1 systemd System and Service Manager
5 * journalctl gained a new --cursor= switch to show entries
6 starting from the specified location in the journal.
8 * We now enforce a size limit on journal entry fields exported
9 with "-o json" in journalctl. Fields larger than 4K will be
10 assigned null. This can be turned off with --all.
12 * An (optional) journal gateway daemon is now available as
13 "systemd-journal-gatewayd.service". This service provides
14 access to the journal via HTTP and JSON. This functionality
15 will be used to implement live log synchronization in both
16 pull and push modes, but has various other users too, such
17 as easy log access for debugging of embedded devices. Right
18 now it is already useful to retrieve the journal via HTTP:
20 # systemctl start systemd-journal-gatewayd.service
21 # wget http://localhost:19531/entries
23 This will download the journal contents in a
24 /var/log/messages compatible format. The same as JSON:
26 # curl -H"Accept: application/json" http://localhost:19531/entries
28 This service is also accessible via a web browser where a
29 single static HTML5 app is served that uses the JSON logic
30 to enable the user to do some basic browsing of the
31 journal. This will be extended later on. Here's an example
32 screenshot of this app in its current state:
34 http://0pointer.de/public/journal-gatewayd
36 Contributions from: Kay Sievers, Lennart Poettering, Robert
37 Milasan, Tom Gundersen
41 * The bash completion logic is now available for journalctl
44 * We don't mount the "cpuset" controller anymore together with
45 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
46 started if no parameters are assigned to it. "cpuset" hence
47 broke code that assumed it it could create "cpu" groups and
50 * journalctl -f will now subscribe to terminal size changes,
51 and line break accordingly.
53 Contributions from: Dave Reisner, Kay Sievers, Lennart
54 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
58 * nspawn will now create a symlink /etc/localtime in the
59 container environment, copying the host's timezone
60 setting. Previously this has been done via a bind mount, but
61 since symlinks cannot be bind mounted this has now been
62 changed to create/update the appropriate symlink.
64 * journalctl -n's line number argument is now optional, and
65 will default to 10 if omitted.
67 * journald will now log the maximum size the journal files may
68 take up on disk. This is particularly useful if the default
69 built-in logic of determining this parameter from the file
70 system size is used. Use "systemctl status
71 systemd-journald.service" to see this information.
73 * The multi-seat X wrapper tool has been stripped down. As X
74 is now capable of enumerating graphics devices via udev in a
75 seat-aware way the wrapper is not strictly necessary
76 anymore. A stripped down temporary stop-gap is still shipped
77 until the upstream display managers have been updated to
78 fully support the new X logic. Expect this wrapper to be
79 removed entirely in one of the next releases.
81 * HandleSleepKey= in logind.conf has been split up into
82 HandleSuspendKey= and HandleHibernateKey=. The old setting
83 is not available anymore. X11 and the kernel are
84 distuingishing between these keys and we should too. This
85 also means the inhibition lock for these keys has been split
88 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
89 Poettering, Lukas Nykryn, Václav Pavlín
93 * Whenever a unit changes state we'll now log this to the
94 journal and show along the unit's own log output in
97 * ConditionPathIsMountPoint= can now properly detect bind
98 mount points too. (Previously, a bind mount of one file
99 system to another place in the same file system could not be
100 detected as mount, since they shared struct stat's st_dev
103 * We will now mount the cgroup controllers cpu, cpuacct,
104 cpuset and the controllers net_cls, net_prio together by
107 * nspawn containers will now have a virtualized boot
108 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
109 over with a randomized ID at container initialization). This
110 has the effect of making "journalctl -b" do the right thing
113 * The JSON output journal serialization has been updated not
114 to generate "endless" list objects anymore, but rather one
115 JSON object per line. This is more in line how most JSON
116 parsers expect JSON objects. The new output mode
117 "json-pretty" has been added to provide similar output, but
118 neatly aligned for readability by humans.
120 * We dropped all explicit sync() invocations in the shutdown
121 code. The kernel does this implicitly anyway in the kernel
122 reboot() syscall. halt(8)'s -n option is now a compatibility
125 * We now support virtualized reboot() in containers, as
126 supported by newer kernels. We will fall back to exit() if
127 CAP_SYS_REBOOT is not available to the container. Also,
128 nspawn makes use of this now and will actually reboot the
129 container if the containerized OS asks for that.
131 * journalctl will only show local log output by default
132 now. Use --merge (-m) to show remote log output, too.
134 * libsystemd-journal gained the new sd_journal_get_usage()
135 call to determine the current disk usage of all journal
136 files. This is exposed in the new "journalctl --disk-usage"
139 * journald gained a new configuration setting SplitMode= in
140 journald.conf which may be used to control how user journals
141 are split off. See journald.conf(5) for details.
143 * A new condition type ConditionFileNotEmpty= has been added.
145 * tmpfiles' "w" lines now support file globbing, to write
146 multiple files at once.
148 * We added Python bindings for the journal submission
149 APIs. More Python APIs for a number of selected APIs will
150 likely follow. Note that we intend to add native bindings
151 only for the Python language, as we consider it common
152 enough to deserve bindings shipped within systemd. There are
153 various projects outside of systemd that provide bindings
154 for languages such as PHP or Lua.
156 * Many conditions will now resolve specifiers such as %i. In
157 addition, PathChanged= and related directives of .path units
158 now support specifiers as well.
160 * There's now a new RPM macro definition for the system preset
163 * journald will now warn if it can't foward a message to the
164 syslog daemon because it's socket is full.
166 * timedated will no longer write or process /etc/timezone,
167 except on Debian. As we do not support late mounted /usr
168 anymore /etc/localtime always being a symlink is now safe,
169 and hence the information in /etc/timezone is not necessary
172 * logind will now always reserve one VT for a text getty (VT6
173 by default). Previously if more than 6 X sessions where
174 started they took up all the VTs with auto-spawned gettys,
175 so that no text gettys were available anymore.
177 * udev will now automatically inform the btrfs kernel logic
178 about btrfs RAID components showing up. This should make
179 simple hotplug based btrfs RAID assembly work.
181 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
182 (but not for its children which will stay at the kernel
183 default). This should allow setups with a lot more listening
186 * systemd will now always pass the configured timezone to the
187 kernel at boot. timedated will do the same when the timezone
190 * logind's inhibition logic has been updated. By default,
191 logind will now handle the lid switch, the power and sleep
192 keys all the time, even in graphical sessions. If DEs want
193 to handle these events on their own they should take the new
194 handle-power-key, handle-sleep-key and handle-lid-switch
195 inhibitors during their runtime. A simple way to achiveve
196 that is to invoke the DE wrapped in an invocation of:
198 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
200 * Access to unit operations is now checked via SELinux taking
201 the unit file label and client process label into account.
203 * systemd will now notify the administrator in the journal
204 when he over-mounts a non-empty directory.
206 * There are new specifiers that are resolved in unit files,
207 for the host name (%H), the machine ID (%m) and the boot ID
210 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
211 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
212 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
213 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
214 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
215 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
216 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
220 * Support for reading structured kernel messages from
221 /dev/kmsg has now been added and is enabled by default.
223 * Support for reading kernel messages from /proc/kmsg has now
224 been removed. If you want kernel messages in the journal
225 make sure to run a recent kernel (>= 3.5) that supports
226 reading structured messages from /dev/kmsg (see
227 above). /proc/kmsg is now exclusive property of classic
228 syslog daemons again.
230 * The libudev API gained the new
231 udev_device_new_from_device_id() call.
233 * The logic for file system namespace (ReadOnlyDirectory=,
234 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
235 require pivot_root() anymore. This means fewer temporary
236 directories are created below /tmp for this feature.
238 * nspawn containers will now see and receive all submounts
239 made on the host OS below the root file system of the
242 * Forward Secure Sealing is now supported for Journal files,
243 which provide cryptographical sealing of journal files so
244 that attackers cannot alter log history anymore without this
245 being detectable. Lennart will soon post a blog story about
246 this explaining it in more detail.
248 * There are two new service settings RestartPreventExitStatus=
249 and SuccessExitStatus= which allow configuration of exit
250 status (exit code or signal) which will be excepted from the
251 restart logic, resp. consider successful.
253 * journalctl gained the new --verify switch that can be used
254 to check the integrity of the structure of journal files and
255 (if Forward Secure Sealing is enabled) the contents of
258 * nspawn containers will now be run with /dev/stdin, /dev/fd/
259 and similar symlinks pre-created. This makes running shells
260 as container init process a lot more fun.
262 * The fstab support can now handle PARTUUID= and PARTLABEL=
265 * A new ConditionHost= condition has been added to match
266 against the hostname (with globs) and machine ID. This is
267 useful for clusters where a single OS image is used to
268 provision a large number of hosts which shall run slightly
269 different sets of services.
271 * Services which hit the restart limit will now be placed in a
274 Contributions from: Bertram Poettering, Dave Reisner, Huang
275 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
276 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
280 * When running in --user mode systemd will now become a
281 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
282 tree a lot more organized.
284 * A new PartOf= unit dependency type has been introduced that
285 may be used to group services in a natural way.
287 * "systemctl enable" may now be used to enable instances of
290 * journalctl now prints error log levels in red, and
291 warning/notice log levels in bright white. It also supports
292 filtering by log level now.
294 * cgtop gained a new -n switch (similar to top), to configure
295 the maximum number of iterations to run for. It also gained
296 -b, to run in batch mode (accepting no input).
298 * The suffix ".service" may now be ommited on most systemctl
299 command lines involving service unit names.
301 * There's a new bus call in logind to lock all sessions, as
302 well as a loginctl verb for it "lock-sessions".
304 * libsystemd-logind.so gained a new call sd_journal_perror()
305 that works similar to libc perror() but logs to the journal
306 and encodes structured information about the error number.
308 * /etc/crypttab entries now understand the new keyfile-size=
311 * shutdown(8) now can send a (configurable) wall message when
312 a shutdown is cancelled.
314 * The mount propagation mode for the root file system will now
315 default to "shared", which is useful to make containers work
316 nicely out-of-the-box so that they receive new mounts from
317 the host. This can be undone locally by running "mount
318 --make-rprivate /" if needed.
320 * The prefdm.service file has been removed. Distributions
321 should maintain this unit downstream if they intend to keep
322 it around. However, we recommend writing normal unit files
323 for display managers instead.
325 * Since systemd is a crucial part of the OS we will now
326 default to a number of compiler switches that improve
327 security (hardening) such as read-only relocations, stack
328 protection, and suchlike.
330 * The TimeoutSec= setting for services is now split into
331 TimeoutStartSec= and TimeoutStopSec= to allow configuration
332 of individual time outs for the start and the stop phase of
335 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
336 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
337 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
338 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
339 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
340 Gundersen, Zbigniew Jędrzejewski-Szmek
344 * The journal and id128 C APIs are now fully documented as man
347 * Extra safety checks have been added when transitioning from
348 the initial RAM disk to the main system to avoid accidental
351 * /etc/crypttab entries now understand the new keyfile-offset=
354 * systemctl -t can now be used to filter by unit load state.
356 * The journal C API gained the new sd_journal_wait() call to
357 make writing synchronous journal clients easier.
359 * journalctl gained the new -D switch to show journals from a
362 * journalctl now displays a special marker between log
363 messages of two different boots.
365 * The journal is now explicitly flushed to /var via a service
366 systemd-journal-flush.service, rather than implicitly simply
367 by seeing /var/log/journal to be writable.
369 * journalctl (and the journal C APIs) can now match for much
370 more complex expressions, with alternatives and
373 * When transitioning from the initial RAM disk to the main
374 system we will now kill all processes in a killing spree to
375 ensure no processes stay around by accident.
377 * Three new specifiers may be used in unit files: %u, %h, %s
378 resolve to the user name, user home directory resp. user
379 shell. This is useful for running systemd user instances.
381 * We now automatically rotate journal files if their data
382 object hash table gets a fill level > 75%. We also size the
383 hash table based on the configured maximum file size. This
384 together should lower hash collisions drastically and thus
385 speed things up a bit.
387 * journalctl gained the new "--header" switch to introspect
388 header data of journal files.
390 * A new setting SystemCallFilters= has been added to services
391 which may be used to apply blacklists or whitelists to
392 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
394 * nspawn gained a new --link-journal= switch (and quicker: -j)
395 to link the container journal with the host. This makes it
396 very easy to centralize log viewing on the host for all
397 guests while still keeping the journal files separated.
399 * Many bugfixes and optimizations
401 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
402 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
403 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
408 * Several tools now understand kernel command line arguments,
409 which are only read when run in an initial RAM disk. They
410 usually follow closely their normal counterparts, but are
413 * There's a new tool to analyze the readahead files that are
414 automatically generated at boot. Use:
416 /usr/lib/systemd/systemd-readahead analyze /.readahead
418 * We now provide an early debug shell on tty9 if this enabled. Use:
420 systemctl enable debug-shell.service
422 * All plymouth related units have been moved into the Plymouth
423 package. Please make sure to upgrade your Plymouth version
426 * systemd-tmpfiles now supports getting passed the basename of
427 a configuration file only, in which case it will look for it
428 in all appropriate directories automatically.
430 * udevadm info now takes a /dev or /sys path as argument, and
431 does the right thing. Example:
433 udevadm info /dev/sda
434 udevadm info /sys/class/block/sda
436 * systemctl now prints a warning if a unit is stopped but a
437 unit that might trigger it continues to run. Example: a
438 service is stopped but the socket that activates it is left
441 * "systemctl status" will now mention if the log output was
442 shortened due to rotation since a service has been started.
444 * The journal API now exposes functions to determine the
445 "cutoff" times due to rotation.
447 * journald now understands SIGUSR1 and SIGUSR2 for triggering
448 immediately flushing of runtime logs to /var if possible,
449 resp. for triggering immediate rotation of the journal
452 * It is now considered an error if a service is attempted to
453 be stopped that is not loaded.
455 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
457 * systemd-analyze now supports Python 3
459 * tmpfiles now supports cleaning up directories via aging
460 where the first level dirs are always kept around but
461 directories beneath it automatically aged. This is enabled
462 by prefixing the age field with '~'.
464 * Seat objects now expose CanGraphical, CanTTY properties
465 which is required to deal with very fast bootups where the
466 display manager might be running before the graphics drivers
467 completed initialization.
469 * Seat objects now expose a State property.
471 * We now include RPM macros for service enabling/disabling
472 based on the preset logic. We recommend RPM based
473 distributions to make use of these macros if possible. This
474 makes it simpler to reuse RPM spec files across
477 * We now make sure that the collected systemd unit name is
478 always valid when services log to the journal via
481 * There's a new man page kernel-command-line(7) detailing all
482 command line options we understand.
484 * The fstab generator may now be disabled at boot by passing
485 fstab=0 on the kernel command line.
487 * A new kernel command line option modules-load= is now understood
488 to load a specific kernel module statically, early at boot.
490 * Unit names specified on the systemctl command line are now
491 automatically escaped as needed. Also, if file system or
492 device paths are specified they are automatically turned
493 into the appropriate mount or device unit names. Example:
495 systemctl status /home
496 systemctl status /dev/sda
498 * The SysVConsole= configuration option has been removed from
501 * The SysV search path is no longer exported on the D-Bus
504 * The Names= option is been removed from unit file parsing.
506 * There's a new man page bootup(7) detailing the boot process.
508 * Every unit and every generator we ship with systemd now
509 comes with full documentation. The self-explanatory boot is
512 * A couple of services gained "systemd-" prefixes in their
513 name if they wrap systemd code, rather than only external
514 code. Among them fsck@.service which is now
515 systemd-fsck@.service.
517 * The HaveWatchdog property has been removed from the D-Bus
520 * systemd.confirm_spawn= on the kernel command line should now
523 * There's a new man page crypttab(5) which details all options
524 we actually understand.
526 * systemd-nspawn gained a new --capability= switch to pass
527 additional capabilities to the container.
529 * timedated will now read known NTP implementation unit names
530 from /usr/lib/systemd/ntp-units.d/*.list,
531 systemd-timedated-ntp.target has been removed.
533 * journalctl gained a new switch "-b" that lists log data of
534 the current boot only.
536 * The notify socket is in the abstract namespace again, in
537 order to support daemons which chroot() at start-up.
539 * There is a new Storage= configuration option for journald
540 which allows configuration of where log data should go. This
541 also provides a way to disable journal logging entirely, so
542 that data collected is only forwarded to the console, the
543 kernel log buffer or another syslog implementation.
545 * Many bugfixes and optimizations
547 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
548 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
549 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
550 Shawn Landden, Tom Gundersen
554 * "systemctl help <unit>" now shows the man page if one is
557 * Several new man pages have been added.
559 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
560 MaxLevelConsole= can now be specified in
561 journald.conf. These options allow reducing the amount of
562 data stored on disk or forwarded by the log level.
564 * TimerSlackNSec= can now be specified in system.conf for
565 PID1. This allows system-wide power savings.
567 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
568 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
573 * logind is now capable of (optionally) handling power and
574 sleep keys as well as the lid switch.
576 * journalctl now understands the syntax "journalctl
577 /usr/bin/avahi-daemon" to get all log output of a specific
580 * CapabilityBoundingSet= in system.conf now also influences
581 the capability bound set of usermode helpers of the kernel.
583 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
584 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
585 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
586 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
590 * Note that we skipped 139 releases here in order to set the
591 new version to something that is greater than both udev's
592 and systemd's most recent version number.
594 * udev: all udev sources are merged into the systemd source tree now.
595 All future udev development will happen in the systemd tree. It
596 is still fully supported to use the udev daemon and tools without
597 systemd running, like in initramfs or other init systems. Building
598 udev though, will require the *build* of the systemd tree, but
599 udev can be properly *run* without systemd.
601 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
602 should be used to create dead device nodes as workarounds for broken
605 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
606 no longer supported. udev_monitor_new_from_netlink() needs to be
607 used to subscribe to events.
609 * udev: when udevd is started by systemd, processes which are left
610 behind by forking them off of udev rules, are unconditionally cleaned
611 up and killed now after the event handling has finished. Services or
612 daemons must be started as systemd services. Services can be
613 pulled-in by udev to get started, but they can no longer be directly
614 forked by udev rules.
616 * udev: the daemon binary is called systemd-udevd now and installed
617 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
618 to adapt to that, create symlink, or rename the binary after building
621 * libudev no longer provides these symbols:
622 udev_monitor_from_socket()
623 udev_queue_get_failed_list_entry()
624 udev_get_{dev,sys,run}_path()
625 The versions number was bumped and symbol versioning introduced.
627 * systemd-loginctl and systemd-journalctl have been renamed
628 to loginctl and journalctl to match systemctl.
630 * The config files: /etc/systemd/systemd-logind.conf and
631 /etc/systemd/systemd-journald.conf have been renamed to
632 logind.conf and journald.conf. Package updates should rename
633 the files to the new names on upgrade.
635 * For almost all files the license is now LGPL2.1+, changed
636 from the previous GPL2.0+. Exceptions are some minor stuff
637 of udev (which will be changed to LGPL2.1 eventually, too),
638 and the MIT licensed sd-daemon.[ch] library that is suitable
639 to be used as drop-in files.
641 * systemd and logind now handle system sleep states, in
642 particular suspending and hibernating.
644 * logind now implements a sleep/shutdown/idle inhibiting logic
645 suitable for a variety of uses. Soonishly Lennart will blog
646 about this in more detail.
648 * var-run.mount and var-lock.mount are no longer provided
649 (which prevously bind mounted these directories to their new
650 places). Distributions which have not converted these
651 directories to symlinks should consider stealing these files
652 from git history and add them downstream.
654 * We introduced the Documentation= field for units and added
655 this to all our shipped units. This is useful to make it
656 easier to explore the boot and the purpose of the various
659 * All smaller setup units (such as
660 systemd-vconsole-setup.service) now detect properly if they
661 are run in a container and are skipped when
662 appropriate. This guarantees an entirely noise-free boot in
663 Linux container environments such as systemd-nspawn.
665 * A framework for implementing offline system updates is now
666 integrated, for details see:
667 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
669 * A new service type Type=idle is available now which helps us
670 avoiding ugly interleaving of getty output and boot status
673 * There's now a system-wide CapabilityBoundingSet= option to
674 globally reduce the set of capabilities for the
675 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
676 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
677 even CAP_NET_ADMIN system-wide for secure systems.
679 * There are now system-wide DefaultLimitXXX= options to
680 globally change the defaults of the various resource limits
681 for all units started by PID 1.
683 * Harald Hoyer's systemd test suite has been integrated into
684 systemd which allows easy testing of systemd builds in qemu
685 and nspawn. (This is really awesome! Ask us for details!)
687 * The fstab parser is now implemented as generator, not inside
690 * systemctl will now warn you if .mount units generated from
691 /etc/fstab are out of date due to changes in fstab that
692 haven't been read by systemd yet.
694 * systemd is now suitable for usage in initrds. Dracut has
695 already been updated to make use of this. With this in place
696 initrds get a slight bit faster but primarily are much
697 easier to introspect and debug since "systemctl status" in
698 the host system can be used to introspect initrd services,
699 and the journal from the initrd is kept around too.
701 * systemd-delta has been added, a tool to explore differences
702 between user/admin configuration and vendor defaults.
704 * PrivateTmp= now affects both /tmp and /var/tmp.
706 * Boot time status messages are now much prettier and feature
707 proper english language. Booting up systemd has never been
710 * Read-ahead pack files now include the inode number of all
711 files to pre-cache. When the inode changes the pre-caching
712 is not attempted. This should be nicer to deal with updated
713 packages which might result in changes of read-ahead
716 * We now temporaritly lower the kernel's read_ahead_kb variable
717 when collecting read-ahead data to ensure the kernel's
718 built-in read-ahead does not add noise to our measurements
719 of necessary blocks to pre-cache.
721 * There's now RequiresMountsFor= to add automatic dependencies
722 for all mounts necessary for a specific file system path.
724 * MountAuto= and SwapAuto= have been removed from
725 system.conf. Mounting file systems at boot has to take place
728 * nspawn now learned a new switch --uuid= to set the machine
729 ID on the command line.
731 * nspawn now learned the -b switch to automatically search
734 * vt102 is now the default TERM for serial TTYs, upgraded from
737 * systemd-logind now works on VT-less systems.
739 * The build tree has been reorganized. The individual
740 components now have directories of their own.
742 * A new condition type ConditionPathIsReadWrite= is now available.
744 * nspawn learned the new -C switch to create cgroups for the
745 container in other hierarchies.
747 * We now have support for hardware watchdogs, configurable in
750 * The scheduled shutdown logic now has a public API.
752 * We now mount /tmp as tmpfs by default, but this can be
753 masked and /etc/fstab can override it.
755 * Since udisks doesn't make use of /media anymore we are not
756 mounting a tmpfs on it anymore.
758 * journalctl gained a new --local switch to only interleave
759 locally generated journal files.
761 * We can now load the IMA policy at boot automatically.
763 * The GTK tools have been split off into a systemd-ui.
765 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
766 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
767 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
768 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
769 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
770 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
771 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
772 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
773 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
778 * This is mostly a bugfix release
780 * Support optional initialization of the machine ID from the
781 KVM or container configured UUID.
783 * Support immediate reboots with "systemctl reboot -ff"
785 * Show /etc/os-release data in systemd-analyze output
787 * Many bugfixes for the journal, including endianess fixes and
788 ensuring that disk space enforcement works
790 * sd-login.h is C++ comptaible again
792 * Extend the /etc/os-release format on request of the Debian
795 * We now refuse non-UTF8 strings used in various configuration
796 and unit files. This is done to ensure we don't pass invalid
797 data over D-Bus or expose it elsewhere.
799 * Register Mimo USB Screens as suitable for automatic seat
802 * Read SELinux client context from journal clients in a race
805 * Reorder configuration file lookup order. /etc now always
806 overrides /run in order to allow the administrator to always
807 and unconditionally override vendor supplied or
808 automatically generated data.
810 * The various user visible bits of the journal now have man
811 pages. We still lack man pages for the journal API calls
814 * We now ship all man pages in HTML format again in the
817 Contributions from: Dave Reisner, Dirk Eibach, Frederic
818 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
819 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
824 * This is mostly a bugfix release
826 * systems lacking /etc/os-release are no longer supported.
828 * Various functionality updates to libsystemd-login.so
830 * Track class of PAM logins to distuingish greeters from
833 Contributions from: Kay Sievers, Lennart Poettering, Michael
838 * This is an important bugfix release for v41.
840 * Building man pages is now optional which should be useful
841 for those building systemd from git but unwilling to install
844 * Watchdog support for supervising services is now usable. In
845 a future release support for hardware watchdogs
846 (i.e. /dev/watchdog) will be added building on this.
848 * Service start rate limiting is now configurable and can be
849 turned off per service. When a start rate limit is hit a
850 reboot can automatically be triggered.
852 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
854 Contributions from: Benjamin Franzke, Bill Nottingham,
855 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
856 Schmidt, Michał Górny, Piotr Drąg
860 * The systemd binary is installed /usr/lib/systemd/systemd now;
861 An existing /sbin/init symlink needs to be adapted with the
864 * The code that loads kernel modules has been ported to invoke
865 libkmod directly, instead of modprobe. This means we do not
866 support systems with module-init-tools anymore.
868 * Watchdog support is now already useful, but still not
871 * A new kernel command line option systemd.setenv= is
872 understood to set system wide environment variables
875 * We now limit the set of capabilities of systemd-journald.
877 * We now set SIGPIPE to ignore by default, since it only is
878 useful in shell pipelines, and has little use in general
879 code. This can be disabled with IgnoreSIPIPE=no in unit
882 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
883 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
888 * This is mostly a bugfix release
890 * We now expose the reason why a service failed in the
891 "Result" D-Bus property.
893 * Rudimentary service watchdog support (will be completed over
894 the next few releases.)
896 * When systemd forks off in order execute some service we will
897 now immediately changes its argv[0] to reflect which process
898 it will execute. This is useful to minimize the time window
899 with a generic argv[0], which makes bootcharts more useful
901 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
902 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
903 Mike Kazantsev, Ray Strode
907 * This is mostly a test release, but incorporates many
910 * New systemd-cgtop tool to show control groups by their
913 * Linking against libacl for ACLs is optional again. If
914 disabled, support tracking device access for active logins
915 goes becomes unavailable, and so does access to the user
916 journals by the respective users.
918 * If a group "adm" exists, journal files are automatically
919 owned by them, thus allow members of this group full access
920 to the system journal as well as all user journals.
922 * The journal now stores the SELinux context of the logging
923 client for all entries.
925 * Add C++ inclusion guards to all public headers
927 * New output mode "cat" in the journal to print only text
928 messages, without any meta data like date or time.
930 * Include tiny X server wrapper as a temporary stop-gap to
931 teach XOrg udev display enumeration. This is used by display
932 managers such as gdm, and will go away as soon as XOrg
933 learned native udev hotplugging for display devices.
935 * Add new systemd-cat tool for executing arbitrary programs
936 with STDERR/STDOUT connected to the journal. Can also act as
937 BSD logger replacement, and does so by default.
939 * Optionally store all locally generated coredumps in the
940 journal along with meta data.
942 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
943 writing short strings to files (for usage for /sys), and for
944 creating symlinks, character and block device nodes.
946 * New unit file option ControlGroupPersistent= to make cgroups
947 persistent, following the mechanisms outlined in
948 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
950 * Support multiple local RTCs in a sane way
952 * No longer monopolize IO when replaying readahead data on
953 rotating disks, since we might starve non-file-system IO to
954 death, since fanotify() will not see accesses done by blkid,
957 * Don't show kernel threads in systemd-cgls anymore, unless
958 requested with new -k switch.
960 Contributions from: Dan Horák, Kay Sievers, Lennart
961 Poettering, Michal Schmidt
965 * This is mostly a test release, but incorporates many
968 * The git repository moved to:
969 git://anongit.freedesktop.org/systemd/systemd
970 ssh://git.freedesktop.org/git/systemd/systemd
972 * First release with the journal
973 http://0pointer.de/blog/projects/the-journal.html
975 * The journal replaces both systemd-kmsg-syslogd and
976 systemd-stdout-bridge.
978 * New sd_pid_get_unit() API call in libsystemd-logind
980 * Many systemadm clean-ups
982 * Introduce remote-fs-pre.target which is ordered before all
983 remote mounts and may be used to start services before all
986 * Added Mageia support
988 * Add bash completion for systemd-loginctl
990 * Actively monitor PID file creation for daemons which exit in
991 the parent process before having finished writing the PID
992 file in the daemon process. Daemons which do this need to be
993 fixed (i.e. PID file creation must have finished before the
994 parent exits), but we now react a bit more gracefully to them.
996 * Add colourful boot output, mimicking the well-known output
997 of existing distributions.
999 * New option PassCredentials= for socket units, for
1000 compatibility with a recent kernel ABI breakage.
1002 * /etc/rc.local is now hooked in via a generator binary, and
1003 thus will no longer act as synchronization point during
1006 * systemctl list-unit-files now supports --root=.
1008 * systemd-tmpfiles now understands two new commands: z, Z for
1009 relabelling files according to the SELinux database. This is
1010 useful to apply SELinux labels to specific files in /sys,
1013 * Output of SysV services is now forwarded to both the console
1014 and the journal by default, not only just the console.
1016 * New man pages for all APIs from libsystemd-login.
1018 * The build tree got reorganized and a the build system is a
1019 lot more modular allowing embedded setups to specifically
1020 select the components of systemd they are interested in.
1022 * Support for Linux systems lacking the kernel VT subsystem is
1025 * configure's --with-rootdir= got renamed to
1026 --with-rootprefix= to follow the naming used by udev and
1029 * Unless specified otherwise we'll now install to /usr instead
1030 of /usr/local by default.
1032 * Processes with '@' in argv[0][0] are now excluded from the
1033 final shut-down killing spree, following the logic explained
1035 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1037 * All processes remaining in a service cgroup when we enter
1038 the START or START_PRE states are now killed with
1039 SIGKILL. That means it is no longer possible to spawn
1040 background processes from ExecStart= lines (which was never
1041 supported anyway, and bad style).
1043 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1044 reloading of units together.
1046 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1047 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1048 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1049 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1050 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek