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