1 systemd System and Service Manager
5 * If /etc/vconsole.conf is non-existent or empty we will no
6 longer load any console font or key map at boot by
7 default. Instead the kernel defaults will be left
8 intact. This is definitely the right thing to do, as no
9 configuration should mean no configuration, and hard-coding
10 font names that are different on all archs is probably a bad
11 idea. Also, the kernel default key map and font should be
12 good enough for most cases anyway, and mostly identical to
13 the userspace fonts/key maps we previously overloaded them
14 with. If distributions want to continue to default to a
15 non-kernel font or key map they should ship a default
16 /etc/vconsole.conf with the appropriate contents.
18 Contributions from: Colin Walters, Daniel J Walsh, Dave
19 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
20 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
24 * journalctl gained a new --cursor= switch to show entries
25 starting from the specified location in the journal.
27 * We now enforce a size limit on journal entry fields exported
28 with "-o json" in journalctl. Fields larger than 4K will be
29 assigned null. This can be turned off with --all.
31 * An (optional) journal gateway daemon is now available as
32 "systemd-journal-gatewayd.service". This service provides
33 access to the journal via HTTP and JSON. This functionality
34 will be used to implement live log synchronization in both
35 pull and push modes, but has various other users too, such
36 as easy log access for debugging of embedded devices. Right
37 now it is already useful to retrieve the journal via HTTP:
39 # systemctl start systemd-journal-gatewayd.service
40 # wget http://localhost:19531/entries
42 This will download the journal contents in a
43 /var/log/messages compatible format. The same as JSON:
45 # curl -H"Accept: application/json" http://localhost:19531/entries
47 This service is also accessible via a web browser where a
48 single static HTML5 app is served that uses the JSON logic
49 to enable the user to do some basic browsing of the
50 journal. This will be extended later on. Here's an example
51 screenshot of this app in its current state:
53 http://0pointer.de/public/journal-gatewayd
55 Contributions from: Kay Sievers, Lennart Poettering, Robert
56 Milasan, Tom Gundersen
60 * The bash completion logic is now available for journalctl
63 * We don't mount the "cpuset" controller anymore together with
64 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
65 started if no parameters are assigned to it. "cpuset" hence
66 broke code that assumed it it could create "cpu" groups and
69 * journalctl -f will now subscribe to terminal size changes,
70 and line break accordingly.
72 Contributions from: Dave Reisner, Kay Sievers, Lennart
73 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
77 * nspawn will now create a symlink /etc/localtime in the
78 container environment, copying the host's timezone
79 setting. Previously this has been done via a bind mount, but
80 since symlinks cannot be bind mounted this has now been
81 changed to create/update the appropriate symlink.
83 * journalctl -n's line number argument is now optional, and
84 will default to 10 if omitted.
86 * journald will now log the maximum size the journal files may
87 take up on disk. This is particularly useful if the default
88 built-in logic of determining this parameter from the file
89 system size is used. Use "systemctl status
90 systemd-journald.service" to see this information.
92 * The multi-seat X wrapper tool has been stripped down. As X
93 is now capable of enumerating graphics devices via udev in a
94 seat-aware way the wrapper is not strictly necessary
95 anymore. A stripped down temporary stop-gap is still shipped
96 until the upstream display managers have been updated to
97 fully support the new X logic. Expect this wrapper to be
98 removed entirely in one of the next releases.
100 * HandleSleepKey= in logind.conf has been split up into
101 HandleSuspendKey= and HandleHibernateKey=. The old setting
102 is not available anymore. X11 and the kernel are
103 distuingishing between these keys and we should too. This
104 also means the inhibition lock for these keys has been split
107 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
108 Poettering, Lukas Nykryn, Václav Pavlín
112 * Whenever a unit changes state we'll now log this to the
113 journal and show along the unit's own log output in
116 * ConditionPathIsMountPoint= can now properly detect bind
117 mount points too. (Previously, a bind mount of one file
118 system to another place in the same file system could not be
119 detected as mount, since they shared struct stat's st_dev
122 * We will now mount the cgroup controllers cpu, cpuacct,
123 cpuset and the controllers net_cls, net_prio together by
126 * nspawn containers will now have a virtualized boot
127 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
128 over with a randomized ID at container initialization). This
129 has the effect of making "journalctl -b" do the right thing
132 * The JSON output journal serialization has been updated not
133 to generate "endless" list objects anymore, but rather one
134 JSON object per line. This is more in line how most JSON
135 parsers expect JSON objects. The new output mode
136 "json-pretty" has been added to provide similar output, but
137 neatly aligned for readability by humans.
139 * We dropped all explicit sync() invocations in the shutdown
140 code. The kernel does this implicitly anyway in the kernel
141 reboot() syscall. halt(8)'s -n option is now a compatibility
144 * We now support virtualized reboot() in containers, as
145 supported by newer kernels. We will fall back to exit() if
146 CAP_SYS_REBOOT is not available to the container. Also,
147 nspawn makes use of this now and will actually reboot the
148 container if the containerized OS asks for that.
150 * journalctl will only show local log output by default
151 now. Use --merge (-m) to show remote log output, too.
153 * libsystemd-journal gained the new sd_journal_get_usage()
154 call to determine the current disk usage of all journal
155 files. This is exposed in the new "journalctl --disk-usage"
158 * journald gained a new configuration setting SplitMode= in
159 journald.conf which may be used to control how user journals
160 are split off. See journald.conf(5) for details.
162 * A new condition type ConditionFileNotEmpty= has been added.
164 * tmpfiles' "w" lines now support file globbing, to write
165 multiple files at once.
167 * We added Python bindings for the journal submission
168 APIs. More Python APIs for a number of selected APIs will
169 likely follow. Note that we intend to add native bindings
170 only for the Python language, as we consider it common
171 enough to deserve bindings shipped within systemd. There are
172 various projects outside of systemd that provide bindings
173 for languages such as PHP or Lua.
175 * Many conditions will now resolve specifiers such as %i. In
176 addition, PathChanged= and related directives of .path units
177 now support specifiers as well.
179 * There's now a new RPM macro definition for the system preset
182 * journald will now warn if it can't foward a message to the
183 syslog daemon because it's socket is full.
185 * timedated will no longer write or process /etc/timezone,
186 except on Debian. As we do not support late mounted /usr
187 anymore /etc/localtime always being a symlink is now safe,
188 and hence the information in /etc/timezone is not necessary
191 * logind will now always reserve one VT for a text getty (VT6
192 by default). Previously if more than 6 X sessions where
193 started they took up all the VTs with auto-spawned gettys,
194 so that no text gettys were available anymore.
196 * udev will now automatically inform the btrfs kernel logic
197 about btrfs RAID components showing up. This should make
198 simple hotplug based btrfs RAID assembly work.
200 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
201 (but not for its children which will stay at the kernel
202 default). This should allow setups with a lot more listening
205 * systemd will now always pass the configured timezone to the
206 kernel at boot. timedated will do the same when the timezone
209 * logind's inhibition logic has been updated. By default,
210 logind will now handle the lid switch, the power and sleep
211 keys all the time, even in graphical sessions. If DEs want
212 to handle these events on their own they should take the new
213 handle-power-key, handle-sleep-key and handle-lid-switch
214 inhibitors during their runtime. A simple way to achiveve
215 that is to invoke the DE wrapped in an invocation of:
217 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
219 * Access to unit operations is now checked via SELinux taking
220 the unit file label and client process label into account.
222 * systemd will now notify the administrator in the journal
223 when he over-mounts a non-empty directory.
225 * There are new specifiers that are resolved in unit files,
226 for the host name (%H), the machine ID (%m) and the boot ID
229 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
230 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
231 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
232 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
233 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
234 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
235 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
239 * Support for reading structured kernel messages from
240 /dev/kmsg has now been added and is enabled by default.
242 * Support for reading kernel messages from /proc/kmsg has now
243 been removed. If you want kernel messages in the journal
244 make sure to run a recent kernel (>= 3.5) that supports
245 reading structured messages from /dev/kmsg (see
246 above). /proc/kmsg is now exclusive property of classic
247 syslog daemons again.
249 * The libudev API gained the new
250 udev_device_new_from_device_id() call.
252 * The logic for file system namespace (ReadOnlyDirectory=,
253 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
254 require pivot_root() anymore. This means fewer temporary
255 directories are created below /tmp for this feature.
257 * nspawn containers will now see and receive all submounts
258 made on the host OS below the root file system of the
261 * Forward Secure Sealing is now supported for Journal files,
262 which provide cryptographical sealing of journal files so
263 that attackers cannot alter log history anymore without this
264 being detectable. Lennart will soon post a blog story about
265 this explaining it in more detail.
267 * There are two new service settings RestartPreventExitStatus=
268 and SuccessExitStatus= which allow configuration of exit
269 status (exit code or signal) which will be excepted from the
270 restart logic, resp. consider successful.
272 * journalctl gained the new --verify switch that can be used
273 to check the integrity of the structure of journal files and
274 (if Forward Secure Sealing is enabled) the contents of
277 * nspawn containers will now be run with /dev/stdin, /dev/fd/
278 and similar symlinks pre-created. This makes running shells
279 as container init process a lot more fun.
281 * The fstab support can now handle PARTUUID= and PARTLABEL=
284 * A new ConditionHost= condition has been added to match
285 against the hostname (with globs) and machine ID. This is
286 useful for clusters where a single OS image is used to
287 provision a large number of hosts which shall run slightly
288 different sets of services.
290 * Services which hit the restart limit will now be placed in a
293 Contributions from: Bertram Poettering, Dave Reisner, Huang
294 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
295 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
299 * When running in --user mode systemd will now become a
300 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
301 tree a lot more organized.
303 * A new PartOf= unit dependency type has been introduced that
304 may be used to group services in a natural way.
306 * "systemctl enable" may now be used to enable instances of
309 * journalctl now prints error log levels in red, and
310 warning/notice log levels in bright white. It also supports
311 filtering by log level now.
313 * cgtop gained a new -n switch (similar to top), to configure
314 the maximum number of iterations to run for. It also gained
315 -b, to run in batch mode (accepting no input).
317 * The suffix ".service" may now be ommited on most systemctl
318 command lines involving service unit names.
320 * There's a new bus call in logind to lock all sessions, as
321 well as a loginctl verb for it "lock-sessions".
323 * libsystemd-logind.so gained a new call sd_journal_perror()
324 that works similar to libc perror() but logs to the journal
325 and encodes structured information about the error number.
327 * /etc/crypttab entries now understand the new keyfile-size=
330 * shutdown(8) now can send a (configurable) wall message when
331 a shutdown is cancelled.
333 * The mount propagation mode for the root file system will now
334 default to "shared", which is useful to make containers work
335 nicely out-of-the-box so that they receive new mounts from
336 the host. This can be undone locally by running "mount
337 --make-rprivate /" if needed.
339 * The prefdm.service file has been removed. Distributions
340 should maintain this unit downstream if they intend to keep
341 it around. However, we recommend writing normal unit files
342 for display managers instead.
344 * Since systemd is a crucial part of the OS we will now
345 default to a number of compiler switches that improve
346 security (hardening) such as read-only relocations, stack
347 protection, and suchlike.
349 * The TimeoutSec= setting for services is now split into
350 TimeoutStartSec= and TimeoutStopSec= to allow configuration
351 of individual time outs for the start and the stop phase of
354 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
355 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
356 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
357 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
358 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
359 Gundersen, Zbigniew Jędrzejewski-Szmek
363 * The journal and id128 C APIs are now fully documented as man
366 * Extra safety checks have been added when transitioning from
367 the initial RAM disk to the main system to avoid accidental
370 * /etc/crypttab entries now understand the new keyfile-offset=
373 * systemctl -t can now be used to filter by unit load state.
375 * The journal C API gained the new sd_journal_wait() call to
376 make writing synchronous journal clients easier.
378 * journalctl gained the new -D switch to show journals from a
381 * journalctl now displays a special marker between log
382 messages of two different boots.
384 * The journal is now explicitly flushed to /var via a service
385 systemd-journal-flush.service, rather than implicitly simply
386 by seeing /var/log/journal to be writable.
388 * journalctl (and the journal C APIs) can now match for much
389 more complex expressions, with alternatives and
392 * When transitioning from the initial RAM disk to the main
393 system we will now kill all processes in a killing spree to
394 ensure no processes stay around by accident.
396 * Three new specifiers may be used in unit files: %u, %h, %s
397 resolve to the user name, user home directory resp. user
398 shell. This is useful for running systemd user instances.
400 * We now automatically rotate journal files if their data
401 object hash table gets a fill level > 75%. We also size the
402 hash table based on the configured maximum file size. This
403 together should lower hash collisions drastically and thus
404 speed things up a bit.
406 * journalctl gained the new "--header" switch to introspect
407 header data of journal files.
409 * A new setting SystemCallFilters= has been added to services
410 which may be used to apply blacklists or whitelists to
411 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
413 * nspawn gained a new --link-journal= switch (and quicker: -j)
414 to link the container journal with the host. This makes it
415 very easy to centralize log viewing on the host for all
416 guests while still keeping the journal files separated.
418 * Many bugfixes and optimizations
420 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
421 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
422 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
427 * Several tools now understand kernel command line arguments,
428 which are only read when run in an initial RAM disk. They
429 usually follow closely their normal counterparts, but are
432 * There's a new tool to analyze the readahead files that are
433 automatically generated at boot. Use:
435 /usr/lib/systemd/systemd-readahead analyze /.readahead
437 * We now provide an early debug shell on tty9 if this enabled. Use:
439 systemctl enable debug-shell.service
441 * All plymouth related units have been moved into the Plymouth
442 package. Please make sure to upgrade your Plymouth version
445 * systemd-tmpfiles now supports getting passed the basename of
446 a configuration file only, in which case it will look for it
447 in all appropriate directories automatically.
449 * udevadm info now takes a /dev or /sys path as argument, and
450 does the right thing. Example:
452 udevadm info /dev/sda
453 udevadm info /sys/class/block/sda
455 * systemctl now prints a warning if a unit is stopped but a
456 unit that might trigger it continues to run. Example: a
457 service is stopped but the socket that activates it is left
460 * "systemctl status" will now mention if the log output was
461 shortened due to rotation since a service has been started.
463 * The journal API now exposes functions to determine the
464 "cutoff" times due to rotation.
466 * journald now understands SIGUSR1 and SIGUSR2 for triggering
467 immediately flushing of runtime logs to /var if possible,
468 resp. for triggering immediate rotation of the journal
471 * It is now considered an error if a service is attempted to
472 be stopped that is not loaded.
474 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
476 * systemd-analyze now supports Python 3
478 * tmpfiles now supports cleaning up directories via aging
479 where the first level dirs are always kept around but
480 directories beneath it automatically aged. This is enabled
481 by prefixing the age field with '~'.
483 * Seat objects now expose CanGraphical, CanTTY properties
484 which is required to deal with very fast bootups where the
485 display manager might be running before the graphics drivers
486 completed initialization.
488 * Seat objects now expose a State property.
490 * We now include RPM macros for service enabling/disabling
491 based on the preset logic. We recommend RPM based
492 distributions to make use of these macros if possible. This
493 makes it simpler to reuse RPM spec files across
496 * We now make sure that the collected systemd unit name is
497 always valid when services log to the journal via
500 * There's a new man page kernel-command-line(7) detailing all
501 command line options we understand.
503 * The fstab generator may now be disabled at boot by passing
504 fstab=0 on the kernel command line.
506 * A new kernel command line option modules-load= is now understood
507 to load a specific kernel module statically, early at boot.
509 * Unit names specified on the systemctl command line are now
510 automatically escaped as needed. Also, if file system or
511 device paths are specified they are automatically turned
512 into the appropriate mount or device unit names. Example:
514 systemctl status /home
515 systemctl status /dev/sda
517 * The SysVConsole= configuration option has been removed from
520 * The SysV search path is no longer exported on the D-Bus
523 * The Names= option is been removed from unit file parsing.
525 * There's a new man page bootup(7) detailing the boot process.
527 * Every unit and every generator we ship with systemd now
528 comes with full documentation. The self-explanatory boot is
531 * A couple of services gained "systemd-" prefixes in their
532 name if they wrap systemd code, rather than only external
533 code. Among them fsck@.service which is now
534 systemd-fsck@.service.
536 * The HaveWatchdog property has been removed from the D-Bus
539 * systemd.confirm_spawn= on the kernel command line should now
542 * There's a new man page crypttab(5) which details all options
543 we actually understand.
545 * systemd-nspawn gained a new --capability= switch to pass
546 additional capabilities to the container.
548 * timedated will now read known NTP implementation unit names
549 from /usr/lib/systemd/ntp-units.d/*.list,
550 systemd-timedated-ntp.target has been removed.
552 * journalctl gained a new switch "-b" that lists log data of
553 the current boot only.
555 * The notify socket is in the abstract namespace again, in
556 order to support daemons which chroot() at start-up.
558 * There is a new Storage= configuration option for journald
559 which allows configuration of where log data should go. This
560 also provides a way to disable journal logging entirely, so
561 that data collected is only forwarded to the console, the
562 kernel log buffer or another syslog implementation.
564 * Many bugfixes and optimizations
566 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
567 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
568 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
569 Shawn Landden, Tom Gundersen
573 * "systemctl help <unit>" now shows the man page if one is
576 * Several new man pages have been added.
578 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
579 MaxLevelConsole= can now be specified in
580 journald.conf. These options allow reducing the amount of
581 data stored on disk or forwarded by the log level.
583 * TimerSlackNSec= can now be specified in system.conf for
584 PID1. This allows system-wide power savings.
586 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
587 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
592 * logind is now capable of (optionally) handling power and
593 sleep keys as well as the lid switch.
595 * journalctl now understands the syntax "journalctl
596 /usr/bin/avahi-daemon" to get all log output of a specific
599 * CapabilityBoundingSet= in system.conf now also influences
600 the capability bound set of usermode helpers of the kernel.
602 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
603 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
604 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
605 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
609 * Note that we skipped 139 releases here in order to set the
610 new version to something that is greater than both udev's
611 and systemd's most recent version number.
613 * udev: all udev sources are merged into the systemd source tree now.
614 All future udev development will happen in the systemd tree. It
615 is still fully supported to use the udev daemon and tools without
616 systemd running, like in initramfs or other init systems. Building
617 udev though, will require the *build* of the systemd tree, but
618 udev can be properly *run* without systemd.
620 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
621 should be used to create dead device nodes as workarounds for broken
624 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
625 no longer supported. udev_monitor_new_from_netlink() needs to be
626 used to subscribe to events.
628 * udev: when udevd is started by systemd, processes which are left
629 behind by forking them off of udev rules, are unconditionally cleaned
630 up and killed now after the event handling has finished. Services or
631 daemons must be started as systemd services. Services can be
632 pulled-in by udev to get started, but they can no longer be directly
633 forked by udev rules.
635 * udev: the daemon binary is called systemd-udevd now and installed
636 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
637 to adapt to that, create symlink, or rename the binary after building
640 * libudev no longer provides these symbols:
641 udev_monitor_from_socket()
642 udev_queue_get_failed_list_entry()
643 udev_get_{dev,sys,run}_path()
644 The versions number was bumped and symbol versioning introduced.
646 * systemd-loginctl and systemd-journalctl have been renamed
647 to loginctl and journalctl to match systemctl.
649 * The config files: /etc/systemd/systemd-logind.conf and
650 /etc/systemd/systemd-journald.conf have been renamed to
651 logind.conf and journald.conf. Package updates should rename
652 the files to the new names on upgrade.
654 * For almost all files the license is now LGPL2.1+, changed
655 from the previous GPL2.0+. Exceptions are some minor stuff
656 of udev (which will be changed to LGPL2.1 eventually, too),
657 and the MIT licensed sd-daemon.[ch] library that is suitable
658 to be used as drop-in files.
660 * systemd and logind now handle system sleep states, in
661 particular suspending and hibernating.
663 * logind now implements a sleep/shutdown/idle inhibiting logic
664 suitable for a variety of uses. Soonishly Lennart will blog
665 about this in more detail.
667 * var-run.mount and var-lock.mount are no longer provided
668 (which prevously bind mounted these directories to their new
669 places). Distributions which have not converted these
670 directories to symlinks should consider stealing these files
671 from git history and add them downstream.
673 * We introduced the Documentation= field for units and added
674 this to all our shipped units. This is useful to make it
675 easier to explore the boot and the purpose of the various
678 * All smaller setup units (such as
679 systemd-vconsole-setup.service) now detect properly if they
680 are run in a container and are skipped when
681 appropriate. This guarantees an entirely noise-free boot in
682 Linux container environments such as systemd-nspawn.
684 * A framework for implementing offline system updates is now
685 integrated, for details see:
686 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
688 * A new service type Type=idle is available now which helps us
689 avoiding ugly interleaving of getty output and boot status
692 * There's now a system-wide CapabilityBoundingSet= option to
693 globally reduce the set of capabilities for the
694 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
695 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
696 even CAP_NET_ADMIN system-wide for secure systems.
698 * There are now system-wide DefaultLimitXXX= options to
699 globally change the defaults of the various resource limits
700 for all units started by PID 1.
702 * Harald Hoyer's systemd test suite has been integrated into
703 systemd which allows easy testing of systemd builds in qemu
704 and nspawn. (This is really awesome! Ask us for details!)
706 * The fstab parser is now implemented as generator, not inside
709 * systemctl will now warn you if .mount units generated from
710 /etc/fstab are out of date due to changes in fstab that
711 haven't been read by systemd yet.
713 * systemd is now suitable for usage in initrds. Dracut has
714 already been updated to make use of this. With this in place
715 initrds get a slight bit faster but primarily are much
716 easier to introspect and debug since "systemctl status" in
717 the host system can be used to introspect initrd services,
718 and the journal from the initrd is kept around too.
720 * systemd-delta has been added, a tool to explore differences
721 between user/admin configuration and vendor defaults.
723 * PrivateTmp= now affects both /tmp and /var/tmp.
725 * Boot time status messages are now much prettier and feature
726 proper english language. Booting up systemd has never been
729 * Read-ahead pack files now include the inode number of all
730 files to pre-cache. When the inode changes the pre-caching
731 is not attempted. This should be nicer to deal with updated
732 packages which might result in changes of read-ahead
735 * We now temporaritly lower the kernel's read_ahead_kb variable
736 when collecting read-ahead data to ensure the kernel's
737 built-in read-ahead does not add noise to our measurements
738 of necessary blocks to pre-cache.
740 * There's now RequiresMountsFor= to add automatic dependencies
741 for all mounts necessary for a specific file system path.
743 * MountAuto= and SwapAuto= have been removed from
744 system.conf. Mounting file systems at boot has to take place
747 * nspawn now learned a new switch --uuid= to set the machine
748 ID on the command line.
750 * nspawn now learned the -b switch to automatically search
753 * vt102 is now the default TERM for serial TTYs, upgraded from
756 * systemd-logind now works on VT-less systems.
758 * The build tree has been reorganized. The individual
759 components now have directories of their own.
761 * A new condition type ConditionPathIsReadWrite= is now available.
763 * nspawn learned the new -C switch to create cgroups for the
764 container in other hierarchies.
766 * We now have support for hardware watchdogs, configurable in
769 * The scheduled shutdown logic now has a public API.
771 * We now mount /tmp as tmpfs by default, but this can be
772 masked and /etc/fstab can override it.
774 * Since udisks doesn't make use of /media anymore we are not
775 mounting a tmpfs on it anymore.
777 * journalctl gained a new --local switch to only interleave
778 locally generated journal files.
780 * We can now load the IMA policy at boot automatically.
782 * The GTK tools have been split off into a systemd-ui.
784 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
785 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
786 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
787 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
788 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
789 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
790 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
791 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
792 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
797 * This is mostly a bugfix release
799 * Support optional initialization of the machine ID from the
800 KVM or container configured UUID.
802 * Support immediate reboots with "systemctl reboot -ff"
804 * Show /etc/os-release data in systemd-analyze output
806 * Many bugfixes for the journal, including endianess fixes and
807 ensuring that disk space enforcement works
809 * sd-login.h is C++ comptaible again
811 * Extend the /etc/os-release format on request of the Debian
814 * We now refuse non-UTF8 strings used in various configuration
815 and unit files. This is done to ensure we don't pass invalid
816 data over D-Bus or expose it elsewhere.
818 * Register Mimo USB Screens as suitable for automatic seat
821 * Read SELinux client context from journal clients in a race
824 * Reorder configuration file lookup order. /etc now always
825 overrides /run in order to allow the administrator to always
826 and unconditionally override vendor supplied or
827 automatically generated data.
829 * The various user visible bits of the journal now have man
830 pages. We still lack man pages for the journal API calls
833 * We now ship all man pages in HTML format again in the
836 Contributions from: Dave Reisner, Dirk Eibach, Frederic
837 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
838 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
843 * This is mostly a bugfix release
845 * systems lacking /etc/os-release are no longer supported.
847 * Various functionality updates to libsystemd-login.so
849 * Track class of PAM logins to distuingish greeters from
852 Contributions from: Kay Sievers, Lennart Poettering, Michael
857 * This is an important bugfix release for v41.
859 * Building man pages is now optional which should be useful
860 for those building systemd from git but unwilling to install
863 * Watchdog support for supervising services is now usable. In
864 a future release support for hardware watchdogs
865 (i.e. /dev/watchdog) will be added building on this.
867 * Service start rate limiting is now configurable and can be
868 turned off per service. When a start rate limit is hit a
869 reboot can automatically be triggered.
871 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
873 Contributions from: Benjamin Franzke, Bill Nottingham,
874 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
875 Schmidt, Michał Górny, Piotr Drąg
879 * The systemd binary is installed /usr/lib/systemd/systemd now;
880 An existing /sbin/init symlink needs to be adapted with the
883 * The code that loads kernel modules has been ported to invoke
884 libkmod directly, instead of modprobe. This means we do not
885 support systems with module-init-tools anymore.
887 * Watchdog support is now already useful, but still not
890 * A new kernel command line option systemd.setenv= is
891 understood to set system wide environment variables
894 * We now limit the set of capabilities of systemd-journald.
896 * We now set SIGPIPE to ignore by default, since it only is
897 useful in shell pipelines, and has little use in general
898 code. This can be disabled with IgnoreSIPIPE=no in unit
901 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
902 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
907 * This is mostly a bugfix release
909 * We now expose the reason why a service failed in the
910 "Result" D-Bus property.
912 * Rudimentary service watchdog support (will be completed over
913 the next few releases.)
915 * When systemd forks off in order execute some service we will
916 now immediately changes its argv[0] to reflect which process
917 it will execute. This is useful to minimize the time window
918 with a generic argv[0], which makes bootcharts more useful
920 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
921 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
922 Mike Kazantsev, Ray Strode
926 * This is mostly a test release, but incorporates many
929 * New systemd-cgtop tool to show control groups by their
932 * Linking against libacl for ACLs is optional again. If
933 disabled, support tracking device access for active logins
934 goes becomes unavailable, and so does access to the user
935 journals by the respective users.
937 * If a group "adm" exists, journal files are automatically
938 owned by them, thus allow members of this group full access
939 to the system journal as well as all user journals.
941 * The journal now stores the SELinux context of the logging
942 client for all entries.
944 * Add C++ inclusion guards to all public headers
946 * New output mode "cat" in the journal to print only text
947 messages, without any meta data like date or time.
949 * Include tiny X server wrapper as a temporary stop-gap to
950 teach XOrg udev display enumeration. This is used by display
951 managers such as gdm, and will go away as soon as XOrg
952 learned native udev hotplugging for display devices.
954 * Add new systemd-cat tool for executing arbitrary programs
955 with STDERR/STDOUT connected to the journal. Can also act as
956 BSD logger replacement, and does so by default.
958 * Optionally store all locally generated coredumps in the
959 journal along with meta data.
961 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
962 writing short strings to files (for usage for /sys), and for
963 creating symlinks, character and block device nodes.
965 * New unit file option ControlGroupPersistent= to make cgroups
966 persistent, following the mechanisms outlined in
967 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
969 * Support multiple local RTCs in a sane way
971 * No longer monopolize IO when replaying readahead data on
972 rotating disks, since we might starve non-file-system IO to
973 death, since fanotify() will not see accesses done by blkid,
976 * Don't show kernel threads in systemd-cgls anymore, unless
977 requested with new -k switch.
979 Contributions from: Dan Horák, Kay Sievers, Lennart
980 Poettering, Michal Schmidt
984 * This is mostly a test release, but incorporates many
987 * The git repository moved to:
988 git://anongit.freedesktop.org/systemd/systemd
989 ssh://git.freedesktop.org/git/systemd/systemd
991 * First release with the journal
992 http://0pointer.de/blog/projects/the-journal.html
994 * The journal replaces both systemd-kmsg-syslogd and
995 systemd-stdout-bridge.
997 * New sd_pid_get_unit() API call in libsystemd-logind
999 * Many systemadm clean-ups
1001 * Introduce remote-fs-pre.target which is ordered before all
1002 remote mounts and may be used to start services before all
1005 * Added Mageia support
1007 * Add bash completion for systemd-loginctl
1009 * Actively monitor PID file creation for daemons which exit in
1010 the parent process before having finished writing the PID
1011 file in the daemon process. Daemons which do this need to be
1012 fixed (i.e. PID file creation must have finished before the
1013 parent exits), but we now react a bit more gracefully to them.
1015 * Add colourful boot output, mimicking the well-known output
1016 of existing distributions.
1018 * New option PassCredentials= for socket units, for
1019 compatibility with a recent kernel ABI breakage.
1021 * /etc/rc.local is now hooked in via a generator binary, and
1022 thus will no longer act as synchronization point during
1025 * systemctl list-unit-files now supports --root=.
1027 * systemd-tmpfiles now understands two new commands: z, Z for
1028 relabelling files according to the SELinux database. This is
1029 useful to apply SELinux labels to specific files in /sys,
1032 * Output of SysV services is now forwarded to both the console
1033 and the journal by default, not only just the console.
1035 * New man pages for all APIs from libsystemd-login.
1037 * The build tree got reorganized and a the build system is a
1038 lot more modular allowing embedded setups to specifically
1039 select the components of systemd they are interested in.
1041 * Support for Linux systems lacking the kernel VT subsystem is
1044 * configure's --with-rootdir= got renamed to
1045 --with-rootprefix= to follow the naming used by udev and
1048 * Unless specified otherwise we'll now install to /usr instead
1049 of /usr/local by default.
1051 * Processes with '@' in argv[0][0] are now excluded from the
1052 final shut-down killing spree, following the logic explained
1054 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1056 * All processes remaining in a service cgroup when we enter
1057 the START or START_PRE states are now killed with
1058 SIGKILL. That means it is no longer possible to spawn
1059 background processes from ExecStart= lines (which was never
1060 supported anyway, and bad style).
1062 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1063 reloading of units together.
1065 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1066 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1067 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1068 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1069 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek