1 systemd System and Service Manager
5 * The documentation has been updated to cover the various new
6 concepts introduced with 205.
8 * Unit files now understand the new %v specifier which
9 resolves to the kernel version string as returned by "uname
12 * systemctl now supports filtering the unit list output by
13 load state, active state and sub state, using the new
16 * "systemctl status" will now show the results of the
17 condition checks (like ConditionPathExists= and similar) of
18 the last start attempts of the unit. They are also logged to
21 * "journalctl -b" may now be used to look for boot output of a
22 specific boot. Try "journalctl -b -1" for the previous boot,
23 but the syntax is substantially more powerful.
25 * "journalctl --show-cursor" has been added which prints the
26 cursor string the last shown log line. This may then be used
27 with the new "journalctl --after-cursor=" switch to continue
28 browsing logs from that point on.
30 * "journalctl --force" may now be used to force regeneration
33 * Device ACLs may now be applied to "dead" devices nodes too,
34 i.e. device nodes that are created based on kernel module
35 information of not yet loaded kernel modules.
37 * A new RPM macro has been added that may be used to apply
38 tmpfiles configuration during package installation.
40 * systemd-detect-virt and ConditionVirtualization= now can
41 detect User-Mode-Linux machines (UML).
43 * journald will now log the effective capabilities set of
44 processes in the message metadata.
46 * systemd-cryptsetup has gained support for TrueCrypt volumes.
48 * The initrd interface has been simplified (more specifically,
49 support for passing performance data via environment
50 variables and fsck results via files in /run has been
51 removed). These features were non-essential, and are
52 nowadays available in a much nicer way by having systemd in
53 the initrd serialize its state and have the hosts systemd
56 * The udev "keymap" data files and tools to apply keyboard
57 specific mappings of scan to key codes, and force-release
58 scan code lists have been entirely replaced by a udev
59 "keyboard" builtin and a hwdb data file.
63 * Two new unit types have been introduced:
65 Scope units are very similar to service units, however, are
66 created out of pre-existing processes -- instead of PID 1
67 forking off the processes. By using scope units it is
68 possible for system services and applications to group their
69 own child processes (worker processes) in a powerful way
70 which then maybe used to organize them, or kill them
71 together, or apply resource limits on them.
73 Slice units may be used to partition system resources in an
74 hierarchial fashion and then assign other units to them. By
75 default there are now three slices: system.slice (for all
76 system services), user.slice (for all user sessions),
77 machine.slice (for VMs and containers).
79 Slices and scopes have been introduced primarily in
80 context of the work to move cgroup handling to a
81 single-writer scheme, where only PID 1
82 creates/removes/manages cgroups.
84 * There's a new concept of "transient" units. In contrast to
85 normal units these units are created via an API at runtime,
86 not from configuration from disk. More specifically this
87 means it is now possible to run arbitrary programs as
88 independent services, with all execution parameters passed
89 in via bus APIs rather than read from disk. Transient units
90 make systemd substantially more dynamic then it ever was,
91 and useful as a general batch manager.
93 * logind has been updated to make use of scope and slice units
94 for managing user sessions. As a user logs in he will get
95 his own private slice unit, to which all sessions are added
96 as scope units. We also added support for automatically
97 adding an instance of user@.service for the user into the
98 slice. Effectively logind will no longer create cgroup
99 hierarchies on its own now, it will defer entirely to PID 1
100 for this by means of scope, service and slice units. Since
101 user sessions this way become entities managed by PID 1
102 the output of "systemctl" is now a lot more comprehensive.
104 * A new mini-daemon "systemd-machined" has been added which
105 may be used by virtualization managers to register local
106 VMs/containers. nspawn has been updated accordingly, and
107 libvirt will be updated shortly. machined will collect a bit
108 of meta information about the VMs/containers, and assign
109 them their own scope unit (see above). The collected
110 meta-data is then made available via the "machinectl" tool,
111 and exposed in "ps" and similar tools. machined/machinectl
112 is compile-time optional.
114 * As discussed earlier, the low-level cgroup configuration
115 options ControlGroup=, ControlGroupModify=,
116 ControlGroupPersistent=, ControlGroupAttribute= have been
117 removed. Please use high-level attribute settings instead as
120 * A new bus call SetUnitProperties() has been added to alter
121 various runtime parameters of a unit. This is primarily
122 useful to alter cgroup parameters dynamically in a nice way,
123 but will be extended later on to make more properties
124 modifiable at runtime. systemctl gained a new set-properties
125 command that wraps this call.
127 * A new tool "systemd-run" has been added which can be used to
128 run arbitrary command lines as transient services or scopes,
129 while configuring a number of settings via the command
130 line. This tool is currently very basic, however already
131 very useful. We plan to extend this tool to even allow
132 queuing of execution jobs with time triggers from the
133 command line, similar in fashion to "at".
135 * nspawn will now inform the user explicitly that kernels with
136 audit enabled break containers, and suggest the user to turn
139 * Support for detecting the IMA and AppArmor security
140 frameworks with ConditionSecurity= has been added.
142 * journalctl gained a new "-k" switch for showing only kernel
143 messages, mimicking dmesg output; in addition to "--user"
144 and "--system" switches for showing only user's own logs
147 * systemd-delta can now show information about drop-in
148 snippets extending unit files.
150 * libsystemd-bus has been substantially updated but is still
151 not available as public API.
153 * systemd will now look for the "debug" argument on the kernel
154 command line and enable debug logging, similar to
155 "systemd.log_level=debug" already did before.
157 * "systemctl set-default", "systemctl get-default" has been
158 added to configure the default.target symlink, which
159 controls what to boot into by default.
161 * "systemctl set-log-level" has been added as a convenient
162 way to raise and lower systemd logging threshold.
164 * "systemd-analyze plot" will now show the time the various
165 generators needed for execution, as well as information
166 about the unit file loading.
168 * libsystemd-journal gained a new sd_journal_open_files() call
169 for opening specific journal files. journactl also gained a
170 new switch to expose this new functionality. Previously we
171 only supported opening all files from a directory, or all
172 files from the system, as opening individual files only is
173 racy due to journal file rotation.
175 * systemd gained the new DefaultEnvironment= setting in
176 /etc/systemd/system.conf to set environment variables for
179 * If a privileged process logs a journal message with the
180 OBJECT_PID= field set, then journald will automatically
181 augment this with additional OBJECT_UID=, OBJECT_GID=,
182 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
183 system services want to log events about specific client
184 processes. journactl/systemctl has been updated to make use
185 of this information if all log messages regarding a specific
188 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
189 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
190 Reisner, David Coppa, David King, David Strauss, Eelco
191 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
192 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
193 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
194 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
195 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
196 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
197 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
198 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
199 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
204 * The Python bindings gained some minimal support for the APIs
205 exposed by libsystemd-logind.
207 * ConditionSecurity= gained support for detecting SMACK. Since
208 this condition already supports SELinux and AppArmor we only
209 miss IMA for this. Patches welcome!
211 Contributions from: Karol Lewandowski, Lennart Poettering,
212 Zbigniew Jędrzejewski-Szmek
216 * systemd-nspawn will now create /etc/resolv.conf if
217 necessary, before bind-mounting the host's file onto it.
219 * systemd-nspawn will now store meta information about a
220 container on the container's cgroup as extended attribute
221 fields, including the root directory.
223 * The cgroup hierarchy has been reworked in many ways. All
224 objects any of the components systemd creates in the cgroup
225 tree are now suffixed. More specifically, user sessions are
226 now placed in cgroups suffixed with ".session", users in
227 cgroups suffixed with ".user", and nspawn containers in
228 cgroups suffixed with ".nspawn". Furthermore, all cgroup
229 names are now escaped in a simple scheme to avoid collision
230 of userspace object names with kernel filenames. This work
231 is preparation for making these objects relocatable in the
232 cgroup tree, in order to allow easy resource partitioning of
233 these objects without causing naming conflicts.
235 * systemctl list-dependencies gained the new switches
236 --plain, --reverse, --after and --before.
238 * systemd-inhibit now shows the process name of processes that
239 have taken an inhibitor lock.
241 * nss-myhostname will now also resolve "localhost"
242 implicitly. This makes /etc/hosts an optional file and
243 nicely handles that on IPv6 ::1 maps to both "localhost" and
246 * libsystemd-logind.so gained a new call
247 sd_get_machine_names() to enumerate running containers and
248 VMs (currently only supported by very new libvirt and
249 nspawn). sd_login_monitor can now be used to watch
250 VMs/containers coming and going.
252 * .include is not allowed recursively anymore, and only in
253 unit files. Usually it is better to use drop-in snippets in
254 .d/*.conf anyway, as introduced with systemd 198.
256 * systemd-analyze gained a new "critical-chain" command that
257 determines the slowest chain of units run during system
258 boot-up. It is very useful for tracking down where
259 optimizing boot time is the most beneficial.
261 * systemd will no longer allow manipulating service paths in
262 the name=systemd:/system cgroup tree using ControlGroup= in
263 units. (But is still fine with it in all other dirs.)
265 * There's a new systemd-nspawn@.service service file that may
266 be used to easily run nspawn containers as system
267 services. With the container's root directory in
268 /var/lib/container/foobar it is now sufficient to run
269 "systemctl start systemd-nspawn@foobar.service" to boot it.
271 * systemd-cgls gained a new parameter "--machine" to list only
272 the processes within a certain container.
274 * ConditionSecurity= now can check for "apparmor". We still
275 are lacking checks for SMACK and IMA for this condition
276 check though. Patches welcome!
278 * A new configuration file /etc/systemd/sleep.conf has been
279 added that may be used to configure which kernel operation
280 systemd is supposed to execute when "suspend", "hibernate"
281 or "hybrid-sleep" is requested. This makes the new kernel
282 "freeze" state accessible to the user.
284 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
285 the passed argument if applicable.
287 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
288 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
289 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
290 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
291 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
292 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
293 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
298 * The output of 'systemctl list-jobs' got some polishing. The
299 '--type=' argument may now be passed more than once. A new
300 command 'systemctl list-sockets' has been added which shows
301 a list of kernel sockets systemd is listening on with the
302 socket units they belong to, plus the units these socket
305 * The experimental libsystemd-bus library got substantial
306 updates to work in conjunction with the (also experimental)
307 kdbus kernel project. It works well enough to exchange
308 messages with some sophistication. Note that kdbus is not
309 ready yet, and the library is mostly an elaborate test case
310 for now, and not installable.
312 * systemd gained a new unit 'systemd-static-nodes.service'
313 that generates static device nodes earlier during boot, and
314 can run in conjunction with udev.
316 * libsystemd-login gained a new call sd_pid_get_user_unit()
317 to retrieve the user systemd unit a process is running
318 in. This is useful for systems where systemd is used as
321 * systemd-nspawn now places all containers in the new /machine
322 top-level cgroup directory in the name=systemd
323 hierarchy. libvirt will soon do the same, so that we get a
324 uniform separation of /system, /user and /machine for system
325 services, user processes and containers/virtual
326 machines. This new cgroup hierarchy is also useful to stick
327 stable names to specific container instances, which can be
328 recognized later this way (this name may be controlled
329 via systemd-nspawn's new -M switch). libsystemd-login also
330 gained a new call sd_pid_get_machine_name() to retrieve the
331 name of the container/VM a specific process belongs to.
333 * bootchart can now store its data in the journal.
335 * libsystemd-journal gained a new call
336 sd_journal_add_conjunction() for AND expressions to the
337 matching logic. This can be used to express more complex
340 * journactl can now take multiple --unit= and --user-unit=
343 * The cryptsetup logic now understands the "luks.key=" kernel
344 command line switch for specifying a file to read the
345 decryption key from. Also, if a configured key file is not
346 found the tool will now automatically fall back to prompting
349 * Python systemd.journal module was updated to wrap recently
350 added functions from libsystemd-journal. The interface was
351 changed to bring the low level interface in s.j._Reader
352 closer to the C API, and the high level interface in
353 s.j.Reader was updated to wrap and convert all data about
356 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
357 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
358 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
359 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
360 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
361 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
365 * journalctl --update-catalog now understands a new --root=
366 option to operate on catalogs found in a different root
369 * During shutdown after systemd has terminated all running
370 services a final killing loop kills all remaining left-over
371 processes. We will now print the name of these processes
372 when we send SIGKILL to them, since this usually indicates a
375 * If /etc/crypttab refers to password files stored on
376 configured mount points automatic dependencies will now be
377 generated to ensure the specific mount is established first
378 before the key file is attempted to be read.
380 * 'systemctl status' will now show information about the
381 network sockets a socket unit is listening on.
383 * 'systemctl status' will also shown information about any
384 drop-in configuration file for units. (Drop-In configuration
385 files in this context are files such as
386 /etc/systemd/systemd/foobar.service.d/*.conf)
388 * systemd-cgtop now optionally shows summed up CPU times of
389 cgroups. Press '%' while running cgtop to switch between
390 percentage and absolute mode. This is useful to determine
391 which cgroups use up the most CPU time over the entire
392 runtime of the system. systemd-cgtop has also been updated
393 to be 'pipeable' for processing with further shell tools.
395 * 'hostnamectl set-hostname' will now allow setting of FQDN
398 * The formatting and parsing of time span values has been
399 changed. The parser now understands fractional expressions
400 such as "5.5h". The formatter will now output fractional
401 expressions for all time spans under 1min, i.e. "5.123456s"
402 rather than "5s 123ms 456us". For time spans under 1s
403 millisecond values are shown, for those under 1ms
404 microsecond values are shown. This should greatly improve
405 all time-related output of systemd.
407 * libsystemd-login and libsystemd-journal gained new
408 functions for querying the poll() events mask and poll()
409 timeout value for integration into arbitrary event
412 * localectl gained the ability to list available X11 keymaps
413 (models, layouts, variants, options).
415 * 'systemd-analyze dot' gained the ability to filter for
416 specific units via shell-style globs, to create smaller,
417 more useful graphs. I.e. it's now possible to create simple
418 graphs of all the dependencies between only target units, or
419 of all units that Avahi has dependencies with.
421 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
422 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
423 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
424 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
425 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
426 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
427 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
431 * The boot-time readahead implementation for rotating media
432 will now read the read-ahead data in multiple passes which
433 consist of all read requests made in equidistant time
434 intervals. This means instead of strictly reading read-ahead
435 data in its physical order on disk we now try to find a
436 middle ground between physical and access time order.
438 * /etc/os-release files gained a new BUILD_ID= field for usage
439 on operating systems that provide continuous builds of OS
442 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
443 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
444 William Douglas, Zbigniew Jędrzejewski-Szmek
448 * systemd-python gained an API exposing libsystemd-daemon.
450 * The SMACK setup logic gained support for uploading CIPSO
453 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
454 ReadOnlyDirectories= and InaccessibleDirectories= has
455 changed. The private /tmp and /var/tmp directories are now
456 shared by all processes of a service (which means
457 ExecStartPre= may now leave data in /tmp that ExecStart= of
458 the same service can still access). When a service is
459 stopped its temporary directories are immediately deleted
460 (normal clean-up with tmpfiles is still done in addition to
463 * By default, systemd will now set a couple of sysctl
464 variables in the kernel: the safe sysrq options are turned
465 on, IP route verification is turned on, and source routing
466 disabled. The recently added hardlink and softlink
467 protection of the kernel is turned on. These settings should
468 be reasonably safe, and good defaults for all new systems.
470 * The predictable network naming logic may now be turned off
471 with a new kernel command line switch: net.ifnames=0.
473 * A new libsystemd-bus module has been added that implements a
474 pretty complete D-Bus client library. For details see:
476 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
478 * journald will now explicitly flush the journal files to disk
479 at the latest 5min after each write. The file will then also
480 be marked offline until the next write. This should increase
481 reliability in case of a crash. The synchronization delay
482 can be configured via SyncIntervalSec= in journald.conf.
484 * There's a new remote-fs-setup.target unit that can be used
485 to pull in specific services when at least one remote file
486 system is to be mounted.
488 * There are new targets timers.target and paths.target as
489 canonical targets to pull user timer and path units in
490 from. This complements sockets.target with a similar
491 purpose for socket units.
493 * libudev gained a new call udev_device_set_attribute_value()
494 to set sysfs attributes of a device.
496 * The udev daemon now sets the default number of worker
497 processes executed in parallel based on the number of available
498 CPUs instead of the amount of available RAM. This is supposed
499 to provide a more reliable default and limit a too aggressive
500 paralellism for setups with 1000s of devices connected.
502 Contributions from: Auke Kok, Colin Walters, Cristian
503 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
504 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
505 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
506 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
507 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
508 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
509 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
510 Zbigniew Jędrzejewski-Szmek
514 * Configuration of unit files may now be extended via drop-in
515 files without having to edit/override the unit files
516 themselves. More specifically, if the administrator wants to
517 change one value for a service file foobar.service he can
518 now do so by dropping in a configuration snippet into
519 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
520 will load all these snippets and apply them on top of the
521 main unit configuration file, possibly extending or
522 overriding its settings. Using these drop-in snippets is
523 generally nicer than the two earlier options for changing
524 unit files locally: copying the files from
525 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
526 them there; or creating a new file in /etc/systemd/system/
527 that incorporates the original one via ".include". Drop-in
528 snippets into these .d/ directories can be placed in any
529 directory systemd looks for units in, and the usual
530 overriding semantics between /usr/lib, /etc and /run apply
533 * Most unit file settings which take lists of items can now be
534 reset by assigning the empty string to them. For example,
535 normally, settings such as Environment=FOO=BAR append a new
536 environment variable assignment to the environment block,
537 each time they are used. By assigning Environment= the empty
538 string the environment block can be reset to empty. This is
539 particularly useful with the .d/*.conf drop-in snippets
540 mentioned above, since this adds the ability to reset list
541 settings from vendor unit files via these drop-ins.
543 * systemctl gained a new "list-dependencies" command for
544 listing the dependencies of a unit recursively.
546 * Inhibitors are now honored and listed by "systemctl
547 suspend", "systemctl poweroff" (and similar) too, not only
548 GNOME. These commands will also list active sessions by
551 * Resource limits (as exposed by the various control group
552 controllers) can now be controlled dynamically at runtime
553 for all units. More specifically, you can now use a command
554 like "systemctl set-cgroup-attr foobar.service cpu.shares
555 2000" to alter the CPU shares a specific service gets. These
556 settings are stored persistently on disk, and thus allow the
557 administrator to easily adjust the resource usage of
558 services with a few simple commands. This dynamic resource
559 management logic is also available to other programs via the
560 bus. Almost any kernel cgroup attribute and controller is
563 * systemd-vconsole-setup will now copy all font settings to
564 all allocated VTs, where it previously applied them only to
567 * libsystemd-login gained the new sd_session_get_tty() API
570 * This release drops support for a few legacy or
571 distribution-specific LSB facility names when parsing init
572 scripts: $x-display-manager, $mail-transfer-agent,
573 $mail-transport-agent, $mail-transfer-agent, $smtp,
574 $null. Also, the mail-transfer-agent.target unit backing
575 this has been removed. Distributions which want to retain
576 compatibility with this should carry the burden for
577 supporting this themselves and patch support for these back
578 in, if they really need to. Also, the facilities $syslog and
579 $local_fs are now ignored, since systemd does not support
580 early-boot LSB init scripts anymore, and these facilities
581 are implied anyway for normal services. syslog.target has
584 * There are new bus calls on PID1's Manager object for
585 cancelling jobs, and removing snapshot units. Previously,
586 both calls were only available on the Job and Snapshot
589 * systemd-journal-gatewayd gained SSL support.
591 * The various "environment" files, such as /etc/locale.conf
592 now support continuation lines with a backslash ("\") as
593 last character in the line, similar in style (but different)
594 to how this is supported in shells.
596 * For normal user processes the _SYSTEMD_USER_UNIT= field is
597 now implicitly appended to every log entry logged. systemctl
598 has been updated to filter by this field when operating on a
599 user systemd instance.
601 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
602 CAP_AUDIT_CONTROL capabilities to the capabilities set for
603 the container. This makes it easier to boot unmodified
604 Fedora systems in a container, which however still requires
605 audit=0 to be passed on the kernel command line. Auditing in
606 kernel and userspace is unfortunately still too broken in
607 context of containers, hence we recommend compiling it out
608 of the kernel or using audit=0. Hopefully this will be fixed
609 one day for good in the kernel.
611 * nspawn gained the new --bind= and --bind-ro= parameters to
612 bind mount specific directories from the host into the
615 * nspawn will now mount its own devpts file system instance
616 into the container, in order not to leak pty devices from
617 the host into the container.
619 * systemd will now read the firmware boot time performance
620 information from the EFI variables, if the used boot loader
621 supports this, and takes it into account for boot performance
622 analysis via "systemd-analyze". This is currently supported
623 only in conjunction with Gummiboot, but could be supported
624 by other boot loaders too. For details see:
626 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
628 * A new generator has been added that automatically mounts the
629 EFI System Partition (ESP) to /boot, if that directory
630 exists, is empty, and no other file system has been
631 configured to be mounted there.
633 * logind will now send out PrepareForSleep(false) out
634 unconditionally, after coming back from suspend. This may be
635 used by applications as asynchronous notification for
636 system resume events.
638 * "systemctl unlock-sessions" has been added, that allows
639 unlocking the screens of all user sessions at once, similar
640 how "systemctl lock-sessions" already locked all users
641 sessions. This is backed by a new D-Bus call UnlockSessions().
643 * "loginctl seat-status" will now show the master device of a
644 seat. (i.e. the device of a seat that needs to be around for
645 the seat to be considered available, usually the graphics
648 * tmpfiles gained a new "X" line type, that allows
649 configuration of files and directories (with wildcards) that
650 shall be excluded from automatic cleanup ("aging").
652 * udev default rules set the device node permissions now only
653 at "add" events, and do not change them any longer with a
654 later "change" event.
656 * The log messages for lid events and power/sleep keypresses
657 now carry a message ID.
659 * We now have a substantially larger unit test suite, but this
660 continues to be work in progress.
662 * udevadm hwdb gained a new --root= parameter to change the
663 root directory to operate relative to.
665 * logind will now issue a background sync() request to the kernel
666 early at shutdown, so that dirty buffers are flushed to disk early
667 instead of at the last moment, in order to optimize shutdown
670 * A new bootctl tool has been added that is an interface for
671 certain boot loader operations. This is currently a preview
672 and is likely to be extended into a small mechanism daemon
673 like timedated, localed, hostnamed, and can be used by
674 graphical UIs to enumerate available boot options, and
675 request boot into firmware operations.
677 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
678 the rest of the package. It also has been updated to work
679 correctly in initrds.
681 * Policykit previously has been runtime optional, and is now
682 also compile time optional via a configure switch.
684 * systemd-analyze has been reimplemented in C. Also "systemctl
685 dot" has moved into systemd-analyze.
687 * "systemctl status" with no further parameters will now print
688 the status of all active or failed units.
690 * Operations such as "systemctl start" can now be executed
691 with a new mode "--irreversible" which may be used to queue
692 operations that cannot accidentally be reversed by a later
693 job queuing. This is by default used to make shutdown
694 requests more robust.
696 * The Python API of systemd now gained a new module for
697 reading journal files.
699 * A new tool kernel-install has been added that can install
700 kernel images according to the Boot Loader Specification:
702 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
704 * Boot time console output has been improved to provide
705 animated boot time output for hanging jobs.
707 * A new tool systemd-activate has been added which can be used
708 to test socket activation with, directly from the command
709 line. This should make it much easier to test and debug
710 socket activation in daemons.
712 * journalctl gained a new "--reverse" (or -r) option to show
713 journal output in reverse order (i.e. newest line first).
715 * journalctl gained a new "--pager-end" (or -e) option to jump
716 to immediately jump to the end of the journal in the
717 pager. This is only supported in conjunction with "less".
719 * journalctl gained a new "--user-unit=" option, that works
720 similar to "--unit=" but filters for user units rather than
723 * A number of unit files to ease adoption of systemd in
724 initrds has been added. This moves some minimal logic from
725 the various initrd implementations into systemd proper.
727 * The journal files are now owned by a new group
728 "systemd-journal", which exists specifically to allow access
729 to the journal, and nothing else. Previously, we used the
730 "adm" group for that, which however possibly covers more
731 than just journal/log file access. This new group is now
732 already used by systemd-journal-gatewayd to ensure this
733 daemon gets access to the journal files and as little else
734 as possible. Note that "make install" will also set FS ACLs
735 up for /var/log/journal to give "adm" and "wheel" read
736 access to it, in addition to "systemd-journal" which owns
737 the journal files. We recommend that packaging scripts also
738 add read access to "adm" + "wheel" to /var/log/journal, and
739 all existing/future journal files. To normal users and
740 administrators little changes, however packagers need to
741 ensure to create the "systemd-journal" system group at
742 package installation time.
744 * The systemd-journal-gatewayd now runs as unprivileged user
745 systemd-journal-gateway:systemd-journal-gateway. Packaging
746 scripts need to create these system user/group at
749 * timedated now exposes a new boolean property CanNTP that
750 indicates whether a local NTP service is available or not.
752 * systemd-detect-virt will now also detect xen PVs
754 * The pstore file system is now mounted by default, if it is
757 * In addition to the SELinux and IMA policies we will now also
758 load SMACK policies at early boot.
760 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
761 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
762 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
763 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
764 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
765 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
766 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
767 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
768 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
769 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
770 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
771 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
772 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
773 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
777 * Timer units now support calendar time events in addition to
778 monotonic time events. That means you can now trigger a unit
779 based on a calendar time specification such as "Thu,Fri
780 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
781 or fifth day of any month of the year 2013, given that it is
782 a thursday or friday. This brings timer event support
783 considerably closer to cron's capabilities. For details on
784 the supported calendar time specification language see
787 * udev now supports a number of different naming policies for
788 network interfaces for predictable names, and a combination
789 of these policies is now the default. Please see this wiki
790 document for details:
792 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
794 * Auke Kok's bootchart implementation has been added to the
795 systemd tree. It's an optional component that can graph the
796 boot in quite some detail. It's one of the best bootchart
797 implementations around and minimal in its code and
800 * nss-myhostname has been integrated into the systemd source
801 tree. nss-myhostname guarantees that the local hostname
802 always stays resolvable via NSS. It has been a weak
803 requirement of systemd-hostnamed since a long time, and
804 since its code is actually trivial we decided to just
805 include it in systemd's source tree. It can be turned off
806 with a configure switch.
808 * The read-ahead logic is now capable of properly detecting
809 whether a btrfs file system is on SSD or rotating media, in
810 order to optimize the read-ahead scheme. Previously, it was
811 only capable of detecting this on traditional file systems
814 * In udev, additional device properties are now read from the
815 IAB in addition to the OUI database. Also, Bluetooth company
816 identities are attached to the devices as well.
818 * In service files %U may be used as specifier that is
819 replaced by the configured user name of the service.
821 * nspawn may now be invoked without a controlling TTY. This
822 makes it suitable for invocation as its own service. This
823 may be used to set up a simple containerized server system
824 using only core OS tools.
826 * systemd and nspawn can now accept socket file descriptors
827 when they are started for socket activation. This enables
828 implementation of socket activated nspawn
829 containers. i.e. think about autospawning an entire OS image
830 when the first SSH or HTTP connection is received. We expect
831 that similar functionality will also be added to libvirt-lxc
834 * journalctl will now suppress ANSI color codes when
837 * systemctl will no longer show control group information for
838 a unit if a the control group is empty anyway.
840 * logind can now automatically suspend/hibernate/shutdown the
843 * /etc/machine-info and hostnamed now also expose the chassis
844 type of the system. This can be used to determine whether
845 the local system is a laptop, desktop, handset or
846 tablet. This information may either be configured by the
847 user/vendor or is automatically determined from ACPI and DMI
848 information if possible.
850 * A number of PolicyKit actions are now bound together with
851 "imply" rules. This should simplify creating UIs because
852 many actions will now authenticate similar ones as well.
854 * Unit files learnt a new condition ConditionACPower= which
855 may be used to conditionalize a unit depending on whether an
856 AC power source is connected or not, of whether the system
857 is running on battery power.
859 * systemctl gained a new "is-failed" verb that may be used in
860 shell scripts and suchlike to check whether a specific unit
861 is in the "failed" state.
863 * The EnvironmentFile= setting in unit files now supports file
864 globbing, and can hence be used to easily read a number of
865 environment files at once.
867 * systemd will no longer detect and recognize specific
868 distributions. All distribution-specific #ifdeffery has been
869 removed, systemd is now fully generic and
870 distribution-agnostic. Effectively, not too much is lost as
871 a lot of the code is still accessible via explicit configure
872 switches. However, support for some distribution specific
873 legacy configuration file formats has been dropped. We
874 recommend distributions to simply adopt the configuration
875 files everybody else uses now and convert the old
876 configuration from packaging scripts. Most distributions
877 already did that. If that's not possible or desirable,
878 distributions are welcome to forward port the specific
879 pieces of code locally from the git history.
881 * When logging a message about a unit systemd will now always
882 log the unit name in the message meta data.
884 * localectl will now also discover system locale data that is
885 not stored in locale archives, but directly unpacked.
887 * logind will no longer unconditionally use framebuffer
888 devices as seat masters, i.e. as devices that are required
889 to be existing before a seat is considered preset. Instead,
890 it will now look for all devices that are tagged as
891 "seat-master" in udev. By default framebuffer devices will
892 be marked as such, but depending on local systems other
893 devices might be marked as well. This may be used to
894 integrate graphics cards using closed source drivers (such
895 as NVidia ones) more nicely into logind. Note however, that
896 we recommend using the open source NVidia drivers instead,
897 and no udev rules for the closed-source drivers will be
898 shipped from us upstream.
900 Contributions from: Adam Williamson, Alessandro Crismani, Auke
901 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
902 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
903 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
904 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
905 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
906 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
907 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
908 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
909 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
910 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
911 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
916 * udev gained support for loading additional device properties
917 from an indexed database that is keyed by vendor/product IDs
918 and similar device identifiers. For the beginning this
919 "hwdb" is populated with data from the well-known PCI and
920 USB database, but also includes PNP, ACPI and OID data. In
921 the longer run this indexed database shall grow into
922 becoming the one central database for non-essential
923 userspace device metadata. Previously, data from the PCI/USB
924 database was only attached to select devices, since the
925 lookup was a relatively expensive operation due to O(n) time
926 complexity (with n being the number of entries in the
927 database). Since this is now O(1), we decided to add in this
928 data for all devices where this is available, by
929 default. Note that the indexed database needs to be rebuilt
930 when new data files are installed. To achieve this you need
931 to update your packaging scripts to invoke "udevadm hwdb
932 --update" after installation of hwdb data files. For
933 RPM-based distributions we introduced the new
934 %udev_hwdb_update macro for this purpose.
936 * The Journal gained support for the "Message Catalog", an
937 indexed database to link up additional information with
938 journal entries. For further details please check:
940 http://www.freedesktop.org/wiki/Software/systemd/catalog
942 The indexed message catalog database also needs to be
943 rebuilt after installation of message catalog files. Use
944 "journalctl --update-catalog" for this. For RPM-based
945 distributions we introduced the %journal_catalog_update
946 macro for this purpose.
948 * The Python Journal bindings gained support for the standard
949 Python logging framework.
951 * The Journal API gained new functions for checking whether
952 the underlying file system of a journal file is capable of
953 properly reporting file change notifications, or whether
954 applications that want to reflect journal changes "live"
955 need to recheck journal files continuously in appropriate
958 * It is now possible to set the "age" field for tmpfiles
959 entries to 0, indicating that files matching this entry
960 shall always be removed when the directories are cleaned up.
962 * coredumpctl gained a new "gdb" verb which invokes gdb
963 right-away on the selected coredump.
965 * There's now support for "hybrid sleep" on kernels that
966 support this, in addition to "suspend" and "hibernate". Use
967 "systemctl hybrid-sleep" to make use of this.
969 * logind's HandleSuspendKey= setting (and related settings)
970 now gained support for a new "lock" setting to simply
971 request the screen lock on all local sessions, instead of
972 actually executing a suspend or hibernation.
974 * systemd will now mount the EFI variables file system by
977 * Socket units now gained support for configuration of the
978 SMACK security label.
980 * timedatectl will now output the time of the last and next
981 daylight saving change.
983 * We dropped support for various legacy and distro-specific
984 concepts, such as insserv, early-boot SysV services
985 (i.e. those for non-standard runlevels such as 'b' or 'S')
986 or ArchLinux /etc/rc.conf support. We recommend the
987 distributions who still need support this to either continue
988 to maintain the necessary patches downstream, or find a
989 different solution. (Talk to us if you have questions!)
991 * Various systemd components will now bypass PolicyKit checks
992 for root and otherwise handle properly if PolicyKit is not
993 found to be around. This should fix most issues for
994 PolicyKit-less systems. Quite frankly this should have been
995 this way since day one. It is absolutely our intention to
996 make systemd work fine on PolicyKit-less systems, and we
997 consider it a bug if something doesn't work as it should if
998 PolicyKit is not around.
1000 * For embedded systems it is now possible to build udev and
1001 systemd without blkid and/or kmod support.
1003 * "systemctl switch-root" is now capable of switching root
1004 more than once. I.e. in addition to transitions from the
1005 initrd to the host OS it is now possible to transition to
1006 further OS images from the host. This is useful to implement
1007 offline updating tools.
1009 * Various other additions have been made to the RPM macros
1010 shipped with systemd. Use %udev_rules_update() after
1011 installing new udev rules files. %_udevhwdbdir,
1012 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
1013 %_sysctldir are now available which resolve to the right
1014 directories for packages to place various data files in.
1016 * journalctl gained the new --full switch (in addition to
1017 --all, to disable ellipsation for long messages.
1019 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
1020 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
1021 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
1022 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
1023 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
1024 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
1025 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
1026 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
1027 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
1031 * journalctl gained new --since= and --until= switches to
1032 filter by time. It also now supports nice filtering for
1033 units via --unit=/-u.
1035 * Type=oneshot services may use ExecReload= and do the
1038 * The journal daemon now supports time-based rotation and
1039 vacuuming, in addition to the usual disk-space based
1042 * The journal will now index the available field values for
1043 each field name. This enables clients to show pretty drop
1044 downs of available match values when filtering. The bash
1045 completion of journalctl has been updated
1046 accordingly. journalctl gained a new switch -F to list all
1047 values a certain field takes in the journal database.
1049 * More service events are now written as structured messages
1050 to the journal, and made recognizable via message IDs.
1052 * The timedated, localed and hostnamed mini-services which
1053 previously only provided support for changing time, locale
1054 and hostname settings from graphical DEs such as GNOME now
1055 also have a minimal (but very useful) text-based client
1056 utility each. This is probably the nicest way to changing
1057 these settings from the command line now, especially since
1058 it lists available options and is fully integrated with bash
1061 * There's now a new tool "systemd-coredumpctl" to list and
1062 extract coredumps from the journal.
1064 * We now install a README each in /var/log/ and
1065 /etc/rc.d/init.d explaining where the system logs and init
1066 scripts went. This hopefully should help folks who go to
1067 that dirs and look into the otherwise now empty void and
1068 scratch their heads.
1070 * When user-services are invoked (by systemd --user) the
1071 $MANAGERPID env var is set to the PID of systemd.
1073 * SIGRTMIN+24 when sent to a --user instance will now result
1074 in immediate termination of systemd.
1076 * gatewayd received numerous feature additions such as a
1077 "follow" mode, for live syncing and filtering.
1079 * browse.html now allows filtering and showing detailed
1080 information on specific entries. Keyboard navigation and
1081 mouse screen support has been added.
1083 * gatewayd/journalctl now supports HTML5/JSON
1084 Server-Sent-Events as output.
1086 * The SysV init script compatibility logic will now
1087 heuristically determine whether a script supports the
1088 "reload" verb, and only then make this available as
1091 * "systemctl status --follow" has been removed, use "journalctl
1094 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
1095 have been removed since they are hardly useful to be
1098 * And I'd like to take the opportunity to specifically mention
1099 Zbigniew for his great contributions. Zbigniew, you rock!
1101 Contributions from: Andrew Eikum, Christian Hesse, Colin
1102 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
1103 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
1104 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
1105 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
1106 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
1107 Jędrzejewski-Szmek, Сковорода Никита Андреевич
1111 * If /etc/vconsole.conf is non-existent or empty we will no
1112 longer load any console font or key map at boot by
1113 default. Instead the kernel defaults will be left
1114 intact. This is definitely the right thing to do, as no
1115 configuration should mean no configuration, and hard-coding
1116 font names that are different on all archs is probably a bad
1117 idea. Also, the kernel default key map and font should be
1118 good enough for most cases anyway, and mostly identical to
1119 the userspace fonts/key maps we previously overloaded them
1120 with. If distributions want to continue to default to a
1121 non-kernel font or key map they should ship a default
1122 /etc/vconsole.conf with the appropriate contents.
1124 Contributions from: Colin Walters, Daniel J Walsh, Dave
1125 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
1126 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1130 * journalctl gained a new --cursor= switch to show entries
1131 starting from the specified location in the journal.
1133 * We now enforce a size limit on journal entry fields exported
1134 with "-o json" in journalctl. Fields larger than 4K will be
1135 assigned null. This can be turned off with --all.
1137 * An (optional) journal gateway daemon is now available as
1138 "systemd-journal-gatewayd.service". This service provides
1139 access to the journal via HTTP and JSON. This functionality
1140 will be used to implement live log synchronization in both
1141 pull and push modes, but has various other users too, such
1142 as easy log access for debugging of embedded devices. Right
1143 now it is already useful to retrieve the journal via HTTP:
1145 # systemctl start systemd-journal-gatewayd.service
1146 # wget http://localhost:19531/entries
1148 This will download the journal contents in a
1149 /var/log/messages compatible format. The same as JSON:
1151 # curl -H"Accept: application/json" http://localhost:19531/entries
1153 This service is also accessible via a web browser where a
1154 single static HTML5 app is served that uses the JSON logic
1155 to enable the user to do some basic browsing of the
1156 journal. This will be extended later on. Here's an example
1157 screenshot of this app in its current state:
1159 http://0pointer.de/public/journal-gatewayd
1161 Contributions from: Kay Sievers, Lennart Poettering, Robert
1162 Milasan, Tom Gundersen
1166 * The bash completion logic is now available for journalctl
1169 * We don't mount the "cpuset" controller anymore together with
1170 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
1171 started if no parameters are assigned to it. "cpuset" hence
1172 broke code that assumed it it could create "cpu" groups and
1175 * journalctl -f will now subscribe to terminal size changes,
1176 and line break accordingly.
1178 Contributions from: Dave Reisner, Kay Sievers, Lennart
1179 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
1183 * nspawn will now create a symlink /etc/localtime in the
1184 container environment, copying the host's timezone
1185 setting. Previously this has been done via a bind mount, but
1186 since symlinks cannot be bind mounted this has now been
1187 changed to create/update the appropriate symlink.
1189 * journalctl -n's line number argument is now optional, and
1190 will default to 10 if omitted.
1192 * journald will now log the maximum size the journal files may
1193 take up on disk. This is particularly useful if the default
1194 built-in logic of determining this parameter from the file
1195 system size is used. Use "systemctl status
1196 systemd-journald.service" to see this information.
1198 * The multi-seat X wrapper tool has been stripped down. As X
1199 is now capable of enumerating graphics devices via udev in a
1200 seat-aware way the wrapper is not strictly necessary
1201 anymore. A stripped down temporary stop-gap is still shipped
1202 until the upstream display managers have been updated to
1203 fully support the new X logic. Expect this wrapper to be
1204 removed entirely in one of the next releases.
1206 * HandleSleepKey= in logind.conf has been split up into
1207 HandleSuspendKey= and HandleHibernateKey=. The old setting
1208 is not available anymore. X11 and the kernel are
1209 distuingishing between these keys and we should too. This
1210 also means the inhibition lock for these keys has been split
1213 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
1214 Poettering, Lukas Nykryn, Václav Pavlín
1218 * Whenever a unit changes state we'll now log this to the
1219 journal and show along the unit's own log output in
1222 * ConditionPathIsMountPoint= can now properly detect bind
1223 mount points too. (Previously, a bind mount of one file
1224 system to another place in the same file system could not be
1225 detected as mount, since they shared struct stat's st_dev
1228 * We will now mount the cgroup controllers cpu, cpuacct,
1229 cpuset and the controllers net_cls, net_prio together by
1232 * nspawn containers will now have a virtualized boot
1233 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
1234 over with a randomized ID at container initialization). This
1235 has the effect of making "journalctl -b" do the right thing
1238 * The JSON output journal serialization has been updated not
1239 to generate "endless" list objects anymore, but rather one
1240 JSON object per line. This is more in line how most JSON
1241 parsers expect JSON objects. The new output mode
1242 "json-pretty" has been added to provide similar output, but
1243 neatly aligned for readability by humans.
1245 * We dropped all explicit sync() invocations in the shutdown
1246 code. The kernel does this implicitly anyway in the kernel
1247 reboot() syscall. halt(8)'s -n option is now a compatibility
1250 * We now support virtualized reboot() in containers, as
1251 supported by newer kernels. We will fall back to exit() if
1252 CAP_SYS_REBOOT is not available to the container. Also,
1253 nspawn makes use of this now and will actually reboot the
1254 container if the containerized OS asks for that.
1256 * journalctl will only show local log output by default
1257 now. Use --merge (-m) to show remote log output, too.
1259 * libsystemd-journal gained the new sd_journal_get_usage()
1260 call to determine the current disk usage of all journal
1261 files. This is exposed in the new "journalctl --disk-usage"
1264 * journald gained a new configuration setting SplitMode= in
1265 journald.conf which may be used to control how user journals
1266 are split off. See journald.conf(5) for details.
1268 * A new condition type ConditionFileNotEmpty= has been added.
1270 * tmpfiles' "w" lines now support file globbing, to write
1271 multiple files at once.
1273 * We added Python bindings for the journal submission
1274 APIs. More Python APIs for a number of selected APIs will
1275 likely follow. Note that we intend to add native bindings
1276 only for the Python language, as we consider it common
1277 enough to deserve bindings shipped within systemd. There are
1278 various projects outside of systemd that provide bindings
1279 for languages such as PHP or Lua.
1281 * Many conditions will now resolve specifiers such as %i. In
1282 addition, PathChanged= and related directives of .path units
1283 now support specifiers as well.
1285 * There's now a new RPM macro definition for the system preset
1288 * journald will now warn if it can't forward a message to the
1289 syslog daemon because it's socket is full.
1291 * timedated will no longer write or process /etc/timezone,
1292 except on Debian. As we do not support late mounted /usr
1293 anymore /etc/localtime always being a symlink is now safe,
1294 and hence the information in /etc/timezone is not necessary
1297 * logind will now always reserve one VT for a text getty (VT6
1298 by default). Previously if more than 6 X sessions where
1299 started they took up all the VTs with auto-spawned gettys,
1300 so that no text gettys were available anymore.
1302 * udev will now automatically inform the btrfs kernel logic
1303 about btrfs RAID components showing up. This should make
1304 simple hotplug based btrfs RAID assembly work.
1306 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1307 (but not for its children which will stay at the kernel
1308 default). This should allow setups with a lot more listening
1311 * systemd will now always pass the configured timezone to the
1312 kernel at boot. timedated will do the same when the timezone
1315 * logind's inhibition logic has been updated. By default,
1316 logind will now handle the lid switch, the power and sleep
1317 keys all the time, even in graphical sessions. If DEs want
1318 to handle these events on their own they should take the new
1319 handle-power-key, handle-sleep-key and handle-lid-switch
1320 inhibitors during their runtime. A simple way to achiveve
1321 that is to invoke the DE wrapped in an invocation of:
1323 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1325 * Access to unit operations is now checked via SELinux taking
1326 the unit file label and client process label into account.
1328 * systemd will now notify the administrator in the journal
1329 when he over-mounts a non-empty directory.
1331 * There are new specifiers that are resolved in unit files,
1332 for the host name (%H), the machine ID (%m) and the boot ID
1335 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
1336 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
1337 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
1338 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
1339 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
1340 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
1341 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
1345 * Support for reading structured kernel messages from
1346 /dev/kmsg has now been added and is enabled by default.
1348 * Support for reading kernel messages from /proc/kmsg has now
1349 been removed. If you want kernel messages in the journal
1350 make sure to run a recent kernel (>= 3.5) that supports
1351 reading structured messages from /dev/kmsg (see
1352 above). /proc/kmsg is now exclusive property of classic
1353 syslog daemons again.
1355 * The libudev API gained the new
1356 udev_device_new_from_device_id() call.
1358 * The logic for file system namespace (ReadOnlyDirectory=,
1359 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
1360 require pivot_root() anymore. This means fewer temporary
1361 directories are created below /tmp for this feature.
1363 * nspawn containers will now see and receive all submounts
1364 made on the host OS below the root file system of the
1367 * Forward Secure Sealing is now supported for Journal files,
1368 which provide cryptographical sealing of journal files so
1369 that attackers cannot alter log history anymore without this
1370 being detectable. Lennart will soon post a blog story about
1371 this explaining it in more detail.
1373 * There are two new service settings RestartPreventExitStatus=
1374 and SuccessExitStatus= which allow configuration of exit
1375 status (exit code or signal) which will be excepted from the
1376 restart logic, resp. consider successful.
1378 * journalctl gained the new --verify switch that can be used
1379 to check the integrity of the structure of journal files and
1380 (if Forward Secure Sealing is enabled) the contents of
1383 * nspawn containers will now be run with /dev/stdin, /dev/fd/
1384 and similar symlinks pre-created. This makes running shells
1385 as container init process a lot more fun.
1387 * The fstab support can now handle PARTUUID= and PARTLABEL=
1390 * A new ConditionHost= condition has been added to match
1391 against the hostname (with globs) and machine ID. This is
1392 useful for clusters where a single OS image is used to
1393 provision a large number of hosts which shall run slightly
1394 different sets of services.
1396 * Services which hit the restart limit will now be placed in a
1399 Contributions from: Bertram Poettering, Dave Reisner, Huang
1400 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
1401 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
1405 * When running in --user mode systemd will now become a
1406 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
1407 tree a lot more organized.
1409 * A new PartOf= unit dependency type has been introduced that
1410 may be used to group services in a natural way.
1412 * "systemctl enable" may now be used to enable instances of
1415 * journalctl now prints error log levels in red, and
1416 warning/notice log levels in bright white. It also supports
1417 filtering by log level now.
1419 * cgtop gained a new -n switch (similar to top), to configure
1420 the maximum number of iterations to run for. It also gained
1421 -b, to run in batch mode (accepting no input).
1423 * The suffix ".service" may now be omitted on most systemctl
1424 command lines involving service unit names.
1426 * There's a new bus call in logind to lock all sessions, as
1427 well as a loginctl verb for it "lock-sessions".
1429 * libsystemd-logind.so gained a new call sd_journal_perror()
1430 that works similar to libc perror() but logs to the journal
1431 and encodes structured information about the error number.
1433 * /etc/crypttab entries now understand the new keyfile-size=
1436 * shutdown(8) now can send a (configurable) wall message when
1437 a shutdown is cancelled.
1439 * The mount propagation mode for the root file system will now
1440 default to "shared", which is useful to make containers work
1441 nicely out-of-the-box so that they receive new mounts from
1442 the host. This can be undone locally by running "mount
1443 --make-rprivate /" if needed.
1445 * The prefdm.service file has been removed. Distributions
1446 should maintain this unit downstream if they intend to keep
1447 it around. However, we recommend writing normal unit files
1448 for display managers instead.
1450 * Since systemd is a crucial part of the OS we will now
1451 default to a number of compiler switches that improve
1452 security (hardening) such as read-only relocations, stack
1453 protection, and suchlike.
1455 * The TimeoutSec= setting for services is now split into
1456 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1457 of individual time outs for the start and the stop phase of
1460 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1461 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1462 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1463 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1464 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1465 Gundersen, Zbigniew Jędrzejewski-Szmek
1469 * The journal and id128 C APIs are now fully documented as man
1472 * Extra safety checks have been added when transitioning from
1473 the initial RAM disk to the main system to avoid accidental
1476 * /etc/crypttab entries now understand the new keyfile-offset=
1479 * systemctl -t can now be used to filter by unit load state.
1481 * The journal C API gained the new sd_journal_wait() call to
1482 make writing synchronous journal clients easier.
1484 * journalctl gained the new -D switch to show journals from a
1487 * journalctl now displays a special marker between log
1488 messages of two different boots.
1490 * The journal is now explicitly flushed to /var via a service
1491 systemd-journal-flush.service, rather than implicitly simply
1492 by seeing /var/log/journal to be writable.
1494 * journalctl (and the journal C APIs) can now match for much
1495 more complex expressions, with alternatives and
1498 * When transitioning from the initial RAM disk to the main
1499 system we will now kill all processes in a killing spree to
1500 ensure no processes stay around by accident.
1502 * Three new specifiers may be used in unit files: %u, %h, %s
1503 resolve to the user name, user home directory resp. user
1504 shell. This is useful for running systemd user instances.
1506 * We now automatically rotate journal files if their data
1507 object hash table gets a fill level > 75%. We also size the
1508 hash table based on the configured maximum file size. This
1509 together should lower hash collisions drastically and thus
1510 speed things up a bit.
1512 * journalctl gained the new "--header" switch to introspect
1513 header data of journal files.
1515 * A new setting SystemCallFilters= has been added to services
1516 which may be used to apply blacklists or whitelists to
1517 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1519 * nspawn gained a new --link-journal= switch (and quicker: -j)
1520 to link the container journal with the host. This makes it
1521 very easy to centralize log viewing on the host for all
1522 guests while still keeping the journal files separated.
1524 * Many bugfixes and optimizations
1526 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1527 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1528 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1533 * Several tools now understand kernel command line arguments,
1534 which are only read when run in an initial RAM disk. They
1535 usually follow closely their normal counterparts, but are
1538 * There's a new tool to analyze the readahead files that are
1539 automatically generated at boot. Use:
1541 /usr/lib/systemd/systemd-readahead analyze /.readahead
1543 * We now provide an early debug shell on tty9 if this enabled. Use:
1545 systemctl enable debug-shell.service
1547 * All plymouth related units have been moved into the Plymouth
1548 package. Please make sure to upgrade your Plymouth version
1551 * systemd-tmpfiles now supports getting passed the basename of
1552 a configuration file only, in which case it will look for it
1553 in all appropriate directories automatically.
1555 * udevadm info now takes a /dev or /sys path as argument, and
1556 does the right thing. Example:
1558 udevadm info /dev/sda
1559 udevadm info /sys/class/block/sda
1561 * systemctl now prints a warning if a unit is stopped but a
1562 unit that might trigger it continues to run. Example: a
1563 service is stopped but the socket that activates it is left
1566 * "systemctl status" will now mention if the log output was
1567 shortened due to rotation since a service has been started.
1569 * The journal API now exposes functions to determine the
1570 "cutoff" times due to rotation.
1572 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1573 immediately flushing of runtime logs to /var if possible,
1574 resp. for triggering immediate rotation of the journal
1577 * It is now considered an error if a service is attempted to
1578 be stopped that is not loaded.
1580 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1582 * systemd-analyze now supports Python 3
1584 * tmpfiles now supports cleaning up directories via aging
1585 where the first level dirs are always kept around but
1586 directories beneath it automatically aged. This is enabled
1587 by prefixing the age field with '~'.
1589 * Seat objects now expose CanGraphical, CanTTY properties
1590 which is required to deal with very fast bootups where the
1591 display manager might be running before the graphics drivers
1592 completed initialization.
1594 * Seat objects now expose a State property.
1596 * We now include RPM macros for service enabling/disabling
1597 based on the preset logic. We recommend RPM based
1598 distributions to make use of these macros if possible. This
1599 makes it simpler to reuse RPM spec files across
1602 * We now make sure that the collected systemd unit name is
1603 always valid when services log to the journal via
1606 * There's a new man page kernel-command-line(7) detailing all
1607 command line options we understand.
1609 * The fstab generator may now be disabled at boot by passing
1610 fstab=0 on the kernel command line.
1612 * A new kernel command line option modules-load= is now understood
1613 to load a specific kernel module statically, early at boot.
1615 * Unit names specified on the systemctl command line are now
1616 automatically escaped as needed. Also, if file system or
1617 device paths are specified they are automatically turned
1618 into the appropriate mount or device unit names. Example:
1620 systemctl status /home
1621 systemctl status /dev/sda
1623 * The SysVConsole= configuration option has been removed from
1624 system.conf parsing.
1626 * The SysV search path is no longer exported on the D-Bus
1629 * The Names= option is been removed from unit file parsing.
1631 * There's a new man page bootup(7) detailing the boot process.
1633 * Every unit and every generator we ship with systemd now
1634 comes with full documentation. The self-explanatory boot is
1637 * A couple of services gained "systemd-" prefixes in their
1638 name if they wrap systemd code, rather than only external
1639 code. Among them fsck@.service which is now
1640 systemd-fsck@.service.
1642 * The HaveWatchdog property has been removed from the D-Bus
1645 * systemd.confirm_spawn= on the kernel command line should now
1648 * There's a new man page crypttab(5) which details all options
1649 we actually understand.
1651 * systemd-nspawn gained a new --capability= switch to pass
1652 additional capabilities to the container.
1654 * timedated will now read known NTP implementation unit names
1655 from /usr/lib/systemd/ntp-units.d/*.list,
1656 systemd-timedated-ntp.target has been removed.
1658 * journalctl gained a new switch "-b" that lists log data of
1659 the current boot only.
1661 * The notify socket is in the abstract namespace again, in
1662 order to support daemons which chroot() at start-up.
1664 * There is a new Storage= configuration option for journald
1665 which allows configuration of where log data should go. This
1666 also provides a way to disable journal logging entirely, so
1667 that data collected is only forwarded to the console, the
1668 kernel log buffer or another syslog implementation.
1670 * Many bugfixes and optimizations
1672 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1673 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1674 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1675 Shawn Landden, Tom Gundersen
1679 * "systemctl help <unit>" now shows the man page if one is
1682 * Several new man pages have been added.
1684 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1685 MaxLevelConsole= can now be specified in
1686 journald.conf. These options allow reducing the amount of
1687 data stored on disk or forwarded by the log level.
1689 * TimerSlackNSec= can now be specified in system.conf for
1690 PID1. This allows system-wide power savings.
1692 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1693 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1698 * logind is now capable of (optionally) handling power and
1699 sleep keys as well as the lid switch.
1701 * journalctl now understands the syntax "journalctl
1702 /usr/bin/avahi-daemon" to get all log output of a specific
1705 * CapabilityBoundingSet= in system.conf now also influences
1706 the capability bound set of usermode helpers of the kernel.
1708 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1709 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1710 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1711 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1715 * Note that we skipped 139 releases here in order to set the
1716 new version to something that is greater than both udev's
1717 and systemd's most recent version number.
1719 * udev: all udev sources are merged into the systemd source tree now.
1720 All future udev development will happen in the systemd tree. It
1721 is still fully supported to use the udev daemon and tools without
1722 systemd running, like in initramfs or other init systems. Building
1723 udev though, will require the *build* of the systemd tree, but
1724 udev can be properly *run* without systemd.
1726 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1727 should be used to create dead device nodes as workarounds for broken
1730 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1731 no longer supported. udev_monitor_new_from_netlink() needs to be
1732 used to subscribe to events.
1734 * udev: when udevd is started by systemd, processes which are left
1735 behind by forking them off of udev rules, are unconditionally cleaned
1736 up and killed now after the event handling has finished. Services or
1737 daemons must be started as systemd services. Services can be
1738 pulled-in by udev to get started, but they can no longer be directly
1739 forked by udev rules.
1741 * udev: the daemon binary is called systemd-udevd now and installed
1742 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1743 to adapt to that, create symlink, or rename the binary after building
1746 * libudev no longer provides these symbols:
1747 udev_monitor_from_socket()
1748 udev_queue_get_failed_list_entry()
1749 udev_get_{dev,sys,run}_path()
1750 The versions number was bumped and symbol versioning introduced.
1752 * systemd-loginctl and systemd-journalctl have been renamed
1753 to loginctl and journalctl to match systemctl.
1755 * The config files: /etc/systemd/systemd-logind.conf and
1756 /etc/systemd/systemd-journald.conf have been renamed to
1757 logind.conf and journald.conf. Package updates should rename
1758 the files to the new names on upgrade.
1760 * For almost all files the license is now LGPL2.1+, changed
1761 from the previous GPL2.0+. Exceptions are some minor stuff
1762 of udev (which will be changed to LGPL2.1 eventually, too),
1763 and the MIT licensed sd-daemon.[ch] library that is suitable
1764 to be used as drop-in files.
1766 * systemd and logind now handle system sleep states, in
1767 particular suspending and hibernating.
1769 * logind now implements a sleep/shutdown/idle inhibiting logic
1770 suitable for a variety of uses. Soonishly Lennart will blog
1771 about this in more detail.
1773 * var-run.mount and var-lock.mount are no longer provided
1774 (which prevously bind mounted these directories to their new
1775 places). Distributions which have not converted these
1776 directories to symlinks should consider stealing these files
1777 from git history and add them downstream.
1779 * We introduced the Documentation= field for units and added
1780 this to all our shipped units. This is useful to make it
1781 easier to explore the boot and the purpose of the various
1784 * All smaller setup units (such as
1785 systemd-vconsole-setup.service) now detect properly if they
1786 are run in a container and are skipped when
1787 appropriate. This guarantees an entirely noise-free boot in
1788 Linux container environments such as systemd-nspawn.
1790 * A framework for implementing offline system updates is now
1791 integrated, for details see:
1792 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1794 * A new service type Type=idle is available now which helps us
1795 avoiding ugly interleaving of getty output and boot status
1798 * There's now a system-wide CapabilityBoundingSet= option to
1799 globally reduce the set of capabilities for the
1800 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1801 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1802 even CAP_NET_ADMIN system-wide for secure systems.
1804 * There are now system-wide DefaultLimitXXX= options to
1805 globally change the defaults of the various resource limits
1806 for all units started by PID 1.
1808 * Harald Hoyer's systemd test suite has been integrated into
1809 systemd which allows easy testing of systemd builds in qemu
1810 and nspawn. (This is really awesome! Ask us for details!)
1812 * The fstab parser is now implemented as generator, not inside
1815 * systemctl will now warn you if .mount units generated from
1816 /etc/fstab are out of date due to changes in fstab that
1817 haven't been read by systemd yet.
1819 * systemd is now suitable for usage in initrds. Dracut has
1820 already been updated to make use of this. With this in place
1821 initrds get a slight bit faster but primarily are much
1822 easier to introspect and debug since "systemctl status" in
1823 the host system can be used to introspect initrd services,
1824 and the journal from the initrd is kept around too.
1826 * systemd-delta has been added, a tool to explore differences
1827 between user/admin configuration and vendor defaults.
1829 * PrivateTmp= now affects both /tmp and /var/tmp.
1831 * Boot time status messages are now much prettier and feature
1832 proper english language. Booting up systemd has never been
1835 * Read-ahead pack files now include the inode number of all
1836 files to pre-cache. When the inode changes the pre-caching
1837 is not attempted. This should be nicer to deal with updated
1838 packages which might result in changes of read-ahead
1841 * We now temporaritly lower the kernel's read_ahead_kb variable
1842 when collecting read-ahead data to ensure the kernel's
1843 built-in read-ahead does not add noise to our measurements
1844 of necessary blocks to pre-cache.
1846 * There's now RequiresMountsFor= to add automatic dependencies
1847 for all mounts necessary for a specific file system path.
1849 * MountAuto= and SwapAuto= have been removed from
1850 system.conf. Mounting file systems at boot has to take place
1853 * nspawn now learned a new switch --uuid= to set the machine
1854 ID on the command line.
1856 * nspawn now learned the -b switch to automatically search
1859 * vt102 is now the default TERM for serial TTYs, upgraded from
1862 * systemd-logind now works on VT-less systems.
1864 * The build tree has been reorganized. The individual
1865 components now have directories of their own.
1867 * A new condition type ConditionPathIsReadWrite= is now available.
1869 * nspawn learned the new -C switch to create cgroups for the
1870 container in other hierarchies.
1872 * We now have support for hardware watchdogs, configurable in
1875 * The scheduled shutdown logic now has a public API.
1877 * We now mount /tmp as tmpfs by default, but this can be
1878 masked and /etc/fstab can override it.
1880 * Since udisks doesn't make use of /media anymore we are not
1881 mounting a tmpfs on it anymore.
1883 * journalctl gained a new --local switch to only interleave
1884 locally generated journal files.
1886 * We can now load the IMA policy at boot automatically.
1888 * The GTK tools have been split off into a systemd-ui.
1890 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1891 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1892 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1893 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1894 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1895 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1896 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1897 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1898 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1903 * This is mostly a bugfix release
1905 * Support optional initialization of the machine ID from the
1906 KVM or container configured UUID.
1908 * Support immediate reboots with "systemctl reboot -ff"
1910 * Show /etc/os-release data in systemd-analyze output
1912 * Many bugfixes for the journal, including endianness fixes and
1913 ensuring that disk space enforcement works
1915 * sd-login.h is C++ comptaible again
1917 * Extend the /etc/os-release format on request of the Debian
1920 * We now refuse non-UTF8 strings used in various configuration
1921 and unit files. This is done to ensure we don't pass invalid
1922 data over D-Bus or expose it elsewhere.
1924 * Register Mimo USB Screens as suitable for automatic seat
1927 * Read SELinux client context from journal clients in a race
1930 * Reorder configuration file lookup order. /etc now always
1931 overrides /run in order to allow the administrator to always
1932 and unconditionally override vendor supplied or
1933 automatically generated data.
1935 * The various user visible bits of the journal now have man
1936 pages. We still lack man pages for the journal API calls
1939 * We now ship all man pages in HTML format again in the
1942 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1943 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1944 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1949 * This is mostly a bugfix release
1951 * systems lacking /etc/os-release are no longer supported.
1953 * Various functionality updates to libsystemd-login.so
1955 * Track class of PAM logins to distuingish greeters from
1958 Contributions from: Kay Sievers, Lennart Poettering, Michael
1963 * This is an important bugfix release for v41.
1965 * Building man pages is now optional which should be useful
1966 for those building systemd from git but unwilling to install
1969 * Watchdog support for supervising services is now usable. In
1970 a future release support for hardware watchdogs
1971 (i.e. /dev/watchdog) will be added building on this.
1973 * Service start rate limiting is now configurable and can be
1974 turned off per service. When a start rate limit is hit a
1975 reboot can automatically be triggered.
1977 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1979 Contributions from: Benjamin Franzke, Bill Nottingham,
1980 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1981 Schmidt, Michał Górny, Piotr Drąg
1985 * The systemd binary is installed /usr/lib/systemd/systemd now;
1986 An existing /sbin/init symlink needs to be adapted with the
1989 * The code that loads kernel modules has been ported to invoke
1990 libkmod directly, instead of modprobe. This means we do not
1991 support systems with module-init-tools anymore.
1993 * Watchdog support is now already useful, but still not
1996 * A new kernel command line option systemd.setenv= is
1997 understood to set system wide environment variables
1998 dynamically at boot.
2000 * We now limit the set of capabilities of systemd-journald.
2002 * We now set SIGPIPE to ignore by default, since it only is
2003 useful in shell pipelines, and has little use in general
2004 code. This can be disabled with IgnoreSIPIPE=no in unit
2007 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
2008 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
2013 * This is mostly a bugfix release
2015 * We now expose the reason why a service failed in the
2016 "Result" D-Bus property.
2018 * Rudimentary service watchdog support (will be completed over
2019 the next few releases.)
2021 * When systemd forks off in order execute some service we will
2022 now immediately changes its argv[0] to reflect which process
2023 it will execute. This is useful to minimize the time window
2024 with a generic argv[0], which makes bootcharts more useful
2026 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
2027 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
2028 Mike Kazantsev, Ray Strode
2032 * This is mostly a test release, but incorporates many
2035 * New systemd-cgtop tool to show control groups by their
2038 * Linking against libacl for ACLs is optional again. If
2039 disabled, support tracking device access for active logins
2040 goes becomes unavailable, and so does access to the user
2041 journals by the respective users.
2043 * If a group "adm" exists, journal files are automatically
2044 owned by them, thus allow members of this group full access
2045 to the system journal as well as all user journals.
2047 * The journal now stores the SELinux context of the logging
2048 client for all entries.
2050 * Add C++ inclusion guards to all public headers
2052 * New output mode "cat" in the journal to print only text
2053 messages, without any meta data like date or time.
2055 * Include tiny X server wrapper as a temporary stop-gap to
2056 teach XOrg udev display enumeration. This is used by display
2057 managers such as gdm, and will go away as soon as XOrg
2058 learned native udev hotplugging for display devices.
2060 * Add new systemd-cat tool for executing arbitrary programs
2061 with STDERR/STDOUT connected to the journal. Can also act as
2062 BSD logger replacement, and does so by default.
2064 * Optionally store all locally generated coredumps in the
2065 journal along with meta data.
2067 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
2068 writing short strings to files (for usage for /sys), and for
2069 creating symlinks, character and block device nodes.
2071 * New unit file option ControlGroupPersistent= to make cgroups
2072 persistent, following the mechanisms outlined in
2073 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
2075 * Support multiple local RTCs in a sane way
2077 * No longer monopolize IO when replaying readahead data on
2078 rotating disks, since we might starve non-file-system IO to
2079 death, since fanotify() will not see accesses done by blkid,
2082 * Don't show kernel threads in systemd-cgls anymore, unless
2083 requested with new -k switch.
2085 Contributions from: Dan Horák, Kay Sievers, Lennart
2086 Poettering, Michal Schmidt
2090 * This is mostly a test release, but incorporates many
2093 * The git repository moved to:
2094 git://anongit.freedesktop.org/systemd/systemd
2095 ssh://git.freedesktop.org/git/systemd/systemd
2097 * First release with the journal
2098 http://0pointer.de/blog/projects/the-journal.html
2100 * The journal replaces both systemd-kmsg-syslogd and
2101 systemd-stdout-bridge.
2103 * New sd_pid_get_unit() API call in libsystemd-logind
2105 * Many systemadm clean-ups
2107 * Introduce remote-fs-pre.target which is ordered before all
2108 remote mounts and may be used to start services before all
2111 * Added Mageia support
2113 * Add bash completion for systemd-loginctl
2115 * Actively monitor PID file creation for daemons which exit in
2116 the parent process before having finished writing the PID
2117 file in the daemon process. Daemons which do this need to be
2118 fixed (i.e. PID file creation must have finished before the
2119 parent exits), but we now react a bit more gracefully to them.
2121 * Add colourful boot output, mimicking the well-known output
2122 of existing distributions.
2124 * New option PassCredentials= for socket units, for
2125 compatibility with a recent kernel ABI breakage.
2127 * /etc/rc.local is now hooked in via a generator binary, and
2128 thus will no longer act as synchronization point during
2131 * systemctl list-unit-files now supports --root=.
2133 * systemd-tmpfiles now understands two new commands: z, Z for
2134 relabelling files according to the SELinux database. This is
2135 useful to apply SELinux labels to specific files in /sys,
2138 * Output of SysV services is now forwarded to both the console
2139 and the journal by default, not only just the console.
2141 * New man pages for all APIs from libsystemd-login.
2143 * The build tree got reorganized and a the build system is a
2144 lot more modular allowing embedded setups to specifically
2145 select the components of systemd they are interested in.
2147 * Support for Linux systems lacking the kernel VT subsystem is
2150 * configure's --with-rootdir= got renamed to
2151 --with-rootprefix= to follow the naming used by udev and
2154 * Unless specified otherwise we'll now install to /usr instead
2155 of /usr/local by default.
2157 * Processes with '@' in argv[0][0] are now excluded from the
2158 final shut-down killing spree, following the logic explained
2160 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
2162 * All processes remaining in a service cgroup when we enter
2163 the START or START_PRE states are now killed with
2164 SIGKILL. That means it is no longer possible to spawn
2165 background processes from ExecStart= lines (which was never
2166 supported anyway, and bad style).
2168 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
2169 reloading of units together.
2171 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
2172 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
2173 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
2174 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
2175 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek