-data is then made available via the "machinectl" tool,
53 and exposed in "ps" and similar tools. machined/machinectl
54 is compile-time optional.
56 * As discussed earlier, the low-level cgroup configuration
57 options ControlGroup=, ControlGroupModify=,
58 ControlGroupPersistent=, ControlGroupAttribute= have been
59 removed. Please use high-level attribute settings instead as
62 * A new bus call SetUnitProperties() has been added to alter
63 various runtime parameters of a unit. This is primarily
64 useful to alter cgroup parameters dynamically in a nice way,
65 but will be extended later on to make more properties
66 modifiable at runtime. systemctl gained a new set-properties
67 command that wraps this call.
69 * A new tool "systemd-run" has been added which can be used to
70 run arbitrary command lines as transient services or scopes,
71 while configuring a number of settings via the command
72 line. This tool is currently very basic, however already
73 very useful. We plan to extend this tool to even allow
74 queuing of execution jobs with time triggers from the
75 command line, similar in fashion to "at".
77 * nspawn will now inform the user explicitly that kernels with
78 audit enabled break containers, and suggest the user to turn
81 * Support for detecting the IMA and AppArmor security
82 frameworks with ConditionSecurity= has been added.
84 * journalctl gained a new "-k" switch for showing only kernel
85 messages, mimicking dmesg output; in addition to "--user"
86 and "--system" switches for showing only user's own logs
89 * systemd-delta can now show information about drop-in
90 snippets extending unit files.
92 * libsystemd-bus has been substantially updated but is still
93 not available as public API.
95 * systemd will now look for the "debug" argument on the kernel
96 command line and enable debug logging, similar to
97 "systemd.log_level=debug" already did before.
99 * "systemctl set-default", "systemctl get-default" has been
100 added to configure the default.target symlink, which
101 controls what to boot into by default.
103 * "systemctl set-log-level" has been added as a convenient
104 way to raise and lower systemd logging threshold.
106 * "systemd-analyze plot" will now show the time the various
107 generators needed for execution, as well as information
108 about the unit file loading.
110 * journalctl gained new "--user" and "--system" switches to
111 only show user/system logs.
113 * libsystemd-journal gained a new sd_journal_open_files() call
114 for opening specific journal files. journactl also gained a
115 new switch to expose this new functionality. Previously we
116 only supported opening all files from a directory, or all
117 files from the system, as opening individual files only is
118 racy due to journal file rotation.
120 * systemd gained the new DefaultEnvironment= setting in
121 /etc/systemd/system.conf to set environment variables for
124 * If a privileged process logs a journal message with the
125 OBJECT_PID= field set, then journald will automatically
126 augment this with additional OBJECT_UID=, OBJECT_GID=,
127 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
128 system services want to log events about specific client
129 processes. journactl/systemctl has been updated to make use
130 of this information if all log messages regarding a specific
133 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
134 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
135 Reisner, David Coppa, David King, David Strauss, Eelco
136 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
137 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
138 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
139 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
140 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
141 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
142 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
143 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
144 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
149 * The Python bindings gained some minimal support for the APIs
150 exposed by libsystemd-logind.
152 * ConditionSecurity= gained support for detecting SMACK. Since
153 this condition already supports SELinux and AppArmor we only
154 miss IMA for this. Patches welcome!
156 Contributions from: Karol Lewandowski, Lennart Poettering,
157 Zbigniew Jędrzejewski-Szmek
161 * systemd-nspawn will now create /etc/resolv.conf if
162 necessary, before bind-mounting the host's file onto it.
164 * systemd-nspawn will now store meta information about a
165 container on the container's cgroup as extended attribute
166 fields, including the root directory.
168 * The cgroup hierarchy has been reworked in many ways. All
169 objects any of the components systemd creates in the cgroup
170 tree are now suffixed. More specifically, user sessions are
171 now placed in cgroups suffixed with ".session", users in
172 cgroups suffixed with ".user", and nspawn containers in
173 cgroups suffixed with ".nspawn". Furthermore, all cgroup
174 names are now escaped in a simple scheme to avoid collision
175 of userspace object names with kernel filenames. This work
176 is preparation for making these objects relocatable in the
177 cgroup tree, in order to allow easy resource partitioning of
178 these objects without causing naming conflicts.
180 * systemctl list-dependencies gained the new switches
181 --plain, --reverse, --after and --before.
183 * systemd-inhibit now shows the process name of processes that
184 have taken an inhibitor lock.
186 * nss-myhostname will now also resolve "localhost"
187 implicitly. This makes /etc/hosts an optional file and
188 nicely handles that on IPv6 ::1 maps to both "localhost" and
191 * libsystemd-logind.so gained a new call
192 sd_get_machine_names() to enumerate running containers and
193 VMs (currently only supported by very new libvirt and
194 nspawn). sd_login_monitor can now be used to watch
195 VMs/containers coming and going.
197 * .include is not allowed recursively anymore, and only in
198 unit files. Usually it is better to use drop-in snippets in
199 .d/*.conf anyway, as introduced with systemd 198.
201 * systemd-analyze gained a new "critical-chain" command that
202 determines the slowest chain of units run during system
203 boot-up. It is very useful for tracking down where
204 optimizing boot time is the most beneficial.
206 * systemd will no longer allow manipulating service paths in
207 the name=systemd:/system cgroup tree using ControlGroup= in
208 units. (But is still fine with it in all other dirs.)
210 * There's a new systemd-nspawn@.service service file that may
211 be used to easily run nspawn containers as system
212 services. With the container's root directory in
213 /var/lib/container/foobar it is now sufficient to run
214 "systemctl start systemd-nspawn@foobar.service" to boot it.
216 * systemd-cgls gained a new parameter "--machine" to list only
217 the processes within a certain container.
219 * ConditionSecurity= now can check for "apparmor". We still
220 are lacking checks for SMACK and IMA for this condition
221 check though. Patches welcome!
223 * A new configuration file /etc/systemd/sleep.conf has been
224 added that may be used to configure which kernel operation
225 systemd is supposed to execute when "suspend", "hibernate"
226 or "hybrid-sleep" is requested. This makes the new kernel
227 "freeze" state accessible to the user.
229 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
230 the passed argument if applicable.
232 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
233 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
234 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
235 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
236 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
237 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
238 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
243 * The output of 'systemctl list-jobs' got some polishing. The
244 '--type=' argument may now be passed more than once. A new
245 command 'systemctl list-sockets' has been added which shows
246 a list of kernel sockets systemd is listening on with the
247 socket units they belong to, plus the units these socket
250 * The experimental libsystemd-bus library got substantial
251 updates to work in conjunction with the (also experimental)
252 kdbus kernel project. It works well enough to exchange
253 messages with some sophistication. Note that kdbus is not
254 ready yet, and the library is mostly an elaborate test case
255 for now, and not installable.
257 * systemd gained a new unit 'systemd-static-nodes.service'
258 that generates static device nodes earlier during boot, and
259 can run in conjunction with udev.
261 * libsystemd-login gained a new call sd_pid_get_user_unit()
262 to retrieve the user systemd unit a process is running
263 in. This is useful for systems where systemd is used as
266 * systemd-nspawn now places all containers in the new /machine
267 top-level cgroup directory in the name=systemd
268 hierarchy. libvirt will soon do the same, so that we get a
269 uniform separation of /system, /user and /machine for system
270 services, user processes and containers/virtual
271 machines. This new cgroup hierarchy is also useful to stick
272 stable names to specific container instances, which can be
273 recognized later this way (this name may be controlled
274 via systemd-nspawn's new -M switch). libsystemd-login also
275 gained a new call sd_pid_get_machine_name() to retrieve the
276 name of the container/VM a specific process belongs to.
278 * bootchart can now store its data in the journal.
280 * libsystemd-journal gained a new call
281 sd_journal_add_conjunction() for AND expressions to the
282 matching logic. This can be used to express more complex
285 * journactl can now take multiple --unit= and --user-unit=
288 * The cryptsetup logic now understands the "luks.key=" kernel
289 command line switch for specifying a file to read the
290 decryption key from. Also, if a configured key file is not
291 found the tool will now automatically fall back to prompting
294 * Python systemd.journal module was updated to wrap recently
295 added functions from libsystemd-journal. The interface was
296 changed to bring the low level interface in s.j._Reader
297 closer to the C API, and the high level interface in
298 s.j.Reader was updated to wrap and convert all data about
301 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
302 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
303 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
304 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
305 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
306 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
310 * journalctl --update-catalog now understands a new --root=
311 option to operate on catalogs found in a different root
314 * During shutdown after systemd has terminated all running
315 services a final killing loop kills all remaining left-over
316 processes. We will now print the name of these processes
317 when we send SIGKILL to them, since this usually indicates a
320 * If /etc/crypttab refers to password files stored on
321 configured mount points automatic dependencies will now be
322 generated to ensure the specific mount is established first
323 before the key file is attempted to be read.
325 * 'systemctl status' will now show information about the
326 network sockets a socket unit is listening on.
328 * 'systemctl status' will also shown information about any
329 drop-in configuration file for units. (Drop-In configuration
330 files in this context are files such as
331 /etc/systemd/systemd/foobar.service.d/*.conf)
333 * systemd-cgtop now optionally shows summed up CPU times of
334 cgroups. Press '%' while running cgtop to switch between
335 percentage and absolute mode. This is useful to determine
336 which cgroups use up the most CPU time over the entire
337 runtime of the system. systemd-cgtop has also been updated
338 to be 'pipeable' for processing with further shell tools.
340 * 'hostnamectl set-hostname' will now allow setting of FQDN
343 * The formatting and parsing of time span values has been
344 changed. The parser now understands fractional expressions
345 such as "5.5h". The formatter will now output fractional
346 expressions for all time spans under 1min, i.e. "5.123456s"
347 rather than "5s 123ms 456us". For time spans under 1s
348 millisecond values are shown, for those under 1ms
349 microsecond values are shown. This should greatly improve
350 all time-related output of systemd.
352 * libsystemd-login and libsystemd-journal gained new
353 functions for querying the poll() events mask and poll()
354 timeout value for integration into arbitrary event
357 * localectl gained the ability to list available X11 keymaps
358 (models, layouts, variants, options).
360 * 'systemd-analyze dot' gained the ability to filter for
361 specific units via shell-style globs, to create smaller,
362 more useful graphs. I.e. it's now possible to create simple
363 graphs of all the dependencies between only target units, or
364 of all units that Avahi has dependencies with.
366 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
367 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
368 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
369 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
370 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
371 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
372 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
376 * The boot-time readahead implementation for rotating media
377 will now read the read-ahead data in multiple passes which
378 consist of all read requests made in equidistant time
379 intervals. This means instead of strictly reading read-ahead
380 data in its physical order on disk we now try to find a
381 middle ground between physical and access time order.
383 * /etc/os-release files gained a new BUILD_ID= field for usage
384 on operating systems that provide continuous builds of OS
387 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
388 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
389 William Douglas, Zbigniew Jędrzejewski-Szmek
393 * systemd-python gained an API exposing libsystemd-daemon.
395 * The SMACK setup logic gained support for uploading CIPSO
398 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
399 ReadOnlyDirectories= and InaccessibleDirectories= has
400 changed. The private /tmp and /var/tmp directories are now
401 shared by all processes of a service (which means
402 ExecStartPre= may now leave data in /tmp that ExecStart= of
403 the same service can still access). When a service is
404 stopped its temporary directories are immediately deleted
405 (normal clean-up with tmpfiles is still done in addition to
408 * By default, systemd will now set a couple of sysctl
409 variables in the kernel: the safe sysrq options are turned
410 on, IP route verification is turned on, and source routing
411 disabled. The recently added hardlink and softlink
412 protection of the kernel is turned on. These settings should
413 be reasonably safe, and good defaults for all new systems.
415 * The predictable network naming logic may now be turned off
416 with a new kernel command line switch: net.ifnames=0.
418 * A new libsystemd-bus module has been added that implements a
419 pretty complete D-Bus client library. For details see:
421 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
423 * journald will now explicitly flush the journal files to disk
424 at the latest 5min after each write. The file will then also
425 be marked offline until the next write. This should increase
426 reliability in case of a crash. The synchronization delay
427 can be configured via SyncIntervalSec= in journald.conf.
429 * There's a new remote-fs-setup.target unit that can be used
430 to pull in specific services when at least one remote file
431 system is to be mounted.
433 * There are new targets timers.target and paths.target as
434 canonical targets to pull user timer and path units in
435 from. This complements sockets.target with a similar
436 purpose for socket units.
438 * libudev gained a new call udev_device_set_attribute_value()
439 to set sysfs attributes of a device.
441 * The udev daemon now sets the default number of worker
442 processes executed in parallel based on the number of available
443 CPUs instead of the amount of available RAM. This is supposed
444 to provide a more reliable default and limit a too aggressive
445 paralellism for setups with 1000s of devices connected.
447 Contributions from: Auke Kok, Colin Walters, Cristian
448 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
449 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
450 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
451 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
452 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
453 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
454 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
455 Zbigniew Jędrzejewski-Szmek
459 * Configuration of unit files may now be extended via drop-in
460 files without having to edit/override the unit files
461 themselves. More specifically, if the administrator wants to
462 change one value for a service file foobar.service he can
463 now do so by dropping in a configuration snippet into
464 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
465 will load all these snippets and apply them on top of the
466 main unit configuration file, possibly extending or
467 overriding its settings. Using these drop-in snippets is
468 generally nicer than the two earlier options for changing
469 unit files locally: copying the files from
470 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
471 them there; or creating a new file in /etc/systemd/system/
472 that incorporates the original one via ".include". Drop-in
473 snippets into these .d/ directories can be placed in any
474 directory systemd looks for units in, and the usual
475 overriding semantics between /usr/lib, /etc and /run apply
478 * Most unit file settings which take lists of items can now be
479 reset by assigning the empty string to them. For example,
480 normally, settings such as Environment=FOO=BAR append a new
481 environment variable assignment to the environment block,
482 each time they are used. By assigning Environment= the empty
483 string the environment block can be reset to empty. This is
484 particularly useful with the .d/*.conf drop-in snippets
485 mentioned above, since this adds the ability to reset list
486 settings from vendor unit files via these drop-ins.
488 * systemctl gained a new "list-dependencies" command for
489 listing the dependencies of a unit recursively.
491 * Inhibitors are now honored and listed by "systemctl
492 suspend", "systemctl poweroff" (and similar) too, not only
493 GNOME. These commands will also list active sessions by
496 * Resource limits (as exposed by the various control group
497 controllers) can now be controlled dynamically at runtime
498 for all units. More specifically, you can now use a command
499 like "systemctl set-cgroup-attr foobar.service cpu.shares
500 2000" to alter the CPU shares a specific service gets. These
501 settings are stored persistently on disk, and thus allow the
502 administrator to easily adjust the resource usage of
503 services with a few simple commands. This dynamic resource
504 management logic is also available to other programs via the
505 bus. Almost any kernel cgroup attribute and controller is
508 * systemd-vconsole-setup will now copy all font settings to
509 all allocated VTs, where it previously applied them only to
512 * libsystemd-login gained the new sd_session_get_tty() API
515 * This release drops support for a few legacy or
516 distribution-specific LSB facility names when parsing init
517 scripts: $x-display-manager, $mail-transfer-agent,
518 $mail-transport-agent, $mail-transfer-agent, $smtp,
519 $null. Also, the mail-transfer-agent.target unit backing
520 this has been removed. Distributions which want to retain
521 compatibility with this should carry the burden for
522 supporting this themselves and patch support for these back
523 in, if they really need to. Also, the facilities $syslog and
524 $local_fs are now ignored, since systemd does not support
525 early-boot LSB init scripts anymore, and these facilities
526 are implied anyway for normal services. syslog.target has
529 * There are new bus calls on PID1's Manager object for
530 cancelling jobs, and removing snapshot units. Previously,
531 both calls were only available on the Job and Snapshot
534 * systemd-journal-gatewayd gained SSL support.
536 * The various "environment" files, such as /etc/locale.conf
537 now support continuation lines with a backslash ("\") as
538 last character in the line, similar in style (but different)
539 to how this is supported in shells.
541 * For normal user processes the _SYSTEMD_USER_UNIT= field is
542 now implicitly appended to every log entry logged. systemctl
543 has been updated to filter by this field when operating on a
544 user systemd instance.
546 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
547 CAP_AUDIT_CONTROL capabilities to the capabilities set for
548 the container. This makes it easier to boot unmodified
549 Fedora systems in a container, which however still requires
550 audit=0 to be passed on the kernel command line. Auditing in
551 kernel and userspace is unfortunately still too broken in
552 context of containers, hence we recommend compiling it out
553 of the kernel or using audit=0. Hopefully this will be fixed
554 one day for good in the kernel.
556 * nspawn gained the new --bind= and --bind-ro= parameters to
557 bind mount specific directories from the host into the
560 * nspawn will now mount its own devpts file system instance
561 into the container, in order not to leak pty devices from
562 the host into the container.
564 * systemd will now read the firmware boot time performance
565 information from the EFI variables, if the used boot loader
566 supports this, and takes it into account for boot performance
567 analysis via "systemd-analyze". This is currently supported
568 only in conjunction with Gummiboot, but could be supported
569 by other boot loaders too. For details see:
571 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
573 * A new generator has been added that automatically mounts the
574 EFI System Partition (ESP) to /boot, if that directory
575 exists, is empty, and no other file system has been
576 configured to be mounted there.
578 * logind will now send out PrepareForSleep(false) out
579 unconditionally, after coming back from suspend. This may be
580 used by applications as asynchronous notification for
581 system resume events.
583 * "systemctl unlock-sessions" has been added, that allows
584 unlocking the screens of all user sessions at once, similar
585 how "systemctl lock-sessions" already locked all users
586 sessions. This is backed by a new D-Bus call UnlockSessions().
588 * "loginctl seat-status" will now show the master device of a
589 seat. (i.e. the device of a seat that needs to be around for
590 the seat to be considered available, usually the graphics
593 * tmpfiles gained a new "X" line type, that allows
594 configuration of files and directories (with wildcards) that
595 shall be excluded from automatic cleanup ("aging").
597 * udev default rules set the device node permissions now only
598 at "add" events, and do not change them any longer with a
599 later "change" event.
601 * The log messages for lid events and power/sleep keypresses
602 now carry a message ID.
604 * We now have a substantially larger unit test suite, but this
605 continues to be work in progress.
607 * udevadm hwdb gained a new --root= parameter to change the
608 root directory to operate relative to.
610 * logind will now issue a background sync() request to the kernel
611 early at shutdown, so that dirty buffers are flushed to disk early
612 instead of at the last moment, in order to optimize shutdown
615 * A new bootctl tool has been added that is an interface for
616 certain boot loader operations. This is currently a preview
617 and is likely to be extended into a small mechanism daemon
618 like timedated, localed, hostnamed, and can be used by
619 graphical UIs to enumerate available boot options, and
620 request boot into firmware operations.
622 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
623 the rest of the package. It also has been updated to work
624 correctly in initrds.
626 * Policykit previously has been runtime optional, and is now
627 also compile time optional via a configure switch.
629 * systemd-analyze has been reimplemented in C. Also "systemctl
630 dot" has moved into systemd-analyze.
632 * "systemctl status" with no further parameters will now print
633 the status of all active or failed units.
635 * Operations such as "systemctl start" can now be executed
636 with a new mode "--irreversible" which may be used to queue
637 operations that cannot accidentally be reversed by a later
638 job queuing. This is by default used to make shutdown
639 requests more robust.
641 * The Python API of systemd now gained a new module for
642 reading journal files.
644 * A new tool kernel-install has been added that can install
645 kernel images according to the Boot Loader Specification:
647 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
649 * Boot time console output has been improved to provide
650 animated boot time output for hanging jobs.
652 * A new tool systemd-activate has been added which can be used
653 to test socket activation with, directly from the command
654 line. This should make it much easier to test and debug
655 socket activation in daemons.
657 * journalctl gained a new "--reverse" (or -r) option to show
658 journal output in reverse order (i.e. newest line first).
660 * journalctl gained a new "--pager-end" (or -e) option to jump
661 to immediately jump to the end of the journal in the
662 pager. This is only supported in conjunction with "less".
664 * journalctl gained a new "--user-unit=" option, that works
665 similar to "--unit=" but filters for user units rather than
668 * A number of unit files to ease adoption of systemd in
669 initrds has been added. This moves some minimal logic from
670 the various initrd implementations into systemd proper.
672 * The journal files are now owned by a new group
673 "systemd-journal", which exists specifically to allow access
674 to the journal, and nothing else. Previously, we used the
675 "adm" group for that, which however possibly covers more
676 than just journal/log file access. This new group is now
677 already used by systemd-journal-gatewayd to ensure this
678 daemon gets access to the journal files and as little else
679 as possible. Note that "make install" will also set FS ACLs
680 up for /var/log/journal to give "adm" and "wheel" read
681 access to it, in addition to "systemd-journal" which owns
682 the journal files. We recommend that packaging scripts also
683 add read access to "adm" + "wheel" to /var/log/journal, and
684 all existing/future journal files. To normal users and
685 administrators little changes, however packagers need to
686 ensure to create the "systemd-journal" system group at
687 package installation time.
689 * The systemd-journal-gatewayd now runs as unprivileged user
690 systemd-journal-gateway:systemd-journal-gateway. Packaging
691 scripts need to create these system user/group at
694 * timedated now exposes a new boolean property CanNTP that
695 indicates whether a local NTP service is available or not.
697 * systemd-detect-virt will now also detect xen PVs
699 * The pstore file system is now mounted by default, if it is
702 * In addition to the SELinux and IMA policies we will now also
703 load SMACK policies at early boot.
705 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
706 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
707 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
708 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
709 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
710 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
711 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
712 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
713 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
714 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
715 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
716 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
717 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
718 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
722 * Timer units now support calendar time events in addition to
723 monotonic time events. That means you can now trigger a unit
724 based on a calendar time specification such as "Thu,Fri
725 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
726 or fifth day of any month of the year 2013, given that it is
727 a thursday or friday. This brings timer event support
728 considerably closer to cron's capabilities. For details on
729 the supported calendar time specification language see
732 * udev now supports a number of different naming policies for
733 network interfaces for predictable names, and a combination
734 of these policies is now the default. Please see this wiki
735 document for details:
737 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
739 * Auke Kok's bootchart implementation has been added to the
740 systemd tree. It's an optional component that can graph the
741 boot in quite some detail. It's one of the best bootchart
742 implementations around and minimal in its code and
745 * nss-myhostname has been integrated into the systemd source
746 tree. nss-myhostname guarantees that the local hostname
747 always stays resolvable via NSS. It has been a weak
748 requirement of systemd-hostnamed since a long time, and
749 since its code is actually trivial we decided to just
750 include it in systemd's source tree. It can be turned off
751 with a configure switch.
753 * The read-ahead logic is now capable of properly detecting
754 whether a btrfs file system is on SSD or rotating media, in
755 order to optimize the read-ahead scheme. Previously, it was
756 only capable of detecting this on traditional file systems
759 * In udev, additional device properties are now read from the
760 IAB in addition to the OUI database. Also, Bluetooth company
761 identities are attached to the devices as well.
763 * In service files %U may be used as specifier that is
764 replaced by the configured user name of the service.
766 * nspawn may now be invoked without a controlling TTY. This
767 makes it suitable for invocation as its own service. This
768 may be used to set up a simple containerized server system
769 using only core OS tools.
771 * systemd and nspawn can now accept socket file descriptors
772 when they are started for socket activation. This enables
773 implementation of socket activated nspawn
774 containers. i.e. think about autospawning an entire OS image
775 when the first SSH or HTTP connection is received. We expect
776 that similar functionality will also be added to libvirt-lxc
779 * journalctl will now suppress ANSI color codes when
782 * systemctl will no longer show control group information for
783 a unit if a the control group is empty anyway.
785 * logind can now automatically suspend/hibernate/shutdown the
788 * /etc/machine-info and hostnamed now also expose the chassis
789 type of the system. This can be used to determine whether
790 the local system is a laptop, desktop, handset or
791 tablet. This information may either be configured by the
792 user/vendor or is automatically determined from ACPI and DMI
793 information if possible.
795 * A number of PolicyKit actions are now bound together with
796 "imply" rules. This should simplify creating UIs because
797 many actions will now authenticate similar ones as well.
799 * Unit files learnt a new condition ConditionACPower= which
800 may be used to conditionalize a unit depending on whether an
801 AC power source is connected or not, of whether the system
802 is running on battery power.
804 * systemctl gained a new "is-failed" verb that may be used in
805 shell scripts and suchlike to check whether a specific unit
806 is in the "failed" state.
808 * The EnvironmentFile= setting in unit files now supports file
809 globbing, and can hence be used to easily read a number of
810 environment files at once.
812 * systemd will no longer detect and recognize specific
813 distributions. All distribution-specific #ifdeffery has been
814 removed, systemd is now fully generic and
815 distribution-agnostic. Effectively, not too much is lost as
816 a lot of the code is still accessible via explicit configure
817 switches. However, support for some distribution specific
818 legacy configuration file formats has been dropped. We
819 recommend distributions to simply adopt the configuration
820 files everybody else uses now and convert the old
821 configuration from packaging scripts. Most distributions
822 already did that. If that's not possible or desirable,
823 distributions are welcome to forward port the specific
824 pieces of code locally from the git history.
826 * When logging a message about a unit systemd will now always
827 log the unit name in the message meta data.
829 * localectl will now also discover system locale data that is
830 not stored in locale archives, but directly unpacked.
832 * logind will no longer unconditionally use framebuffer
833 devices as seat masters, i.e. as devices that are required
834 to be existing before a seat is considered preset. Instead,
835 it will now look for all devices that are tagged as
836 "seat-master" in udev. By default framebuffer devices will
837 be marked as such, but depending on local systems other
838 devices might be marked as well. This may be used to
839 integrate graphics cards using closed source drivers (such
840 as NVidia ones) more nicely into logind. Note however, that
841 we recommend using the open source NVidia drivers instead,
842 and no udev rules for the closed-source drivers will be
843 shipped from us upstream.
845 Contributions from: Adam Williamson, Alessandro Crismani, Auke
846 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
847 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
848 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
849 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
850 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
851 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
852 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
853 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
854 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
855 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
856 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
861 * udev gained support for loading additional device properties
862 from an indexed database that is keyed by vendor/product IDs
863 and similar device identifiers. For the beginning this
864 "hwdb" is populated with data from the well-known PCI and
865 USB database, but also includes PNP, ACPI and OID data. In
866 the longer run this indexed database shall grow into
867 becoming the one central database for non-essential
868 userspace device metadata. Previously, data from the PCI/USB
869 database was only attached to select devices, since the
870 lookup was a relatively expensive operation due to O(n) time
871 complexity (with n being the number of entries in the
872 database). Since this is now O(1), we decided to add in this
873 data for all devices where this is available, by
874 default. Note that the indexed database needs to be rebuilt
875 when new data files are installed. To achieve this you need
876 to update your packaging scripts to invoke "udevadm hwdb
877 --update" after installation of hwdb data files. For
878 RPM-based distributions we introduced the new
879 %udev_hwdb_update macro for this purpose.
881 * The Journal gained support for the "Message Catalog", an
882 indexed database to link up additional information with
883 journal entries. For further details please check:
885 http://www.freedesktop.org/wiki/Software/systemd/catalog
887 The indexed message catalog database also needs to be
888 rebuilt after installation of message catalog files. Use
889 "journalctl --update-catalog" for this. For RPM-based
890 distributions we introduced the %journal_catalog_update
891 macro for this purpose.
893 * The Python Journal bindings gained support for the standard
894 Python logging framework.
896 * The Journal API gained new functions for checking whether
897 the underlying file system of a journal file is capable of
898 properly reporting file change notifications, or whether
899 applications that want to reflect journal changes "live"
900 need to recheck journal files continuously in appropriate
903 * It is now possible to set the "age" field for tmpfiles
904 entries to 0, indicating that files matching this entry
905 shall always be removed when the directories are cleaned up.
907 * coredumpctl gained a new "gdb" verb which invokes gdb
908 right-away on the selected coredump.
910 * There's now support for "hybrid sleep" on kernels that
911 support this, in addition to "suspend" and "hibernate". Use
912 "systemctl hybrid-sleep" to make use of this.
914 * logind's HandleSuspendKey= setting (and related settings)
915 now gained support for a new "lock" setting to simply
916 request the screen lock on all local sessions, instead of
917 actually executing a suspend or hibernation.
919 * systemd will now mount the EFI variables file system by
922 * Socket units now gained support for configuration of the
923 SMACK security label.
925 * timedatectl will now output the time of the last and next
926 daylight saving change.
928 * We dropped support for various legacy and distro-specific
929 concepts, such as insserv, early-boot SysV services
930 (i.e. those for non-standard runlevels such as 'b' or 'S')
931 or ArchLinux /etc/rc.conf support. We recommend the
932 distributions who still need support this to either continue
933 to maintain the necessary patches downstream, or find a
934 different solution. (Talk to us if you have questions!)
936 * Various systemd components will now bypass PolicyKit checks
937 for root and otherwise handle properly if PolicyKit is not
938 found to be around. This should fix most issues for
939 PolicyKit-less systems. Quite frankly this should have been
940 this way since day one. It is absolutely our intention to
941 make systemd work fine on PolicyKit-less systems, and we
942 consider it a bug if something doesn't work as it should if
943 PolicyKit is not around.
945 * For embedded systems it is now possible to build udev and
946 systemd without blkid and/or kmod support.
948 * "systemctl switch-root" is now capable of switching root
949 more than once. I.e. in addition to transitions from the
950 initrd to the host OS it is now possible to transition to
951 further OS images from the host. This is useful to implement
952 offline updating tools.
954 * Various other additions have been made to the RPM macros
955 shipped with systemd. Use %udev_rules_update() after
956 installing new udev rules files. %_udevhwdbdir,
957 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
958 %_sysctldir are now available which resolve to the right
959 directories for packages to place various data files in.
961 * journalctl gained the new --full switch (in addition to
962 --all, to disable ellipsation for long messages.
964 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
965 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
966 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
967 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
968 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
969 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
970 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
971 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
972 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
976 * journalctl gained new --since= and --until= switches to
977 filter by time. It also now supports nice filtering for
978 units via --unit=/-u.
980 * Type=oneshot services may use ExecReload= and do the
983 * The journal daemon now supports time-based rotation and
984 vacuuming, in addition to the usual disk-space based
987 * The journal will now index the available field values for
988 each field name. This enables clients to show pretty drop
989 downs of available match values when filtering. The bash
990 completion of journalctl has been updated
991 accordingly. journalctl gained a new switch -F to list all
992 values a certain field takes in the journal database.
994 * More service events are now written as structured messages
995 to the journal, and made recognizable via message IDs.
997 * The timedated, localed and hostnamed mini-services which
998 previously only provided support for changing time, locale
999 and hostname settings from graphical DEs such as GNOME now
1000 also have a minimal (but very useful) text-based client
1001 utility each. This is probably the nicest way to changing
1002 these settings from the command line now, especially since
1003 it lists available options and is fully integrated with bash
1006 * There's now a new tool "systemd-coredumpctl" to list and
1007 extract coredumps from the journal.
1009 * We now install a README each in /var/log/ and
1010 /etc/rc.d/init.d explaining where the system logs and init
1011 scripts went. This hopefully should help folks who go to
1012 that dirs and look into the otherwise now empty void and
1013 scratch their heads.
1015 * When user-services are invoked (by systemd --user) the
1016 $MANAGERPID env var is set to the PID of systemd.
1018 * SIGRTMIN+24 when sent to a --user instance will now result
1019 in immediate termination of systemd.
1021 * gatewayd received numerous feature additions such as a
1022 "follow" mode, for live syncing and filtering.
1024 * browse.html now allows filtering and showing detailed
1025 information on specific entries. Keyboard navigation and
1026 mouse screen support has been added.
1028 * gatewayd/journalctl now supports HTML5/JSON
1029 Server-Sent-Events as output.
1031 * The SysV init script compatibility logic will now
1032 heuristically determine whether a script supports the
1033 "reload" verb, and only then make this available as
1036 * "systemctl status --follow" has been removed, use "journalctl
1039 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
1040 have been removed since they are hardly useful to be
1043 * And I'd like to take the opportunity to specifically mention
1044 Zbigniew for his great contributions. Zbigniew, you rock!
1046 Contributions from: Andrew Eikum, Christian Hesse, Colin
1047 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
1048 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
1049 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
1050 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
1051 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
1052 Jędrzejewski-Szmek, Сковорода Никита Андреевич
1056 * If /etc/vconsole.conf is non-existent or empty we will no
1057 longer load any console font or key map at boot by
1058 default. Instead the kernel defaults will be left
1059 intact. This is definitely the right thing to do, as no
1060 configuration should mean no configuration, and hard-coding
1061 font names that are different on all archs is probably a bad
1062 idea. Also, the kernel default key map and font should be
1063 good enough for most cases anyway, and mostly identical to
1064 the userspace fonts/key maps we previously overloaded them
1065 with. If distributions want to continue to default to a
1066 non-kernel font or key map they should ship a default
1067 /etc/vconsole.conf with the appropriate contents.
1069 Contributions from: Colin Walters, Daniel J Walsh, Dave
1070 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
1071 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1075 * journalctl gained a new --cursor= switch to show entries
1076 starting from the specified location in the journal.
1078 * We now enforce a size limit on journal entry fields exported
1079 with "-o json" in journalctl. Fields larger than 4K will be
1080 assigned null. This can be turned off with --all.
1082 * An (optional) journal gateway daemon is now available as
1083 "systemd-journal-gatewayd.service". This service provides
1084 access to the journal via HTTP and JSON. This functionality
1085 will be used to implement live log synchronization in both
1086 pull and push modes, but has various other users too, such
1087 as easy log access for debugging of embedded devices. Right
1088 now it is already useful to retrieve the journal via HTTP:
1090 # systemctl start systemd-journal-gatewayd.service
1091 # wget http://localhost:19531/entries
1093 This will download the journal contents in a
1094 /var/log/messages compatible format. The same as JSON:
1096 # curl -H"Accept: application/json" http://localhost:19531/entries
1098 This service is also accessible via a web browser where a
1099 single static HTML5 app is served that uses the JSON logic
1100 to enable the user to do some basic browsing of the
1101 journal. This will be extended later on. Here's an example
1102 screenshot of this app in its current state:
1104 http://0pointer.de/public/journal-gatewayd
1106 Contributions from: Kay Sievers, Lennart Poettering, Robert
1107 Milasan, Tom Gundersen
1111 * The bash completion logic is now available for journalctl
1114 * We don't mount the "cpuset" controller anymore together with
1115 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
1116 started if no parameters are assigned to it. "cpuset" hence
1117 broke code that assumed it it could create "cpu" groups and
1120 * journalctl -f will now subscribe to terminal size changes,
1121 and line break accordingly.
1123 Contributions from: Dave Reisner, Kay Sievers, Lennart
1124 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
1128 * nspawn will now create a symlink /etc/localtime in the
1129 container environment, copying the host's timezone
1130 setting. Previously this has been done via a bind mount, but
1131 since symlinks cannot be bind mounted this has now been
1132 changed to create/update the appropriate symlink.
1134 * journalctl -n's line number argument is now optional, and
1135 will default to 10 if omitted.
1137 * journald will now log the maximum size the journal files may
1138 take up on disk. This is particularly useful if the default
1139 built-in logic of determining this parameter from the file
1140 system size is used. Use "systemctl status
1141 systemd-journald.service" to see this information.
1143 * The multi-seat X wrapper tool has been stripped down. As X
1144 is now capable of enumerating graphics devices via udev in a
1145 seat-aware way the wrapper is not strictly necessary
1146 anymore. A stripped down temporary stop-gap is still shipped
1147 until the upstream display managers have been updated to
1148 fully support the new X logic. Expect this wrapper to be
1149 removed entirely in one of the next releases.
1151 * HandleSleepKey= in logind.conf has been split up into
1152 HandleSuspendKey= and HandleHibernateKey=. The old setting
1153 is not available anymore. X11 and the kernel are
1154 distuingishing between these keys and we should too. This
1155 also means the inhibition lock for these keys has been split
1158 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
1159 Poettering, Lukas Nykryn, Václav Pavlín
1163 * Whenever a unit changes state we'll now log this to the
1164 journal and show along the unit's own log output in
1167 * ConditionPathIsMountPoint= can now properly detect bind
1168 mount points too. (Previously, a bind mount of one file
1169 system to another place in the same file system could not be
1170 detected as mount, since they shared struct stat's st_dev
1173 * We will now mount the cgroup controllers cpu, cpuacct,
1174 cpuset and the controllers net_cls, net_prio together by
1177 * nspawn containers will now have a virtualized boot
1178 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
1179 over with a randomized ID at container initialization). This
1180 has the effect of making "journalctl -b" do the right thing
1183 * The JSON output journal serialization has been updated not
1184 to generate "endless" list objects anymore, but rather one
1185 JSON object per line. This is more in line how most JSON
1186 parsers expect JSON objects. The new output mode
1187 "json-pretty" has been added to provide similar output, but
1188 neatly aligned for readability by humans.
1190 * We dropped all explicit sync() invocations in the shutdown
1191 code. The kernel does this implicitly anyway in the kernel
1192 reboot() syscall. halt(8)'s -n option is now a compatibility
1195 * We now support virtualized reboot() in containers, as
1196 supported by newer kernels. We will fall back to exit() if
1197 CAP_SYS_REBOOT is not available to the container. Also,
1198 nspawn makes use of this now and will actually reboot the
1199 container if the containerized OS asks for that.
1201 * journalctl will only show local log output by default
1202 now. Use --merge (-m) to show remote log output, too.
1204 * libsystemd-journal gained the new sd_journal_get_usage()
1205 call to determine the current disk usage of all journal
1206 files. This is exposed in the new "journalctl --disk-usage"
1209 * journald gained a new configuration setting SplitMode= in
1210 journald.conf which may be used to control how user journals
1211 are split off. See journald.conf(5) for details.
1213 * A new condition type ConditionFileNotEmpty= has been added.
1215 * tmpfiles' "w" lines now support file globbing, to write
1216 multiple files at once.
1218 * We added Python bindings for the journal submission
1219 APIs. More Python APIs for a number of selected APIs will
1220 likely follow. Note that we intend to add native bindings
1221 only for the Python language, as we consider it common
1222 enough to deserve bindings shipped within systemd. There are
1223 various projects outside of systemd that provide bindings
1224 for languages such as PHP or Lua.
1226 * Many conditions will now resolve specifiers such as %i. In
1227 addition, PathChanged= and related directives of .path units
1228 now support specifiers as well.
1230 * There's now a new RPM macro definition for the system preset
1233 * journald will now warn if it can't forward a message to the
1234 syslog daemon because it's socket is full.
1236 * timedated will no longer write or process /etc/timezone,
1237 except on Debian. As we do not support late mounted /usr
1238 anymore /etc/localtime always being a symlink is now safe,
1239 and hence the information in /etc/timezone is not necessary
1242 * logind will now always reserve one VT for a text getty (VT6
1243 by default). Previously if more than 6 X sessions where
1244 started they took up all the VTs with auto-spawned gettys,
1245 so that no text gettys were available anymore.
1247 * udev will now automatically inform the btrfs kernel logic
1248 about btrfs RAID components showing up. This should make
1249 simple hotplug based btrfs RAID assembly work.
1251 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1252 (but not for its children which will stay at the kernel
1253 default). This should allow setups with a lot more listening
1256 * systemd will now always pass the configured timezone to the
1257 kernel at boot. timedated will do the same when the timezone
1260 * logind's inhibition logic has been updated. By default,
1261 logind will now handle the lid switch, the power and sleep
1262 keys all the time, even in graphical sessions. If DEs want
1263 to handle these events on their own they should take the new
1264 handle-power-key, handle-sleep-key and handle-lid-switch
1265 inhibitors during their runtime. A simple way to achiveve
1266 that is to invoke the DE wrapped in an invocation of:
1268 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1270 * Access to unit operations is now checked via SELinux taking
1271 the unit file label and client process label into account.
1273 * systemd will now notify the administrator in the journal
1274 when he over-mounts a non-empty directory.
1276 * There are new specifiers that are resolved in unit files,
1277 for the host name (%H), the machine ID (%m) and the boot ID
1280 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
1281 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
1282 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
1283 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
1284 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
1285 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
1286 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
1290 * Support for reading structured kernel messages from
1291 /dev/kmsg has now been added and is enabled by default.
1293 * Support for reading kernel messages from /proc/kmsg has now
1294 been removed. If you want kernel messages in the journal
1295 make sure to run a recent kernel (>= 3.5) that supports
1296 reading structured messages from /dev/kmsg (see
1297 above). /proc/kmsg is now exclusive property of classic
1298 syslog daemons again.
1300 * The libudev API gained the new
1301 udev_device_new_from_device_id() call.
1303 * The logic for file system namespace (ReadOnlyDirectory=,
1304 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
1305 require pivot_root() anymore. This means fewer temporary
1306 directories are created below /tmp for this feature.
1308 * nspawn containers will now see and receive all submounts
1309 made on the host OS below the root file system of the
1312 * Forward Secure Sealing is now supported for Journal files,
1313 which provide cryptographical sealing of journal files so
1314 that attackers cannot alter log history anymore without this
1315 being detectable. Lennart will soon post a blog story about
1316 this explaining it in more detail.
1318 * There are two new service settings RestartPreventExitStatus=
1319 and SuccessExitStatus= which allow configuration of exit
1320 status (exit code or signal) which will be excepted from the
1321 restart logic, resp. consider successful.
1323 * journalctl gained the new --verify switch that can be used
1324 to check the integrity of the structure of journal files and
1325 (if Forward Secure Sealing is enabled) the contents of
1328 * nspawn containers will now be run with /dev/stdin, /dev/fd/
1329 and similar symlinks pre-created. This makes running shells
1330 as container init process a lot more fun.
1332 * The fstab support can now handle PARTUUID= and PARTLABEL=
1335 * A new ConditionHost= condition has been added to match
1336 against the hostname (with globs) and machine ID. This is
1337 useful for clusters where a single OS image is used to
1338 provision a large number of hosts which shall run slightly
1339 different sets of services.
1341 * Services which hit the restart limit will now be placed in a
1344 Contributions from: Bertram Poettering, Dave Reisner, Huang
1345 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
1346 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
1350 * When running in --user mode systemd will now become a
1351 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
1352 tree a lot more organized.
1354 * A new PartOf= unit dependency type has been introduced that
1355 may be used to group services in a natural way.
1357 * "systemctl enable" may now be used to enable instances of
1360 * journalctl now prints error log levels in red, and
1361 warning/notice log levels in bright white. It also supports
1362 filtering by log level now.
1364 * cgtop gained a new -n switch (similar to top), to configure
1365 the maximum number of iterations to run for. It also gained
1366 -b, to run in batch mode (accepting no input).
1368 * The suffix ".service" may now be omitted on most systemctl
1369 command lines involving service unit names.
1371 * There's a new bus call in logind to lock all sessions, as
1372 well as a loginctl verb for it "lock-sessions".
1374 * libsystemd-logind.so gained a new call sd_journal_perror()
1375 that works similar to libc perror() but logs to the journal
1376 and encodes structured information about the error number.
1378 * /etc/crypttab entries now understand the new keyfile-size=
1381 * shutdown(8) now can send a (configurable) wall message when
1382 a shutdown is cancelled.
1384 * The mount propagation mode for the root file system will now
1385 default to "shared", which is useful to make containers work
1386 nicely out-of-the-box so that they receive new mounts from
1387 the host. This can be undone locally by running "mount
1388 --make-rprivate /" if needed.
1390 * The prefdm.service file has been removed. Distributions
1391 should maintain this unit downstream if they intend to keep
1392 it around. However, we recommend writing normal unit files
1393 for display managers instead.
1395 * Since systemd is a crucial part of the OS we will now
1396 default to a number of compiler switches that improve
1397 security (hardening) such as read-only relocations, stack
1398 protection, and suchlike.
1400 * The TimeoutSec= setting for services is now split into
1401 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1402 of individual time outs for the start and the stop phase of
1405 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1406 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1407 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1408 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1409 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1410 Gundersen, Zbigniew Jędrzejewski-Szmek
1414 * The journal and id128 C APIs are now fully documented as man
1417 * Extra safety checks have been added when transitioning from
1418 the initial RAM disk to the main system to avoid accidental
1421 * /etc/crypttab entries now understand the new keyfile-offset=
1424 * systemctl -t can now be used to filter by unit load state.
1426 * The journal C API gained the new sd_journal_wait() call to
1427 make writing synchronous journal clients easier.
1429 * journalctl gained the new -D switch to show journals from a
1432 * journalctl now displays a special marker between log
1433 messages of two different boots.
1435 * The journal is now explicitly flushed to /var via a service
1436 systemd-journal-flush.service, rather than implicitly simply
1437 by seeing /var/log/journal to be writable.
1439 * journalctl (and the journal C APIs) can now match for much
1440 more complex expressions, with alternatives and
1443 * When transitioning from the initial RAM disk to the main
1444 system we will now kill all processes in a killing spree to
1445 ensure no processes stay around by accident.
1447 * Three new specifiers may be used in unit files: %u, %h, %s
1448 resolve to the user name, user home directory resp. user
1449 shell. This is useful for running systemd user instances.
1451 * We now automatically rotate journal files if their data
1452 object hash table gets a fill level > 75%. We also size the
1453 hash table based on the configured maximum file size. This
1454 together should lower hash collisions drastically and thus
1455 speed things up a bit.
1457 * journalctl gained the new "--header" switch to introspect
1458 header data of journal files.
1460 * A new setting SystemCallFilters= has been added to services
1461 which may be used to apply blacklists or whitelists to
1462 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1464 * nspawn gained a new --link-journal= switch (and quicker: -j)
1465 to link the container journal with the host. This makes it
1466 very easy to centralize log viewing on the host for all
1467 guests while still keeping the journal files separated.
1469 * Many bugfixes and optimizations
1471 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1472 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1473 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1478 * Several tools now understand kernel command line arguments,
1479 which are only read when run in an initial RAM disk. They
1480 usually follow closely their normal counterparts, but are
1483 * There's a new tool to analyze the readahead files that are
1484 automatically generated at boot. Use:
1486 /usr/lib/systemd/systemd-readahead analyze /.readahead
1488 * We now provide an early debug shell on tty9 if this enabled. Use:
1490 systemctl enable debug-shell.service
1492 * All plymouth related units have been moved into the Plymouth
1493 package. Please make sure to upgrade your Plymouth version
1496 * systemd-tmpfiles now supports getting passed the basename of
1497 a configuration file only, in which case it will look for it
1498 in all appropriate directories automatically.
1500 * udevadm info now takes a /dev or /sys path as argument, and
1501 does the right thing. Example:
1503 udevadm info /dev/sda
1504 udevadm info /sys/class/block/sda
1506 * systemctl now prints a warning if a unit is stopped but a
1507 unit that might trigger it continues to run. Example: a
1508 service is stopped but the socket that activates it is left
1511 * "systemctl status" will now mention if the log output was
1512 shortened due to rotation since a service has been started.
1514 * The journal API now exposes functions to determine the
1515 "cutoff" times due to rotation.
1517 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1518 immediately flushing of runtime logs to /var if possible,
1519 resp. for triggering immediate rotation of the journal
1522 * It is now considered an error if a service is attempted to
1523 be stopped that is not loaded.
1525 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1527 * systemd-analyze now supports Python 3
1529 * tmpfiles now supports cleaning up directories via aging
1530 where the first level dirs are always kept around but
1531 directories beneath it automatically aged. This is enabled
1532 by prefixing the age field with '~'.
1534 * Seat objects now expose CanGraphical, CanTTY properties
1535 which is required to deal with very fast bootups where the
1536 display manager might be running before the graphics drivers
1537 completed initialization.
1539 * Seat objects now expose a State property.
1541 * We now include RPM macros for service enabling/disabling
1542 based on the preset logic. We recommend RPM based
1543 distributions to make use of these macros if possible. This
1544 makes it simpler to reuse RPM spec files across
1547 * We now make sure that the collected systemd unit name is
1548 always valid when services log to the journal via
1551 * There's a new man page kernel-command-line(7) detailing all
1552 command line options we understand.
1554 * The fstab generator may now be disabled at boot by passing
1555 fstab=0 on the kernel command line.
1557 * A new kernel command line option modules-load= is now understood
1558 to load a specific kernel module statically, early at boot.
1560 * Unit names specified on the systemctl command line are now
1561 automatically escaped as needed. Also, if file system or
1562 device paths are specified they are automatically turned
1563 into the appropriate mount or device unit names. Example:
1565 systemctl status /home
1566 systemctl status /dev/sda
1568 * The SysVConsole= configuration option has been removed from
1569 system.conf parsing.
1571 * The SysV search path is no longer exported on the D-Bus
1574 * The Names= option is been removed from unit file parsing.
1576 * There's a new man page bootup(7) detailing the boot process.
1578 * Every unit and every generator we ship with systemd now
1579 comes with full documentation. The self-explanatory boot is
1582 * A couple of services gained "systemd-" prefixes in their
1583 name if they wrap systemd code, rather than only external
1584 code. Among them fsck@.service which is now
1585 systemd-fsck@.service.
1587 * The HaveWatchdog property has been removed from the D-Bus
1590 * systemd.confirm_spawn= on the kernel command line should now
1593 * There's a new man page crypttab(5) which details all options
1594 we actually understand.
1596 * systemd-nspawn gained a new --capability= switch to pass
1597 additional capabilities to the container.
1599 * timedated will now read known NTP implementation unit names
1600 from /usr/lib/systemd/ntp-units.d/*.list,
1601 systemd-timedated-ntp.target has been removed.
1603 * journalctl gained a new switch "-b" that lists log data of
1604 the current boot only.
1606 * The notify socket is in the abstract namespace again, in
1607 order to support daemons which chroot() at start-up.
1609 * There is a new Storage= configuration option for journald
1610 which allows configuration of where log data should go. This
1611 also provides a way to disable journal logging entirely, so
1612 that data collected is only forwarded to the console, the
1613 kernel log buffer or another syslog implementation.
1615 * Many bugfixes and optimizations
1617 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1618 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1619 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1620 Shawn Landden, Tom Gundersen
1624 * "systemctl help <unit>" now shows the man page if one is
1627 * Several new man pages have been added.
1629 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1630 MaxLevelConsole= can now be specified in
1631 journald.conf. These options allow reducing the amount of
1632 data stored on disk or forwarded by the log level.
1634 * TimerSlackNSec= can now be specified in system.conf for
1635 PID1. This allows system-wide power savings.
1637 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1638 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1643 * logind is now capable of (optionally) handling power and
1644 sleep keys as well as the lid switch.
1646 * journalctl now understands the syntax "journalctl
1647 /usr/bin/avahi-daemon" to get all log output of a specific
1650 * CapabilityBoundingSet= in system.conf now also influences
1651 the capability bound set of usermode helpers of the kernel.
1653 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1654 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1655 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1656 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1660 * Note that we skipped 139 releases here in order to set the
1661 new version to something that is greater than both udev's
1662 and systemd's most recent version number.
1664 * udev: all udev sources are merged into the systemd source tree now.
1665 All future udev development will happen in the systemd tree. It
1666 is still fully supported to use the udev daemon and tools without
1667 systemd running, like in initramfs or other init systems. Building
1668 udev though, will require the *build* of the systemd tree, but
1669 udev can be properly *run* without systemd.
1671 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1672 should be used to create dead device nodes as workarounds for broken
1675 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1676 no longer supported. udev_monitor_new_from_netlink() needs to be
1677 used to subscribe to events.
1679 * udev: when udevd is started by systemd, processes which are left
1680 behind by forking them off of udev rules, are unconditionally cleaned
1681 up and killed now after the event handling has finished. Services or
1682 daemons must be started as systemd services. Services can be
1683 pulled-in by udev to get started, but they can no longer be directly
1684 forked by udev rules.
1686 * udev: the daemon binary is called systemd-udevd now and installed
1687 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1688 to adapt to that, create symlink, or rename the binary after building
1691 * libudev no longer provides these symbols:
1692 udev_monitor_from_socket()
1693 udev_queue_get_failed_list_entry()
1694 udev_get_{dev,sys,run}_path()
1695 The versions number was bumped and symbol versioning introduced.
1697 * systemd-loginctl and systemd-journalctl have been renamed
1698 to loginctl and journalctl to match systemctl.
1700 * The config files: /etc/systemd/systemd-logind.conf and
1701 /etc/systemd/systemd-journald.conf have been renamed to
1702 logind.conf and journald.conf. Package updates should rename
1703 the files to the new names on upgrade.
1705 * For almost all files the license is now LGPL2.1+, changed
1706 from the previous GPL2.0+. Exceptions are some minor stuff
1707 of udev (which will be changed to LGPL2.1 eventually, too),
1708 and the MIT licensed sd-daemon.[ch] library that is suitable
1709 to be used as drop-in files.
1711 * systemd and logind now handle system sleep states, in
1712 particular suspending and hibernating.
1714 * logind now implements a sleep/shutdown/idle inhibiting logic
1715 suitable for a variety of uses. Soonishly Lennart will blog
1716 about this in more detail.
1718 * var-run.mount and var-lock.mount are no longer provided
1719 (which prevously bind mounted these directories to their new
1720 places). Distributions which have not converted these
1721 directories to symlinks should consider stealing these files
1722 from git history and add them downstream.
1724 * We introduced the Documentation= field for units and added
1725 this to all our shipped units. This is useful to make it
1726 easier to explore the boot and the purpose of the various
1729 * All smaller setup units (such as
1730 systemd-vconsole-setup.service) now detect properly if they
1731 are run in a container and are skipped when
1732 appropriate. This guarantees an entirely noise-free boot in
1733 Linux container environments such as systemd-nspawn.
1735 * A framework for implementing offline system updates is now
1736 integrated, for details see:
1737 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1739 * A new service type Type=idle is available now which helps us
1740 avoiding ugly interleaving of getty output and boot status
1743 * There's now a system-wide CapabilityBoundingSet= option to
1744 globally reduce the set of capabilities for the
1745 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1746 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1747 even CAP_NET_ADMIN system-wide for secure systems.
1749 * There are now system-wide DefaultLimitXXX= options to
1750 globally change the defaults of the various resource limits
1751 for all units started by PID 1.
1753 * Harald Hoyer's systemd test suite has been integrated into
1754 systemd which allows easy testing of systemd builds in qemu
1755 and nspawn. (This is really awesome! Ask us for details!)
1757 * The fstab parser is now implemented as generator, not inside
1760 * systemctl will now warn you if .mount units generated from
1761 /etc/fstab are out of date due to changes in fstab that
1762 haven't been read by systemd yet.
1764 * systemd is now suitable for usage in initrds. Dracut has
1765 already been updated to make use of this. With this in place
1766 initrds get a slight bit faster but primarily are much
1767 easier to introspect and debug since "systemctl status" in
1768 the host system can be used to introspect initrd services,
1769 and the journal from the initrd is kept around too.
1771 * systemd-delta has been added, a tool to explore differences
1772 between user/admin configuration and vendor defaults.
1774 * PrivateTmp= now affects both /tmp and /var/tmp.
1776 * Boot time status messages are now much prettier and feature
1777 proper english language. Booting up systemd has never been
1780 * Read-ahead pack files now include the inode number of all
1781 files to pre-cache. When the inode changes the pre-caching
1782 is not attempted. This should be nicer to deal with updated
1783 packages which might result in changes of read-ahead
1786 * We now temporaritly lower the kernel's read_ahead_kb variable
1787 when collecting read-ahead data to ensure the kernel's
1788 built-in read-ahead does not add noise to our measurements
1789 of necessary blocks to pre-cache.
1791 * There's now RequiresMountsFor= to add automatic dependencies
1792 for all mounts necessary for a specific file system path.
1794 * MountAuto= and SwapAuto= have been removed from
1795 system.conf. Mounting file systems at boot has to take place
1798 * nspawn now learned a new switch --uuid= to set the machine
1799 ID on the command line.
1801 * nspawn now learned the -b switch to automatically search
1804 * vt102 is now the default TERM for serial TTYs, upgraded from
1807 * systemd-logind now works on VT-less systems.
1809 * The build tree has been reorganized. The individual
1810 components now have directories of their own.
1812 * A new condition type ConditionPathIsReadWrite= is now available.
1814 * nspawn learned the new -C switch to create cgroups for the
1815 container in other hierarchies.
1817 * We now have support for hardware watchdogs, configurable in
1820 * The scheduled shutdown logic now has a public API.
1822 * We now mount /tmp as tmpfs by default, but this can be
1823 masked and /etc/fstab can override it.
1825 * Since udisks doesn't make use of /media anymore we are not
1826 mounting a tmpfs on it anymore.
1828 * journalctl gained a new --local switch to only interleave
1829 locally generated journal files.
1831 * We can now load the IMA policy at boot automatically.
1833 * The GTK tools have been split off into a systemd-ui.
1835 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1836 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1837 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1838 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1839 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1840 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1841 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1842 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1843 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1848 * This is mostly a bugfix release
1850 * Support optional initialization of the machine ID from the
1851 KVM or container configured UUID.
1853 * Support immediate reboots with "systemctl reboot -ff"
1855 * Show /etc/os-release data in systemd-analyze output
1857 * Many bugfixes for the journal, including endianness fixes and
1858 ensuring that disk space enforcement works
1860 * sd-login.h is C++ comptaible again
1862 * Extend the /etc/os-release format on request of the Debian
1865 * We now refuse non-UTF8 strings used in various configuration
1866 and unit files. This is done to ensure we don't pass invalid
1867 data over D-Bus or expose it elsewhere.
1869 * Register Mimo USB Screens as suitable for automatic seat
1872 * Read SELinux client context from journal clients in a race
1875 * Reorder configuration file lookup order. /etc now always
1876 overrides /run in order to allow the administrator to always
1877 and unconditionally override vendor supplied or
1878 automatically generated data.
1880 * The various user visible bits of the journal now have man
1881 pages. We still lack man pages for the journal API calls
1884 * We now ship all man pages in HTML format again in the
1887 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1888 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1889 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1894 * This is mostly a bugfix release
1896 * systems lacking /etc/os-release are no longer supported.
1898 * Various functionality updates to libsystemd-login.so
1900 * Track class of PAM logins to distuingish greeters from
1903 Contributions from: Kay Sievers, Lennart Poettering, Michael
1908 * This is an important bugfix release for v41.
1910 * Building man pages is now optional which should be useful
1911 for those building systemd from git but unwilling to install
1914 * Watchdog support for supervising services is now usable. In
1915 a future release support for hardware watchdogs
1916 (i.e. /dev/watchdog) will be added building on this.
1918 * Service start rate limiting is now configurable and can be
1919 turned off per service. When a start rate limit is hit a
1920 reboot can automatically be triggered.
1922 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1924 Contributions from: Benjamin Franzke, Bill Nottingham,
1925 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1926 Schmidt, Michał Górny, Piotr Drąg
1930 * The systemd binary is installed /usr/lib/systemd/systemd now;
1931 An existing /sbin/init symlink needs to be adapted with the
1934 * The code that loads kernel modules has been ported to invoke
1935 libkmod directly, instead of modprobe. This means we do not
1936 support systems with module-init-tools anymore.
1938 * Watchdog support is now already useful, but still not
1941 * A new kernel command line option systemd.setenv= is
1942 understood to set system wide environment variables
1943 dynamically at boot.
1945 * We now limit the set of capabilities of systemd-journald.
1947 * We now set SIGPIPE to ignore by default, since it only is
1948 useful in shell pipelines, and has little use in general
1949 code. This can be disabled with IgnoreSIPIPE=no in unit
1952 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1953 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1958 * This is mostly a bugfix release
1960 * We now expose the reason why a service failed in the
1961 "Result" D-Bus property.
1963 * Rudimentary service watchdog support (will be completed over
1964 the next few releases.)
1966 * When systemd forks off in order execute some service we will
1967 now immediately changes its argv[0] to reflect which process
1968 it will execute. This is useful to minimize the time window
1969 with a generic argv[0], which makes bootcharts more useful
1971 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1972 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1973 Mike Kazantsev, Ray Strode
1977 * This is mostly a test release, but incorporates many
1980 * New systemd-cgtop tool to show control groups by their
1983 * Linking against libacl for ACLs is optional again. If
1984 disabled, support tracking device access for active logins
1985 goes becomes unavailable, and so does access to the user
1986 journals by the respective users.
1988 * If a group "adm" exists, journal files are automatically
1989 owned by them, thus allow members of this group full access
1990 to the system journal as well as all user journals.
1992 * The journal now stores the SELinux context of the logging
1993 client for all entries.
1995 * Add C++ inclusion guards to all public headers
1997 * New output mode "cat" in the journal to print only text
1998 messages, without any meta data like date or time.
2000 * Include tiny X server wrapper as a temporary stop-gap to
2001 teach XOrg udev display enumeration. This is used by display
2002 managers such as gdm, and will go away as soon as XOrg
2003 learned native udev hotplugging for display devices.
2005 * Add new systemd-cat tool for executing arbitrary programs
2006 with STDERR/STDOUT connected to the journal. Can also act as
2007 BSD logger replacement, and does so by default.
2009 * Optionally store all locally generated coredumps in the
2010 journal along with meta data.
2012 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
2013 writing short strings to files (for usage for /sys), and for
2014 creating symlinks, character and block device nodes.
2016 * New unit file option ControlGroupPersistent= to make cgroups
2017 persistent, following the mechanisms outlined in
2018 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
2020 * Support multiple local RTCs in a sane way
2022 * No longer monopolize IO when replaying readahead data on
2023 rotating disks, since we might starve non-file-system IO to
2024 death, since fanotify() will not see accesses done by blkid,
2027 * Don't show kernel threads in systemd-cgls anymore, unless
2028 requested with new -k switch.
2030 Contributions from: Dan Horák, Kay Sievers, Lennart
2031 Poettering, Michal Schmidt
2035 * This is mostly a test release, but incorporates many
2038 * The git repository moved to:
2039 git://anongit.freedesktop.org/systemd/systemd
2040 ssh://git.freedesktop.org/git/systemd/systemd
2042 * First release with the journal
2043 http://0pointer.de/blog/projects/the-journal.html
2045 * The journal replaces both systemd-kmsg-syslogd and
2046 systemd-stdout-bridge.
2048 * New sd_pid_get_unit() API call in libsystemd-logind
2050 * Many systemadm clean-ups
2052 * Introduce remote-fs-pre.target which is ordered before all
2053 remote mounts and may be used to start services before all
2056 * Added Mageia support
2058 * Add bash completion for systemd-loginctl
2060 * Actively monitor PID file creation for daemons which exit in
2061 the parent process before having finished writing the PID
2062 file in the daemon process. Daemons which do this need to be
2063 fixed (i.e. PID file creation must have finished before the
2064 parent exits), but we now react a bit more gracefully to them.
2066 * Add colourful boot output, mimicking the well-known output
2067 of existing distributions.
2069 * New option PassCredentials= for socket units, for
2070 compatibility with a recent kernel ABI breakage.
2072 * /etc/rc.local is now hooked in via a generator binary, and
2073 thus will no longer act as synchronization point during
2076 * systemctl list-unit-files now supports --root=.
2078 * systemd-tmpfiles now understands two new commands: z, Z for
2079 relabelling files according to the SELinux database. This is
2080 useful to apply SELinux labels to specific files in /sys,
2083 * Output of SysV services is now forwarded to both the console
2084 and the journal by default, not only just the console.
2086 * New man pages for all APIs from libsystemd-login.
2088 * The build tree got reorganized and a the build system is a
2089 lot more modular allowing embedded setups to specifically
2090 select the components of systemd they are interested in.
2092 * Support for Linux systems lacking the kernel VT subsystem is
2095 * configure's --with-rootdir= got renamed to
2096 --with-rootprefix= to follow the naming used by udev and
2099 * Unless specified otherwise we'll now install to /usr instead
2100 of /usr/local by default.
2102 * Processes with '@' in argv[0][0] are now excluded from the
2103 final shut-down killing spree, following the logic explained
2105 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
2107 * All processes remaining in a service cgroup when we enter
2108 the START or START_PRE states are now killed with
2109 SIGKILL. That means it is no longer possible to spawn
2110 background processes from ExecStart= lines (which was never
2111 supported anyway, and bad style).
2113 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
2114 reloading of units together.
2116 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
2117 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
2118 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
2119 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
2120 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek