1 systemd System and Service Manager
5 * journalctl gained the new --since= and --until= switches to
6 filter by time. It also now supports nice filtering for
9 * Type=oneshot services not may use ExecReload= and do the
12 * The journal daemon now supports time-based rotation and
13 vacuuming, in addition to the usual disk-space based
16 * The journal will now index the available field values for
17 each field name. This enables clients to show pretty drop
18 downs of available match values when filtering. The bash
19 completion of journalctl has been updated
20 accordingly. journalctl gained a new switch -F to list all
21 values a certain field takes in the journal database.
23 * More service events are now written as structured messages
24 to the journal, and made recognizable via message IDs.
26 * The timedated, localed and hostnamed mini-services which
27 previously only provided support for changing time, locale
28 and hostname settings from graphical DEs such as GNOME now
29 also have a minimal (but very useful) text-based client
30 utility each. This is probably the nicest way to changing
31 these settings from the command line now, especially since
32 it lists available options and is fully integrated with bash
35 * There's now a new tool "systemd-coredumpctl" to list and
36 extract coredumps from the journal.
38 * We now install a README each in /var/log/ and
39 /etc/rc.d/init.d explaining where the system logs and init
40 scripts went. This hopefully should help folks who go to
41 that dirs and look into the otherwise now empty void and
44 * When user-services are invoked (by systemd --user) the
45 $MANAGERPID env var is set to the PID of systemd.
47 * SIGRTMIN+24 when sent to a --user instance will now result
48 in immediate termination of systemd.
50 * gatewayd received numerous feature additions such as a
51 "follow" mode, for live syncing and filtering.
53 * browse.html now allows filtering and showing detailed
54 information on specific entries. Keyboard navigation and
55 mouse screen support has been added.
57 * gatewayd/journalctl now supports HTML5/JSON
58 Server-Sent-Events as output.
60 * The SysV init script compatibility logic will now
61 heuristically determine whether a script supports the
62 "reload" verb, and only then make this available as
65 * "systemctl status --follow" has been removed, use "journal
68 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
69 have been removed since they are hardly useful to be
72 * And I'd like to take the opportunity to specifically mention
73 Zbigniew for his great contributions. Zbigniew, you rock!
75 Contributions from: Andrew Eikum, Christian Hesse, Colin
76 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
77 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
78 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
79 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
80 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
81 Jędrzejewski-Szmek, Сковорода Никита Андреевич
85 * If /etc/vconsole.conf is non-existent or empty we will no
86 longer load any console font or key map at boot by
87 default. Instead the kernel defaults will be left
88 intact. This is definitely the right thing to do, as no
89 configuration should mean no configuration, and hard-coding
90 font names that are different on all archs is probably a bad
91 idea. Also, the kernel default key map and font should be
92 good enough for most cases anyway, and mostly identical to
93 the userspace fonts/key maps we previously overloaded them
94 with. If distributions want to continue to default to a
95 non-kernel font or key map they should ship a default
96 /etc/vconsole.conf with the appropriate contents.
98 Contributions from: Colin Walters, Daniel J Walsh, Dave
99 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
100 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
104 * journalctl gained a new --cursor= switch to show entries
105 starting from the specified location in the journal.
107 * We now enforce a size limit on journal entry fields exported
108 with "-o json" in journalctl. Fields larger than 4K will be
109 assigned null. This can be turned off with --all.
111 * An (optional) journal gateway daemon is now available as
112 "systemd-journal-gatewayd.service". This service provides
113 access to the journal via HTTP and JSON. This functionality
114 will be used to implement live log synchronization in both
115 pull and push modes, but has various other users too, such
116 as easy log access for debugging of embedded devices. Right
117 now it is already useful to retrieve the journal via HTTP:
119 # systemctl start systemd-journal-gatewayd.service
120 # wget http://localhost:19531/entries
122 This will download the journal contents in a
123 /var/log/messages compatible format. The same as JSON:
125 # curl -H"Accept: application/json" http://localhost:19531/entries
127 This service is also accessible via a web browser where a
128 single static HTML5 app is served that uses the JSON logic
129 to enable the user to do some basic browsing of the
130 journal. This will be extended later on. Here's an example
131 screenshot of this app in its current state:
133 http://0pointer.de/public/journal-gatewayd
135 Contributions from: Kay Sievers, Lennart Poettering, Robert
136 Milasan, Tom Gundersen
140 * The bash completion logic is now available for journalctl
143 * We don't mount the "cpuset" controller anymore together with
144 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
145 started if no parameters are assigned to it. "cpuset" hence
146 broke code that assumed it it could create "cpu" groups and
149 * journalctl -f will now subscribe to terminal size changes,
150 and line break accordingly.
152 Contributions from: Dave Reisner, Kay Sievers, Lennart
153 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
157 * nspawn will now create a symlink /etc/localtime in the
158 container environment, copying the host's timezone
159 setting. Previously this has been done via a bind mount, but
160 since symlinks cannot be bind mounted this has now been
161 changed to create/update the appropriate symlink.
163 * journalctl -n's line number argument is now optional, and
164 will default to 10 if omitted.
166 * journald will now log the maximum size the journal files may
167 take up on disk. This is particularly useful if the default
168 built-in logic of determining this parameter from the file
169 system size is used. Use "systemctl status
170 systemd-journald.service" to see this information.
172 * The multi-seat X wrapper tool has been stripped down. As X
173 is now capable of enumerating graphics devices via udev in a
174 seat-aware way the wrapper is not strictly necessary
175 anymore. A stripped down temporary stop-gap is still shipped
176 until the upstream display managers have been updated to
177 fully support the new X logic. Expect this wrapper to be
178 removed entirely in one of the next releases.
180 * HandleSleepKey= in logind.conf has been split up into
181 HandleSuspendKey= and HandleHibernateKey=. The old setting
182 is not available anymore. X11 and the kernel are
183 distuingishing between these keys and we should too. This
184 also means the inhibition lock for these keys has been split
187 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
188 Poettering, Lukas Nykryn, Václav Pavlín
192 * Whenever a unit changes state we'll now log this to the
193 journal and show along the unit's own log output in
196 * ConditionPathIsMountPoint= can now properly detect bind
197 mount points too. (Previously, a bind mount of one file
198 system to another place in the same file system could not be
199 detected as mount, since they shared struct stat's st_dev
202 * We will now mount the cgroup controllers cpu, cpuacct,
203 cpuset and the controllers net_cls, net_prio together by
206 * nspawn containers will now have a virtualized boot
207 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
208 over with a randomized ID at container initialization). This
209 has the effect of making "journalctl -b" do the right thing
212 * The JSON output journal serialization has been updated not
213 to generate "endless" list objects anymore, but rather one
214 JSON object per line. This is more in line how most JSON
215 parsers expect JSON objects. The new output mode
216 "json-pretty" has been added to provide similar output, but
217 neatly aligned for readability by humans.
219 * We dropped all explicit sync() invocations in the shutdown
220 code. The kernel does this implicitly anyway in the kernel
221 reboot() syscall. halt(8)'s -n option is now a compatibility
224 * We now support virtualized reboot() in containers, as
225 supported by newer kernels. We will fall back to exit() if
226 CAP_SYS_REBOOT is not available to the container. Also,
227 nspawn makes use of this now and will actually reboot the
228 container if the containerized OS asks for that.
230 * journalctl will only show local log output by default
231 now. Use --merge (-m) to show remote log output, too.
233 * libsystemd-journal gained the new sd_journal_get_usage()
234 call to determine the current disk usage of all journal
235 files. This is exposed in the new "journalctl --disk-usage"
238 * journald gained a new configuration setting SplitMode= in
239 journald.conf which may be used to control how user journals
240 are split off. See journald.conf(5) for details.
242 * A new condition type ConditionFileNotEmpty= has been added.
244 * tmpfiles' "w" lines now support file globbing, to write
245 multiple files at once.
247 * We added Python bindings for the journal submission
248 APIs. More Python APIs for a number of selected APIs will
249 likely follow. Note that we intend to add native bindings
250 only for the Python language, as we consider it common
251 enough to deserve bindings shipped within systemd. There are
252 various projects outside of systemd that provide bindings
253 for languages such as PHP or Lua.
255 * Many conditions will now resolve specifiers such as %i. In
256 addition, PathChanged= and related directives of .path units
257 now support specifiers as well.
259 * There's now a new RPM macro definition for the system preset
262 * journald will now warn if it can't foward a message to the
263 syslog daemon because it's socket is full.
265 * timedated will no longer write or process /etc/timezone,
266 except on Debian. As we do not support late mounted /usr
267 anymore /etc/localtime always being a symlink is now safe,
268 and hence the information in /etc/timezone is not necessary
271 * logind will now always reserve one VT for a text getty (VT6
272 by default). Previously if more than 6 X sessions where
273 started they took up all the VTs with auto-spawned gettys,
274 so that no text gettys were available anymore.
276 * udev will now automatically inform the btrfs kernel logic
277 about btrfs RAID components showing up. This should make
278 simple hotplug based btrfs RAID assembly work.
280 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
281 (but not for its children which will stay at the kernel
282 default). This should allow setups with a lot more listening
285 * systemd will now always pass the configured timezone to the
286 kernel at boot. timedated will do the same when the timezone
289 * logind's inhibition logic has been updated. By default,
290 logind will now handle the lid switch, the power and sleep
291 keys all the time, even in graphical sessions. If DEs want
292 to handle these events on their own they should take the new
293 handle-power-key, handle-sleep-key and handle-lid-switch
294 inhibitors during their runtime. A simple way to achiveve
295 that is to invoke the DE wrapped in an invocation of:
297 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
299 * Access to unit operations is now checked via SELinux taking
300 the unit file label and client process label into account.
302 * systemd will now notify the administrator in the journal
303 when he over-mounts a non-empty directory.
305 * There are new specifiers that are resolved in unit files,
306 for the host name (%H), the machine ID (%m) and the boot ID
309 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
310 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
311 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
312 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
313 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
314 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
315 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
319 * Support for reading structured kernel messages from
320 /dev/kmsg has now been added and is enabled by default.
322 * Support for reading kernel messages from /proc/kmsg has now
323 been removed. If you want kernel messages in the journal
324 make sure to run a recent kernel (>= 3.5) that supports
325 reading structured messages from /dev/kmsg (see
326 above). /proc/kmsg is now exclusive property of classic
327 syslog daemons again.
329 * The libudev API gained the new
330 udev_device_new_from_device_id() call.
332 * The logic for file system namespace (ReadOnlyDirectory=,
333 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
334 require pivot_root() anymore. This means fewer temporary
335 directories are created below /tmp for this feature.
337 * nspawn containers will now see and receive all submounts
338 made on the host OS below the root file system of the
341 * Forward Secure Sealing is now supported for Journal files,
342 which provide cryptographical sealing of journal files so
343 that attackers cannot alter log history anymore without this
344 being detectable. Lennart will soon post a blog story about
345 this explaining it in more detail.
347 * There are two new service settings RestartPreventExitStatus=
348 and SuccessExitStatus= which allow configuration of exit
349 status (exit code or signal) which will be excepted from the
350 restart logic, resp. consider successful.
352 * journalctl gained the new --verify switch that can be used
353 to check the integrity of the structure of journal files and
354 (if Forward Secure Sealing is enabled) the contents of
357 * nspawn containers will now be run with /dev/stdin, /dev/fd/
358 and similar symlinks pre-created. This makes running shells
359 as container init process a lot more fun.
361 * The fstab support can now handle PARTUUID= and PARTLABEL=
364 * A new ConditionHost= condition has been added to match
365 against the hostname (with globs) and machine ID. This is
366 useful for clusters where a single OS image is used to
367 provision a large number of hosts which shall run slightly
368 different sets of services.
370 * Services which hit the restart limit will now be placed in a
373 Contributions from: Bertram Poettering, Dave Reisner, Huang
374 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
375 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
379 * When running in --user mode systemd will now become a
380 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
381 tree a lot more organized.
383 * A new PartOf= unit dependency type has been introduced that
384 may be used to group services in a natural way.
386 * "systemctl enable" may now be used to enable instances of
389 * journalctl now prints error log levels in red, and
390 warning/notice log levels in bright white. It also supports
391 filtering by log level now.
393 * cgtop gained a new -n switch (similar to top), to configure
394 the maximum number of iterations to run for. It also gained
395 -b, to run in batch mode (accepting no input).
397 * The suffix ".service" may now be ommited on most systemctl
398 command lines involving service unit names.
400 * There's a new bus call in logind to lock all sessions, as
401 well as a loginctl verb for it "lock-sessions".
403 * libsystemd-logind.so gained a new call sd_journal_perror()
404 that works similar to libc perror() but logs to the journal
405 and encodes structured information about the error number.
407 * /etc/crypttab entries now understand the new keyfile-size=
410 * shutdown(8) now can send a (configurable) wall message when
411 a shutdown is cancelled.
413 * The mount propagation mode for the root file system will now
414 default to "shared", which is useful to make containers work
415 nicely out-of-the-box so that they receive new mounts from
416 the host. This can be undone locally by running "mount
417 --make-rprivate /" if needed.
419 * The prefdm.service file has been removed. Distributions
420 should maintain this unit downstream if they intend to keep
421 it around. However, we recommend writing normal unit files
422 for display managers instead.
424 * Since systemd is a crucial part of the OS we will now
425 default to a number of compiler switches that improve
426 security (hardening) such as read-only relocations, stack
427 protection, and suchlike.
429 * The TimeoutSec= setting for services is now split into
430 TimeoutStartSec= and TimeoutStopSec= to allow configuration
431 of individual time outs for the start and the stop phase of
434 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
435 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
436 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
437 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
438 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
439 Gundersen, Zbigniew Jędrzejewski-Szmek
443 * The journal and id128 C APIs are now fully documented as man
446 * Extra safety checks have been added when transitioning from
447 the initial RAM disk to the main system to avoid accidental
450 * /etc/crypttab entries now understand the new keyfile-offset=
453 * systemctl -t can now be used to filter by unit load state.
455 * The journal C API gained the new sd_journal_wait() call to
456 make writing synchronous journal clients easier.
458 * journalctl gained the new -D switch to show journals from a
461 * journalctl now displays a special marker between log
462 messages of two different boots.
464 * The journal is now explicitly flushed to /var via a service
465 systemd-journal-flush.service, rather than implicitly simply
466 by seeing /var/log/journal to be writable.
468 * journalctl (and the journal C APIs) can now match for much
469 more complex expressions, with alternatives and
472 * When transitioning from the initial RAM disk to the main
473 system we will now kill all processes in a killing spree to
474 ensure no processes stay around by accident.
476 * Three new specifiers may be used in unit files: %u, %h, %s
477 resolve to the user name, user home directory resp. user
478 shell. This is useful for running systemd user instances.
480 * We now automatically rotate journal files if their data
481 object hash table gets a fill level > 75%. We also size the
482 hash table based on the configured maximum file size. This
483 together should lower hash collisions drastically and thus
484 speed things up a bit.
486 * journalctl gained the new "--header" switch to introspect
487 header data of journal files.
489 * A new setting SystemCallFilters= has been added to services
490 which may be used to apply blacklists or whitelists to
491 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
493 * nspawn gained a new --link-journal= switch (and quicker: -j)
494 to link the container journal with the host. This makes it
495 very easy to centralize log viewing on the host for all
496 guests while still keeping the journal files separated.
498 * Many bugfixes and optimizations
500 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
501 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
502 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
507 * Several tools now understand kernel command line arguments,
508 which are only read when run in an initial RAM disk. They
509 usually follow closely their normal counterparts, but are
512 * There's a new tool to analyze the readahead files that are
513 automatically generated at boot. Use:
515 /usr/lib/systemd/systemd-readahead analyze /.readahead
517 * We now provide an early debug shell on tty9 if this enabled. Use:
519 systemctl enable debug-shell.service
521 * All plymouth related units have been moved into the Plymouth
522 package. Please make sure to upgrade your Plymouth version
525 * systemd-tmpfiles now supports getting passed the basename of
526 a configuration file only, in which case it will look for it
527 in all appropriate directories automatically.
529 * udevadm info now takes a /dev or /sys path as argument, and
530 does the right thing. Example:
532 udevadm info /dev/sda
533 udevadm info /sys/class/block/sda
535 * systemctl now prints a warning if a unit is stopped but a
536 unit that might trigger it continues to run. Example: a
537 service is stopped but the socket that activates it is left
540 * "systemctl status" will now mention if the log output was
541 shortened due to rotation since a service has been started.
543 * The journal API now exposes functions to determine the
544 "cutoff" times due to rotation.
546 * journald now understands SIGUSR1 and SIGUSR2 for triggering
547 immediately flushing of runtime logs to /var if possible,
548 resp. for triggering immediate rotation of the journal
551 * It is now considered an error if a service is attempted to
552 be stopped that is not loaded.
554 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
556 * systemd-analyze now supports Python 3
558 * tmpfiles now supports cleaning up directories via aging
559 where the first level dirs are always kept around but
560 directories beneath it automatically aged. This is enabled
561 by prefixing the age field with '~'.
563 * Seat objects now expose CanGraphical, CanTTY properties
564 which is required to deal with very fast bootups where the
565 display manager might be running before the graphics drivers
566 completed initialization.
568 * Seat objects now expose a State property.
570 * We now include RPM macros for service enabling/disabling
571 based on the preset logic. We recommend RPM based
572 distributions to make use of these macros if possible. This
573 makes it simpler to reuse RPM spec files across
576 * We now make sure that the collected systemd unit name is
577 always valid when services log to the journal via
580 * There's a new man page kernel-command-line(7) detailing all
581 command line options we understand.
583 * The fstab generator may now be disabled at boot by passing
584 fstab=0 on the kernel command line.
586 * A new kernel command line option modules-load= is now understood
587 to load a specific kernel module statically, early at boot.
589 * Unit names specified on the systemctl command line are now
590 automatically escaped as needed. Also, if file system or
591 device paths are specified they are automatically turned
592 into the appropriate mount or device unit names. Example:
594 systemctl status /home
595 systemctl status /dev/sda
597 * The SysVConsole= configuration option has been removed from
600 * The SysV search path is no longer exported on the D-Bus
603 * The Names= option is been removed from unit file parsing.
605 * There's a new man page bootup(7) detailing the boot process.
607 * Every unit and every generator we ship with systemd now
608 comes with full documentation. The self-explanatory boot is
611 * A couple of services gained "systemd-" prefixes in their
612 name if they wrap systemd code, rather than only external
613 code. Among them fsck@.service which is now
614 systemd-fsck@.service.
616 * The HaveWatchdog property has been removed from the D-Bus
619 * systemd.confirm_spawn= on the kernel command line should now
622 * There's a new man page crypttab(5) which details all options
623 we actually understand.
625 * systemd-nspawn gained a new --capability= switch to pass
626 additional capabilities to the container.
628 * timedated will now read known NTP implementation unit names
629 from /usr/lib/systemd/ntp-units.d/*.list,
630 systemd-timedated-ntp.target has been removed.
632 * journalctl gained a new switch "-b" that lists log data of
633 the current boot only.
635 * The notify socket is in the abstract namespace again, in
636 order to support daemons which chroot() at start-up.
638 * There is a new Storage= configuration option for journald
639 which allows configuration of where log data should go. This
640 also provides a way to disable journal logging entirely, so
641 that data collected is only forwarded to the console, the
642 kernel log buffer or another syslog implementation.
644 * Many bugfixes and optimizations
646 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
647 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
648 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
649 Shawn Landden, Tom Gundersen
653 * "systemctl help <unit>" now shows the man page if one is
656 * Several new man pages have been added.
658 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
659 MaxLevelConsole= can now be specified in
660 journald.conf. These options allow reducing the amount of
661 data stored on disk or forwarded by the log level.
663 * TimerSlackNSec= can now be specified in system.conf for
664 PID1. This allows system-wide power savings.
666 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
667 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
672 * logind is now capable of (optionally) handling power and
673 sleep keys as well as the lid switch.
675 * journalctl now understands the syntax "journalctl
676 /usr/bin/avahi-daemon" to get all log output of a specific
679 * CapabilityBoundingSet= in system.conf now also influences
680 the capability bound set of usermode helpers of the kernel.
682 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
683 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
684 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
685 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
689 * Note that we skipped 139 releases here in order to set the
690 new version to something that is greater than both udev's
691 and systemd's most recent version number.
693 * udev: all udev sources are merged into the systemd source tree now.
694 All future udev development will happen in the systemd tree. It
695 is still fully supported to use the udev daemon and tools without
696 systemd running, like in initramfs or other init systems. Building
697 udev though, will require the *build* of the systemd tree, but
698 udev can be properly *run* without systemd.
700 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
701 should be used to create dead device nodes as workarounds for broken
704 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
705 no longer supported. udev_monitor_new_from_netlink() needs to be
706 used to subscribe to events.
708 * udev: when udevd is started by systemd, processes which are left
709 behind by forking them off of udev rules, are unconditionally cleaned
710 up and killed now after the event handling has finished. Services or
711 daemons must be started as systemd services. Services can be
712 pulled-in by udev to get started, but they can no longer be directly
713 forked by udev rules.
715 * udev: the daemon binary is called systemd-udevd now and installed
716 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
717 to adapt to that, create symlink, or rename the binary after building
720 * libudev no longer provides these symbols:
721 udev_monitor_from_socket()
722 udev_queue_get_failed_list_entry()
723 udev_get_{dev,sys,run}_path()
724 The versions number was bumped and symbol versioning introduced.
726 * systemd-loginctl and systemd-journalctl have been renamed
727 to loginctl and journalctl to match systemctl.
729 * The config files: /etc/systemd/systemd-logind.conf and
730 /etc/systemd/systemd-journald.conf have been renamed to
731 logind.conf and journald.conf. Package updates should rename
732 the files to the new names on upgrade.
734 * For almost all files the license is now LGPL2.1+, changed
735 from the previous GPL2.0+. Exceptions are some minor stuff
736 of udev (which will be changed to LGPL2.1 eventually, too),
737 and the MIT licensed sd-daemon.[ch] library that is suitable
738 to be used as drop-in files.
740 * systemd and logind now handle system sleep states, in
741 particular suspending and hibernating.
743 * logind now implements a sleep/shutdown/idle inhibiting logic
744 suitable for a variety of uses. Soonishly Lennart will blog
745 about this in more detail.
747 * var-run.mount and var-lock.mount are no longer provided
748 (which prevously bind mounted these directories to their new
749 places). Distributions which have not converted these
750 directories to symlinks should consider stealing these files
751 from git history and add them downstream.
753 * We introduced the Documentation= field for units and added
754 this to all our shipped units. This is useful to make it
755 easier to explore the boot and the purpose of the various
758 * All smaller setup units (such as
759 systemd-vconsole-setup.service) now detect properly if they
760 are run in a container and are skipped when
761 appropriate. This guarantees an entirely noise-free boot in
762 Linux container environments such as systemd-nspawn.
764 * A framework for implementing offline system updates is now
765 integrated, for details see:
766 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
768 * A new service type Type=idle is available now which helps us
769 avoiding ugly interleaving of getty output and boot status
772 * There's now a system-wide CapabilityBoundingSet= option to
773 globally reduce the set of capabilities for the
774 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
775 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
776 even CAP_NET_ADMIN system-wide for secure systems.
778 * There are now system-wide DefaultLimitXXX= options to
779 globally change the defaults of the various resource limits
780 for all units started by PID 1.
782 * Harald Hoyer's systemd test suite has been integrated into
783 systemd which allows easy testing of systemd builds in qemu
784 and nspawn. (This is really awesome! Ask us for details!)
786 * The fstab parser is now implemented as generator, not inside
789 * systemctl will now warn you if .mount units generated from
790 /etc/fstab are out of date due to changes in fstab that
791 haven't been read by systemd yet.
793 * systemd is now suitable for usage in initrds. Dracut has
794 already been updated to make use of this. With this in place
795 initrds get a slight bit faster but primarily are much
796 easier to introspect and debug since "systemctl status" in
797 the host system can be used to introspect initrd services,
798 and the journal from the initrd is kept around too.
800 * systemd-delta has been added, a tool to explore differences
801 between user/admin configuration and vendor defaults.
803 * PrivateTmp= now affects both /tmp and /var/tmp.
805 * Boot time status messages are now much prettier and feature
806 proper english language. Booting up systemd has never been
809 * Read-ahead pack files now include the inode number of all
810 files to pre-cache. When the inode changes the pre-caching
811 is not attempted. This should be nicer to deal with updated
812 packages which might result in changes of read-ahead
815 * We now temporaritly lower the kernel's read_ahead_kb variable
816 when collecting read-ahead data to ensure the kernel's
817 built-in read-ahead does not add noise to our measurements
818 of necessary blocks to pre-cache.
820 * There's now RequiresMountsFor= to add automatic dependencies
821 for all mounts necessary for a specific file system path.
823 * MountAuto= and SwapAuto= have been removed from
824 system.conf. Mounting file systems at boot has to take place
827 * nspawn now learned a new switch --uuid= to set the machine
828 ID on the command line.
830 * nspawn now learned the -b switch to automatically search
833 * vt102 is now the default TERM for serial TTYs, upgraded from
836 * systemd-logind now works on VT-less systems.
838 * The build tree has been reorganized. The individual
839 components now have directories of their own.
841 * A new condition type ConditionPathIsReadWrite= is now available.
843 * nspawn learned the new -C switch to create cgroups for the
844 container in other hierarchies.
846 * We now have support for hardware watchdogs, configurable in
849 * The scheduled shutdown logic now has a public API.
851 * We now mount /tmp as tmpfs by default, but this can be
852 masked and /etc/fstab can override it.
854 * Since udisks doesn't make use of /media anymore we are not
855 mounting a tmpfs on it anymore.
857 * journalctl gained a new --local switch to only interleave
858 locally generated journal files.
860 * We can now load the IMA policy at boot automatically.
862 * The GTK tools have been split off into a systemd-ui.
864 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
865 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
866 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
867 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
868 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
869 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
870 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
871 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
872 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
877 * This is mostly a bugfix release
879 * Support optional initialization of the machine ID from the
880 KVM or container configured UUID.
882 * Support immediate reboots with "systemctl reboot -ff"
884 * Show /etc/os-release data in systemd-analyze output
886 * Many bugfixes for the journal, including endianess fixes and
887 ensuring that disk space enforcement works
889 * sd-login.h is C++ comptaible again
891 * Extend the /etc/os-release format on request of the Debian
894 * We now refuse non-UTF8 strings used in various configuration
895 and unit files. This is done to ensure we don't pass invalid
896 data over D-Bus or expose it elsewhere.
898 * Register Mimo USB Screens as suitable for automatic seat
901 * Read SELinux client context from journal clients in a race
904 * Reorder configuration file lookup order. /etc now always
905 overrides /run in order to allow the administrator to always
906 and unconditionally override vendor supplied or
907 automatically generated data.
909 * The various user visible bits of the journal now have man
910 pages. We still lack man pages for the journal API calls
913 * We now ship all man pages in HTML format again in the
916 Contributions from: Dave Reisner, Dirk Eibach, Frederic
917 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
918 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
923 * This is mostly a bugfix release
925 * systems lacking /etc/os-release are no longer supported.
927 * Various functionality updates to libsystemd-login.so
929 * Track class of PAM logins to distuingish greeters from
932 Contributions from: Kay Sievers, Lennart Poettering, Michael
937 * This is an important bugfix release for v41.
939 * Building man pages is now optional which should be useful
940 for those building systemd from git but unwilling to install
943 * Watchdog support for supervising services is now usable. In
944 a future release support for hardware watchdogs
945 (i.e. /dev/watchdog) will be added building on this.
947 * Service start rate limiting is now configurable and can be
948 turned off per service. When a start rate limit is hit a
949 reboot can automatically be triggered.
951 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
953 Contributions from: Benjamin Franzke, Bill Nottingham,
954 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
955 Schmidt, Michał Górny, Piotr Drąg
959 * The systemd binary is installed /usr/lib/systemd/systemd now;
960 An existing /sbin/init symlink needs to be adapted with the
963 * The code that loads kernel modules has been ported to invoke
964 libkmod directly, instead of modprobe. This means we do not
965 support systems with module-init-tools anymore.
967 * Watchdog support is now already useful, but still not
970 * A new kernel command line option systemd.setenv= is
971 understood to set system wide environment variables
974 * We now limit the set of capabilities of systemd-journald.
976 * We now set SIGPIPE to ignore by default, since it only is
977 useful in shell pipelines, and has little use in general
978 code. This can be disabled with IgnoreSIPIPE=no in unit
981 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
982 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
987 * This is mostly a bugfix release
989 * We now expose the reason why a service failed in the
990 "Result" D-Bus property.
992 * Rudimentary service watchdog support (will be completed over
993 the next few releases.)
995 * When systemd forks off in order execute some service we will
996 now immediately changes its argv[0] to reflect which process
997 it will execute. This is useful to minimize the time window
998 with a generic argv[0], which makes bootcharts more useful
1000 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1001 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1002 Mike Kazantsev, Ray Strode
1006 * This is mostly a test release, but incorporates many
1009 * New systemd-cgtop tool to show control groups by their
1012 * Linking against libacl for ACLs is optional again. If
1013 disabled, support tracking device access for active logins
1014 goes becomes unavailable, and so does access to the user
1015 journals by the respective users.
1017 * If a group "adm" exists, journal files are automatically
1018 owned by them, thus allow members of this group full access
1019 to the system journal as well as all user journals.
1021 * The journal now stores the SELinux context of the logging
1022 client for all entries.
1024 * Add C++ inclusion guards to all public headers
1026 * New output mode "cat" in the journal to print only text
1027 messages, without any meta data like date or time.
1029 * Include tiny X server wrapper as a temporary stop-gap to
1030 teach XOrg udev display enumeration. This is used by display
1031 managers such as gdm, and will go away as soon as XOrg
1032 learned native udev hotplugging for display devices.
1034 * Add new systemd-cat tool for executing arbitrary programs
1035 with STDERR/STDOUT connected to the journal. Can also act as
1036 BSD logger replacement, and does so by default.
1038 * Optionally store all locally generated coredumps in the
1039 journal along with meta data.
1041 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
1042 writing short strings to files (for usage for /sys), and for
1043 creating symlinks, character and block device nodes.
1045 * New unit file option ControlGroupPersistent= to make cgroups
1046 persistent, following the mechanisms outlined in
1047 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
1049 * Support multiple local RTCs in a sane way
1051 * No longer monopolize IO when replaying readahead data on
1052 rotating disks, since we might starve non-file-system IO to
1053 death, since fanotify() will not see accesses done by blkid,
1056 * Don't show kernel threads in systemd-cgls anymore, unless
1057 requested with new -k switch.
1059 Contributions from: Dan Horák, Kay Sievers, Lennart
1060 Poettering, Michal Schmidt
1064 * This is mostly a test release, but incorporates many
1067 * The git repository moved to:
1068 git://anongit.freedesktop.org/systemd/systemd
1069 ssh://git.freedesktop.org/git/systemd/systemd
1071 * First release with the journal
1072 http://0pointer.de/blog/projects/the-journal.html
1074 * The journal replaces both systemd-kmsg-syslogd and
1075 systemd-stdout-bridge.
1077 * New sd_pid_get_unit() API call in libsystemd-logind
1079 * Many systemadm clean-ups
1081 * Introduce remote-fs-pre.target which is ordered before all
1082 remote mounts and may be used to start services before all
1085 * Added Mageia support
1087 * Add bash completion for systemd-loginctl
1089 * Actively monitor PID file creation for daemons which exit in
1090 the parent process before having finished writing the PID
1091 file in the daemon process. Daemons which do this need to be
1092 fixed (i.e. PID file creation must have finished before the
1093 parent exits), but we now react a bit more gracefully to them.
1095 * Add colourful boot output, mimicking the well-known output
1096 of existing distributions.
1098 * New option PassCredentials= for socket units, for
1099 compatibility with a recent kernel ABI breakage.
1101 * /etc/rc.local is now hooked in via a generator binary, and
1102 thus will no longer act as synchronization point during
1105 * systemctl list-unit-files now supports --root=.
1107 * systemd-tmpfiles now understands two new commands: z, Z for
1108 relabelling files according to the SELinux database. This is
1109 useful to apply SELinux labels to specific files in /sys,
1112 * Output of SysV services is now forwarded to both the console
1113 and the journal by default, not only just the console.
1115 * New man pages for all APIs from libsystemd-login.
1117 * The build tree got reorganized and a the build system is a
1118 lot more modular allowing embedded setups to specifically
1119 select the components of systemd they are interested in.
1121 * Support for Linux systems lacking the kernel VT subsystem is
1124 * configure's --with-rootdir= got renamed to
1125 --with-rootprefix= to follow the naming used by udev and
1128 * Unless specified otherwise we'll now install to /usr instead
1129 of /usr/local by default.
1131 * Processes with '@' in argv[0][0] are now excluded from the
1132 final shut-down killing spree, following the logic explained
1134 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1136 * All processes remaining in a service cgroup when we enter
1137 the START or START_PRE states are now killed with
1138 SIGKILL. That means it is no longer possible to spawn
1139 background processes from ExecStart= lines (which was never
1140 supported anyway, and bad style).
1142 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1143 reloading of units together.
1145 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1146 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1147 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1148 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1149 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek