1 systemd System and Service Manager
5 * Support for reading structured kernel messages from
6 /dev/kmsg has now been added and is enabled by default.
8 * Support for reading kernel messages from /proc/kmsg has now
9 been removed. If you want kernel messages in the journal
10 make sure to run a recent kernel (>= 3.5) that supports
11 reading structured messages from /dev/kmsg (see
12 above). /proc/kmsg is now exclusive property of classic
15 * The libudev API gained the new
16 udev_device_new_from_device_id() call.
18 * The logic for file system namespace (ReadOnlyDirectory=,
19 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
20 require pivot_root() anymore. This means fewer temporary
21 directories are created below /tmp for this feature.
23 * nspawn containers will now see and receive all submounts
24 made on the host OS below the root file system of the
27 * Forward Secure Sealing is now supported for Journal files,
28 which provide cryptographical sealing of journal files so
29 that attackers cannot alter log history anymore without this
30 being detectable. Lennart will soon post a blog story about
31 this explaining it in more detail.
33 * There are two new service settings RestartPreventExitStatus=
34 and SuccessExitStatus= which allow configuration of exit
35 status (exit code or signal) which will be excepted from the
36 restart logic, resp. consider successful.
38 * journalctl gained the new --verify switch that can be used
39 to check the integrity of the structure of journal files and
40 (if Forward Secure Sealing is enabled) the contents of
43 * nspawn containers will now be run with /dev/stdin, /dev/fd/
44 and similar symlinks pre-created. This makes running shells
45 as container init process a lot more fun.
47 * The fstab support can now handle PARTUUID= and PARTLABEL=
50 * A new ConditionHost= condition has been added to match
51 against the hostname (with globs) and machine ID. This is
52 useful for clusters where a single OS image is used to
53 provision a large number of hosts which shall run slightly
54 different sets of services.
56 * Services which hit the restart limit will now be placed in a
59 Contributions from Bertram Poettering, Dave Reisner, Huang
60 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
61 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
65 * When running in --user mode systemd will now become a
66 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
67 tree a lot more organized.
69 * A new PartOf= unit dependency type has been introduced that
70 may be used to group services in a natural way.
72 * "systemctl enable" may now be used to enable instances of
75 * journalctl now prints error log levels in red, and
76 warning/notice log levels in bright white. It also supports
77 filtering by log level now.
79 * cgtop gained a new -n switch (similar to top), to configure
80 the maximum number of iterations to run for. It also gained
81 -b, to run in batch mode (accepting no input).
83 * The suffix ".service" may now be ommited on most systemctl
84 command lines involving service unit names.
86 * There's a new bus call in logind to lock all sessions, as
87 well as a loginctl verb for it "lock-sessions".
89 * libsystemd-logind.so gained a new call sd_journal_perror()
90 that works similar to libc perror() but logs to the journal
91 and encodes structured information about the error number.
93 * /etc/crypttab entries now understand the new keyfile-size=
96 * shutdown(8) now can send a (configurable) wall message when
97 a shutdown is cancelled.
99 * The mount propagation mode for the root file system will now
100 default to "shared", which is useful to make containers work
101 nicely out-of-the-box so that they receive new mounts from
102 the host. This can be undone locally by running "mount
103 --make-rprivate /" if needed.
105 * The prefdm.service file has been removed. Distributions
106 should maintain this unit downstream if they intend to keep
107 it around. However, we recommend writing normal unit files
108 for display managers instead.
110 * Since systemd is a crucial part of the OS we will now
111 default to a number of compiler switches that improve
112 security (hardening) such as read-only relocations, stack
113 protection, and suchlike.
115 * The TimeoutSec= setting for services is now split into
116 TimeoutStartSec= and TimeoutStopSec= to allow configuration
117 of individual time outs for the start and the stop phase of
120 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
121 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
122 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
123 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
124 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
125 Gundersen, Zbigniew Jędrzejewski-Szmek
129 * The journal and id128 C APIs are now fully documented as man
132 * Extra safety checks have been added when transitioning from
133 the initial RAM disk to the main system to avoid accidental
136 * /etc/crypttab entries now understand the new keyfile-offset=
139 * systemctl -t can now be used to filter by unit load state.
141 * The journal C API gained the new sd_journal_wait() call to
142 make writing synchronous journal clients easier.
144 * journalctl gained the new -D switch to show journals from a
147 * journalctl now displays a special marker between log
148 messages of two different boots.
150 * The journal is now explicitly flushed to /var via a service
151 systemd-journal-flush.service, rather than implicitly simply
152 by seeing /var/log/journal to be writable.
154 * journalctl (and the journal C APIs) can now match for much
155 more complex expressions, with alternatives and
158 * When transitioning from the initial RAM disk to the main
159 system we will now kill all processes in a killing spree to
160 ensure no processes stay around by accident.
162 * Three new specifiers may be used in unit files: %u, %h, %s
163 resolve to the user name, user home directory resp. user
164 shell. This is useful for running systemd user instances.
166 * We now automatically rotate journal files if their data
167 object hash table gets a fill level > 75%. We also size the
168 hash table based on the configured maximum file size. This
169 together should lower hash collisions drastically and thus
170 speed things up a bit.
172 * journalctl gained the new "--header" switch to introspect
173 header data of journal files.
175 * A new setting SystemCallFilters= has been added to services
176 which may be used to apply blacklists or whitelists to
177 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
179 * nspawn gained a new --link-journal= switch (and quicker: -j)
180 to link the container journal with the host. This makes it
181 very easy to centralize log viewing on the host for all
182 guests while still keeping the journal files separated.
184 * Many bugfixes and optimizations
186 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
187 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
188 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
193 * Several tools now understand kernel command line arguments,
194 which are only read when run in an initial RAM disk. They
195 usually follow closely their normal counterparts, but are
198 * There's a new tool to analyze the readahead files that are
199 automatically generated at boot. Use:
201 /usr/lib/systemd/systemd-readahead analyze /.readahead
203 * We now provide an early debug shell on tty9 if this enabled. Use:
205 systemctl enable debug-shell.service
207 * All plymouth related units have been moved into the Plymouth
208 package. Please make sure to upgrade your Plymouth version
211 * systemd-tmpfiles now supports getting passed the basename of
212 a configuration file only, in which case it will look for it
213 in all appropriate directories automatically.
215 * udevadm info now takes a /dev or /sys path as argument, and
216 does the right thing. Example:
218 udevadm info /dev/sda
219 udevadm info /sys/class/block/sda
221 * systemctl now prints a warning if a unit is stopped but a
222 unit that might trigger it continues to run. Example: a
223 service is stopped but the socket that activates it is left
226 * "systemctl status" will now mention if the log output was
227 shortened due to rotation since a service has been started.
229 * The journal API now exposes functions to determine the
230 "cutoff" times due to rotation.
232 * journald now understands SIGUSR1 and SIGUSR2 for triggering
233 immediately flushing of runtime logs to /var if possible,
234 resp. for triggering immediate rotation of the journal
237 * It is now considered an error if a service is attempted to
238 be stopped that is not loaded.
240 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
242 * systemd-analyze now supports Python 3
244 * tmpfiles now supports cleaning up directories via aging
245 where the first level dirs are always kept around but
246 directories beneath it automatically aged. This is enabled
247 by prefixing the age field with '~'.
249 * Seat objects now expose CanGraphical, CanTTY properties
250 which is required to deal with very fast bootups where the
251 display manager might be running before the graphics drivers
252 completed initialization.
254 * Seat objects now expose a State property.
256 * We now include RPM macros for service enabling/disabling
257 based on the preset logic. We recommend RPM based
258 distributions to make use of these macros if possible. This
259 makes it simpler to reuse RPM spec files across
262 * We now make sure that the collected systemd unit name is
263 always valid when services log to the journal via
266 * There's a new man page kernel-command-line(7) detailing all
267 command line options we understand.
269 * The fstab generator may now be disabled at boot by passing
270 fstab=0 on the kernel command line.
272 * A new kernel command line option modules-load= is now understood
273 to load a specific kernel module statically, early at boot.
275 * Unit names specified on the systemctl command line are now
276 automatically escaped as needed. Also, if file system or
277 device paths are specified they are automatically turned
278 into the appropriate mount or device unit names. Example:
280 systemctl status /home
281 systemctl status /dev/sda
283 * The SysVConsole= configuration option has been removed from
286 * The SysV search path is no longer exported on the D-Bus
289 * The Names= option is been removed from unit file parsing.
291 * There's a new man page bootup(7) detailing the boot process.
293 * Every unit and every generator we ship with systemd now
294 comes with full documentation. The self-explanatory boot is
297 * A couple of services gained "systemd-" prefixes in their
298 name if they wrap systemd code, rather than only external
299 code. Among them fsck@.service which is now
300 systemd-fsck@.service.
302 * The HaveWatchdog property has been removed from the D-Bus
305 * systemd.confirm_spawn= on the kernel command line should now
308 * There's a new man page crypttab(5) which details all options
309 we actually understand.
311 * systemd-nspawn gained a new --capability= switch to pass
312 additional capabilities to the container.
314 * timedated will now read known NTP implementation unit names
315 from /usr/lib/systemd/ntp-units.d/*.list,
316 systemd-timedated-ntp.target has been removed.
318 * journalctl gained a new switch "-b" that lists log data of
319 the current boot only.
321 * The notify socket is in the abstract namespace again, in
322 order to support daemons which chroot() at start-up.
324 * There is a new Storage= configuration option for journald
325 which allows configuration of where log data should go. This
326 also provides a way to disable journal logging entirely, so
327 that data collected is only forwarded to the console, the
328 kernel log buffer or another syslog implementation.
330 * Many bugfixes and optimizations
332 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
333 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
334 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
335 Shawn Landden, Tom Gundersen
338 * "systemctl help <unit>" now shows the man page if one is
341 * Several new man pages have been added.
343 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
344 MaxLevelConsole= can now be specified in
345 journald.conf. These options allow reducing the amount of
346 data stored on disk or forwarded by the log level.
348 * TimerSlackNSec= can now be specified in system.conf for
349 PID1. This allows system-wide power savings.
351 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
352 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
356 * logind is now capable of (optionally) handling power and
357 sleep keys as well as the lid switch.
359 * journalctl now understands the syntax "journalctl
360 /usr/bin/avahi-daemon" to get all log output of a specific
363 * CapabilityBoundingSet= in system.conf now also influences
364 the capability bound set of usermode helpers of the kernel.
366 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
367 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
368 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
369 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
372 * Note that we skipped 139 releases here in order to set the
373 new version to something that is greater than both udev's
374 and systemd's most recent version number.
376 * udev: all udev sources are merged into the systemd source tree now.
377 All future udev development will happen in the systemd tree. It
378 is still fully supported to use the udev daemon and tools without
379 systemd running, like in initramfs or other init systems. Building
380 udev though, will require the *build* of the systemd tree, but
381 udev can be properly *run* without systemd.
383 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
384 should be used to create dead device nodes as workarounds for broken
387 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
388 no longer supported. udev_monitor_new_from_netlink() needs to be
389 used to subscribe to events.
391 * udev: when udevd is started by systemd, processes which are left
392 behind by forking them off of udev rules, are unconditionally cleaned
393 up and killed now after the event handling has finished. Services or
394 daemons must be started as systemd services. Services can be
395 pulled-in by udev to get started, but they can no longer be directly
396 forked by udev rules.
398 * udev: the daemon binary is called systemd-udevd now and installed
399 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
400 to adapt to that, create symlink, or rename the binary after building
403 * libudev no longer provides these symbols:
404 udev_monitor_from_socket()
405 udev_queue_get_failed_list_entry()
406 udev_get_{dev,sys,run}_path()
407 The versions number was bumped and symbol versioning introduced.
409 * systemd-loginctl and systemd-journalctl have been renamed
410 to loginctl and journalctl to match systemctl.
412 * The config files: /etc/systemd/systemd-logind.conf and
413 /etc/systemd/systemd-journald.conf have been renamed to
414 logind.conf and journald.conf. Package updates should rename
415 the files to the new names on upgrade.
417 * For almost all files the license is now LGPL2.1+, changed
418 from the previous GPL2.0+. Exceptions are some minor stuff
419 of udev (which will be changed to LGPL2.1 eventually, too),
420 and the MIT licensed sd-daemon.[ch] library that is suitable
421 to be used as drop-in files.
423 * systemd and logind now handle system sleep states, in
424 particular suspending and hibernating.
426 * logind now implements a sleep/shutdown/idle inhibiting logic
427 suitable for a variety of uses. Soonishly Lennart will blog
428 about this in more detail.
430 * var-run.mount and var-lock.mount are no longer provided
431 (which prevously bind mounted these directories to their new
432 places). Distributions which have not converted these
433 directories to symlinks should consider stealing these files
434 from git history and add them downstream.
436 * We introduced the Documentation= field for units and added
437 this to all our shipped units. This is useful to make it
438 easier to explore the boot and the purpose of the various
441 * All smaller setup units (such as
442 systemd-vconsole-setup.service) now detect properly if they
443 are run in a container and are skipped when
444 appropriate. This guarantees an entirely noise-free boot in
445 Linux container environments such as systemd-nspawn.
447 * A framework for implementing offline system updates is now
448 integrated, for details see:
449 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
451 * A new service type Type=idle is available now which helps us
452 avoiding ugly interleaving of getty output and boot status
455 * There's now a system-wide CapabilityBoundingSet= option to
456 globally reduce the set of capabilities for the
457 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
458 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
459 even CAP_NET_ADMIN system-wide for secure systems.
461 * There are now system-wide DefaultLimitXXX= options to
462 globally change the defaults of the various resource limits
463 for all units started by PID 1.
465 * Harald Hoyer's systemd test suite has been integrated into
466 systemd which allows easy testing of systemd builds in qemu
467 and nspawn. (This is really awesome! Ask us for details!)
469 * The fstab parser is now implemented as generator, not inside
472 * systemctl will now warn you if .mount units generated from
473 /etc/fstab are out of date due to changes in fstab that
474 haven't been read by systemd yet.
476 * systemd is now suitable for usage in initrds. Dracut has
477 already been updated to make use of this. With this in place
478 initrds get a slight bit faster but primarily are much
479 easier to introspect and debug since "systemctl status" in
480 the host system can be used to introspect initrd services,
481 and the journal from the initrd is kept around too.
483 * systemd-delta has been added, a tool to explore differences
484 between user/admin configuration and vendor defaults.
486 * PrivateTmp= now affects both /tmp and /var/tmp.
488 * Boot time status messages are now much prettier and feature
489 proper english language. Booting up systemd has never been
492 * Read-ahead pack files now include the inode number of all
493 files to pre-cache. When the inode changes the pre-caching
494 is not attempted. This should be nicer to deal with updated
495 packages which might result in changes of read-ahead
498 * We now temporaritly lower the kernel's read_ahead_kb variable
499 when collecting read-ahead data to ensure the kernel's
500 built-in read-ahead does not add noise to our measurements
501 of necessary blocks to pre-cache.
503 * There's now RequiresMountsFor= to add automatic dependencies
504 for all mounts necessary for a specific file system path.
506 * MountAuto= and SwapAuto= have been removed from
507 system.conf. Mounting file systems at boot has to take place
510 * nspawn now learned a new switch --uuid= to set the machine
511 ID on the command line.
513 * nspawn now learned the -b switch to automatically search
516 * vt102 is now the default TERM for serial TTYs, upgraded from
519 * systemd-logind now works on VT-less systems.
521 * The build tree has been reorganized. The individual
522 components now have directories of their own.
524 * A new condition type ConditionPathIsReadWrite= is now available.
526 * nspawn learned the new -C switch to create cgroups for the
527 container in other hierarchies.
529 * We now have support for hardware watchdogs, configurable in
532 * The scheduled shutdown logic now has a public API.
534 * We now mount /tmp as tmpfs by default, but this can be
535 masked and /etc/fstab can override it.
537 * Since udisks doesn't make use of /media anymore we are not
538 mounting a tmpfs on it anymore.
540 * journalctl gained a new --local switch to only interleave
541 locally generated journal files.
543 * We can now load the IMA policy at boot automatically.
545 * The GTK tools have been split off into a systemd-ui.
547 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
548 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
549 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
550 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
551 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
552 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
553 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
554 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
555 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
559 * This is mostly a bugfix release
561 * Support optional initialization of the machine ID from the
562 KVM or container configured UUID.
564 * Support immediate reboots with "systemctl reboot -ff"
566 * Show /etc/os-release data in systemd-analyze output
568 * Many bugfixes for the journal, including endianess fixes and
569 ensuring that disk space enforcement works
571 * sd-login.h is C++ comptaible again
573 * Extend the /etc/os-release format on request of the Debian
576 * We now refuse non-UTF8 strings used in various configuration
577 and unit files. This is done to ensure we don't pass invalid
578 data over D-Bus or expose it elsewhere.
580 * Register Mimo USB Screens as suitable for automatic seat
583 * Read SELinux client context from journal clients in a race
586 * Reorder configuration file lookup order. /etc now always
587 overrides /run in order to allow the administrator to always
588 and unconditionally override vendor supplied or
589 automatically generated data.
591 * The various user visible bits of the journal now have man
592 pages. We still lack man pages for the journal API calls
595 * We now ship all man pages in HTML format again in the
598 Contributions from: Dave Reisner, Dirk Eibach, Frederic
599 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
600 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
604 * This is mostly a bugfix release
606 * systems lacking /etc/os-release are no longer supported.
608 * Various functionality updates to libsystemd-login.so
610 * Track class of PAM logins to distuingish greeters from
613 Contributions from: Kay Sievers, Lennart Poettering, Michael
617 * This is an important bugfix release for v41.
619 * Building man pages is now optional which should be useful
620 for those building systemd from git but unwilling to install
623 * Watchdog support for supervising services is now usable. In
624 a future release support for hardware watchdogs
625 (i.e. /dev/watchdog) will be added building on this.
627 * Service start rate limiting is now configurable and can be
628 turned off per service. When a start rate limit is hit a
629 reboot can automatically be triggered.
631 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
633 Contributions from: Benjamin Franzke, Bill Nottingham,
634 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
635 Schmidt, Michał Górny, Piotr Drąg
638 * The systemd binary is installed /usr/lib/systemd/systemd now;
639 An existing /sbin/init symlink needs to be adapted with the
642 * The code that loads kernel modules has been ported to invoke
643 libkmod directly, instead of modprobe. This means we do not
644 support systems with module-init-tools anymore.
646 * Watchdog support is now already useful, but still not
649 * A new kernel command line option systemd.setenv= is
650 understood to set system wide environment variables
653 * We now limit the set of capabilities of systemd-journald.
655 * We now set SIGPIPE to ignore by default, since it only is
656 useful in shell pipelines, and has little use in general
657 code. This can be disabled with IgnoreSIPIPE=no in unit
660 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
661 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
665 * This is mostly a bugfix release
667 * We now expose the reason why a service failed in the
668 "Result" D-Bus property.
670 * Rudimentary service watchdog support (will be completed over
671 the next few releases.)
673 * When systemd forks off in order execute some service we will
674 now immediately changes its argv[0] to reflect which process
675 it will execute. This is useful to minimize the time window
676 with a generic argv[0], which makes bootcharts more useful
678 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
679 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
680 Mike Kazantsev, Ray Strode
683 * This is mostly a test release, but incorporates many
686 * New systemd-cgtop tool to show control groups by their
689 * Linking against libacl for ACLs is optional again. If
690 disabled, support tracking device access for active logins
691 goes becomes unavailable, and so does access to the user
692 journals by the respective users.
694 * If a group "adm" exists, journal files are automatically
695 owned by them, thus allow members of this group full access
696 to the system journal as well as all user journals.
698 * The journal now stores the SELinux context of the logging
699 client for all entries.
701 * Add C++ inclusion guards to all public headers
703 * New output mode "cat" in the journal to print only text
704 messages, without any meta data like date or time.
706 * Include tiny X server wrapper as a temporary stop-gap to
707 teach XOrg udev display enumeration. This is used by display
708 managers such as gdm, and will go away as soon as XOrg
709 learned native udev hotplugging for display devices.
711 * Add new systemd-cat tool for executing arbitrary programs
712 with STDERR/STDOUT connected to the journal. Can also act as
713 BSD logger replacement, and does so by default.
715 * Optionally store all locally generated coredumps in the
716 journal along with meta data.
718 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
719 writing short strings to files (for usage for /sys), and for
720 creating symlinks, character and block device nodes.
722 * New unit file option ControlGroupPersistent= to make cgroups
723 persistent, following the mechanisms outlined in
724 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
726 * Support multiple local RTCs in a sane way
728 * No longer monopolize IO when replaying readahead data on
729 rotating disks, since we might starve non-file-system IO to
730 death, since fanotify() will not see accesses done by blkid,
733 * Don't show kernel threads in systemd-cgls anymore, unless
734 requested with new -k switch.
736 Contributions from: Dan Horák, Kay Sievers, Lennart
737 Poettering, Michal Schmidt
740 * This is mostly a test release, but incorporates many
743 * The git repository moved to:
744 git://anongit.freedesktop.org/systemd/systemd
745 ssh://git.freedesktop.org/git/systemd/systemd
747 * First release with the journal
748 http://0pointer.de/blog/projects/the-journal.html
750 * The journal replaces both systemd-kmsg-syslogd and
751 systemd-stdout-bridge.
753 * New sd_pid_get_unit() API call in libsystemd-logind
755 * Many systemadm clean-ups
757 * Introduce remote-fs-pre.target which is ordered before all
758 remote mounts and may be used to start services before all
761 * Added Mageia support
763 * Add bash completion for systemd-loginctl
765 * Actively monitor PID file creation for daemons which exit in
766 the parent process before having finished writing the PID
767 file in the daemon process. Daemons which do this need to be
768 fixed (i.e. PID file creation must have finished before the
769 parent exits), but we now react a bit more gracefully to them.
771 * Add colourful boot output, mimicking the well-known output
772 of existing distributions.
774 * New option PassCredentials= for socket units, for
775 compatibility with a recent kernel ABI breakage.
777 * /etc/rc.local is now hooked in via a generator binary, and
778 thus will no longer act as synchronization point during
781 * systemctl list-unit-files now supports --root=.
783 * systemd-tmpfiles now understands two new commands: z, Z for
784 relabelling files according to the SELinux database. This is
785 useful to apply SELinux labels to specific files in /sys,
788 * Output of SysV services is now forwarded to both the console
789 and the journal by default, not only just the console.
791 * New man pages for all APIs from libsystemd-login.
793 * The build tree got reorganized and a the build system is a
794 lot more modular allowing embedded setups to specifically
795 select the components of systemd they are interested in.
797 * Support for Linux systems lacking the kernel VT subsystem is
800 * configure's --with-rootdir= got renamed to
801 --with-rootprefix= to follow the naming used by udev and
804 * Unless specified otherwise we'll now install to /usr instead
805 of /usr/local by default.
807 * Processes with '@' in argv[0][0] are now excluded from the
808 final shut-down killing spree, following the logic explained
810 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
812 * All processes remaining in a service cgroup when we enter
813 the START or START_PRE states are now killed with
814 SIGKILL. That means it is no longer possible to spawn
815 background processes from ExecStart= lines (which was never
816 supported anyway, and bad style).
818 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
819 reloading of units together.
821 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
822 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
823 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
824 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
825 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek