1 systemd System and Service Manager
5 * The bash completion logic is now available for journalctl
8 * We don't mount the "cpuset" controller anymore together with
9 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
10 started if no parameters are assigned to it. "cpuset" hence
11 broke code that assumed it it could create "cpu" groups and
14 * journalctl -f will now subscribe to terminal size changes,
15 and line break accordingly.
17 Changes from: Dave Reisner, Kay Sievers, Lennart Poettering,
18 Lukas Nykrynm, Mirco Tischler, Václav Pavlín
22 * nspawn will now create a symlink /etc/localtime in the
23 container environment, copying the host's timezone
24 setting. Previously this has been done via a bind mount, but
25 since symlinks cannot be bind mounted this has now been
26 changed to create/update the appropriate symlink.
28 * journalctl -n's line number argument is now optional, and
29 will default to 10 if omitted.
31 * journald will now log the maximum size the journal files may
32 take up on disk. This is particularly useful if the default
33 built-in logic of determining this parameter from the file
34 system size is used. Use "systemctl status
35 systemd-journald.service" to see this information.
37 * The multi-seat X wrapper tool has been stripped down. As X
38 is now capable of enumerating graphics devices via udev in a
39 seat-aware way the wrapper is not strictly necessary
40 anymore. A stripped down temporary stop-gap is still shipped
41 until the upstream display managers have been updated to
42 fully support the new X logic. Expect this wrapper to be
43 removed entirely in one of the next releases.
45 * HandleSleepKey= in logind.conf has been split up into
46 HandleSuspendKey= and HandleHibernateKey=. The old setting
47 is not available anymore. X11 and the kernel are
48 distuingishing between these keys and we should too. This
49 also means the inhibition lock for these keys has been split
52 Changes from: Dave Airlie, Eelco Dolstra, Lennart Poettering,
53 Lukas Nykryn, Václav Pavlín
57 * Whenever a unit changes state we'll now log this to the
58 journal and show along the unit's own log output in
61 * ConditionPathIsMountPoint= can now properly detect bind
62 mount points too. (Previously, a bind mount of one file
63 system to another place in the same file system could not be
64 detected as mount, since they shared struct stat's st_dev
67 * We will now mount the cgroup controllers cpu, cpuacct,
68 cpuset and the controllers net_cls, net_prio together by
71 * nspawn containers will now have a virtualized boot
72 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
73 over with a randomized ID at container initialization). This
74 has the effect of making "journalctl -b" do the right thing
77 * The JSON output journal serialization has been updated not
78 to generate "endless" list objects anymore, but rather one
79 JSON object per line. This is more in line how most JSON
80 parsers expect JSON objects. The new output mode
81 "json-pretty" has been added to provide similar output, but
82 neatly aligned for readability by humans.
84 * We dropped all explicit sync() invocations in the shutdown
85 code. The kernel does this implicitly anyway in the kernel
86 reboot() syscall. halt(8)'s -n option is now a compatibility
89 * We now support virtualized reboot() in containers, as
90 supported by newer kernels. We will fall back to exit() if
91 CAP_SYS_REBOOT is not available to the container. Also,
92 nspawn makes use of this now and will actually reboot the
93 container if the containerized OS asks for that.
95 * journalctl will only show local log output by default
96 now. Use --merge (-m) to show remote log output, too.
98 * libsystemd-journal gained the new sd_journal_get_usage()
99 call to determine the current disk usage of all journal
100 files. This is exposed in the new "journalctl --disk-usage"
103 * journald gained a new configuration setting SplitMode= in
104 journald.conf which may be used to control how user journals
105 are split off. See journald.conf(5) for details.
107 * A new condition type ConditionFileNotEmpty= has been added.
109 * tmpfiles' "w" lines now support file globbing, to write
110 multiple files at once.
112 * We added Python bindings for the journal submission
113 APIs. More Python APIs for a number of selected APIs will
114 likely follow. Note that we intend to add native bindings
115 only for the Python language, as we consider it common
116 enough to deserve bindings shipped within systemd. There are
117 various projects outside of systemd that provide bindings
118 for languages such as PHP or Lua.
120 * Many conditions will now resolve specifiers such as %i. In
121 addition, PathChanged= and related directives of .path units
122 now support specifiers as well.
124 * There's now a new RPM macro definition for the system preset
127 * journald will now warn if it can't foward a message to the
128 syslog daemon because it's socket is full.
130 * timedated will no longer write or process /etc/timezone,
131 except on Debian. As we do not support late mounted /usr
132 anymore /etc/localtime always being a symlink is now safe,
133 and hence the information in /etc/timezone is not necessary
136 * logind will now always reserve one VT for a text getty (VT6
137 by default). Previously if more than 6 X sessions where
138 started they took up all the VTs with auto-spawned gettys,
139 so that no text gettys were available anymore.
141 * udev will now automatically inform the btrfs kernel logic
142 about btrfs RAID components showing up. This should make
143 simple hotplug based btrfs RAID assembly work.
145 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
146 (but not for its children which will stay at the kernel
147 default). This should allow setups with a lot more listening
150 * systemd will now always pass the configured timezone to the
151 kernel at boot. timedated will do the same when the timezone
154 * logind's inhibition logic has been updated. By default,
155 logind will now handle the lid switch, the power and sleep
156 keys all the time, even in graphical sessions. If DEs want
157 to handle these events on their own they should take the new
158 handle-power-key, handle-sleep-key and handle-lid-switch
159 inhibitors during their runtime. A simple way to achiveve
160 that is to invoke the DE wrapped in an invocation of:
162 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
164 * Access to unit operations is now checked via SELinux taking
165 the unit file label and client process label into account.
167 * systemd will now notify the administrator in the journal
168 when he over-mounts a non-empty directory.
170 * There are new specifiers that are resolved in unit files,
171 for the host name (%H), the machine ID (%m) and the boot ID
174 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
175 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
176 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
177 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
178 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
179 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
180 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
184 * Support for reading structured kernel messages from
185 /dev/kmsg has now been added and is enabled by default.
187 * Support for reading kernel messages from /proc/kmsg has now
188 been removed. If you want kernel messages in the journal
189 make sure to run a recent kernel (>= 3.5) that supports
190 reading structured messages from /dev/kmsg (see
191 above). /proc/kmsg is now exclusive property of classic
192 syslog daemons again.
194 * The libudev API gained the new
195 udev_device_new_from_device_id() call.
197 * The logic for file system namespace (ReadOnlyDirectory=,
198 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
199 require pivot_root() anymore. This means fewer temporary
200 directories are created below /tmp for this feature.
202 * nspawn containers will now see and receive all submounts
203 made on the host OS below the root file system of the
206 * Forward Secure Sealing is now supported for Journal files,
207 which provide cryptographical sealing of journal files so
208 that attackers cannot alter log history anymore without this
209 being detectable. Lennart will soon post a blog story about
210 this explaining it in more detail.
212 * There are two new service settings RestartPreventExitStatus=
213 and SuccessExitStatus= which allow configuration of exit
214 status (exit code or signal) which will be excepted from the
215 restart logic, resp. consider successful.
217 * journalctl gained the new --verify switch that can be used
218 to check the integrity of the structure of journal files and
219 (if Forward Secure Sealing is enabled) the contents of
222 * nspawn containers will now be run with /dev/stdin, /dev/fd/
223 and similar symlinks pre-created. This makes running shells
224 as container init process a lot more fun.
226 * The fstab support can now handle PARTUUID= and PARTLABEL=
229 * A new ConditionHost= condition has been added to match
230 against the hostname (with globs) and machine ID. This is
231 useful for clusters where a single OS image is used to
232 provision a large number of hosts which shall run slightly
233 different sets of services.
235 * Services which hit the restart limit will now be placed in a
238 Contributions from: Bertram Poettering, Dave Reisner, Huang
239 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
240 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
244 * When running in --user mode systemd will now become a
245 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
246 tree a lot more organized.
248 * A new PartOf= unit dependency type has been introduced that
249 may be used to group services in a natural way.
251 * "systemctl enable" may now be used to enable instances of
254 * journalctl now prints error log levels in red, and
255 warning/notice log levels in bright white. It also supports
256 filtering by log level now.
258 * cgtop gained a new -n switch (similar to top), to configure
259 the maximum number of iterations to run for. It also gained
260 -b, to run in batch mode (accepting no input).
262 * The suffix ".service" may now be ommited on most systemctl
263 command lines involving service unit names.
265 * There's a new bus call in logind to lock all sessions, as
266 well as a loginctl verb for it "lock-sessions".
268 * libsystemd-logind.so gained a new call sd_journal_perror()
269 that works similar to libc perror() but logs to the journal
270 and encodes structured information about the error number.
272 * /etc/crypttab entries now understand the new keyfile-size=
275 * shutdown(8) now can send a (configurable) wall message when
276 a shutdown is cancelled.
278 * The mount propagation mode for the root file system will now
279 default to "shared", which is useful to make containers work
280 nicely out-of-the-box so that they receive new mounts from
281 the host. This can be undone locally by running "mount
282 --make-rprivate /" if needed.
284 * The prefdm.service file has been removed. Distributions
285 should maintain this unit downstream if they intend to keep
286 it around. However, we recommend writing normal unit files
287 for display managers instead.
289 * Since systemd is a crucial part of the OS we will now
290 default to a number of compiler switches that improve
291 security (hardening) such as read-only relocations, stack
292 protection, and suchlike.
294 * The TimeoutSec= setting for services is now split into
295 TimeoutStartSec= and TimeoutStopSec= to allow configuration
296 of individual time outs for the start and the stop phase of
299 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
300 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
301 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
302 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
303 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
304 Gundersen, Zbigniew Jędrzejewski-Szmek
308 * The journal and id128 C APIs are now fully documented as man
311 * Extra safety checks have been added when transitioning from
312 the initial RAM disk to the main system to avoid accidental
315 * /etc/crypttab entries now understand the new keyfile-offset=
318 * systemctl -t can now be used to filter by unit load state.
320 * The journal C API gained the new sd_journal_wait() call to
321 make writing synchronous journal clients easier.
323 * journalctl gained the new -D switch to show journals from a
326 * journalctl now displays a special marker between log
327 messages of two different boots.
329 * The journal is now explicitly flushed to /var via a service
330 systemd-journal-flush.service, rather than implicitly simply
331 by seeing /var/log/journal to be writable.
333 * journalctl (and the journal C APIs) can now match for much
334 more complex expressions, with alternatives and
337 * When transitioning from the initial RAM disk to the main
338 system we will now kill all processes in a killing spree to
339 ensure no processes stay around by accident.
341 * Three new specifiers may be used in unit files: %u, %h, %s
342 resolve to the user name, user home directory resp. user
343 shell. This is useful for running systemd user instances.
345 * We now automatically rotate journal files if their data
346 object hash table gets a fill level > 75%. We also size the
347 hash table based on the configured maximum file size. This
348 together should lower hash collisions drastically and thus
349 speed things up a bit.
351 * journalctl gained the new "--header" switch to introspect
352 header data of journal files.
354 * A new setting SystemCallFilters= has been added to services
355 which may be used to apply blacklists or whitelists to
356 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
358 * nspawn gained a new --link-journal= switch (and quicker: -j)
359 to link the container journal with the host. This makes it
360 very easy to centralize log viewing on the host for all
361 guests while still keeping the journal files separated.
363 * Many bugfixes and optimizations
365 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
366 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
367 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
372 * Several tools now understand kernel command line arguments,
373 which are only read when run in an initial RAM disk. They
374 usually follow closely their normal counterparts, but are
377 * There's a new tool to analyze the readahead files that are
378 automatically generated at boot. Use:
380 /usr/lib/systemd/systemd-readahead analyze /.readahead
382 * We now provide an early debug shell on tty9 if this enabled. Use:
384 systemctl enable debug-shell.service
386 * All plymouth related units have been moved into the Plymouth
387 package. Please make sure to upgrade your Plymouth version
390 * systemd-tmpfiles now supports getting passed the basename of
391 a configuration file only, in which case it will look for it
392 in all appropriate directories automatically.
394 * udevadm info now takes a /dev or /sys path as argument, and
395 does the right thing. Example:
397 udevadm info /dev/sda
398 udevadm info /sys/class/block/sda
400 * systemctl now prints a warning if a unit is stopped but a
401 unit that might trigger it continues to run. Example: a
402 service is stopped but the socket that activates it is left
405 * "systemctl status" will now mention if the log output was
406 shortened due to rotation since a service has been started.
408 * The journal API now exposes functions to determine the
409 "cutoff" times due to rotation.
411 * journald now understands SIGUSR1 and SIGUSR2 for triggering
412 immediately flushing of runtime logs to /var if possible,
413 resp. for triggering immediate rotation of the journal
416 * It is now considered an error if a service is attempted to
417 be stopped that is not loaded.
419 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
421 * systemd-analyze now supports Python 3
423 * tmpfiles now supports cleaning up directories via aging
424 where the first level dirs are always kept around but
425 directories beneath it automatically aged. This is enabled
426 by prefixing the age field with '~'.
428 * Seat objects now expose CanGraphical, CanTTY properties
429 which is required to deal with very fast bootups where the
430 display manager might be running before the graphics drivers
431 completed initialization.
433 * Seat objects now expose a State property.
435 * We now include RPM macros for service enabling/disabling
436 based on the preset logic. We recommend RPM based
437 distributions to make use of these macros if possible. This
438 makes it simpler to reuse RPM spec files across
441 * We now make sure that the collected systemd unit name is
442 always valid when services log to the journal via
445 * There's a new man page kernel-command-line(7) detailing all
446 command line options we understand.
448 * The fstab generator may now be disabled at boot by passing
449 fstab=0 on the kernel command line.
451 * A new kernel command line option modules-load= is now understood
452 to load a specific kernel module statically, early at boot.
454 * Unit names specified on the systemctl command line are now
455 automatically escaped as needed. Also, if file system or
456 device paths are specified they are automatically turned
457 into the appropriate mount or device unit names. Example:
459 systemctl status /home
460 systemctl status /dev/sda
462 * The SysVConsole= configuration option has been removed from
465 * The SysV search path is no longer exported on the D-Bus
468 * The Names= option is been removed from unit file parsing.
470 * There's a new man page bootup(7) detailing the boot process.
472 * Every unit and every generator we ship with systemd now
473 comes with full documentation. The self-explanatory boot is
476 * A couple of services gained "systemd-" prefixes in their
477 name if they wrap systemd code, rather than only external
478 code. Among them fsck@.service which is now
479 systemd-fsck@.service.
481 * The HaveWatchdog property has been removed from the D-Bus
484 * systemd.confirm_spawn= on the kernel command line should now
487 * There's a new man page crypttab(5) which details all options
488 we actually understand.
490 * systemd-nspawn gained a new --capability= switch to pass
491 additional capabilities to the container.
493 * timedated will now read known NTP implementation unit names
494 from /usr/lib/systemd/ntp-units.d/*.list,
495 systemd-timedated-ntp.target has been removed.
497 * journalctl gained a new switch "-b" that lists log data of
498 the current boot only.
500 * The notify socket is in the abstract namespace again, in
501 order to support daemons which chroot() at start-up.
503 * There is a new Storage= configuration option for journald
504 which allows configuration of where log data should go. This
505 also provides a way to disable journal logging entirely, so
506 that data collected is only forwarded to the console, the
507 kernel log buffer or another syslog implementation.
509 * Many bugfixes and optimizations
511 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
512 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
513 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
514 Shawn Landden, Tom Gundersen
518 * "systemctl help <unit>" now shows the man page if one is
521 * Several new man pages have been added.
523 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
524 MaxLevelConsole= can now be specified in
525 journald.conf. These options allow reducing the amount of
526 data stored on disk or forwarded by the log level.
528 * TimerSlackNSec= can now be specified in system.conf for
529 PID1. This allows system-wide power savings.
531 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
532 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
537 * logind is now capable of (optionally) handling power and
538 sleep keys as well as the lid switch.
540 * journalctl now understands the syntax "journalctl
541 /usr/bin/avahi-daemon" to get all log output of a specific
544 * CapabilityBoundingSet= in system.conf now also influences
545 the capability bound set of usermode helpers of the kernel.
547 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
548 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
549 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
550 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
554 * Note that we skipped 139 releases here in order to set the
555 new version to something that is greater than both udev's
556 and systemd's most recent version number.
558 * udev: all udev sources are merged into the systemd source tree now.
559 All future udev development will happen in the systemd tree. It
560 is still fully supported to use the udev daemon and tools without
561 systemd running, like in initramfs or other init systems. Building
562 udev though, will require the *build* of the systemd tree, but
563 udev can be properly *run* without systemd.
565 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
566 should be used to create dead device nodes as workarounds for broken
569 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
570 no longer supported. udev_monitor_new_from_netlink() needs to be
571 used to subscribe to events.
573 * udev: when udevd is started by systemd, processes which are left
574 behind by forking them off of udev rules, are unconditionally cleaned
575 up and killed now after the event handling has finished. Services or
576 daemons must be started as systemd services. Services can be
577 pulled-in by udev to get started, but they can no longer be directly
578 forked by udev rules.
580 * udev: the daemon binary is called systemd-udevd now and installed
581 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
582 to adapt to that, create symlink, or rename the binary after building
585 * libudev no longer provides these symbols:
586 udev_monitor_from_socket()
587 udev_queue_get_failed_list_entry()
588 udev_get_{dev,sys,run}_path()
589 The versions number was bumped and symbol versioning introduced.
591 * systemd-loginctl and systemd-journalctl have been renamed
592 to loginctl and journalctl to match systemctl.
594 * The config files: /etc/systemd/systemd-logind.conf and
595 /etc/systemd/systemd-journald.conf have been renamed to
596 logind.conf and journald.conf. Package updates should rename
597 the files to the new names on upgrade.
599 * For almost all files the license is now LGPL2.1+, changed
600 from the previous GPL2.0+. Exceptions are some minor stuff
601 of udev (which will be changed to LGPL2.1 eventually, too),
602 and the MIT licensed sd-daemon.[ch] library that is suitable
603 to be used as drop-in files.
605 * systemd and logind now handle system sleep states, in
606 particular suspending and hibernating.
608 * logind now implements a sleep/shutdown/idle inhibiting logic
609 suitable for a variety of uses. Soonishly Lennart will blog
610 about this in more detail.
612 * var-run.mount and var-lock.mount are no longer provided
613 (which prevously bind mounted these directories to their new
614 places). Distributions which have not converted these
615 directories to symlinks should consider stealing these files
616 from git history and add them downstream.
618 * We introduced the Documentation= field for units and added
619 this to all our shipped units. This is useful to make it
620 easier to explore the boot and the purpose of the various
623 * All smaller setup units (such as
624 systemd-vconsole-setup.service) now detect properly if they
625 are run in a container and are skipped when
626 appropriate. This guarantees an entirely noise-free boot in
627 Linux container environments such as systemd-nspawn.
629 * A framework for implementing offline system updates is now
630 integrated, for details see:
631 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
633 * A new service type Type=idle is available now which helps us
634 avoiding ugly interleaving of getty output and boot status
637 * There's now a system-wide CapabilityBoundingSet= option to
638 globally reduce the set of capabilities for the
639 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
640 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
641 even CAP_NET_ADMIN system-wide for secure systems.
643 * There are now system-wide DefaultLimitXXX= options to
644 globally change the defaults of the various resource limits
645 for all units started by PID 1.
647 * Harald Hoyer's systemd test suite has been integrated into
648 systemd which allows easy testing of systemd builds in qemu
649 and nspawn. (This is really awesome! Ask us for details!)
651 * The fstab parser is now implemented as generator, not inside
654 * systemctl will now warn you if .mount units generated from
655 /etc/fstab are out of date due to changes in fstab that
656 haven't been read by systemd yet.
658 * systemd is now suitable for usage in initrds. Dracut has
659 already been updated to make use of this. With this in place
660 initrds get a slight bit faster but primarily are much
661 easier to introspect and debug since "systemctl status" in
662 the host system can be used to introspect initrd services,
663 and the journal from the initrd is kept around too.
665 * systemd-delta has been added, a tool to explore differences
666 between user/admin configuration and vendor defaults.
668 * PrivateTmp= now affects both /tmp and /var/tmp.
670 * Boot time status messages are now much prettier and feature
671 proper english language. Booting up systemd has never been
674 * Read-ahead pack files now include the inode number of all
675 files to pre-cache. When the inode changes the pre-caching
676 is not attempted. This should be nicer to deal with updated
677 packages which might result in changes of read-ahead
680 * We now temporaritly lower the kernel's read_ahead_kb variable
681 when collecting read-ahead data to ensure the kernel's
682 built-in read-ahead does not add noise to our measurements
683 of necessary blocks to pre-cache.
685 * There's now RequiresMountsFor= to add automatic dependencies
686 for all mounts necessary for a specific file system path.
688 * MountAuto= and SwapAuto= have been removed from
689 system.conf. Mounting file systems at boot has to take place
692 * nspawn now learned a new switch --uuid= to set the machine
693 ID on the command line.
695 * nspawn now learned the -b switch to automatically search
698 * vt102 is now the default TERM for serial TTYs, upgraded from
701 * systemd-logind now works on VT-less systems.
703 * The build tree has been reorganized. The individual
704 components now have directories of their own.
706 * A new condition type ConditionPathIsReadWrite= is now available.
708 * nspawn learned the new -C switch to create cgroups for the
709 container in other hierarchies.
711 * We now have support for hardware watchdogs, configurable in
714 * The scheduled shutdown logic now has a public API.
716 * We now mount /tmp as tmpfs by default, but this can be
717 masked and /etc/fstab can override it.
719 * Since udisks doesn't make use of /media anymore we are not
720 mounting a tmpfs on it anymore.
722 * journalctl gained a new --local switch to only interleave
723 locally generated journal files.
725 * We can now load the IMA policy at boot automatically.
727 * The GTK tools have been split off into a systemd-ui.
729 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
730 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
731 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
732 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
733 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
734 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
735 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
736 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
737 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
742 * This is mostly a bugfix release
744 * Support optional initialization of the machine ID from the
745 KVM or container configured UUID.
747 * Support immediate reboots with "systemctl reboot -ff"
749 * Show /etc/os-release data in systemd-analyze output
751 * Many bugfixes for the journal, including endianess fixes and
752 ensuring that disk space enforcement works
754 * sd-login.h is C++ comptaible again
756 * Extend the /etc/os-release format on request of the Debian
759 * We now refuse non-UTF8 strings used in various configuration
760 and unit files. This is done to ensure we don't pass invalid
761 data over D-Bus or expose it elsewhere.
763 * Register Mimo USB Screens as suitable for automatic seat
766 * Read SELinux client context from journal clients in a race
769 * Reorder configuration file lookup order. /etc now always
770 overrides /run in order to allow the administrator to always
771 and unconditionally override vendor supplied or
772 automatically generated data.
774 * The various user visible bits of the journal now have man
775 pages. We still lack man pages for the journal API calls
778 * We now ship all man pages in HTML format again in the
781 Contributions from: Dave Reisner, Dirk Eibach, Frederic
782 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
783 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
788 * This is mostly a bugfix release
790 * systems lacking /etc/os-release are no longer supported.
792 * Various functionality updates to libsystemd-login.so
794 * Track class of PAM logins to distuingish greeters from
797 Contributions from: Kay Sievers, Lennart Poettering, Michael
802 * This is an important bugfix release for v41.
804 * Building man pages is now optional which should be useful
805 for those building systemd from git but unwilling to install
808 * Watchdog support for supervising services is now usable. In
809 a future release support for hardware watchdogs
810 (i.e. /dev/watchdog) will be added building on this.
812 * Service start rate limiting is now configurable and can be
813 turned off per service. When a start rate limit is hit a
814 reboot can automatically be triggered.
816 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
818 Contributions from: Benjamin Franzke, Bill Nottingham,
819 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
820 Schmidt, Michał Górny, Piotr Drąg
824 * The systemd binary is installed /usr/lib/systemd/systemd now;
825 An existing /sbin/init symlink needs to be adapted with the
828 * The code that loads kernel modules has been ported to invoke
829 libkmod directly, instead of modprobe. This means we do not
830 support systems with module-init-tools anymore.
832 * Watchdog support is now already useful, but still not
835 * A new kernel command line option systemd.setenv= is
836 understood to set system wide environment variables
839 * We now limit the set of capabilities of systemd-journald.
841 * We now set SIGPIPE to ignore by default, since it only is
842 useful in shell pipelines, and has little use in general
843 code. This can be disabled with IgnoreSIPIPE=no in unit
846 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
847 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
852 * This is mostly a bugfix release
854 * We now expose the reason why a service failed in the
855 "Result" D-Bus property.
857 * Rudimentary service watchdog support (will be completed over
858 the next few releases.)
860 * When systemd forks off in order execute some service we will
861 now immediately changes its argv[0] to reflect which process
862 it will execute. This is useful to minimize the time window
863 with a generic argv[0], which makes bootcharts more useful
865 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
866 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
867 Mike Kazantsev, Ray Strode
871 * This is mostly a test release, but incorporates many
874 * New systemd-cgtop tool to show control groups by their
877 * Linking against libacl for ACLs is optional again. If
878 disabled, support tracking device access for active logins
879 goes becomes unavailable, and so does access to the user
880 journals by the respective users.
882 * If a group "adm" exists, journal files are automatically
883 owned by them, thus allow members of this group full access
884 to the system journal as well as all user journals.
886 * The journal now stores the SELinux context of the logging
887 client for all entries.
889 * Add C++ inclusion guards to all public headers
891 * New output mode "cat" in the journal to print only text
892 messages, without any meta data like date or time.
894 * Include tiny X server wrapper as a temporary stop-gap to
895 teach XOrg udev display enumeration. This is used by display
896 managers such as gdm, and will go away as soon as XOrg
897 learned native udev hotplugging for display devices.
899 * Add new systemd-cat tool for executing arbitrary programs
900 with STDERR/STDOUT connected to the journal. Can also act as
901 BSD logger replacement, and does so by default.
903 * Optionally store all locally generated coredumps in the
904 journal along with meta data.
906 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
907 writing short strings to files (for usage for /sys), and for
908 creating symlinks, character and block device nodes.
910 * New unit file option ControlGroupPersistent= to make cgroups
911 persistent, following the mechanisms outlined in
912 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
914 * Support multiple local RTCs in a sane way
916 * No longer monopolize IO when replaying readahead data on
917 rotating disks, since we might starve non-file-system IO to
918 death, since fanotify() will not see accesses done by blkid,
921 * Don't show kernel threads in systemd-cgls anymore, unless
922 requested with new -k switch.
924 Contributions from: Dan Horák, Kay Sievers, Lennart
925 Poettering, Michal Schmidt
929 * This is mostly a test release, but incorporates many
932 * The git repository moved to:
933 git://anongit.freedesktop.org/systemd/systemd
934 ssh://git.freedesktop.org/git/systemd/systemd
936 * First release with the journal
937 http://0pointer.de/blog/projects/the-journal.html
939 * The journal replaces both systemd-kmsg-syslogd and
940 systemd-stdout-bridge.
942 * New sd_pid_get_unit() API call in libsystemd-logind
944 * Many systemadm clean-ups
946 * Introduce remote-fs-pre.target which is ordered before all
947 remote mounts and may be used to start services before all
950 * Added Mageia support
952 * Add bash completion for systemd-loginctl
954 * Actively monitor PID file creation for daemons which exit in
955 the parent process before having finished writing the PID
956 file in the daemon process. Daemons which do this need to be
957 fixed (i.e. PID file creation must have finished before the
958 parent exits), but we now react a bit more gracefully to them.
960 * Add colourful boot output, mimicking the well-known output
961 of existing distributions.
963 * New option PassCredentials= for socket units, for
964 compatibility with a recent kernel ABI breakage.
966 * /etc/rc.local is now hooked in via a generator binary, and
967 thus will no longer act as synchronization point during
970 * systemctl list-unit-files now supports --root=.
972 * systemd-tmpfiles now understands two new commands: z, Z for
973 relabelling files according to the SELinux database. This is
974 useful to apply SELinux labels to specific files in /sys,
977 * Output of SysV services is now forwarded to both the console
978 and the journal by default, not only just the console.
980 * New man pages for all APIs from libsystemd-login.
982 * The build tree got reorganized and a the build system is a
983 lot more modular allowing embedded setups to specifically
984 select the components of systemd they are interested in.
986 * Support for Linux systems lacking the kernel VT subsystem is
989 * configure's --with-rootdir= got renamed to
990 --with-rootprefix= to follow the naming used by udev and
993 * Unless specified otherwise we'll now install to /usr instead
994 of /usr/local by default.
996 * Processes with '@' in argv[0][0] are now excluded from the
997 final shut-down killing spree, following the logic explained
999 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1001 * All processes remaining in a service cgroup when we enter
1002 the START or START_PRE states are now killed with
1003 SIGKILL. That means it is no longer possible to spawn
1004 background processes from ExecStart= lines (which was never
1005 supported anyway, and bad style).
1007 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1008 reloading of units together.
1010 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1011 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1012 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1013 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1014 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek