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
87 * systemd-delta can now show information about drop-in
88 snippets extending unit files.
90 * libsystemd-bus has been substantially updated but is still
91 not available as public API.
93 * systemd will now look for the "debug" argument on the kernel
94 command line and enable debug logging, similar to
95 "systemd.log_level=debug" already did before.
97 * "systemctl set-default", "systemctl get-default" has been
98 added to configure the default.target symlink, which
99 controls what to boot into by default.
101 * "systemd-analyze plot" will now show the time the various
102 generators needed for execution, as well as information
103 about the unit file loading.
105 * journalctl gained new "--user" and "--system" switches to
106 only show user/system logs.
108 * libsystemd-journal gained a new sd_journal_open_files() call
109 for opening specific journal files. journactl also gained a
110 new switch to expose this new functionality. Previously we
111 only supported opening all files from a directory, or all
112 files from the system, as opening individual files only is
113 racy due to journal file rotation.
115 * systemd gained the new DefaultEnvironment= setting in
116 /etc/systemd/system.conf to set environment variables for
119 * If a privileged process logs a journal message with the
120 OBJECT_PID= field set, then journald will automatically
121 augment this with additional OBJECT_UID=, OBJECT_GID=,
122 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
123 system services want to log events about specific client
124 processes. journactl/systemctl has been updated to make use
125 of this information if all log messages regarding a specific
128 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
129 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
130 Reisner, David Coppa, David King, David Strauss, Eelco
131 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
132 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
133 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
134 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
135 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
136 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
137 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
138 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
139 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
144 * The Python bindings gained some minimal support for the APIs
145 exposed by libsystemd-logind.
147 * ConditionSecurity= gained support for detecting SMACK. Since
148 this condition already supports SELinux and AppArmor we only
149 miss IMA for this. Patches welcome!
151 Contributions from: Karol Lewandowski, Lennart Poettering,
152 Zbigniew Jędrzejewski-Szmek
156 * systemd-nspawn will now create /etc/resolv.conf if
157 necessary, before bind-mounting the host's file onto it.
159 * systemd-nspawn will now store meta information about a
160 container on the container's cgroup as extended attribute
161 fields, including the root directory.
163 * The cgroup hierarchy has been reworked in many ways. All
164 objects any of the components systemd creates in the cgroup
165 tree are now suffixed. More specifically, user sessions are
166 now placed in cgroups suffixed with ".session", users in
167 cgroups suffixed with ".user", and nspawn containers in
168 cgroups suffixed with ".nspawn". Furthermore, all cgroup
169 names are now escaped in a simple scheme to avoid collision
170 of userspace object names with kernel filenames. This work
171 is preparation for making these objects relocatable in the
172 cgroup tree, in order to allow easy resource partitioning of
173 these objects without causing naming conflicts.
175 * systemctl list-dependencies gained the new switches
176 --plain, --reverse, --after and --before.
178 * systemd-inhibit now shows the process name of processes that
179 have taken an inhibitor lock.
181 * nss-myhostname will now also resolve "localhost"
182 implicitly. This makes /etc/hosts an optional file and
183 nicely handles that on IPv6 ::1 maps to both "localhost" and
186 * libsystemd-logind.so gained a new call
187 sd_get_machine_names() to enumerate running containers and
188 VMs (currently only supported by very new libvirt and
189 nspawn). sd_login_monitor can now be used to watch
190 VMs/containers coming and going.
192 * .include is not allowed recursively anymore, and only in
193 unit files. Usually it is better to use drop-in snippets in
194 .d/*.conf anyway, as introduced with systemd 198.
196 * systemd-analyze gained a new "critical-chain" command that
197 determines the slowest chain of units run during system
198 boot-up. It is very useful for tracking down where
199 optimizing boot time is the most beneficial.
201 * systemd will no longer allow manipulating service paths in
202 the name=systemd:/system cgroup tree using ControlGroup= in
203 units. (But is still fine with it in all other dirs.)
205 * There's a new systemd-nspawn@.service service file that may
206 be used to easily run nspawn containers as system
207 services. With the container's root directory in
208 /var/lib/container/foobar it is now sufficient to run
209 "systemctl start systemd-nspawn@foobar.service" to boot it.
211 * systemd-cgls gained a new parameter "--machine" to list only
212 the processes within a certain container.
214 * ConditionSecurity= now can check for "apparmor". We still
215 are lacking checks for SMACK and IMA for this condition
216 check though. Patches welcome!
218 * A new configuration file /etc/systemd/sleep.conf has been
219 added that may be used to configure which kernel operation
220 systemd is supposed to execute when "suspend", "hibernate"
221 or "hybrid-sleep" is requested. This makes the new kernel
222 "freeze" state accessible to the user.
224 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
225 the passed argument if applicable.
227 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
228 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
229 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
230 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
231 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
232 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
233 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
238 * The output of 'systemctl list-jobs' got some polishing. The
239 '--type=' argument may now be passed more than once. A new
240 command 'systemctl list-sockets' has been added which shows
241 a list of kernel sockets systemd is listening on with the
242 socket units they belong to, plus the units these socket
245 * The experimental libsystemd-bus library got substantial
246 updates to work in conjunction with the (also experimental)
247 kdbus kernel project. It works well enough to exchange
248 messages with some sophistication. Note that kdbus is not
249 ready yet, and the library is mostly an elaborate test case
250 for now, and not installable.
252 * systemd gained a new unit 'systemd-static-nodes.service'
253 that generates static device nodes earlier during boot, and
254 can run in conjunction with udev.
256 * libsystemd-login gained a new call sd_pid_get_user_unit()
257 to retrieve the user systemd unit a process is running
258 in. This is useful for systems where systemd is used as
261 * systemd-nspawn now places all containers in the new /machine
262 top-level cgroup directory in the name=systemd
263 hierarchy. libvirt will soon do the same, so that we get a
264 uniform separation of /system, /user and /machine for system
265 services, user processes and containers/virtual
266 machines. This new cgroup hierarchy is also useful to stick
267 stable names to specific container instances, which can be
268 recognized later this way (this name may be controlled
269 via systemd-nspawn's new -M switch). libsystemd-login also
270 gained a new call sd_pid_get_machine_name() to retrieve the
271 name of the container/VM a specific process belongs to.
273 * bootchart can now store its data in the journal.
275 * libsystemd-journal gained a new call
276 sd_journal_add_conjunction() for AND expressions to the
277 matching logic. This can be used to express more complex
280 * journactl can now take multiple --unit= and --user-unit=
283 * The cryptsetup logic now understands the "luks.key=" kernel
284 command line switch for specifying a file to read the
285 decryption key from. Also, if a configured key file is not
286 found the tool will now automatically fall back to prompting
289 * Python systemd.journal module was updated to wrap recently
290 added functions from libsystemd-journal. The interface was
291 changed to bring the low level interface in s.j._Reader
292 closer to the C API, and the high level interface in
293 s.j.Reader was updated to wrap and convert all data about
296 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
297 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
298 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
299 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
300 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
301 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
305 * journalctl --update-catalog now understands a new --root=
306 option to operate on catalogs found in a different root
309 * During shutdown after systemd has terminated all running
310 services a final killing loop kills all remaining left-over
311 processes. We will now print the name of these processes
312 when we send SIGKILL to them, since this usually indicates a
315 * If /etc/crypttab refers to password files stored on
316 configured mount points automatic dependencies will now be
317 generated to ensure the specific mount is established first
318 before the key file is attempted to be read.
320 * 'systemctl status' will now show information about the
321 network sockets a socket unit is listening on.
323 * 'systemctl status' will also shown information about any
324 drop-in configuration file for units. (Drop-In configuration
325 files in this context are files such as
326 /etc/systemd/systemd/foobar.service.d/*.conf)
328 * systemd-cgtop now optionally shows summed up CPU times of
329 cgroups. Press '%' while running cgtop to switch between
330 percentage and absolute mode. This is useful to determine
331 which cgroups use up the most CPU time over the entire
332 runtime of the system. systemd-cgtop has also been updated
333 to be 'pipeable' for processing with further shell tools.
335 * 'hostnamectl set-hostname' will now allow setting of FQDN
338 * The formatting and parsing of time span values has been
339 changed. The parser now understands fractional expressions
340 such as "5.5h". The formatter will now output fractional
341 expressions for all time spans under 1min, i.e. "5.123456s"
342 rather than "5s 123ms 456us". For time spans under 1s
343 millisecond values are shown, for those under 1ms
344 microsecond values are shown. This should greatly improve
345 all time-related output of systemd.
347 * libsystemd-login and libsystemd-journal gained new
348 functions for querying the poll() events mask and poll()
349 timeout value for integration into arbitrary event
352 * localectl gained the ability to list available X11 keymaps
353 (models, layouts, variants, options).
355 * 'systemd-analyze dot' gained the ability to filter for
356 specific units via shell-style globs, to create smaller,
357 more useful graphs. I.e. it's now possible to create simple
358 graphs of all the dependencies between only target units, or
359 of all units that Avahi has dependencies with.
361 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
362 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
363 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
364 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
365 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
366 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
367 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
371 * The boot-time readahead implementation for rotating media
372 will now read the read-ahead data in multiple passes which
373 consist of all read requests made in equidistant time
374 intervals. This means instead of strictly reading read-ahead
375 data in its physical order on disk we now try to find a
376 middle ground between physical and access time order.
378 * /etc/os-release files gained a new BUILD_ID= field for usage
379 on operating systems that provide continuous builds of OS
382 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
383 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
384 William Douglas, Zbigniew Jędrzejewski-Szmek
388 * systemd-python gained an API exposing libsystemd-daemon.
390 * The SMACK setup logic gained support for uploading CIPSO
393 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
394 ReadOnlyDirectories= and InaccessibleDirectories= has
395 changed. The private /tmp and /var/tmp directories are now
396 shared by all processes of a service (which means
397 ExecStartPre= may now leave data in /tmp that ExecStart= of
398 the same service can still access). When a service is
399 stopped its temporary directories are immediately deleted
400 (normal clean-up with tmpfiles is still done in addition to
403 * By default, systemd will now set a couple of sysctl
404 variables in the kernel: the safe sysrq options are turned
405 on, IP route verification is turned on, and source routing
406 disabled. The recently added hardlink and softlink
407 protection of the kernel is turned on. These settings should
408 be reasonably safe, and good defaults for all new systems.
410 * The predictable network naming logic may now be turned off
411 with a new kernel command line switch: net.ifnames=0.
413 * A new libsystemd-bus module has been added that implements a
414 pretty complete D-Bus client library. For details see:
416 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
418 * journald will now explicitly flush the journal files to disk
419 at the latest 5min after each write. The file will then also
420 be marked offline until the next write. This should increase
421 reliability in case of a crash. The synchronization delay
422 can be configured via SyncIntervalSec= in journald.conf.
424 * There's a new remote-fs-setup.target unit that can be used
425 to pull in specific services when at least one remote file
426 system is to be mounted.
428 * There are new targets timers.target and paths.target as
429 canonical targets to pull user timer and path units in
430 from. This complements sockets.target with a similar
431 purpose for socket units.
433 * libudev gained a new call udev_device_set_attribute_value()
434 to set sysfs attributes of a device.
436 * The udev daemon now sets the default number of worker
437 processes executed in parallel based on the number of available
438 CPUs instead of the amount of available RAM. This is supposed
439 to provide a more reliable default and limit a too aggressive
440 paralellism for setups with 1000s of devices connected.
442 Contributions from: Auke Kok, Colin Walters, Cristian
443 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
444 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
445 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
446 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
447 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
448 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
449 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
450 Zbigniew Jędrzejewski-Szmek
454 * Configuration of unit files may now be extended via drop-in
455 files without having to edit/override the unit files
456 themselves. More specifically, if the administrator wants to
457 change one value for a service file foobar.service he can
458 now do so by dropping in a configuration snippet into
459 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
460 will load all these snippets and apply them on top of the
461 main unit configuration file, possibly extending or
462 overriding its settings. Using these drop-in snippets is
463 generally nicer than the two earlier options for changing
464 unit files locally: copying the files from
465 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
466 them there; or creating a new file in /etc/systemd/system/
467 that incorporates the original one via ".include". Drop-in
468 snippets into these .d/ directories can be placed in any
469 directory systemd looks for units in, and the usual
470 overriding semantics between /usr/lib, /etc and /run apply
473 * Most unit file settings which take lists of items can now be
474 reset by assigning the empty string to them. For example,
475 normally, settings such as Environment=FOO=BAR append a new
476 environment variable assignment to the environment block,
477 each time they are used. By assigning Environment= the empty
478 string the environment block can be reset to empty. This is
479 particularly useful with the .d/*.conf drop-in snippets
480 mentioned above, since this adds the ability to reset list
481 settings from vendor unit files via these drop-ins.
483 * systemctl gained a new "list-dependencies" command for
484 listing the dependencies of a unit recursively.
486 * Inhibitors are now honored and listed by "systemctl
487 suspend", "systemctl poweroff" (and similar) too, not only
488 GNOME. These commands will also list active sessions by
491 * Resource limits (as exposed by the various control group
492 controllers) can now be controlled dynamically at runtime
493 for all units. More specifically, you can now use a command
494 like "systemctl set-cgroup-attr foobar.service cpu.shares
495 2000" to alter the CPU shares a specific service gets. These
496 settings are stored persistently on disk, and thus allow the
497 administrator to easily adjust the resource usage of
498 services with a few simple commands. This dynamic resource
499 management logic is also available to other programs via the
500 bus. Almost any kernel cgroup attribute and controller is
503 * systemd-vconsole-setup will now copy all font settings to
504 all allocated VTs, where it previously applied them only to
507 * libsystemd-login gained the new sd_session_get_tty() API
510 * This release drops support for a few legacy or
511 distribution-specific LSB facility names when parsing init
512 scripts: $x-display-manager, $mail-transfer-agent,
513 $mail-transport-agent, $mail-transfer-agent, $smtp,
514 $null. Also, the mail-transfer-agent.target unit backing
515 this has been removed. Distributions which want to retain
516 compatibility with this should carry the burden for
517 supporting this themselves and patch support for these back
518 in, if they really need to. Also, the facilities $syslog and
519 $local_fs are now ignored, since systemd does not support
520 early-boot LSB init scripts anymore, and these facilities
521 are implied anyway for normal services. syslog.target has
524 * There are new bus calls on PID1's Manager object for
525 cancelling jobs, and removing snapshot units. Previously,
526 both calls were only available on the Job and Snapshot
529 * systemd-journal-gatewayd gained SSL support.
531 * The various "environment" files, such as /etc/locale.conf
532 now support continuation lines with a backslash ("\") as
533 last character in the line, similar in style (but different)
534 to how this is supported in shells.
536 * For normal user processes the _SYSTEMD_USER_UNIT= field is
537 now implicitly appended to every log entry logged. systemctl
538 has been updated to filter by this field when operating on a
539 user systemd instance.
541 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
542 CAP_AUDIT_CONTROL capabilities to the capabilities set for
543 the container. This makes it easier to boot unmodified
544 Fedora systems in a container, which however still requires
545 audit=0 to be passed on the kernel command line. Auditing in
546 kernel and userspace is unfortunately still too broken in
547 context of containers, hence we recommend compiling it out
548 of the kernel or using audit=0. Hopefully this will be fixed
549 one day for good in the kernel.
551 * nspawn gained the new --bind= and --bind-ro= parameters to
552 bind mount specific directories from the host into the
555 * nspawn will now mount its own devpts file system instance
556 into the container, in order not to leak pty devices from
557 the host into the container.
559 * systemd will now read the firmware boot time performance
560 information from the EFI variables, if the used boot loader
561 supports this, and takes it into account for boot performance
562 analysis via "systemd-analyze". This is currently supported
563 only in conjunction with Gummiboot, but could be supported
564 by other boot loaders too. For details see:
566 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
568 * A new generator has been added that automatically mounts the
569 EFI System Partition (ESP) to /boot, if that directory
570 exists, is empty, and no other file system has been
571 configured to be mounted there.
573 * logind will now send out PrepareForSleep(false) out
574 unconditionally, after coming back from suspend. This may be
575 used by applications as asynchronous notification for
576 system resume events.
578 * "systemctl unlock-sessions" has been added, that allows
579 unlocking the screens of all user sessions at once, similar
580 how "systemctl lock-sessions" already locked all users
581 sessions. This is backed by a new D-Bus call UnlockSessions().
583 * "loginctl seat-status" will now show the master device of a
584 seat. (i.e. the device of a seat that needs to be around for
585 the seat to be considered available, usually the graphics
588 * tmpfiles gained a new "X" line type, that allows
589 configuration of files and directories (with wildcards) that
590 shall be excluded from automatic cleanup ("aging").
592 * udev default rules set the device node permissions now only
593 at "add" events, and do not change them any longer with a
594 later "change" event.
596 * The log messages for lid events and power/sleep keypresses
597 now carry a message ID.
599 * We now have a substantially larger unit test suite, but this
600 continues to be work in progress.
602 * udevadm hwdb gained a new --root= parameter to change the
603 root directory to operate relative to.
605 * logind will now issue a background sync() request to the kernel
606 early at shutdown, so that dirty buffers are flushed to disk early
607 instead of at the last moment, in order to optimize shutdown
610 * A new bootctl tool has been added that is an interface for
611 certain boot loader operations. This is currently a preview
612 and is likely to be extended into a small mechanism daemon
613 like timedated, localed, hostnamed, and can be used by
614 graphical UIs to enumerate available boot options, and
615 request boot into firmware operations.
617 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
618 the rest of the package. It also has been updated to work
619 correctly in initrds.
621 * Policykit previously has been runtime optional, and is now
622 also compile time optional via a configure switch.
624 * systemd-analyze has been reimplemented in C. Also "systemctl
625 dot" has moved into systemd-analyze.
627 * "systemctl status" with no further parameters will now print
628 the status of all active or failed units.
630 * Operations such as "systemctl start" can now be executed
631 with a new mode "--irreversible" which may be used to queue
632 operations that cannot accidentally be reversed by a later
633 job queuing. This is by default used to make shutdown
634 requests more robust.
636 * The Python API of systemd now gained a new module for
637 reading journal files.
639 * A new tool kernel-install has been added that can install
640 kernel images according to the Boot Loader Specification:
642 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
644 * Boot time console output has been improved to provide
645 animated boot time output for hanging jobs.
647 * A new tool systemd-activate has been added which can be used
648 to test socket activation with, directly from the command
649 line. This should make it much easier to test and debug
650 socket activation in daemons.
652 * journalctl gained a new "--reverse" (or -r) option to show
653 journal output in reverse order (i.e. newest line first).
655 * journalctl gained a new "--pager-end" (or -e) option to jump
656 to immediately jump to the end of the journal in the
657 pager. This is only supported in conjunction with "less".
659 * journalctl gained a new "--user-unit=" option, that works
660 similar to "--unit=" but filters for user units rather than
663 * A number of unit files to ease adoption of systemd in
664 initrds has been added. This moves some minimal logic from
665 the various initrd implementations into systemd proper.
667 * The journal files are now owned by a new group
668 "systemd-journal", which exists specifically to allow access
669 to the journal, and nothing else. Previously, we used the
670 "adm" group for that, which however possibly covers more
671 than just journal/log file access. This new group is now
672 already used by systemd-journal-gatewayd to ensure this
673 daemon gets access to the journal files and as little else
674 as possible. Note that "make install" will also set FS ACLs
675 up for /var/log/journal to give "adm" and "wheel" read
676 access to it, in addition to "systemd-journal" which owns
677 the journal files. We recommend that packaging scripts also
678 add read access to "adm" + "wheel" to /var/log/journal, and
679 all existing/future journal files. To normal users and
680 administrators little changes, however packagers need to
681 ensure to create the "systemd-journal" system group at
682 package installation time.
684 * The systemd-journal-gatewayd now runs as unprivileged user
685 systemd-journal-gateway:systemd-journal-gateway. Packaging
686 scripts need to create these system user/group at
689 * timedated now exposes a new boolean property CanNTP that
690 indicates whether a local NTP service is available or not.
692 * systemd-detect-virt will now also detect xen PVs
694 * The pstore file system is now mounted by default, if it is
697 * In addition to the SELinux and IMA policies we will now also
698 load SMACK policies at early boot.
700 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
701 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
702 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
703 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
704 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
705 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
706 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
707 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
708 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
709 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
710 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
711 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
712 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
713 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
717 * Timer units now support calendar time events in addition to
718 monotonic time events. That means you can now trigger a unit
719 based on a calendar time specification such as "Thu,Fri
720 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
721 or fifth day of any month of the year 2013, given that it is
722 a thursday or friday. This brings timer event support
723 considerably closer to cron's capabilities. For details on
724 the supported calendar time specification language see
727 * udev now supports a number of different naming policies for
728 network interfaces for predictable names, and a combination
729 of these policies is now the default. Please see this wiki
730 document for details:
732 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
734 * Auke Kok's bootchart implementation has been added to the
735 systemd tree. It's an optional component that can graph the
736 boot in quite some detail. It's one of the best bootchart
737 implementations around and minimal in its code and
740 * nss-myhostname has been integrated into the systemd source
741 tree. nss-myhostname guarantees that the local hostname
742 always stays resolvable via NSS. It has been a weak
743 requirement of systemd-hostnamed since a long time, and
744 since its code is actually trivial we decided to just
745 include it in systemd's source tree. It can be turned off
746 with a configure switch.
748 * The read-ahead logic is now capable of properly detecting
749 whether a btrfs file system is on SSD or rotating media, in
750 order to optimize the read-ahead scheme. Previously, it was
751 only capable of detecting this on traditional file systems
754 * In udev, additional device properties are now read from the
755 IAB in addition to the OUI database. Also, Bluetooth company
756 identities are attached to the devices as well.
758 * In service files %U may be used as specifier that is
759 replaced by the configured user name of the service.
761 * nspawn may now be invoked without a controlling TTY. This
762 makes it suitable for invocation as its own service. This
763 may be used to set up a simple containerized server system
764 using only core OS tools.
766 * systemd and nspawn can now accept socket file descriptors
767 when they are started for socket activation. This enables
768 implementation of socket activated nspawn
769 containers. i.e. think about autospawning an entire OS image
770 when the first SSH or HTTP connection is received. We expect
771 that similar functionality will also be added to libvirt-lxc
774 * journalctl will now suppress ANSI color codes when
777 * systemctl will no longer show control group information for
778 a unit if a the control group is empty anyway.
780 * logind can now automatically suspend/hibernate/shutdown the
783 * /etc/machine-info and hostnamed now also expose the chassis
784 type of the system. This can be used to determine whether
785 the local system is a laptop, desktop, handset or
786 tablet. This information may either be configured by the
787 user/vendor or is automatically determined from ACPI and DMI
788 information if possible.
790 * A number of PolicyKit actions are now bound together with
791 "imply" rules. This should simplify creating UIs because
792 many actions will now authenticate similar ones as well.
794 * Unit files learnt a new condition ConditionACPower= which
795 may be used to conditionalize a unit depending on whether an
796 AC power source is connected or not, of whether the system
797 is running on battery power.
799 * systemctl gained a new "is-failed" verb that may be used in
800 shell scripts and suchlike to check whether a specific unit
801 is in the "failed" state.
803 * The EnvironmentFile= setting in unit files now supports file
804 globbing, and can hence be used to easily read a number of
805 environment files at once.
807 * systemd will no longer detect and recognize specific
808 distributions. All distribution-specific #ifdeffery has been
809 removed, systemd is now fully generic and
810 distribution-agnostic. Effectively, not too much is lost as
811 a lot of the code is still accessible via explicit configure
812 switches. However, support for some distribution specific
813 legacy configuration file formats has been dropped. We
814 recommend distributions to simply adopt the configuration
815 files everybody else uses now and convert the old
816 configuration from packaging scripts. Most distributions
817 already did that. If that's not possible or desirable,
818 distributions are welcome to forward port the specific
819 pieces of code locally from the git history.
821 * When logging a message about a unit systemd will now always
822 log the unit name in the message meta data.
824 * localectl will now also discover system locale data that is
825 not stored in locale archives, but directly unpacked.
827 * logind will no longer unconditionally use framebuffer
828 devices as seat masters, i.e. as devices that are required
829 to be existing before a seat is considered preset. Instead,
830 it will now look for all devices that are tagged as
831 "seat-master" in udev. By default framebuffer devices will
832 be marked as such, but depending on local systems other
833 devices might be marked as well. This may be used to
834 integrate graphics cards using closed source drivers (such
835 as NVidia ones) more nicely into logind. Note however, that
836 we recommend using the open source NVidia drivers instead,
837 and no udev rules for the closed-source drivers will be
838 shipped from us upstream.
840 Contributions from: Adam Williamson, Alessandro Crismani, Auke
841 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
842 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
843 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
844 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
845 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
846 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
847 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
848 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
849 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
850 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
851 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
856 * udev gained support for loading additional device properties
857 from an indexed database that is keyed by vendor/product IDs
858 and similar device identifiers. For the beginning this
859 "hwdb" is populated with data from the well-known PCI and
860 USB database, but also includes PNP, ACPI and OID data. In
861 the longer run this indexed database shall grow into
862 becoming the one central database for non-essential
863 userspace device metadata. Previously, data from the PCI/USB
864 database was only attached to select devices, since the
865 lookup was a relatively expensive operation due to O(n) time
866 complexity (with n being the number of entries in the
867 database). Since this is now O(1), we decided to add in this
868 data for all devices where this is available, by
869 default. Note that the indexed database needs to be rebuilt
870 when new data files are installed. To achieve this you need
871 to update your packaging scripts to invoke "udevadm hwdb
872 --update" after installation of hwdb data files. For
873 RPM-based distributions we introduced the new
874 %udev_hwdb_update macro for this purpose.
876 * The Journal gained support for the "Message Catalog", an
877 indexed database to link up additional information with
878 journal entries. For further details please check:
880 http://www.freedesktop.org/wiki/Software/systemd/catalog
882 The indexed message catalog database also needs to be
883 rebuilt after installation of message catalog files. Use
884 "journalctl --update-catalog" for this. For RPM-based
885 distributions we introduced the %journal_catalog_update
886 macro for this purpose.
888 * The Python Journal bindings gained support for the standard
889 Python logging framework.
891 * The Journal API gained new functions for checking whether
892 the underlying file system of a journal file is capable of
893 properly reporting file change notifications, or whether
894 applications that want to reflect journal changes "live"
895 need to recheck journal files continuously in appropriate
898 * It is now possible to set the "age" field for tmpfiles
899 entries to 0, indicating that files matching this entry
900 shall always be removed when the directories are cleaned up.
902 * coredumpctl gained a new "gdb" verb which invokes gdb
903 right-away on the selected coredump.
905 * There's now support for "hybrid sleep" on kernels that
906 support this, in addition to "suspend" and "hibernate". Use
907 "systemctl hybrid-sleep" to make use of this.
909 * logind's HandleSuspendKey= setting (and related settings)
910 now gained support for a new "lock" setting to simply
911 request the screen lock on all local sessions, instead of
912 actually executing a suspend or hibernation.
914 * systemd will now mount the EFI variables file system by
917 * Socket units now gained support for configuration of the
918 SMACK security label.
920 * timedatectl will now output the time of the last and next
921 daylight saving change.
923 * We dropped support for various legacy and distro-specific
924 concepts, such as insserv, early-boot SysV services
925 (i.e. those for non-standard runlevels such as 'b' or 'S')
926 or ArchLinux /etc/rc.conf support. We recommend the
927 distributions who still need support this to either continue
928 to maintain the necessary patches downstream, or find a
929 different solution. (Talk to us if you have questions!)
931 * Various systemd components will now bypass PolicyKit checks
932 for root and otherwise handle properly if PolicyKit is not
933 found to be around. This should fix most issues for
934 PolicyKit-less systems. Quite frankly this should have been
935 this way since day one. It is absolutely our intention to
936 make systemd work fine on PolicyKit-less systems, and we
937 consider it a bug if something doesn't work as it should if
938 PolicyKit is not around.
940 * For embedded systems it is now possible to build udev and
941 systemd without blkid and/or kmod support.
943 * "systemctl switch-root" is now capable of switching root
944 more than once. I.e. in addition to transitions from the
945 initrd to the host OS it is now possible to transition to
946 further OS images from the host. This is useful to implement
947 offline updating tools.
949 * Various other additions have been made to the RPM macros
950 shipped with systemd. Use %udev_rules_update() after
951 installing new udev rules files. %_udevhwdbdir,
952 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
953 %_sysctldir are now available which resolve to the right
954 directories for packages to place various data files in.
956 * journalctl gained the new --full switch (in addition to
957 --all, to disable ellipsation for long messages.
959 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
960 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
961 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
962 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
963 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
964 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
965 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
966 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
967 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
971 * journalctl gained new --since= and --until= switches to
972 filter by time. It also now supports nice filtering for
973 units via --unit=/-u.
975 * Type=oneshot services may use ExecReload= and do the
978 * The journal daemon now supports time-based rotation and
979 vacuuming, in addition to the usual disk-space based
982 * The journal will now index the available field values for
983 each field name. This enables clients to show pretty drop
984 downs of available match values when filtering. The bash
985 completion of journalctl has been updated
986 accordingly. journalctl gained a new switch -F to list all
987 values a certain field takes in the journal database.
989 * More service events are now written as structured messages
990 to the journal, and made recognizable via message IDs.
992 * The timedated, localed and hostnamed mini-services which
993 previously only provided support for changing time, locale
994 and hostname settings from graphical DEs such as GNOME now
995 also have a minimal (but very useful) text-based client
996 utility each. This is probably the nicest way to changing
997 these settings from the command line now, especially since
998 it lists available options and is fully integrated with bash
1001 * There's now a new tool "systemd-coredumpctl" to list and
1002 extract coredumps from the journal.
1004 * We now install a README each in /var/log/ and
1005 /etc/rc.d/init.d explaining where the system logs and init
1006 scripts went. This hopefully should help folks who go to
1007 that dirs and look into the otherwise now empty void and
1008 scratch their heads.
1010 * When user-services are invoked (by systemd --user) the
1011 $MANAGERPID env var is set to the PID of systemd.
1013 * SIGRTMIN+24 when sent to a --user instance will now result
1014 in immediate termination of systemd.
1016 * gatewayd received numerous feature additions such as a
1017 "follow" mode, for live syncing and filtering.
1019 * browse.html now allows filtering and showing detailed
1020 information on specific entries. Keyboard navigation and
1021 mouse screen support has been added.
1023 * gatewayd/journalctl now supports HTML5/JSON
1024 Server-Sent-Events as output.
1026 * The SysV init script compatibility logic will now
1027 heuristically determine whether a script supports the
1028 "reload" verb, and only then make this available as
1031 * "systemctl status --follow" has been removed, use "journalctl
1034 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
1035 have been removed since they are hardly useful to be
1038 * And I'd like to take the opportunity to specifically mention
1039 Zbigniew for his great contributions. Zbigniew, you rock!
1041 Contributions from: Andrew Eikum, Christian Hesse, Colin
1042 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
1043 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
1044 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
1045 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
1046 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
1047 Jędrzejewski-Szmek, Сковорода Никита Андреевич
1051 * If /etc/vconsole.conf is non-existent or empty we will no
1052 longer load any console font or key map at boot by
1053 default. Instead the kernel defaults will be left
1054 intact. This is definitely the right thing to do, as no
1055 configuration should mean no configuration, and hard-coding
1056 font names that are different on all archs is probably a bad
1057 idea. Also, the kernel default key map and font should be
1058 good enough for most cases anyway, and mostly identical to
1059 the userspace fonts/key maps we previously overloaded them
1060 with. If distributions want to continue to default to a
1061 non-kernel font or key map they should ship a default
1062 /etc/vconsole.conf with the appropriate contents.
1064 Contributions from: Colin Walters, Daniel J Walsh, Dave
1065 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
1066 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1070 * journalctl gained a new --cursor= switch to show entries
1071 starting from the specified location in the journal.
1073 * We now enforce a size limit on journal entry fields exported
1074 with "-o json" in journalctl. Fields larger than 4K will be
1075 assigned null. This can be turned off with --all.
1077 * An (optional) journal gateway daemon is now available as
1078 "systemd-journal-gatewayd.service". This service provides
1079 access to the journal via HTTP and JSON. This functionality
1080 will be used to implement live log synchronization in both
1081 pull and push modes, but has various other users too, such
1082 as easy log access for debugging of embedded devices. Right
1083 now it is already useful to retrieve the journal via HTTP:
1085 # systemctl start systemd-journal-gatewayd.service
1086 # wget http://localhost:19531/entries
1088 This will download the journal contents in a
1089 /var/log/messages compatible format. The same as JSON:
1091 # curl -H"Accept: application/json" http://localhost:19531/entries
1093 This service is also accessible via a web browser where a
1094 single static HTML5 app is served that uses the JSON logic
1095 to enable the user to do some basic browsing of the
1096 journal. This will be extended later on. Here's an example
1097 screenshot of this app in its current state:
1099 http://0pointer.de/public/journal-gatewayd
1101 Contributions from: Kay Sievers, Lennart Poettering, Robert
1102 Milasan, Tom Gundersen
1106 * The bash completion logic is now available for journalctl
1109 * We don't mount the "cpuset" controller anymore together with
1110 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
1111 started if no parameters are assigned to it. "cpuset" hence
1112 broke code that assumed it it could create "cpu" groups and
1115 * journalctl -f will now subscribe to terminal size changes,
1116 and line break accordingly.
1118 Contributions from: Dave Reisner, Kay Sievers, Lennart
1119 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
1123 * nspawn will now create a symlink /etc/localtime in the
1124 container environment, copying the host's timezone
1125 setting. Previously this has been done via a bind mount, but
1126 since symlinks cannot be bind mounted this has now been
1127 changed to create/update the appropriate symlink.
1129 * journalctl -n's line number argument is now optional, and
1130 will default to 10 if omitted.
1132 * journald will now log the maximum size the journal files may
1133 take up on disk. This is particularly useful if the default
1134 built-in logic of determining this parameter from the file
1135 system size is used. Use "systemctl status
1136 systemd-journald.service" to see this information.
1138 * The multi-seat X wrapper tool has been stripped down. As X
1139 is now capable of enumerating graphics devices via udev in a
1140 seat-aware way the wrapper is not strictly necessary
1141 anymore. A stripped down temporary stop-gap is still shipped
1142 until the upstream display managers have been updated to
1143 fully support the new X logic. Expect this wrapper to be
1144 removed entirely in one of the next releases.
1146 * HandleSleepKey= in logind.conf has been split up into
1147 HandleSuspendKey= and HandleHibernateKey=. The old setting
1148 is not available anymore. X11 and the kernel are
1149 distuingishing between these keys and we should too. This
1150 also means the inhibition lock for these keys has been split
1153 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
1154 Poettering, Lukas Nykryn, Václav Pavlín
1158 * Whenever a unit changes state we'll now log this to the
1159 journal and show along the unit's own log output in
1162 * ConditionPathIsMountPoint= can now properly detect bind
1163 mount points too. (Previously, a bind mount of one file
1164 system to another place in the same file system could not be
1165 detected as mount, since they shared struct stat's st_dev
1168 * We will now mount the cgroup controllers cpu, cpuacct,
1169 cpuset and the controllers net_cls, net_prio together by
1172 * nspawn containers will now have a virtualized boot
1173 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
1174 over with a randomized ID at container initialization). This
1175 has the effect of making "journalctl -b" do the right thing
1178 * The JSON output journal serialization has been updated not
1179 to generate "endless" list objects anymore, but rather one
1180 JSON object per line. This is more in line how most JSON
1181 parsers expect JSON objects. The new output mode
1182 "json-pretty" has been added to provide similar output, but
1183 neatly aligned for readability by humans.
1185 * We dropped all explicit sync() invocations in the shutdown
1186 code. The kernel does this implicitly anyway in the kernel
1187 reboot() syscall. halt(8)'s -n option is now a compatibility
1190 * We now support virtualized reboot() in containers, as
1191 supported by newer kernels. We will fall back to exit() if
1192 CAP_SYS_REBOOT is not available to the container. Also,
1193 nspawn makes use of this now and will actually reboot the
1194 container if the containerized OS asks for that.
1196 * journalctl will only show local log output by default
1197 now. Use --merge (-m) to show remote log output, too.
1199 * libsystemd-journal gained the new sd_journal_get_usage()
1200 call to determine the current disk usage of all journal
1201 files. This is exposed in the new "journalctl --disk-usage"
1204 * journald gained a new configuration setting SplitMode= in
1205 journald.conf which may be used to control how user journals
1206 are split off. See journald.conf(5) for details.
1208 * A new condition type ConditionFileNotEmpty= has been added.
1210 * tmpfiles' "w" lines now support file globbing, to write
1211 multiple files at once.
1213 * We added Python bindings for the journal submission
1214 APIs. More Python APIs for a number of selected APIs will
1215 likely follow. Note that we intend to add native bindings
1216 only for the Python language, as we consider it common
1217 enough to deserve bindings shipped within systemd. There are
1218 various projects outside of systemd that provide bindings
1219 for languages such as PHP or Lua.
1221 * Many conditions will now resolve specifiers such as %i. In
1222 addition, PathChanged= and related directives of .path units
1223 now support specifiers as well.
1225 * There's now a new RPM macro definition for the system preset
1228 * journald will now warn if it can't forward a message to the
1229 syslog daemon because it's socket is full.
1231 * timedated will no longer write or process /etc/timezone,
1232 except on Debian. As we do not support late mounted /usr
1233 anymore /etc/localtime always being a symlink is now safe,
1234 and hence the information in /etc/timezone is not necessary
1237 * logind will now always reserve one VT for a text getty (VT6
1238 by default). Previously if more than 6 X sessions where
1239 started they took up all the VTs with auto-spawned gettys,
1240 so that no text gettys were available anymore.
1242 * udev will now automatically inform the btrfs kernel logic
1243 about btrfs RAID components showing up. This should make
1244 simple hotplug based btrfs RAID assembly work.
1246 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1247 (but not for its children which will stay at the kernel
1248 default). This should allow setups with a lot more listening
1251 * systemd will now always pass the configured timezone to the
1252 kernel at boot. timedated will do the same when the timezone
1255 * logind's inhibition logic has been updated. By default,
1256 logind will now handle the lid switch, the power and sleep
1257 keys all the time, even in graphical sessions. If DEs want
1258 to handle these events on their own they should take the new
1259 handle-power-key, handle-sleep-key and handle-lid-switch
1260 inhibitors during their runtime. A simple way to achiveve
1261 that is to invoke the DE wrapped in an invocation of:
1263 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1265 * Access to unit operations is now checked via SELinux taking
1266 the unit file label and client process label into account.
1268 * systemd will now notify the administrator in the journal
1269 when he over-mounts a non-empty directory.
1271 * There are new specifiers that are resolved in unit files,
1272 for the host name (%H), the machine ID (%m) and the boot ID
1275 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
1276 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
1277 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
1278 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
1279 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
1280 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
1281 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
1285 * Support for reading structured kernel messages from
1286 /dev/kmsg has now been added and is enabled by default.
1288 * Support for reading kernel messages from /proc/kmsg has now
1289 been removed. If you want kernel messages in the journal
1290 make sure to run a recent kernel (>= 3.5) that supports
1291 reading structured messages from /dev/kmsg (see
1292 above). /proc/kmsg is now exclusive property of classic
1293 syslog daemons again.
1295 * The libudev API gained the new
1296 udev_device_new_from_device_id() call.
1298 * The logic for file system namespace (ReadOnlyDirectory=,
1299 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
1300 require pivot_root() anymore. This means fewer temporary
1301 directories are created below /tmp for this feature.
1303 * nspawn containers will now see and receive all submounts
1304 made on the host OS below the root file system of the
1307 * Forward Secure Sealing is now supported for Journal files,
1308 which provide cryptographical sealing of journal files so
1309 that attackers cannot alter log history anymore without this
1310 being detectable. Lennart will soon post a blog story about
1311 this explaining it in more detail.
1313 * There are two new service settings RestartPreventExitStatus=
1314 and SuccessExitStatus= which allow configuration of exit
1315 status (exit code or signal) which will be excepted from the
1316 restart logic, resp. consider successful.
1318 * journalctl gained the new --verify switch that can be used
1319 to check the integrity of the structure of journal files and
1320 (if Forward Secure Sealing is enabled) the contents of
1323 * nspawn containers will now be run with /dev/stdin, /dev/fd/
1324 and similar symlinks pre-created. This makes running shells
1325 as container init process a lot more fun.
1327 * The fstab support can now handle PARTUUID= and PARTLABEL=
1330 * A new ConditionHost= condition has been added to match
1331 against the hostname (with globs) and machine ID. This is
1332 useful for clusters where a single OS image is used to
1333 provision a large number of hosts which shall run slightly
1334 different sets of services.
1336 * Services which hit the restart limit will now be placed in a
1339 Contributions from: Bertram Poettering, Dave Reisner, Huang
1340 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
1341 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
1345 * When running in --user mode systemd will now become a
1346 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
1347 tree a lot more organized.
1349 * A new PartOf= unit dependency type has been introduced that
1350 may be used to group services in a natural way.
1352 * "systemctl enable" may now be used to enable instances of
1355 * journalctl now prints error log levels in red, and
1356 warning/notice log levels in bright white. It also supports
1357 filtering by log level now.
1359 * cgtop gained a new -n switch (similar to top), to configure
1360 the maximum number of iterations to run for. It also gained
1361 -b, to run in batch mode (accepting no input).
1363 * The suffix ".service" may now be omitted on most systemctl
1364 command lines involving service unit names.
1366 * There's a new bus call in logind to lock all sessions, as
1367 well as a loginctl verb for it "lock-sessions".
1369 * libsystemd-logind.so gained a new call sd_journal_perror()
1370 that works similar to libc perror() but logs to the journal
1371 and encodes structured information about the error number.
1373 * /etc/crypttab entries now understand the new keyfile-size=
1376 * shutdown(8) now can send a (configurable) wall message when
1377 a shutdown is cancelled.
1379 * The mount propagation mode for the root file system will now
1380 default to "shared", which is useful to make containers work
1381 nicely out-of-the-box so that they receive new mounts from
1382 the host. This can be undone locally by running "mount
1383 --make-rprivate /" if needed.
1385 * The prefdm.service file has been removed. Distributions
1386 should maintain this unit downstream if they intend to keep
1387 it around. However, we recommend writing normal unit files
1388 for display managers instead.
1390 * Since systemd is a crucial part of the OS we will now
1391 default to a number of compiler switches that improve
1392 security (hardening) such as read-only relocations, stack
1393 protection, and suchlike.
1395 * The TimeoutSec= setting for services is now split into
1396 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1397 of individual time outs for the start and the stop phase of
1400 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1401 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1402 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1403 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1404 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1405 Gundersen, Zbigniew Jędrzejewski-Szmek
1409 * The journal and id128 C APIs are now fully documented as man
1412 * Extra safety checks have been added when transitioning from
1413 the initial RAM disk to the main system to avoid accidental
1416 * /etc/crypttab entries now understand the new keyfile-offset=
1419 * systemctl -t can now be used to filter by unit load state.
1421 * The journal C API gained the new sd_journal_wait() call to
1422 make writing synchronous journal clients easier.
1424 * journalctl gained the new -D switch to show journals from a
1427 * journalctl now displays a special marker between log
1428 messages of two different boots.
1430 * The journal is now explicitly flushed to /var via a service
1431 systemd-journal-flush.service, rather than implicitly simply
1432 by seeing /var/log/journal to be writable.
1434 * journalctl (and the journal C APIs) can now match for much
1435 more complex expressions, with alternatives and
1438 * When transitioning from the initial RAM disk to the main
1439 system we will now kill all processes in a killing spree to
1440 ensure no processes stay around by accident.
1442 * Three new specifiers may be used in unit files: %u, %h, %s
1443 resolve to the user name, user home directory resp. user
1444 shell. This is useful for running systemd user instances.
1446 * We now automatically rotate journal files if their data
1447 object hash table gets a fill level > 75%. We also size the
1448 hash table based on the configured maximum file size. This
1449 together should lower hash collisions drastically and thus
1450 speed things up a bit.
1452 * journalctl gained the new "--header" switch to introspect
1453 header data of journal files.
1455 * A new setting SystemCallFilters= has been added to services
1456 which may be used to apply blacklists or whitelists to
1457 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1459 * nspawn gained a new --link-journal= switch (and quicker: -j)
1460 to link the container journal with the host. This makes it
1461 very easy to centralize log viewing on the host for all
1462 guests while still keeping the journal files separated.
1464 * Many bugfixes and optimizations
1466 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1467 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1468 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1473 * Several tools now understand kernel command line arguments,
1474 which are only read when run in an initial RAM disk. They
1475 usually follow closely their normal counterparts, but are
1478 * There's a new tool to analyze the readahead files that are
1479 automatically generated at boot. Use:
1481 /usr/lib/systemd/systemd-readahead analyze /.readahead
1483 * We now provide an early debug shell on tty9 if this enabled. Use:
1485 systemctl enable debug-shell.service
1487 * All plymouth related units have been moved into the Plymouth
1488 package. Please make sure to upgrade your Plymouth version
1491 * systemd-tmpfiles now supports getting passed the basename of
1492 a configuration file only, in which case it will look for it
1493 in all appropriate directories automatically.
1495 * udevadm info now takes a /dev or /sys path as argument, and
1496 does the right thing. Example:
1498 udevadm info /dev/sda
1499 udevadm info /sys/class/block/sda
1501 * systemctl now prints a warning if a unit is stopped but a
1502 unit that might trigger it continues to run. Example: a
1503 service is stopped but the socket that activates it is left
1506 * "systemctl status" will now mention if the log output was
1507 shortened due to rotation since a service has been started.
1509 * The journal API now exposes functions to determine the
1510 "cutoff" times due to rotation.
1512 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1513 immediately flushing of runtime logs to /var if possible,
1514 resp. for triggering immediate rotation of the journal
1517 * It is now considered an error if a service is attempted to
1518 be stopped that is not loaded.
1520 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1522 * systemd-analyze now supports Python 3
1524 * tmpfiles now supports cleaning up directories via aging
1525 where the first level dirs are always kept around but
1526 directories beneath it automatically aged. This is enabled
1527 by prefixing the age field with '~'.
1529 * Seat objects now expose CanGraphical, CanTTY properties
1530 which is required to deal with very fast bootups where the
1531 display manager might be running before the graphics drivers
1532 completed initialization.
1534 * Seat objects now expose a State property.
1536 * We now include RPM macros for service enabling/disabling
1537 based on the preset logic. We recommend RPM based
1538 distributions to make use of these macros if possible. This
1539 makes it simpler to reuse RPM spec files across
1542 * We now make sure that the collected systemd unit name is
1543 always valid when services log to the journal via
1546 * There's a new man page kernel-command-line(7) detailing all
1547 command line options we understand.
1549 * The fstab generator may now be disabled at boot by passing
1550 fstab=0 on the kernel command line.
1552 * A new kernel command line option modules-load= is now understood
1553 to load a specific kernel module statically, early at boot.
1555 * Unit names specified on the systemctl command line are now
1556 automatically escaped as needed. Also, if file system or
1557 device paths are specified they are automatically turned
1558 into the appropriate mount or device unit names. Example:
1560 systemctl status /home
1561 systemctl status /dev/sda
1563 * The SysVConsole= configuration option has been removed from
1564 system.conf parsing.
1566 * The SysV search path is no longer exported on the D-Bus
1569 * The Names= option is been removed from unit file parsing.
1571 * There's a new man page bootup(7) detailing the boot process.
1573 * Every unit and every generator we ship with systemd now
1574 comes with full documentation. The self-explanatory boot is
1577 * A couple of services gained "systemd-" prefixes in their
1578 name if they wrap systemd code, rather than only external
1579 code. Among them fsck@.service which is now
1580 systemd-fsck@.service.
1582 * The HaveWatchdog property has been removed from the D-Bus
1585 * systemd.confirm_spawn= on the kernel command line should now
1588 * There's a new man page crypttab(5) which details all options
1589 we actually understand.
1591 * systemd-nspawn gained a new --capability= switch to pass
1592 additional capabilities to the container.
1594 * timedated will now read known NTP implementation unit names
1595 from /usr/lib/systemd/ntp-units.d/*.list,
1596 systemd-timedated-ntp.target has been removed.
1598 * journalctl gained a new switch "-b" that lists log data of
1599 the current boot only.
1601 * The notify socket is in the abstract namespace again, in
1602 order to support daemons which chroot() at start-up.
1604 * There is a new Storage= configuration option for journald
1605 which allows configuration of where log data should go. This
1606 also provides a way to disable journal logging entirely, so
1607 that data collected is only forwarded to the console, the
1608 kernel log buffer or another syslog implementation.
1610 * Many bugfixes and optimizations
1612 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1613 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1614 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1615 Shawn Landden, Tom Gundersen
1619 * "systemctl help <unit>" now shows the man page if one is
1622 * Several new man pages have been added.
1624 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1625 MaxLevelConsole= can now be specified in
1626 journald.conf. These options allow reducing the amount of
1627 data stored on disk or forwarded by the log level.
1629 * TimerSlackNSec= can now be specified in system.conf for
1630 PID1. This allows system-wide power savings.
1632 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1633 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1638 * logind is now capable of (optionally) handling power and
1639 sleep keys as well as the lid switch.
1641 * journalctl now understands the syntax "journalctl
1642 /usr/bin/avahi-daemon" to get all log output of a specific
1645 * CapabilityBoundingSet= in system.conf now also influences
1646 the capability bound set of usermode helpers of the kernel.
1648 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1649 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1650 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1651 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1655 * Note that we skipped 139 releases here in order to set the
1656 new version to something that is greater than both udev's
1657 and systemd's most recent version number.
1659 * udev: all udev sources are merged into the systemd source tree now.
1660 All future udev development will happen in the systemd tree. It
1661 is still fully supported to use the udev daemon and tools without
1662 systemd running, like in initramfs or other init systems. Building
1663 udev though, will require the *build* of the systemd tree, but
1664 udev can be properly *run* without systemd.
1666 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1667 should be used to create dead device nodes as workarounds for broken
1670 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1671 no longer supported. udev_monitor_new_from_netlink() needs to be
1672 used to subscribe to events.
1674 * udev: when udevd is started by systemd, processes which are left
1675 behind by forking them off of udev rules, are unconditionally cleaned
1676 up and killed now after the event handling has finished. Services or
1677 daemons must be started as systemd services. Services can be
1678 pulled-in by udev to get started, but they can no longer be directly
1679 forked by udev rules.
1681 * udev: the daemon binary is called systemd-udevd now and installed
1682 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1683 to adapt to that, create symlink, or rename the binary after building
1686 * libudev no longer provides these symbols:
1687 udev_monitor_from_socket()
1688 udev_queue_get_failed_list_entry()
1689 udev_get_{dev,sys,run}_path()
1690 The versions number was bumped and symbol versioning introduced.
1692 * systemd-loginctl and systemd-journalctl have been renamed
1693 to loginctl and journalctl to match systemctl.
1695 * The config files: /etc/systemd/systemd-logind.conf and
1696 /etc/systemd/systemd-journald.conf have been renamed to
1697 logind.conf and journald.conf. Package updates should rename
1698 the files to the new names on upgrade.
1700 * For almost all files the license is now LGPL2.1+, changed
1701 from the previous GPL2.0+. Exceptions are some minor stuff
1702 of udev (which will be changed to LGPL2.1 eventually, too),
1703 and the MIT licensed sd-daemon.[ch] library that is suitable
1704 to be used as drop-in files.
1706 * systemd and logind now handle system sleep states, in
1707 particular suspending and hibernating.
1709 * logind now implements a sleep/shutdown/idle inhibiting logic
1710 suitable for a variety of uses. Soonishly Lennart will blog
1711 about this in more detail.
1713 * var-run.mount and var-lock.mount are no longer provided
1714 (which prevously bind mounted these directories to their new
1715 places). Distributions which have not converted these
1716 directories to symlinks should consider stealing these files
1717 from git history and add them downstream.
1719 * We introduced the Documentation= field for units and added
1720 this to all our shipped units. This is useful to make it
1721 easier to explore the boot and the purpose of the various
1724 * All smaller setup units (such as
1725 systemd-vconsole-setup.service) now detect properly if they
1726 are run in a container and are skipped when
1727 appropriate. This guarantees an entirely noise-free boot in
1728 Linux container environments such as systemd-nspawn.
1730 * A framework for implementing offline system updates is now
1731 integrated, for details see:
1732 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1734 * A new service type Type=idle is available now which helps us
1735 avoiding ugly interleaving of getty output and boot status
1738 * There's now a system-wide CapabilityBoundingSet= option to
1739 globally reduce the set of capabilities for the
1740 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1741 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1742 even CAP_NET_ADMIN system-wide for secure systems.
1744 * There are now system-wide DefaultLimitXXX= options to
1745 globally change the defaults of the various resource limits
1746 for all units started by PID 1.
1748 * Harald Hoyer's systemd test suite has been integrated into
1749 systemd which allows easy testing of systemd builds in qemu
1750 and nspawn. (This is really awesome! Ask us for details!)
1752 * The fstab parser is now implemented as generator, not inside
1755 * systemctl will now warn you if .mount units generated from
1756 /etc/fstab are out of date due to changes in fstab that
1757 haven't been read by systemd yet.
1759 * systemd is now suitable for usage in initrds. Dracut has
1760 already been updated to make use of this. With this in place
1761 initrds get a slight bit faster but primarily are much
1762 easier to introspect and debug since "systemctl status" in
1763 the host system can be used to introspect initrd services,
1764 and the journal from the initrd is kept around too.
1766 * systemd-delta has been added, a tool to explore differences
1767 between user/admin configuration and vendor defaults.
1769 * PrivateTmp= now affects both /tmp and /var/tmp.
1771 * Boot time status messages are now much prettier and feature
1772 proper english language. Booting up systemd has never been
1775 * Read-ahead pack files now include the inode number of all
1776 files to pre-cache. When the inode changes the pre-caching
1777 is not attempted. This should be nicer to deal with updated
1778 packages which might result in changes of read-ahead
1781 * We now temporaritly lower the kernel's read_ahead_kb variable
1782 when collecting read-ahead data to ensure the kernel's
1783 built-in read-ahead does not add noise to our measurements
1784 of necessary blocks to pre-cache.
1786 * There's now RequiresMountsFor= to add automatic dependencies
1787 for all mounts necessary for a specific file system path.
1789 * MountAuto= and SwapAuto= have been removed from
1790 system.conf. Mounting file systems at boot has to take place
1793 * nspawn now learned a new switch --uuid= to set the machine
1794 ID on the command line.
1796 * nspawn now learned the -b switch to automatically search
1799 * vt102 is now the default TERM for serial TTYs, upgraded from
1802 * systemd-logind now works on VT-less systems.
1804 * The build tree has been reorganized. The individual
1805 components now have directories of their own.
1807 * A new condition type ConditionPathIsReadWrite= is now available.
1809 * nspawn learned the new -C switch to create cgroups for the
1810 container in other hierarchies.
1812 * We now have support for hardware watchdogs, configurable in
1815 * The scheduled shutdown logic now has a public API.
1817 * We now mount /tmp as tmpfs by default, but this can be
1818 masked and /etc/fstab can override it.
1820 * Since udisks doesn't make use of /media anymore we are not
1821 mounting a tmpfs on it anymore.
1823 * journalctl gained a new --local switch to only interleave
1824 locally generated journal files.
1826 * We can now load the IMA policy at boot automatically.
1828 * The GTK tools have been split off into a systemd-ui.
1830 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1831 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1832 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1833 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1834 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1835 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1836 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1837 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1838 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1843 * This is mostly a bugfix release
1845 * Support optional initialization of the machine ID from the
1846 KVM or container configured UUID.
1848 * Support immediate reboots with "systemctl reboot -ff"
1850 * Show /etc/os-release data in systemd-analyze output
1852 * Many bugfixes for the journal, including endianness fixes and
1853 ensuring that disk space enforcement works
1855 * sd-login.h is C++ comptaible again
1857 * Extend the /etc/os-release format on request of the Debian
1860 * We now refuse non-UTF8 strings used in various configuration
1861 and unit files. This is done to ensure we don't pass invalid
1862 data over D-Bus or expose it elsewhere.
1864 * Register Mimo USB Screens as suitable for automatic seat
1867 * Read SELinux client context from journal clients in a race
1870 * Reorder configuration file lookup order. /etc now always
1871 overrides /run in order to allow the administrator to always
1872 and unconditionally override vendor supplied or
1873 automatically generated data.
1875 * The various user visible bits of the journal now have man
1876 pages. We still lack man pages for the journal API calls
1879 * We now ship all man pages in HTML format again in the
1882 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1883 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1884 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1889 * This is mostly a bugfix release
1891 * systems lacking /etc/os-release are no longer supported.
1893 * Various functionality updates to libsystemd-login.so
1895 * Track class of PAM logins to distuingish greeters from
1898 Contributions from: Kay Sievers, Lennart Poettering, Michael
1903 * This is an important bugfix release for v41.
1905 * Building man pages is now optional which should be useful
1906 for those building systemd from git but unwilling to install
1909 * Watchdog support for supervising services is now usable. In
1910 a future release support for hardware watchdogs
1911 (i.e. /dev/watchdog) will be added building on this.
1913 * Service start rate limiting is now configurable and can be
1914 turned off per service. When a start rate limit is hit a
1915 reboot can automatically be triggered.
1917 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1919 Contributions from: Benjamin Franzke, Bill Nottingham,
1920 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1921 Schmidt, Michał Górny, Piotr Drąg
1925 * The systemd binary is installed /usr/lib/systemd/systemd now;
1926 An existing /sbin/init symlink needs to be adapted with the
1929 * The code that loads kernel modules has been ported to invoke
1930 libkmod directly, instead of modprobe. This means we do not
1931 support systems with module-init-tools anymore.
1933 * Watchdog support is now already useful, but still not
1936 * A new kernel command line option systemd.setenv= is
1937 understood to set system wide environment variables
1938 dynamically at boot.
1940 * We now limit the set of capabilities of systemd-journald.
1942 * We now set SIGPIPE to ignore by default, since it only is
1943 useful in shell pipelines, and has little use in general
1944 code. This can be disabled with IgnoreSIPIPE=no in unit
1947 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1948 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1953 * This is mostly a bugfix release
1955 * We now expose the reason why a service failed in the
1956 "Result" D-Bus property.
1958 * Rudimentary service watchdog support (will be completed over
1959 the next few releases.)
1961 * When systemd forks off in order execute some service we will
1962 now immediately changes its argv[0] to reflect which process
1963 it will execute. This is useful to minimize the time window
1964 with a generic argv[0], which makes bootcharts more useful
1966 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1967 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1968 Mike Kazantsev, Ray Strode
1972 * This is mostly a test release, but incorporates many
1975 * New systemd-cgtop tool to show control groups by their
1978 * Linking against libacl for ACLs is optional again. If
1979 disabled, support tracking device access for active logins
1980 goes becomes unavailable, and so does access to the user
1981 journals by the respective users.
1983 * If a group "adm" exists, journal files are automatically
1984 owned by them, thus allow members of this group full access
1985 to the system journal as well as all user journals.
1987 * The journal now stores the SELinux context of the logging
1988 client for all entries.
1990 * Add C++ inclusion guards to all public headers
1992 * New output mode "cat" in the journal to print only text
1993 messages, without any meta data like date or time.
1995 * Include tiny X server wrapper as a temporary stop-gap to
1996 teach XOrg udev display enumeration. This is used by display
1997 managers such as gdm, and will go away as soon as XOrg
1998 learned native udev hotplugging for display devices.
2000 * Add new systemd-cat tool for executing arbitrary programs
2001 with STDERR/STDOUT connected to the journal. Can also act as
2002 BSD logger replacement, and does so by default.
2004 * Optionally store all locally generated coredumps in the
2005 journal along with meta data.
2007 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
2008 writing short strings to files (for usage for /sys), and for
2009 creating symlinks, character and block device nodes.
2011 * New unit file option ControlGroupPersistent= to make cgroups
2012 persistent, following the mechanisms outlined in
2013 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
2015 * Support multiple local RTCs in a sane way
2017 * No longer monopolize IO when replaying readahead data on
2018 rotating disks, since we might starve non-file-system IO to
2019 death, since fanotify() will not see accesses done by blkid,
2022 * Don't show kernel threads in systemd-cgls anymore, unless
2023 requested with new -k switch.
2025 Contributions from: Dan Horák, Kay Sievers, Lennart
2026 Poettering, Michal Schmidt
2030 * This is mostly a test release, but incorporates many
2033 * The git repository moved to:
2034 git://anongit.freedesktop.org/systemd/systemd
2035 ssh://git.freedesktop.org/git/systemd/systemd
2037 * First release with the journal
2038 http://0pointer.de/blog/projects/the-journal.html
2040 * The journal replaces both systemd-kmsg-syslogd and
2041 systemd-stdout-bridge.
2043 * New sd_pid_get_unit() API call in libsystemd-logind
2045 * Many systemadm clean-ups
2047 * Introduce remote-fs-pre.target which is ordered before all
2048 remote mounts and may be used to start services before all
2051 * Added Mageia support
2053 * Add bash completion for systemd-loginctl
2055 * Actively monitor PID file creation for daemons which exit in
2056 the parent process before having finished writing the PID
2057 file in the daemon process. Daemons which do this need to be
2058 fixed (i.e. PID file creation must have finished before the
2059 parent exits), but we now react a bit more gracefully to them.
2061 * Add colourful boot output, mimicking the well-known output
2062 of existing distributions.
2064 * New option PassCredentials= for socket units, for
2065 compatibility with a recent kernel ABI breakage.
2067 * /etc/rc.local is now hooked in via a generator binary, and
2068 thus will no longer act as synchronization point during
2071 * systemctl list-unit-files now supports --root=.
2073 * systemd-tmpfiles now understands two new commands: z, Z for
2074 relabelling files according to the SELinux database. This is
2075 useful to apply SELinux labels to specific files in /sys,
2078 * Output of SysV services is now forwarded to both the console
2079 and the journal by default, not only just the console.
2081 * New man pages for all APIs from libsystemd-login.
2083 * The build tree got reorganized and a the build system is a
2084 lot more modular allowing embedded setups to specifically
2085 select the components of systemd they are interested in.
2087 * Support for Linux systems lacking the kernel VT subsystem is
2090 * configure's --with-rootdir= got renamed to
2091 --with-rootprefix= to follow the naming used by udev and
2094 * Unless specified otherwise we'll now install to /usr instead
2095 of /usr/local by default.
2097 * Processes with '@' in argv[0][0] are now excluded from the
2098 final shut-down killing spree, following the logic explained
2100 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
2102 * All processes remaining in a service cgroup when we enter
2103 the START or START_PRE states are now killed with
2104 SIGKILL. That means it is no longer possible to spawn
2105 background processes from ExecStart= lines (which was never
2106 supported anyway, and bad style).
2108 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
2109 reloading of units together.
2111 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
2112 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
2113 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
2114 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
2115 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek