1 systemd System and Service Manager
5 * A new component "systemd-networkd" has been added that can
6 be used to configure local network interfaces statically or
7 via DHCP. It is capable up bridges, VLANs and bonding. This
8 currently provides no hookups for interactive network
9 configuration. Use this for your initrd, container, embedded
10 or server setup, if you need a simple, yet powerful network
11 configuration solution. This configuration subsystem is
12 quite nifty as it allows wildcard hotplug matching in
13 interfaces. For example, with a single configuration snippet
14 you can configure that all ethernet interfaces showing up
15 are automatically added to a bridge, or similar. It
16 optionally supports link-sensing and more.
18 * A new tool "systemd-socket-proxyd" has been added which can
19 acts as a bidirectional proxy for TCP sockets. This is
20 useful for adding socket activation support to services that
21 do not actually support socket activation, including virtual
22 machines and suchlike.
24 * Add a new tool to save/restore rfkill state on
27 * Save/restore state of kbd backlights in addition to display
28 backlights on shutdown/boot.
30 * udev learned a new SECLABEL{} construct to label device
31 nodes with a specific security label when they appear. For
32 now only SECLABEL{selinux} is supported, but the syntax is
33 prepared for additional security frameworks.
35 * udev gained a new scheme to configure link-level attributes
36 from files in /etc/systemd/network/*.link. These files can
37 match against MAC address, device path, driver name and type
38 and will apply attributes like the naming policy, link speed
39 MTU, duplex settings, WakeOnLan settings, MAC address, MAC
40 address assignment policy (randomized, ...).
42 * When the User= switch is used in a unit file, also
43 initialize $SHELL= based on user database.
45 * systemd no longer depends on libdbus. All communication is
46 now done with sd-bus, systemd's low-level bus library
49 * kdbus support has been added to PID 1 itself. When kdbus is
50 enabled this causes PID 1 to set up the system bus, and
51 enable support for a new ".busname" unit type that
52 encapsulates bus name activation on kdbus. It works a little
53 bit like ".socket" units, except for bus names. A new
54 generator has been added that converts classic dbus1 service
55 activation files automatically into native systemd .busname
58 * sd-bus: add a light-weight vtable implementation that allows
59 defining objects on the bus with a simple static const
60 vtable array of its methods, signals and properties.
62 * systemd will not generate nor install static dbus
63 introspection data anymore to /usr/share/dbus-1/interfaces,
64 as the precise format of these files are unclear, and
65 nothing makes use of it.
67 * A proxy daemon is now provided to proxy clients connecting
68 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
69 compatibility with classic D-Bus.
71 * A bus driver implementation has been added that supports the
72 classic D-Bus bus driver calls on kdbus, also for
73 compatibility purposes.
75 * A new API "sd-event.h" has been added that implements a
76 minimal event loop API built around epoll. It provides a
77 couple of features that direct epoll usage is lacking:
78 priorization of events, scales to large numbers of timer
79 events, per-event timer slack (accuracy), system-wide
80 coalescing of timer events, exit handlers, watchdog
81 supervision support using systemd's sd_notify() API, child
84 * A new API "sd-rntl.h" has been added that provides an API
85 around the route netlink interface of the kernel, similar in
88 * A new API "sd-dhcp.h" has been added that provides a small
89 DHCPv4 client side implementation. This is used by
92 * There's a new kernel command line option
93 "systemd.restore_state". When set none of the systemd tools
94 will restore saved runtime state to hardware devices. More
95 specifically, the rfkill and backlight states are not
98 * The FsckPassNo= compatibility option in mount/service units
99 has been removed. The fstab generator will now add the
100 necessary dependencies automatically, and does not require
101 PID1's support for that anymore.
103 * journalctl gained a new switch --list-boots, that lists
104 recent boots with their times and boot IDs.
106 * The various tools like systemctl, loginctl, timedatectl,
107 busctl, systemd-run, ... have gained a new switch "-M" to
108 connect to a specific, local OS container (as direct
109 connection, without requiring SSH). This works on any
110 container that is registered with machined, such as those
111 created by libvirt-lxc or nspawn.
113 * systemd-run and systemd-analyze also gained support for "-H"
114 to connect to remote hosts via SSH. This is particular
115 useful for systemd-run since it enables queuing of jobs onto
118 * machinectl gained a new command "login" to open a getty
119 login in any local container. This works with any container
120 that is registered with machined (such as those created by
121 libvirt-lxc or nspawn), and which run systemd inside.
123 * machinectl gained a new "reboot" command that may be used to
124 trigger a reboot on a specific container that is registered
125 with machined. This works on any container that runs an init
128 * systemctl gained a new "list-timers" command to print a nice
129 listing of installed timer units with the times they elapse
132 * Alternative reboot() parameters may now be specified on the
133 "systemctl reboot" command line and are passed to the
134 reboot() system call.
136 * systemctl gained a new --job-mode= switch to configure the
137 mode to queue a job with. This is a more generic version of
138 --fail, --irreversible, --ignore-dependencies which are
139 still available but not advertised anymore.
141 * systemd-activate gained a new --setenv= parameter to specify
142 additional environment variables to pass to the executed
145 * /etc/systemd/system.conf gained new settings to configure
146 various default timeouts of units, as well as the default
147 start limit interval and burst. These may still be overriden
150 * PID1 will now export on the bus profile data of the security
151 policy upload process (such as the SELinux policy upload to
154 * journald: when forwarding logs to the console include
157 * OnCalendar= in timer units now understands the special
158 strings "yearly" and "annually". (Both are equivalent)
160 * The accuracy of timer units is now configurable with the new
161 AccuracySec= setting. It defaults to 1min.
163 * A new dependency type JoinsNamespaceOf= has been added that
164 allows running two services within the same /tmp and network
165 namespace, if PrivateNetwork= or PrivateTmp= are used.
167 * A new command "cat" has been added to systemctl. It outputs
168 the original unit file of a unit, and concatenates the
169 contents of addition "drop-in" unit file snippets to it, so
170 that the full configuration is shown.
172 * systemctl now supports globbing on the various "list-xyz"
173 commands, like "list-units" or "list-sockets", as well as on
174 thsoe commands which take multiple unit names.
176 * All systemd daemons now make use of the watchdog logic so
177 that systemd automatically notices when they hang.
179 * If the $container_ttys environment variable is set
180 getty-generator will automatically spawn a getty for each
181 listed tty. This is useful for container managers to request
182 login gettys to be spawned on as many ttys as needed.
184 * %h, %s, %U specifier support is not available anymore when
185 used in unit files for PID 1. This is because NSS calls are
186 not safe from PID 1. They stay available for --user
187 instances of systemd, and as special case for the root user.
189 * When the kernel command line argument "kdbus" is specified
190 systemd will automatically load the kdbus kernel
191 module. This is useful for testing kdbus without having to
192 turn it on unconditionally.
194 * loginctl gained a new "--no-legend" switch to turn off output
197 * The "sd-login.h" API gained three new calls:
198 sd_session_is_remote(), sd_session_get_remote_user(),
199 sd_session_get_remote_host() to query information about
202 * The udev device database now also carries vendor/product
203 information about SDIO devices.
205 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
206 determine whether watchdog notifications are requested by
209 * "systemd-delta" will now also display changes made via .d/
210 drop-ins for unit files.
212 * Socket-activated per-connection services will now include a
213 short description of the connection parameters in the
216 * tmpfiles gained a new "--boot" option. When this is not used
217 only lines where the command character is not suffixed with
218 "!" are executed. When this option is specified those
219 options are executed too. This is useful to ensure that
220 specific lines are not executed by accident during runtime,
221 and only at boot (for example, a line that creates
224 * A new API "sd-resolv.h" has been added, that provides a
225 simple asynchronous around glibc NSS host name resolution
226 calls, such as getaddrinfo(). In contrast to glibc's
227 getaddrinfo_a() it does not use signals. In contrast to most
228 other asynchronous name resolution libraries this one does
229 not not reimplement DNS, but reused NSS, so that alternative
230 host name resolution systems continue to work, such as mDNS,
231 LDAP, ... This API is based on libasyncns, but has been
232 cleaned up for inclusion in systemd.
234 * journalctl's --unit= switch gained support for globbing.
236 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h" are no
237 longer found in individual libraries libsystemd-journal.so,
238 libsystemd-login.so, libsystemd-id128.so. Instead we have
239 merged them into a single library libsystemd.so which
240 provides all symbols. The reason for this are cyclic
241 dependencies, as these libraries tend to use each other's
242 symbols. So far we maneged to work-around that by linking a
243 copy of a good part of our code into each of these libraries
244 again and again, which however makes certain things hard to
245 do, like sharing static variables. Also, it substantially
246 increases footprint. With this change there's only one
247 library for the basic APIs systemd provides. Also,
248 "sd-bus.h", "sd-memfd.h", "sd-event.h", "sd-rtnl.h",
249 "sd-resolve.h", "sd-utf8.h" are found in this library as
250 well, however are subject to the --enable-kdbus switch (see
251 below). Note that "sd-dhcp.h" and "sd-daemon.h" are not part
252 of this libraries (the former because it only consumes,
253 never provides services of/to other APIs, and the latter
254 because it is completely standalone). To make the transition
255 from the separate libraries to the unified one easy we
256 provide the --enable-compat-libs compile time switch which
257 will generate stub libraries that are compatible with the
258 old ones but redirect all calls to the new one.
260 * All the kdbus logic and the new APIs "sd-bus.h",
261 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
262 "sd-utf8.h" is compile-time optional, via the
263 "--enable-kdbus" switch and is not compiled in by
264 default. To make use of you have to explicitly enable the
265 switch. Note however, that neither the kernel nor the
266 userspace API for all of this is considered stable yet. We
267 want to maintain the freedom to still change the APIs for
268 now. By specifying this build-time switch you acknowledge
269 that you are aware of the instability of the current
270 APIs. Also, note that while kdbus is pretty much complete,
271 it lacks one thing: proper policy support. This means you
272 can build a fully working system with all features, however
273 it will be highly insecure. Policy will be added in one of
274 the next releases, at the same time as we will declare the
277 * systemctl gained a new "import-environment" command which
278 uploads the callers environment (or parts thereof) into the
279 service manager so that it is inherited by services started
280 by the manager. This is useful to upload variables like
281 $DISPLAY into the user service manager.
283 * A new PrivateDevices= switch has been added to service units
284 which allows running a service with a namespaced /dev
285 directory that does not contain any device nodes for
286 physical devices. More specifically it only includes devices
287 such as /dev/null, /dev/urandom and /dev/zero which are API
290 * logind has been extended to support behaviour like VT
291 switching on seats that do not support a VT. This makes
292 multi-session available on seats that are not the first seat
293 (seat0), and on systems where kernel support for VTs has
294 been disabled at compile time.
296 * If a process holds a delay lock for system sleep or shutdown
297 and fails to release it in time we will now log about its
298 identity. This makes it easier to identify processes that
299 cause slow suspends or power-offs.
301 * When parsing /etc/crypttab, support a new key-slot= option
302 as supported by Debian, which allows indicating which LUKS
305 * The boot-time has been improved to show information about
306 timeouts that are expiring as they are expiring.
308 * The sd_journald_sendv() API call has been updated to be
309 async-signal-safe so that it may be invoked from signal
310 handlers for logging purposes.
312 * Boot-time status output is now enabled automatically after a
313 short timeout if boot does not progress, in order to give
314 the user an indication what he is waiting for.
316 * The KillMode= switch in service units gained a new possible
317 value "mixed". If set and the unit is shutdown then the
318 initial SIGTERM signal is sent only to the main daemon
319 process, while the following SIGKILL signal is then sent to
320 all remaining processes of the service.
322 * When a scope unit is registered a new property "Controller"
323 may be set. If set to a valid bus name systemd will send a
324 RequestStop() signal to this name when it would like to shut
325 down the scope. This may be used to hook manager logic into
326 the shutdown logic of scope units. Also, scope units may now
327 be put in a special "abandoned" state in which case the
328 manager process which created them takes no further
329 responsibilities for it.
331 * When reading unit files systemd will now implicitly verify
332 the access mode of these files, and warn about certain
333 suspicious combinations. This has been added to make it
334 easier to track down packaging bugs where unit files are
335 marked executable or world-writable.
337 * systemd-nspawn gained a new "--setenv=" switch to set
338 container-wide environment variables.
340 * systemd-nspawn has been updated to create a new kdbus domain
341 for each container that is invoked, thus allowing each
342 container to have its own set of system and user busses,
343 independently of the host.
345 * systemd-nspawn gained a new --drop-capability= switch to run
346 the container with less capabilities than the default. Both
347 --drop-capability= and --capability= now take the specia
348 string "all" for dropping or keeping all capabilities.
350 * systemd-nspawn gained new switches for executing containers
351 with specific SELinux labels set.
353 * systemd-nspawn gained a new --quiet switch to not generate
354 any additional output but the container's own console
357 * systemd-nspawn gained a new --share-system switch to run a
358 container without PID namespacing enabled.
360 * systemd-nspawn gained a new --register= switch to control
361 whether the container is registered with machined or
362 not. This is useful for containers that do not register full
363 OS images, but only specific apps.
365 * systemd-nspawn gained a new --keep-unit which may be used
366 when invoked as only program from a service unit, and
367 results in registration of the unit service itself in
368 machined, instead of a newly opened scope unit.
370 * systemd-nspawn gained a new --network-interface= switch for
371 moving arbitrary interfaces to the container. The new
372 --network-veth switch creates a virtual ethernet connection
373 between host and container. Thew new --network-bridge=
374 switch then additionally allows assigning the host side of
375 this virtual ethernet connection to a bridge device.
377 * logind will now also track a "Desktop" identifier for each
378 session which encodes the desktop environment of it. This is
379 useful for desktop environments that want to identify
380 multiple running sessions of itself easily.
382 * A new SELinuxContext= setting for service units has been
383 added that allows setting a specific SELinux execution
384 context for a service.
386 * Most systemd client tools will now honour $SYSTEMD_LESS for
387 settings of the "less" pager. By default, these tools will
388 override $LESS to allow certain operations like
389 jump-to-the-end work. With $SYSTEMD_LESS it is possible to
390 influence this logic.
392 * systemd's "seccomp" hook-up has been changed to make use of
393 the libseccomp library instead of using its own
394 implementation. This has benefits for portability among
397 * For usage together with SystemCallFilter= a new
398 SystemCallErrorNumber= setting has been introduce that
399 allows configuration if a system error number to return on
400 filtered syscalls, instead of immediately killing the
401 process. Also, SystemCallArchitectures= has been added to
402 limit access to system calls of a particular architecture
403 (in order to turn off support for unused secondary
404 architectures). There's also a global
405 SystemcallArchitecture= setting in system.conf now to turn
406 off support for non-native system calls system-wide.
408 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
409 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
410 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
411 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
412 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
413 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
414 Elia Pinto, Florian Weimer, George McCollister, Goffredo
415 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
416 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
417 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
418 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
419 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
420 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
421 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
422 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
423 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
424 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
425 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
426 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
427 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
428 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
429 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
430 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
431 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
433 -- Berlin, 2014-02-xx
437 * logind has gained support for facilitating privileged input
438 and drm device access for unprivileged clients. This work is
439 useful to allow Wayland display servers (and similar
440 programs, such as kmscon) to run under the user's ID and
441 access input and drm devices which are normally
442 protected. When this is used (and the kernel is new enough)
443 logind will "mute" IO on the file descriptors passed to
444 Wayland as long as it is in the background and "unmute" it
445 if it returns into the foreground. This allows secure
446 session switching without allowing background sessions to
447 eavesdrop on input and display data. This also introduces
448 session switching support if VT support is turned off in the
449 kernel, and on seats that are not seat0.
451 * A new kernel command line option luks.options= is understood
452 now which allows specifiying LUKS options for usage for LUKS
453 encrypted partitions specified with luks.uuid=.
455 * tmpfiles.d(5) snippets may now use specifier expansion in
456 path names. More specifically %m, %b, %H, %v, are now
457 replaced by the local machine id, boot id, hostname, and
458 kernel version number.
460 * A new tmpfiles.d(5) command "m" has been introduced which
461 may be used to change the owner/group/access mode of a file
462 or directory if it exists, but do nothing if it doesn't.
464 * This release removes high-level support for the
465 MemorySoftLimit= cgroup setting. The underlying kernel
466 cgroup attribute memory.soft_limit= is currently badly
467 designed and likely to be removed from the kernel API in its
468 current form, hence we shouldn't expose it for now.
470 * The memory.use_hierarchy cgroup attribute is now enabled for
471 all cgroups systemd creates in the memory cgroup
472 hierarchy. This option is likely to be come the built-in
473 default in the kernel anyway, and the non-hierarchial mode
474 never made much sense in the intrinsically hierarchial
477 * A new field _SYSTEMD_SLICE= is logged along with all journal
478 messages containing the slice a message was generated
479 from. This is useful to allow easy per-customer filtering of
480 logs among other things.
482 * systemd-journald will no longer adjust the group of journal
483 files it creates to the "systemd-journal" group. Instead we
484 rely on the journal directory to be owned by the
485 "systemd-journal" group, and its setgid bit set, so that the
486 kernel file system layer will automatically enforce that
487 journal files inherit this group assignment. The reason for
488 this change is that we cannot allow NSS look-ups from
489 journald which would be necessary to resolve
490 "systemd-journal" to a numeric GID, because this might
491 create deadlocks if NSS involves synchronous queries to
492 other daemons (such as nscd, or sssd) which in turn are
493 logging clients of journald and might block on it, which
494 would then dead lock. A tmpfiles.d(5) snippet included in
495 systemd will make sure the setgid bit and group are
496 properly set on the journal directory if it exists on every
497 boot. However, we recommend adjusting it manually after
498 upgrades too (or from RPM scriptlets), so that the change is
499 not delayed until next reboot.
501 * Backlight and random seed files in /var/lib/ have moved into
502 the /var/lib/systemd/ directory, in order to centralize all
503 systemd generated files in one directory.
505 * Boot time performance measurements (as displayed by
506 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
507 performance information if that's available to determine how
508 much time BIOS and boot loader initialization required. With
509 a sufficiently new BIOS you hence no longer need to boot
510 with Gummiboot to get access to such information.
512 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
513 Cristian Rodríguez, Dave Reisner, David Herrmann, David
514 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
515 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
516 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
517 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
518 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
520 -- Berlin, 2013-10-02
524 * The Restart= option for services now understands a new
525 on-watchdog setting, which will restart the service
526 automatically if the service stops sending out watchdog keep
527 alive messages (as configured with WatchdogSec=).
529 * The getty generator (which is responsible for bringing up a
530 getty on configured serial consoles) will no longer only
531 start a getty on the primary kernel console but on all
532 others, too. This makes the order in which console= is
533 specified on the kernel command line less important.
535 * libsystemd-logind gained a new sd_session_get_vt() call to
536 retrieve the VT number of a session.
538 * If the option "tries=0" is set for an entry of /etc/crypttab
539 its passphrase is queried indefinitely instead of any
540 maximum number of tries.
542 * If a service with a configure PID file terminates its PID
543 file will now be removed automatically if it still exists
544 afterwards. This should put an end to stale PID files.
546 * systemd-run will now also take relative binary path names
547 for execution and no longer insists on absolute paths.
549 * InaccessibleDirectories= and ReadOnlyDirectories= now take
550 paths that are optionally prefixed with "-" to indicate that
551 it should not be considered a failure if they don't exist.
553 * journalctl -o (and similar commands) now understands a new
554 output mode "short-precise", it is similar to "short" but
555 shows timestamps with usec accuracy.
557 * The option "discard" (as known from Debian) is now
558 synonymous to "allow-discards" in /etc/crypttab. In fact,
559 "discard" is preferred now (since it is easier to remember
562 * Some licensing clean-ups were made, so that more code is now
563 LGPL-2.1 licensed than before.
565 * A minimal tool to save/restore the display backlight
566 brightness across reboots has been added. It will store the
567 backlight setting as late as possible at shutdown, and
568 restore it as early as possible during reboot.
570 * A logic to automatically discover and enable home and swap
571 partitions on GPT disks has been added. With this in place
572 /etc/fstab becomes optional for many setups as systemd can
573 discover certain partitions located on the root disk
574 automatically. Home partitions are recognized under their
575 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
576 partitions are recognized under their GPT type ID
577 0657fd6da4ab43c484e50933c84b4f4f.
579 * systemd will no longer pass any environment from the kernel
580 or initrd to system services. If you want to set an
581 environment for all services, do so via the kernel command
582 line systemd.setenv= assignment.
584 * The systemd-sysctl tool no longer natively reads the file
585 /etc/sysctl.conf. If desired, the file should be symlinked
586 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
587 legacy support by a symlink rather than built-in code, it
588 also makes the otherwise hidden order of application of the
589 different files visible. (Note that this partly reverts to a
590 pre-198 application order of sysctl knobs!)
592 * The "systemctl set-log-level" and "systemctl dump" commands
593 have been moved to systemd-analyze.
595 * systemd-run learned the new --remain-after-exit switch,
596 which causes the scope unit not to be cleaned up
597 automatically after the process terminated.
599 * tmpfiles learned a new --exclude-prefix= switch to exclude
600 certain paths from operation.
602 * journald will now automatically flush all messages to disk
603 as soon as a message of the log priorities CRIT, ALERT or
606 Contributions from: Andrew Cook, Brandon Philips, Christian
607 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
608 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
609 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
610 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
611 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
612 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
613 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
614 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
615 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
616 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
617 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
618 William Giokas, Zbigniew Jędrzejewski-Szmek
620 -- Berlin, 2013-09-13
624 * The documentation has been updated to cover the various new
625 concepts introduced with 205.
627 * Unit files now understand the new %v specifier which
628 resolves to the kernel version string as returned by "uname
631 * systemctl now supports filtering the unit list output by
632 load state, active state and sub state, using the new
635 * "systemctl status" will now show the results of the
636 condition checks (like ConditionPathExists= and similar) of
637 the last start attempts of the unit. They are also logged to
640 * "journalctl -b" may now be used to look for boot output of a
641 specific boot. Try "journalctl -b -1" for the previous boot,
642 but the syntax is substantially more powerful.
644 * "journalctl --show-cursor" has been added which prints the
645 cursor string the last shown log line. This may then be used
646 with the new "journalctl --after-cursor=" switch to continue
647 browsing logs from that point on.
649 * "journalctl --force" may now be used to force regeneration
652 * Creation of "dead" device nodes has been moved from udev
653 into kmod and tmpfiles. Previously, udev would read the kmod
654 databases to pre-generate dead device nodes based on meta
655 information contained in kernel modules, so that these would
656 be auto-loaded on access rather then at boot. As this
657 doesn't really have much to do with the exposing actual
658 kernel devices to userspace this has always been slightly
659 alien in the udev codebase. Following the new scheme kmod
660 will now generate a runtime snippet for tmpfiles from the
661 module meta information and it now is tmpfiles' job to the
662 create the nodes. This also allows overriding access and
663 other parameters for the nodes using the usual tmpfiles
664 facilities. As side effect this allows us to remove the
665 CAP_SYS_MKNOD capability bit from udevd entirely.
667 * logind's device ACLs may now be applied to these "dead"
668 devices nodes too, thus finally allowing managed access to
669 devices such as /dev/snd/sequencer whithout loading the
670 backing module right-away.
672 * A new RPM macro has been added that may be used to apply
673 tmpfiles configuration during package installation.
675 * systemd-detect-virt and ConditionVirtualization= now can
676 detect User-Mode-Linux machines (UML).
678 * journald will now implicitly log the effective capabilities
679 set of processes in the message metadata.
681 * systemd-cryptsetup has gained support for TrueCrypt volumes.
683 * The initrd interface has been simplified (more specifically,
684 support for passing performance data via environment
685 variables and fsck results via files in /run has been
686 removed). These features were non-essential, and are
687 nowadays available in a much nicer way by having systemd in
688 the initrd serialize its state and have the hosts systemd
689 deserialize it again.
691 * The udev "keymap" data files and tools to apply keyboard
692 specific mappings of scan to key codes, and force-release
693 scan code lists have been entirely replaced by a udev
694 "keyboard" builtin and a hwdb data file.
696 * systemd will now honour the kernel's "quiet" command line
697 argument also during late shutdown, resulting in a
698 completely silent shutdown when used.
700 * There's now an option to control the SO_REUSEPORT socket
701 option in .socket units.
703 * Instance units will now automatically get a per-template
704 subslice of system.slice unless something else is explicitly
705 configured. For example, instances of sshd@.service will now
706 implicitly be placed in system-sshd.slice rather than
707 system.slice as before.
709 * Test coverage support may now be enabled at build time.
711 Contributions from: Dave Reisner, Frederic Crozat, Harald
712 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
713 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
714 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
715 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
716 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
717 Giokas, Zbigniew Jędrzejewski-Szmek
719 -- Berlin, 2013-07-23
723 * Two new unit types have been introduced:
725 Scope units are very similar to service units, however, are
726 created out of pre-existing processes -- instead of PID 1
727 forking off the processes. By using scope units it is
728 possible for system services and applications to group their
729 own child processes (worker processes) in a powerful way
730 which then maybe used to organize them, or kill them
731 together, or apply resource limits on them.
733 Slice units may be used to partition system resources in an
734 hierarchial fashion and then assign other units to them. By
735 default there are now three slices: system.slice (for all
736 system services), user.slice (for all user sessions),
737 machine.slice (for VMs and containers).
739 Slices and scopes have been introduced primarily in
740 context of the work to move cgroup handling to a
741 single-writer scheme, where only PID 1
742 creates/removes/manages cgroups.
744 * There's a new concept of "transient" units. In contrast to
745 normal units these units are created via an API at runtime,
746 not from configuration from disk. More specifically this
747 means it is now possible to run arbitrary programs as
748 independent services, with all execution parameters passed
749 in via bus APIs rather than read from disk. Transient units
750 make systemd substantially more dynamic then it ever was,
751 and useful as a general batch manager.
753 * logind has been updated to make use of scope and slice units
754 for managing user sessions. As a user logs in he will get
755 his own private slice unit, to which all sessions are added
756 as scope units. We also added support for automatically
757 adding an instance of user@.service for the user into the
758 slice. Effectively logind will no longer create cgroup
759 hierarchies on its own now, it will defer entirely to PID 1
760 for this by means of scope, service and slice units. Since
761 user sessions this way become entities managed by PID 1
762 the output of "systemctl" is now a lot more comprehensive.
764 * A new mini-daemon "systemd-machined" has been added which
765 may be used by virtualization managers to register local
766 VMs/containers. nspawn has been updated accordingly, and
767 libvirt will be updated shortly. machined will collect a bit
768 of meta information about the VMs/containers, and assign
769 them their own scope unit (see above). The collected
770 meta-data is then made available via the "machinectl" tool,
771 and exposed in "ps" and similar tools. machined/machinectl
772 is compile-time optional.
774 * As discussed earlier, the low-level cgroup configuration
775 options ControlGroup=, ControlGroupModify=,
776 ControlGroupPersistent=, ControlGroupAttribute= have been
777 removed. Please use high-level attribute settings instead as
780 * A new bus call SetUnitProperties() has been added to alter
781 various runtime parameters of a unit. This is primarily
782 useful to alter cgroup parameters dynamically in a nice way,
783 but will be extended later on to make more properties
784 modifiable at runtime. systemctl gained a new set-properties
785 command that wraps this call.
787 * A new tool "systemd-run" has been added which can be used to
788 run arbitrary command lines as transient services or scopes,
789 while configuring a number of settings via the command
790 line. This tool is currently very basic, however already
791 very useful. We plan to extend this tool to even allow
792 queuing of execution jobs with time triggers from the
793 command line, similar in fashion to "at".
795 * nspawn will now inform the user explicitly that kernels with
796 audit enabled break containers, and suggest the user to turn
799 * Support for detecting the IMA and AppArmor security
800 frameworks with ConditionSecurity= has been added.
802 * journalctl gained a new "-k" switch for showing only kernel
803 messages, mimicking dmesg output; in addition to "--user"
804 and "--system" switches for showing only user's own logs
807 * systemd-delta can now show information about drop-in
808 snippets extending unit files.
810 * libsystemd-bus has been substantially updated but is still
811 not available as public API.
813 * systemd will now look for the "debug" argument on the kernel
814 command line and enable debug logging, similar to
815 "systemd.log_level=debug" already did before.
817 * "systemctl set-default", "systemctl get-default" has been
818 added to configure the default.target symlink, which
819 controls what to boot into by default.
821 * "systemctl set-log-level" has been added as a convenient
822 way to raise and lower systemd logging threshold.
824 * "systemd-analyze plot" will now show the time the various
825 generators needed for execution, as well as information
826 about the unit file loading.
828 * libsystemd-journal gained a new sd_journal_open_files() call
829 for opening specific journal files. journactl also gained a
830 new switch to expose this new functionality. Previously we
831 only supported opening all files from a directory, or all
832 files from the system, as opening individual files only is
833 racy due to journal file rotation.
835 * systemd gained the new DefaultEnvironment= setting in
836 /etc/systemd/system.conf to set environment variables for
839 * If a privileged process logs a journal message with the
840 OBJECT_PID= field set, then journald will automatically
841 augment this with additional OBJECT_UID=, OBJECT_GID=,
842 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
843 system services want to log events about specific client
844 processes. journactl/systemctl has been updated to make use
845 of this information if all log messages regarding a specific
848 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
849 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
850 Reisner, David Coppa, David King, David Strauss, Eelco
851 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
852 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
853 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
854 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
855 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
856 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
857 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
858 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
859 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
864 * The Python bindings gained some minimal support for the APIs
865 exposed by libsystemd-logind.
867 * ConditionSecurity= gained support for detecting SMACK. Since
868 this condition already supports SELinux and AppArmor we only
869 miss IMA for this. Patches welcome!
871 Contributions from: Karol Lewandowski, Lennart Poettering,
872 Zbigniew Jędrzejewski-Szmek
876 * systemd-nspawn will now create /etc/resolv.conf if
877 necessary, before bind-mounting the host's file onto it.
879 * systemd-nspawn will now store meta information about a
880 container on the container's cgroup as extended attribute
881 fields, including the root directory.
883 * The cgroup hierarchy has been reworked in many ways. All
884 objects any of the components systemd creates in the cgroup
885 tree are now suffixed. More specifically, user sessions are
886 now placed in cgroups suffixed with ".session", users in
887 cgroups suffixed with ".user", and nspawn containers in
888 cgroups suffixed with ".nspawn". Furthermore, all cgroup
889 names are now escaped in a simple scheme to avoid collision
890 of userspace object names with kernel filenames. This work
891 is preparation for making these objects relocatable in the
892 cgroup tree, in order to allow easy resource partitioning of
893 these objects without causing naming conflicts.
895 * systemctl list-dependencies gained the new switches
896 --plain, --reverse, --after and --before.
898 * systemd-inhibit now shows the process name of processes that
899 have taken an inhibitor lock.
901 * nss-myhostname will now also resolve "localhost"
902 implicitly. This makes /etc/hosts an optional file and
903 nicely handles that on IPv6 ::1 maps to both "localhost" and
906 * libsystemd-logind.so gained a new call
907 sd_get_machine_names() to enumerate running containers and
908 VMs (currently only supported by very new libvirt and
909 nspawn). sd_login_monitor can now be used to watch
910 VMs/containers coming and going.
912 * .include is not allowed recursively anymore, and only in
913 unit files. Usually it is better to use drop-in snippets in
914 .d/*.conf anyway, as introduced with systemd 198.
916 * systemd-analyze gained a new "critical-chain" command that
917 determines the slowest chain of units run during system
918 boot-up. It is very useful for tracking down where
919 optimizing boot time is the most beneficial.
921 * systemd will no longer allow manipulating service paths in
922 the name=systemd:/system cgroup tree using ControlGroup= in
923 units. (But is still fine with it in all other dirs.)
925 * There's a new systemd-nspawn@.service service file that may
926 be used to easily run nspawn containers as system
927 services. With the container's root directory in
928 /var/lib/container/foobar it is now sufficient to run
929 "systemctl start systemd-nspawn@foobar.service" to boot it.
931 * systemd-cgls gained a new parameter "--machine" to list only
932 the processes within a certain container.
934 * ConditionSecurity= now can check for "apparmor". We still
935 are lacking checks for SMACK and IMA for this condition
936 check though. Patches welcome!
938 * A new configuration file /etc/systemd/sleep.conf has been
939 added that may be used to configure which kernel operation
940 systemd is supposed to execute when "suspend", "hibernate"
941 or "hybrid-sleep" is requested. This makes the new kernel
942 "freeze" state accessible to the user.
944 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
945 the passed argument if applicable.
947 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
948 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
949 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
950 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
951 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
952 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
953 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
958 * The output of 'systemctl list-jobs' got some polishing. The
959 '--type=' argument may now be passed more than once. A new
960 command 'systemctl list-sockets' has been added which shows
961 a list of kernel sockets systemd is listening on with the
962 socket units they belong to, plus the units these socket
965 * The experimental libsystemd-bus library got substantial
966 updates to work in conjunction with the (also experimental)
967 kdbus kernel project. It works well enough to exchange
968 messages with some sophistication. Note that kdbus is not
969 ready yet, and the library is mostly an elaborate test case
970 for now, and not installable.
972 * systemd gained a new unit 'systemd-static-nodes.service'
973 that generates static device nodes earlier during boot, and
974 can run in conjunction with udev.
976 * libsystemd-login gained a new call sd_pid_get_user_unit()
977 to retrieve the user systemd unit a process is running
978 in. This is useful for systems where systemd is used as
981 * systemd-nspawn now places all containers in the new /machine
982 top-level cgroup directory in the name=systemd
983 hierarchy. libvirt will soon do the same, so that we get a
984 uniform separation of /system, /user and /machine for system
985 services, user processes and containers/virtual
986 machines. This new cgroup hierarchy is also useful to stick
987 stable names to specific container instances, which can be
988 recognized later this way (this name may be controlled
989 via systemd-nspawn's new -M switch). libsystemd-login also
990 gained a new call sd_pid_get_machine_name() to retrieve the
991 name of the container/VM a specific process belongs to.
993 * bootchart can now store its data in the journal.
995 * libsystemd-journal gained a new call
996 sd_journal_add_conjunction() for AND expressions to the
997 matching logic. This can be used to express more complex
1000 * journactl can now take multiple --unit= and --user-unit=
1003 * The cryptsetup logic now understands the "luks.key=" kernel
1004 command line switch for specifying a file to read the
1005 decryption key from. Also, if a configured key file is not
1006 found the tool will now automatically fall back to prompting
1009 * Python systemd.journal module was updated to wrap recently
1010 added functions from libsystemd-journal. The interface was
1011 changed to bring the low level interface in s.j._Reader
1012 closer to the C API, and the high level interface in
1013 s.j.Reader was updated to wrap and convert all data about
1016 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1017 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1018 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1019 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1020 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1021 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1025 * journalctl --update-catalog now understands a new --root=
1026 option to operate on catalogs found in a different root
1029 * During shutdown after systemd has terminated all running
1030 services a final killing loop kills all remaining left-over
1031 processes. We will now print the name of these processes
1032 when we send SIGKILL to them, since this usually indicates a
1035 * If /etc/crypttab refers to password files stored on
1036 configured mount points automatic dependencies will now be
1037 generated to ensure the specific mount is established first
1038 before the key file is attempted to be read.
1040 * 'systemctl status' will now show information about the
1041 network sockets a socket unit is listening on.
1043 * 'systemctl status' will also shown information about any
1044 drop-in configuration file for units. (Drop-In configuration
1045 files in this context are files such as
1046 /etc/systemd/systemd/foobar.service.d/*.conf)
1048 * systemd-cgtop now optionally shows summed up CPU times of
1049 cgroups. Press '%' while running cgtop to switch between
1050 percentage and absolute mode. This is useful to determine
1051 which cgroups use up the most CPU time over the entire
1052 runtime of the system. systemd-cgtop has also been updated
1053 to be 'pipeable' for processing with further shell tools.
1055 * 'hostnamectl set-hostname' will now allow setting of FQDN
1058 * The formatting and parsing of time span values has been
1059 changed. The parser now understands fractional expressions
1060 such as "5.5h". The formatter will now output fractional
1061 expressions for all time spans under 1min, i.e. "5.123456s"
1062 rather than "5s 123ms 456us". For time spans under 1s
1063 millisecond values are shown, for those under 1ms
1064 microsecond values are shown. This should greatly improve
1065 all time-related output of systemd.
1067 * libsystemd-login and libsystemd-journal gained new
1068 functions for querying the poll() events mask and poll()
1069 timeout value for integration into arbitrary event
1072 * localectl gained the ability to list available X11 keymaps
1073 (models, layouts, variants, options).
1075 * 'systemd-analyze dot' gained the ability to filter for
1076 specific units via shell-style globs, to create smaller,
1077 more useful graphs. I.e. it's now possible to create simple
1078 graphs of all the dependencies between only target units, or
1079 of all units that Avahi has dependencies with.
1081 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1082 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
1083 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
1084 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
1085 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
1086 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
1087 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
1091 * The boot-time readahead implementation for rotating media
1092 will now read the read-ahead data in multiple passes which
1093 consist of all read requests made in equidistant time
1094 intervals. This means instead of strictly reading read-ahead
1095 data in its physical order on disk we now try to find a
1096 middle ground between physical and access time order.
1098 * /etc/os-release files gained a new BUILD_ID= field for usage
1099 on operating systems that provide continuous builds of OS
1102 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
1103 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
1104 William Douglas, Zbigniew Jędrzejewski-Szmek
1108 * systemd-python gained an API exposing libsystemd-daemon.
1110 * The SMACK setup logic gained support for uploading CIPSO
1113 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
1114 ReadOnlyDirectories= and InaccessibleDirectories= has
1115 changed. The private /tmp and /var/tmp directories are now
1116 shared by all processes of a service (which means
1117 ExecStartPre= may now leave data in /tmp that ExecStart= of
1118 the same service can still access). When a service is
1119 stopped its temporary directories are immediately deleted
1120 (normal clean-up with tmpfiles is still done in addition to
1123 * By default, systemd will now set a couple of sysctl
1124 variables in the kernel: the safe sysrq options are turned
1125 on, IP route verification is turned on, and source routing
1126 disabled. The recently added hardlink and softlink
1127 protection of the kernel is turned on. These settings should
1128 be reasonably safe, and good defaults for all new systems.
1130 * The predictable network naming logic may now be turned off
1131 with a new kernel command line switch: net.ifnames=0.
1133 * A new libsystemd-bus module has been added that implements a
1134 pretty complete D-Bus client library. For details see:
1136 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
1138 * journald will now explicitly flush the journal files to disk
1139 at the latest 5min after each write. The file will then also
1140 be marked offline until the next write. This should increase
1141 reliability in case of a crash. The synchronization delay
1142 can be configured via SyncIntervalSec= in journald.conf.
1144 * There's a new remote-fs-setup.target unit that can be used
1145 to pull in specific services when at least one remote file
1146 system is to be mounted.
1148 * There are new targets timers.target and paths.target as
1149 canonical targets to pull user timer and path units in
1150 from. This complements sockets.target with a similar
1151 purpose for socket units.
1153 * libudev gained a new call udev_device_set_attribute_value()
1154 to set sysfs attributes of a device.
1156 * The udev daemon now sets the default number of worker
1157 processes executed in parallel based on the number of available
1158 CPUs instead of the amount of available RAM. This is supposed
1159 to provide a more reliable default and limit a too aggressive
1160 paralellism for setups with 1000s of devices connected.
1162 Contributions from: Auke Kok, Colin Walters, Cristian
1163 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
1164 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
1165 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
1166 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
1167 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
1168 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
1169 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
1170 Zbigniew Jędrzejewski-Szmek
1174 * Configuration of unit files may now be extended via drop-in
1175 files without having to edit/override the unit files
1176 themselves. More specifically, if the administrator wants to
1177 change one value for a service file foobar.service he can
1178 now do so by dropping in a configuration snippet into
1179 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
1180 will load all these snippets and apply them on top of the
1181 main unit configuration file, possibly extending or
1182 overriding its settings. Using these drop-in snippets is
1183 generally nicer than the two earlier options for changing
1184 unit files locally: copying the files from
1185 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
1186 them there; or creating a new file in /etc/systemd/system/
1187 that incorporates the original one via ".include". Drop-in
1188 snippets into these .d/ directories can be placed in any
1189 directory systemd looks for units in, and the usual
1190 overriding semantics between /usr/lib, /etc and /run apply
1193 * Most unit file settings which take lists of items can now be
1194 reset by assigning the empty string to them. For example,
1195 normally, settings such as Environment=FOO=BAR append a new
1196 environment variable assignment to the environment block,
1197 each time they are used. By assigning Environment= the empty
1198 string the environment block can be reset to empty. This is
1199 particularly useful with the .d/*.conf drop-in snippets
1200 mentioned above, since this adds the ability to reset list
1201 settings from vendor unit files via these drop-ins.
1203 * systemctl gained a new "list-dependencies" command for
1204 listing the dependencies of a unit recursively.
1206 * Inhibitors are now honored and listed by "systemctl
1207 suspend", "systemctl poweroff" (and similar) too, not only
1208 GNOME. These commands will also list active sessions by
1211 * Resource limits (as exposed by the various control group
1212 controllers) can now be controlled dynamically at runtime
1213 for all units. More specifically, you can now use a command
1214 like "systemctl set-cgroup-attr foobar.service cpu.shares
1215 2000" to alter the CPU shares a specific service gets. These
1216 settings are stored persistently on disk, and thus allow the
1217 administrator to easily adjust the resource usage of
1218 services with a few simple commands. This dynamic resource
1219 management logic is also available to other programs via the
1220 bus. Almost any kernel cgroup attribute and controller is
1223 * systemd-vconsole-setup will now copy all font settings to
1224 all allocated VTs, where it previously applied them only to
1227 * libsystemd-login gained the new sd_session_get_tty() API
1230 * This release drops support for a few legacy or
1231 distribution-specific LSB facility names when parsing init
1232 scripts: $x-display-manager, $mail-transfer-agent,
1233 $mail-transport-agent, $mail-transfer-agent, $smtp,
1234 $null. Also, the mail-transfer-agent.target unit backing
1235 this has been removed. Distributions which want to retain
1236 compatibility with this should carry the burden for
1237 supporting this themselves and patch support for these back
1238 in, if they really need to. Also, the facilities $syslog and
1239 $local_fs are now ignored, since systemd does not support
1240 early-boot LSB init scripts anymore, and these facilities
1241 are implied anyway for normal services. syslog.target has
1244 * There are new bus calls on PID1's Manager object for
1245 cancelling jobs, and removing snapshot units. Previously,
1246 both calls were only available on the Job and Snapshot
1249 * systemd-journal-gatewayd gained SSL support.
1251 * The various "environment" files, such as /etc/locale.conf
1252 now support continuation lines with a backslash ("\") as
1253 last character in the line, similar in style (but different)
1254 to how this is supported in shells.
1256 * For normal user processes the _SYSTEMD_USER_UNIT= field is
1257 now implicitly appended to every log entry logged. systemctl
1258 has been updated to filter by this field when operating on a
1259 user systemd instance.
1261 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
1262 CAP_AUDIT_CONTROL capabilities to the capabilities set for
1263 the container. This makes it easier to boot unmodified
1264 Fedora systems in a container, which however still requires
1265 audit=0 to be passed on the kernel command line. Auditing in
1266 kernel and userspace is unfortunately still too broken in
1267 context of containers, hence we recommend compiling it out
1268 of the kernel or using audit=0. Hopefully this will be fixed
1269 one day for good in the kernel.
1271 * nspawn gained the new --bind= and --bind-ro= parameters to
1272 bind mount specific directories from the host into the
1275 * nspawn will now mount its own devpts file system instance
1276 into the container, in order not to leak pty devices from
1277 the host into the container.
1279 * systemd will now read the firmware boot time performance
1280 information from the EFI variables, if the used boot loader
1281 supports this, and takes it into account for boot performance
1282 analysis via "systemd-analyze". This is currently supported
1283 only in conjunction with Gummiboot, but could be supported
1284 by other boot loaders too. For details see:
1286 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
1288 * A new generator has been added that automatically mounts the
1289 EFI System Partition (ESP) to /boot, if that directory
1290 exists, is empty, and no other file system has been
1291 configured to be mounted there.
1293 * logind will now send out PrepareForSleep(false) out
1294 unconditionally, after coming back from suspend. This may be
1295 used by applications as asynchronous notification for
1296 system resume events.
1298 * "systemctl unlock-sessions" has been added, that allows
1299 unlocking the screens of all user sessions at once, similar
1300 how "systemctl lock-sessions" already locked all users
1301 sessions. This is backed by a new D-Bus call UnlockSessions().
1303 * "loginctl seat-status" will now show the master device of a
1304 seat. (i.e. the device of a seat that needs to be around for
1305 the seat to be considered available, usually the graphics
1308 * tmpfiles gained a new "X" line type, that allows
1309 configuration of files and directories (with wildcards) that
1310 shall be excluded from automatic cleanup ("aging").
1312 * udev default rules set the device node permissions now only
1313 at "add" events, and do not change them any longer with a
1314 later "change" event.
1316 * The log messages for lid events and power/sleep keypresses
1317 now carry a message ID.
1319 * We now have a substantially larger unit test suite, but this
1320 continues to be work in progress.
1322 * udevadm hwdb gained a new --root= parameter to change the
1323 root directory to operate relative to.
1325 * logind will now issue a background sync() request to the kernel
1326 early at shutdown, so that dirty buffers are flushed to disk early
1327 instead of at the last moment, in order to optimize shutdown
1330 * A new bootctl tool has been added that is an interface for
1331 certain boot loader operations. This is currently a preview
1332 and is likely to be extended into a small mechanism daemon
1333 like timedated, localed, hostnamed, and can be used by
1334 graphical UIs to enumerate available boot options, and
1335 request boot into firmware operations.
1337 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
1338 the rest of the package. It also has been updated to work
1339 correctly in initrds.
1341 * Policykit previously has been runtime optional, and is now
1342 also compile time optional via a configure switch.
1344 * systemd-analyze has been reimplemented in C. Also "systemctl
1345 dot" has moved into systemd-analyze.
1347 * "systemctl status" with no further parameters will now print
1348 the status of all active or failed units.
1350 * Operations such as "systemctl start" can now be executed
1351 with a new mode "--irreversible" which may be used to queue
1352 operations that cannot accidentally be reversed by a later
1353 job queuing. This is by default used to make shutdown
1354 requests more robust.
1356 * The Python API of systemd now gained a new module for
1357 reading journal files.
1359 * A new tool kernel-install has been added that can install
1360 kernel images according to the Boot Loader Specification:
1362 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
1364 * Boot time console output has been improved to provide
1365 animated boot time output for hanging jobs.
1367 * A new tool systemd-activate has been added which can be used
1368 to test socket activation with, directly from the command
1369 line. This should make it much easier to test and debug
1370 socket activation in daemons.
1372 * journalctl gained a new "--reverse" (or -r) option to show
1373 journal output in reverse order (i.e. newest line first).
1375 * journalctl gained a new "--pager-end" (or -e) option to jump
1376 to immediately jump to the end of the journal in the
1377 pager. This is only supported in conjunction with "less".
1379 * journalctl gained a new "--user-unit=" option, that works
1380 similar to "--unit=" but filters for user units rather than
1383 * A number of unit files to ease adoption of systemd in
1384 initrds has been added. This moves some minimal logic from
1385 the various initrd implementations into systemd proper.
1387 * The journal files are now owned by a new group
1388 "systemd-journal", which exists specifically to allow access
1389 to the journal, and nothing else. Previously, we used the
1390 "adm" group for that, which however possibly covers more
1391 than just journal/log file access. This new group is now
1392 already used by systemd-journal-gatewayd to ensure this
1393 daemon gets access to the journal files and as little else
1394 as possible. Note that "make install" will also set FS ACLs
1395 up for /var/log/journal to give "adm" and "wheel" read
1396 access to it, in addition to "systemd-journal" which owns
1397 the journal files. We recommend that packaging scripts also
1398 add read access to "adm" + "wheel" to /var/log/journal, and
1399 all existing/future journal files. To normal users and
1400 administrators little changes, however packagers need to
1401 ensure to create the "systemd-journal" system group at
1402 package installation time.
1404 * The systemd-journal-gatewayd now runs as unprivileged user
1405 systemd-journal-gateway:systemd-journal-gateway. Packaging
1406 scripts need to create these system user/group at
1409 * timedated now exposes a new boolean property CanNTP that
1410 indicates whether a local NTP service is available or not.
1412 * systemd-detect-virt will now also detect xen PVs
1414 * The pstore file system is now mounted by default, if it is
1417 * In addition to the SELinux and IMA policies we will now also
1418 load SMACK policies at early boot.
1420 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
1421 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
1422 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
1423 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
1424 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
1425 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
1426 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
1427 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
1428 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
1429 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
1430 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
1431 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
1432 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
1433 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
1437 * Timer units now support calendar time events in addition to
1438 monotonic time events. That means you can now trigger a unit
1439 based on a calendar time specification such as "Thu,Fri
1440 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
1441 or fifth day of any month of the year 2013, given that it is
1442 a thursday or friday. This brings timer event support
1443 considerably closer to cron's capabilities. For details on
1444 the supported calendar time specification language see
1447 * udev now supports a number of different naming policies for
1448 network interfaces for predictable names, and a combination
1449 of these policies is now the default. Please see this wiki
1450 document for details:
1452 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
1454 * Auke Kok's bootchart implementation has been added to the
1455 systemd tree. It's an optional component that can graph the
1456 boot in quite some detail. It's one of the best bootchart
1457 implementations around and minimal in its code and
1460 * nss-myhostname has been integrated into the systemd source
1461 tree. nss-myhostname guarantees that the local hostname
1462 always stays resolvable via NSS. It has been a weak
1463 requirement of systemd-hostnamed since a long time, and
1464 since its code is actually trivial we decided to just
1465 include it in systemd's source tree. It can be turned off
1466 with a configure switch.
1468 * The read-ahead logic is now capable of properly detecting
1469 whether a btrfs file system is on SSD or rotating media, in
1470 order to optimize the read-ahead scheme. Previously, it was
1471 only capable of detecting this on traditional file systems
1474 * In udev, additional device properties are now read from the
1475 IAB in addition to the OUI database. Also, Bluetooth company
1476 identities are attached to the devices as well.
1478 * In service files %U may be used as specifier that is
1479 replaced by the configured user name of the service.
1481 * nspawn may now be invoked without a controlling TTY. This
1482 makes it suitable for invocation as its own service. This
1483 may be used to set up a simple containerized server system
1484 using only core OS tools.
1486 * systemd and nspawn can now accept socket file descriptors
1487 when they are started for socket activation. This enables
1488 implementation of socket activated nspawn
1489 containers. i.e. think about autospawning an entire OS image
1490 when the first SSH or HTTP connection is received. We expect
1491 that similar functionality will also be added to libvirt-lxc
1494 * journalctl will now suppress ANSI color codes when
1495 presenting log data.
1497 * systemctl will no longer show control group information for
1498 a unit if a the control group is empty anyway.
1500 * logind can now automatically suspend/hibernate/shutdown the
1503 * /etc/machine-info and hostnamed now also expose the chassis
1504 type of the system. This can be used to determine whether
1505 the local system is a laptop, desktop, handset or
1506 tablet. This information may either be configured by the
1507 user/vendor or is automatically determined from ACPI and DMI
1508 information if possible.
1510 * A number of PolicyKit actions are now bound together with
1511 "imply" rules. This should simplify creating UIs because
1512 many actions will now authenticate similar ones as well.
1514 * Unit files learnt a new condition ConditionACPower= which
1515 may be used to conditionalize a unit depending on whether an
1516 AC power source is connected or not, of whether the system
1517 is running on battery power.
1519 * systemctl gained a new "is-failed" verb that may be used in
1520 shell scripts and suchlike to check whether a specific unit
1521 is in the "failed" state.
1523 * The EnvironmentFile= setting in unit files now supports file
1524 globbing, and can hence be used to easily read a number of
1525 environment files at once.
1527 * systemd will no longer detect and recognize specific
1528 distributions. All distribution-specific #ifdeffery has been
1529 removed, systemd is now fully generic and
1530 distribution-agnostic. Effectively, not too much is lost as
1531 a lot of the code is still accessible via explicit configure
1532 switches. However, support for some distribution specific
1533 legacy configuration file formats has been dropped. We
1534 recommend distributions to simply adopt the configuration
1535 files everybody else uses now and convert the old
1536 configuration from packaging scripts. Most distributions
1537 already did that. If that's not possible or desirable,
1538 distributions are welcome to forward port the specific
1539 pieces of code locally from the git history.
1541 * When logging a message about a unit systemd will now always
1542 log the unit name in the message meta data.
1544 * localectl will now also discover system locale data that is
1545 not stored in locale archives, but directly unpacked.
1547 * logind will no longer unconditionally use framebuffer
1548 devices as seat masters, i.e. as devices that are required
1549 to be existing before a seat is considered preset. Instead,
1550 it will now look for all devices that are tagged as
1551 "seat-master" in udev. By default framebuffer devices will
1552 be marked as such, but depending on local systems other
1553 devices might be marked as well. This may be used to
1554 integrate graphics cards using closed source drivers (such
1555 as NVidia ones) more nicely into logind. Note however, that
1556 we recommend using the open source NVidia drivers instead,
1557 and no udev rules for the closed-source drivers will be
1558 shipped from us upstream.
1560 Contributions from: Adam Williamson, Alessandro Crismani, Auke
1561 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
1562 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
1563 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
1564 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
1565 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
1566 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
1567 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
1568 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
1569 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
1570 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
1571 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
1576 * udev gained support for loading additional device properties
1577 from an indexed database that is keyed by vendor/product IDs
1578 and similar device identifiers. For the beginning this
1579 "hwdb" is populated with data from the well-known PCI and
1580 USB database, but also includes PNP, ACPI and OID data. In
1581 the longer run this indexed database shall grow into
1582 becoming the one central database for non-essential
1583 userspace device metadata. Previously, data from the PCI/USB
1584 database was only attached to select devices, since the
1585 lookup was a relatively expensive operation due to O(n) time
1586 complexity (with n being the number of entries in the
1587 database). Since this is now O(1), we decided to add in this
1588 data for all devices where this is available, by
1589 default. Note that the indexed database needs to be rebuilt
1590 when new data files are installed. To achieve this you need
1591 to update your packaging scripts to invoke "udevadm hwdb
1592 --update" after installation of hwdb data files. For
1593 RPM-based distributions we introduced the new
1594 %udev_hwdb_update macro for this purpose.
1596 * The Journal gained support for the "Message Catalog", an
1597 indexed database to link up additional information with
1598 journal entries. For further details please check:
1600 http://www.freedesktop.org/wiki/Software/systemd/catalog
1602 The indexed message catalog database also needs to be
1603 rebuilt after installation of message catalog files. Use
1604 "journalctl --update-catalog" for this. For RPM-based
1605 distributions we introduced the %journal_catalog_update
1606 macro for this purpose.
1608 * The Python Journal bindings gained support for the standard
1609 Python logging framework.
1611 * The Journal API gained new functions for checking whether
1612 the underlying file system of a journal file is capable of
1613 properly reporting file change notifications, or whether
1614 applications that want to reflect journal changes "live"
1615 need to recheck journal files continuously in appropriate
1618 * It is now possible to set the "age" field for tmpfiles
1619 entries to 0, indicating that files matching this entry
1620 shall always be removed when the directories are cleaned up.
1622 * coredumpctl gained a new "gdb" verb which invokes gdb
1623 right-away on the selected coredump.
1625 * There's now support for "hybrid sleep" on kernels that
1626 support this, in addition to "suspend" and "hibernate". Use
1627 "systemctl hybrid-sleep" to make use of this.
1629 * logind's HandleSuspendKey= setting (and related settings)
1630 now gained support for a new "lock" setting to simply
1631 request the screen lock on all local sessions, instead of
1632 actually executing a suspend or hibernation.
1634 * systemd will now mount the EFI variables file system by
1637 * Socket units now gained support for configuration of the
1638 SMACK security label.
1640 * timedatectl will now output the time of the last and next
1641 daylight saving change.
1643 * We dropped support for various legacy and distro-specific
1644 concepts, such as insserv, early-boot SysV services
1645 (i.e. those for non-standard runlevels such as 'b' or 'S')
1646 or ArchLinux /etc/rc.conf support. We recommend the
1647 distributions who still need support this to either continue
1648 to maintain the necessary patches downstream, or find a
1649 different solution. (Talk to us if you have questions!)
1651 * Various systemd components will now bypass PolicyKit checks
1652 for root and otherwise handle properly if PolicyKit is not
1653 found to be around. This should fix most issues for
1654 PolicyKit-less systems. Quite frankly this should have been
1655 this way since day one. It is absolutely our intention to
1656 make systemd work fine on PolicyKit-less systems, and we
1657 consider it a bug if something doesn't work as it should if
1658 PolicyKit is not around.
1660 * For embedded systems it is now possible to build udev and
1661 systemd without blkid and/or kmod support.
1663 * "systemctl switch-root" is now capable of switching root
1664 more than once. I.e. in addition to transitions from the
1665 initrd to the host OS it is now possible to transition to
1666 further OS images from the host. This is useful to implement
1667 offline updating tools.
1669 * Various other additions have been made to the RPM macros
1670 shipped with systemd. Use %udev_rules_update() after
1671 installing new udev rules files. %_udevhwdbdir,
1672 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
1673 %_sysctldir are now available which resolve to the right
1674 directories for packages to place various data files in.
1676 * journalctl gained the new --full switch (in addition to
1677 --all, to disable ellipsation for long messages.
1679 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
1680 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
1681 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
1682 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
1683 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
1684 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
1685 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
1686 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
1687 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
1691 * journalctl gained new --since= and --until= switches to
1692 filter by time. It also now supports nice filtering for
1693 units via --unit=/-u.
1695 * Type=oneshot services may use ExecReload= and do the
1698 * The journal daemon now supports time-based rotation and
1699 vacuuming, in addition to the usual disk-space based
1702 * The journal will now index the available field values for
1703 each field name. This enables clients to show pretty drop
1704 downs of available match values when filtering. The bash
1705 completion of journalctl has been updated
1706 accordingly. journalctl gained a new switch -F to list all
1707 values a certain field takes in the journal database.
1709 * More service events are now written as structured messages
1710 to the journal, and made recognizable via message IDs.
1712 * The timedated, localed and hostnamed mini-services which
1713 previously only provided support for changing time, locale
1714 and hostname settings from graphical DEs such as GNOME now
1715 also have a minimal (but very useful) text-based client
1716 utility each. This is probably the nicest way to changing
1717 these settings from the command line now, especially since
1718 it lists available options and is fully integrated with bash
1721 * There's now a new tool "systemd-coredumpctl" to list and
1722 extract coredumps from the journal.
1724 * We now install a README each in /var/log/ and
1725 /etc/rc.d/init.d explaining where the system logs and init
1726 scripts went. This hopefully should help folks who go to
1727 that dirs and look into the otherwise now empty void and
1728 scratch their heads.
1730 * When user-services are invoked (by systemd --user) the
1731 $MANAGERPID env var is set to the PID of systemd.
1733 * SIGRTMIN+24 when sent to a --user instance will now result
1734 in immediate termination of systemd.
1736 * gatewayd received numerous feature additions such as a
1737 "follow" mode, for live syncing and filtering.
1739 * browse.html now allows filtering and showing detailed
1740 information on specific entries. Keyboard navigation and
1741 mouse screen support has been added.
1743 * gatewayd/journalctl now supports HTML5/JSON
1744 Server-Sent-Events as output.
1746 * The SysV init script compatibility logic will now
1747 heuristically determine whether a script supports the
1748 "reload" verb, and only then make this available as
1751 * "systemctl status --follow" has been removed, use "journalctl
1754 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
1755 have been removed since they are hardly useful to be
1758 * And I'd like to take the opportunity to specifically mention
1759 Zbigniew for his great contributions. Zbigniew, you rock!
1761 Contributions from: Andrew Eikum, Christian Hesse, Colin
1762 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
1763 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
1764 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
1765 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
1766 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
1767 Jędrzejewski-Szmek, Сковорода Никита Андреевич
1771 * If /etc/vconsole.conf is non-existent or empty we will no
1772 longer load any console font or key map at boot by
1773 default. Instead the kernel defaults will be left
1774 intact. This is definitely the right thing to do, as no
1775 configuration should mean no configuration, and hard-coding
1776 font names that are different on all archs is probably a bad
1777 idea. Also, the kernel default key map and font should be
1778 good enough for most cases anyway, and mostly identical to
1779 the userspace fonts/key maps we previously overloaded them
1780 with. If distributions want to continue to default to a
1781 non-kernel font or key map they should ship a default
1782 /etc/vconsole.conf with the appropriate contents.
1784 Contributions from: Colin Walters, Daniel J Walsh, Dave
1785 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
1786 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1790 * journalctl gained a new --cursor= switch to show entries
1791 starting from the specified location in the journal.
1793 * We now enforce a size limit on journal entry fields exported
1794 with "-o json" in journalctl. Fields larger than 4K will be
1795 assigned null. This can be turned off with --all.
1797 * An (optional) journal gateway daemon is now available as
1798 "systemd-journal-gatewayd.service". This service provides
1799 access to the journal via HTTP and JSON. This functionality
1800 will be used to implement live log synchronization in both
1801 pull and push modes, but has various other users too, such
1802 as easy log access for debugging of embedded devices. Right
1803 now it is already useful to retrieve the journal via HTTP:
1805 # systemctl start systemd-journal-gatewayd.service
1806 # wget http://localhost:19531/entries
1808 This will download the journal contents in a
1809 /var/log/messages compatible format. The same as JSON:
1811 # curl -H"Accept: application/json" http://localhost:19531/entries
1813 This service is also accessible via a web browser where a
1814 single static HTML5 app is served that uses the JSON logic
1815 to enable the user to do some basic browsing of the
1816 journal. This will be extended later on. Here's an example
1817 screenshot of this app in its current state:
1819 http://0pointer.de/public/journal-gatewayd
1821 Contributions from: Kay Sievers, Lennart Poettering, Robert
1822 Milasan, Tom Gundersen
1826 * The bash completion logic is now available for journalctl
1829 * We don't mount the "cpuset" controller anymore together with
1830 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
1831 started if no parameters are assigned to it. "cpuset" hence
1832 broke code that assumed it it could create "cpu" groups and
1835 * journalctl -f will now subscribe to terminal size changes,
1836 and line break accordingly.
1838 Contributions from: Dave Reisner, Kay Sievers, Lennart
1839 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
1843 * nspawn will now create a symlink /etc/localtime in the
1844 container environment, copying the host's timezone
1845 setting. Previously this has been done via a bind mount, but
1846 since symlinks cannot be bind mounted this has now been
1847 changed to create/update the appropriate symlink.
1849 * journalctl -n's line number argument is now optional, and
1850 will default to 10 if omitted.
1852 * journald will now log the maximum size the journal files may
1853 take up on disk. This is particularly useful if the default
1854 built-in logic of determining this parameter from the file
1855 system size is used. Use "systemctl status
1856 systemd-journald.service" to see this information.
1858 * The multi-seat X wrapper tool has been stripped down. As X
1859 is now capable of enumerating graphics devices via udev in a
1860 seat-aware way the wrapper is not strictly necessary
1861 anymore. A stripped down temporary stop-gap is still shipped
1862 until the upstream display managers have been updated to
1863 fully support the new X logic. Expect this wrapper to be
1864 removed entirely in one of the next releases.
1866 * HandleSleepKey= in logind.conf has been split up into
1867 HandleSuspendKey= and HandleHibernateKey=. The old setting
1868 is not available anymore. X11 and the kernel are
1869 distuingishing between these keys and we should too. This
1870 also means the inhibition lock for these keys has been split
1873 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
1874 Poettering, Lukas Nykryn, Václav Pavlín
1878 * Whenever a unit changes state we'll now log this to the
1879 journal and show along the unit's own log output in
1882 * ConditionPathIsMountPoint= can now properly detect bind
1883 mount points too. (Previously, a bind mount of one file
1884 system to another place in the same file system could not be
1885 detected as mount, since they shared struct stat's st_dev
1888 * We will now mount the cgroup controllers cpu, cpuacct,
1889 cpuset and the controllers net_cls, net_prio together by
1892 * nspawn containers will now have a virtualized boot
1893 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
1894 over with a randomized ID at container initialization). This
1895 has the effect of making "journalctl -b" do the right thing
1898 * The JSON output journal serialization has been updated not
1899 to generate "endless" list objects anymore, but rather one
1900 JSON object per line. This is more in line how most JSON
1901 parsers expect JSON objects. The new output mode
1902 "json-pretty" has been added to provide similar output, but
1903 neatly aligned for readability by humans.
1905 * We dropped all explicit sync() invocations in the shutdown
1906 code. The kernel does this implicitly anyway in the kernel
1907 reboot() syscall. halt(8)'s -n option is now a compatibility
1910 * We now support virtualized reboot() in containers, as
1911 supported by newer kernels. We will fall back to exit() if
1912 CAP_SYS_REBOOT is not available to the container. Also,
1913 nspawn makes use of this now and will actually reboot the
1914 container if the containerized OS asks for that.
1916 * journalctl will only show local log output by default
1917 now. Use --merge (-m) to show remote log output, too.
1919 * libsystemd-journal gained the new sd_journal_get_usage()
1920 call to determine the current disk usage of all journal
1921 files. This is exposed in the new "journalctl --disk-usage"
1924 * journald gained a new configuration setting SplitMode= in
1925 journald.conf which may be used to control how user journals
1926 are split off. See journald.conf(5) for details.
1928 * A new condition type ConditionFileNotEmpty= has been added.
1930 * tmpfiles' "w" lines now support file globbing, to write
1931 multiple files at once.
1933 * We added Python bindings for the journal submission
1934 APIs. More Python APIs for a number of selected APIs will
1935 likely follow. Note that we intend to add native bindings
1936 only for the Python language, as we consider it common
1937 enough to deserve bindings shipped within systemd. There are
1938 various projects outside of systemd that provide bindings
1939 for languages such as PHP or Lua.
1941 * Many conditions will now resolve specifiers such as %i. In
1942 addition, PathChanged= and related directives of .path units
1943 now support specifiers as well.
1945 * There's now a new RPM macro definition for the system preset
1948 * journald will now warn if it can't forward a message to the
1949 syslog daemon because it's socket is full.
1951 * timedated will no longer write or process /etc/timezone,
1952 except on Debian. As we do not support late mounted /usr
1953 anymore /etc/localtime always being a symlink is now safe,
1954 and hence the information in /etc/timezone is not necessary
1957 * logind will now always reserve one VT for a text getty (VT6
1958 by default). Previously if more than 6 X sessions where
1959 started they took up all the VTs with auto-spawned gettys,
1960 so that no text gettys were available anymore.
1962 * udev will now automatically inform the btrfs kernel logic
1963 about btrfs RAID components showing up. This should make
1964 simple hotplug based btrfs RAID assembly work.
1966 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1967 (but not for its children which will stay at the kernel
1968 default). This should allow setups with a lot more listening
1971 * systemd will now always pass the configured timezone to the
1972 kernel at boot. timedated will do the same when the timezone
1975 * logind's inhibition logic has been updated. By default,
1976 logind will now handle the lid switch, the power and sleep
1977 keys all the time, even in graphical sessions. If DEs want
1978 to handle these events on their own they should take the new
1979 handle-power-key, handle-sleep-key and handle-lid-switch
1980 inhibitors during their runtime. A simple way to achiveve
1981 that is to invoke the DE wrapped in an invocation of:
1983 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1985 * Access to unit operations is now checked via SELinux taking
1986 the unit file label and client process label into account.
1988 * systemd will now notify the administrator in the journal
1989 when he over-mounts a non-empty directory.
1991 * There are new specifiers that are resolved in unit files,
1992 for the host name (%H), the machine ID (%m) and the boot ID
1995 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
1996 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
1997 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
1998 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
1999 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2000 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2001 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2005 * Support for reading structured kernel messages from
2006 /dev/kmsg has now been added and is enabled by default.
2008 * Support for reading kernel messages from /proc/kmsg has now
2009 been removed. If you want kernel messages in the journal
2010 make sure to run a recent kernel (>= 3.5) that supports
2011 reading structured messages from /dev/kmsg (see
2012 above). /proc/kmsg is now exclusive property of classic
2013 syslog daemons again.
2015 * The libudev API gained the new
2016 udev_device_new_from_device_id() call.
2018 * The logic for file system namespace (ReadOnlyDirectory=,
2019 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2020 require pivot_root() anymore. This means fewer temporary
2021 directories are created below /tmp for this feature.
2023 * nspawn containers will now see and receive all submounts
2024 made on the host OS below the root file system of the
2027 * Forward Secure Sealing is now supported for Journal files,
2028 which provide cryptographical sealing of journal files so
2029 that attackers cannot alter log history anymore without this
2030 being detectable. Lennart will soon post a blog story about
2031 this explaining it in more detail.
2033 * There are two new service settings RestartPreventExitStatus=
2034 and SuccessExitStatus= which allow configuration of exit
2035 status (exit code or signal) which will be excepted from the
2036 restart logic, resp. consider successful.
2038 * journalctl gained the new --verify switch that can be used
2039 to check the integrity of the structure of journal files and
2040 (if Forward Secure Sealing is enabled) the contents of
2043 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2044 and similar symlinks pre-created. This makes running shells
2045 as container init process a lot more fun.
2047 * The fstab support can now handle PARTUUID= and PARTLABEL=
2050 * A new ConditionHost= condition has been added to match
2051 against the hostname (with globs) and machine ID. This is
2052 useful for clusters where a single OS image is used to
2053 provision a large number of hosts which shall run slightly
2054 different sets of services.
2056 * Services which hit the restart limit will now be placed in a
2059 Contributions from: Bertram Poettering, Dave Reisner, Huang
2060 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2061 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2065 * When running in --user mode systemd will now become a
2066 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2067 tree a lot more organized.
2069 * A new PartOf= unit dependency type has been introduced that
2070 may be used to group services in a natural way.
2072 * "systemctl enable" may now be used to enable instances of
2075 * journalctl now prints error log levels in red, and
2076 warning/notice log levels in bright white. It also supports
2077 filtering by log level now.
2079 * cgtop gained a new -n switch (similar to top), to configure
2080 the maximum number of iterations to run for. It also gained
2081 -b, to run in batch mode (accepting no input).
2083 * The suffix ".service" may now be omitted on most systemctl
2084 command lines involving service unit names.
2086 * There's a new bus call in logind to lock all sessions, as
2087 well as a loginctl verb for it "lock-sessions".
2089 * libsystemd-logind.so gained a new call sd_journal_perror()
2090 that works similar to libc perror() but logs to the journal
2091 and encodes structured information about the error number.
2093 * /etc/crypttab entries now understand the new keyfile-size=
2096 * shutdown(8) now can send a (configurable) wall message when
2097 a shutdown is cancelled.
2099 * The mount propagation mode for the root file system will now
2100 default to "shared", which is useful to make containers work
2101 nicely out-of-the-box so that they receive new mounts from
2102 the host. This can be undone locally by running "mount
2103 --make-rprivate /" if needed.
2105 * The prefdm.service file has been removed. Distributions
2106 should maintain this unit downstream if they intend to keep
2107 it around. However, we recommend writing normal unit files
2108 for display managers instead.
2110 * Since systemd is a crucial part of the OS we will now
2111 default to a number of compiler switches that improve
2112 security (hardening) such as read-only relocations, stack
2113 protection, and suchlike.
2115 * The TimeoutSec= setting for services is now split into
2116 TimeoutStartSec= and TimeoutStopSec= to allow configuration
2117 of individual time outs for the start and the stop phase of
2120 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
2121 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
2122 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
2123 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
2124 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
2125 Gundersen, Zbigniew Jędrzejewski-Szmek
2129 * The journal and id128 C APIs are now fully documented as man
2132 * Extra safety checks have been added when transitioning from
2133 the initial RAM disk to the main system to avoid accidental
2136 * /etc/crypttab entries now understand the new keyfile-offset=
2139 * systemctl -t can now be used to filter by unit load state.
2141 * The journal C API gained the new sd_journal_wait() call to
2142 make writing synchronous journal clients easier.
2144 * journalctl gained the new -D switch to show journals from a
2147 * journalctl now displays a special marker between log
2148 messages of two different boots.
2150 * The journal is now explicitly flushed to /var via a service
2151 systemd-journal-flush.service, rather than implicitly simply
2152 by seeing /var/log/journal to be writable.
2154 * journalctl (and the journal C APIs) can now match for much
2155 more complex expressions, with alternatives and
2158 * When transitioning from the initial RAM disk to the main
2159 system we will now kill all processes in a killing spree to
2160 ensure no processes stay around by accident.
2162 * Three new specifiers may be used in unit files: %u, %h, %s
2163 resolve to the user name, user home directory resp. user
2164 shell. This is useful for running systemd user instances.
2166 * We now automatically rotate journal files if their data
2167 object hash table gets a fill level > 75%. We also size the
2168 hash table based on the configured maximum file size. This
2169 together should lower hash collisions drastically and thus
2170 speed things up a bit.
2172 * journalctl gained the new "--header" switch to introspect
2173 header data of journal files.
2175 * A new setting SystemCallFilters= has been added to services
2176 which may be used to apply blacklists or whitelists to
2177 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
2179 * nspawn gained a new --link-journal= switch (and quicker: -j)
2180 to link the container journal with the host. This makes it
2181 very easy to centralize log viewing on the host for all
2182 guests while still keeping the journal files separated.
2184 * Many bugfixes and optimizations
2186 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
2187 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
2188 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
2193 * Several tools now understand kernel command line arguments,
2194 which are only read when run in an initial RAM disk. They
2195 usually follow closely their normal counterparts, but are
2198 * There's a new tool to analyze the readahead files that are
2199 automatically generated at boot. Use:
2201 /usr/lib/systemd/systemd-readahead analyze /.readahead
2203 * We now provide an early debug shell on tty9 if this enabled. Use:
2205 systemctl enable debug-shell.service
2207 * All plymouth related units have been moved into the Plymouth
2208 package. Please make sure to upgrade your Plymouth version
2211 * systemd-tmpfiles now supports getting passed the basename of
2212 a configuration file only, in which case it will look for it
2213 in all appropriate directories automatically.
2215 * udevadm info now takes a /dev or /sys path as argument, and
2216 does the right thing. Example:
2218 udevadm info /dev/sda
2219 udevadm info /sys/class/block/sda
2221 * systemctl now prints a warning if a unit is stopped but a
2222 unit that might trigger it continues to run. Example: a
2223 service is stopped but the socket that activates it is left
2226 * "systemctl status" will now mention if the log output was
2227 shortened due to rotation since a service has been started.
2229 * The journal API now exposes functions to determine the
2230 "cutoff" times due to rotation.
2232 * journald now understands SIGUSR1 and SIGUSR2 for triggering
2233 immediately flushing of runtime logs to /var if possible,
2234 resp. for triggering immediate rotation of the journal
2237 * It is now considered an error if a service is attempted to
2238 be stopped that is not loaded.
2240 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
2242 * systemd-analyze now supports Python 3
2244 * tmpfiles now supports cleaning up directories via aging
2245 where the first level dirs are always kept around but
2246 directories beneath it automatically aged. This is enabled
2247 by prefixing the age field with '~'.
2249 * Seat objects now expose CanGraphical, CanTTY properties
2250 which is required to deal with very fast bootups where the
2251 display manager might be running before the graphics drivers
2252 completed initialization.
2254 * Seat objects now expose a State property.
2256 * We now include RPM macros for service enabling/disabling
2257 based on the preset logic. We recommend RPM based
2258 distributions to make use of these macros if possible. This
2259 makes it simpler to reuse RPM spec files across
2262 * We now make sure that the collected systemd unit name is
2263 always valid when services log to the journal via
2266 * There's a new man page kernel-command-line(7) detailing all
2267 command line options we understand.
2269 * The fstab generator may now be disabled at boot by passing
2270 fstab=0 on the kernel command line.
2272 * A new kernel command line option modules-load= is now understood
2273 to load a specific kernel module statically, early at boot.
2275 * Unit names specified on the systemctl command line are now
2276 automatically escaped as needed. Also, if file system or
2277 device paths are specified they are automatically turned
2278 into the appropriate mount or device unit names. Example:
2280 systemctl status /home
2281 systemctl status /dev/sda
2283 * The SysVConsole= configuration option has been removed from
2284 system.conf parsing.
2286 * The SysV search path is no longer exported on the D-Bus
2289 * The Names= option is been removed from unit file parsing.
2291 * There's a new man page bootup(7) detailing the boot process.
2293 * Every unit and every generator we ship with systemd now
2294 comes with full documentation. The self-explanatory boot is
2297 * A couple of services gained "systemd-" prefixes in their
2298 name if they wrap systemd code, rather than only external
2299 code. Among them fsck@.service which is now
2300 systemd-fsck@.service.
2302 * The HaveWatchdog property has been removed from the D-Bus
2305 * systemd.confirm_spawn= on the kernel command line should now
2308 * There's a new man page crypttab(5) which details all options
2309 we actually understand.
2311 * systemd-nspawn gained a new --capability= switch to pass
2312 additional capabilities to the container.
2314 * timedated will now read known NTP implementation unit names
2315 from /usr/lib/systemd/ntp-units.d/*.list,
2316 systemd-timedated-ntp.target has been removed.
2318 * journalctl gained a new switch "-b" that lists log data of
2319 the current boot only.
2321 * The notify socket is in the abstract namespace again, in
2322 order to support daemons which chroot() at start-up.
2324 * There is a new Storage= configuration option for journald
2325 which allows configuration of where log data should go. This
2326 also provides a way to disable journal logging entirely, so
2327 that data collected is only forwarded to the console, the
2328 kernel log buffer or another syslog implementation.
2330 * Many bugfixes and optimizations
2332 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
2333 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
2334 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
2335 Shawn Landden, Tom Gundersen
2339 * "systemctl help <unit>" now shows the man page if one is
2342 * Several new man pages have been added.
2344 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
2345 MaxLevelConsole= can now be specified in
2346 journald.conf. These options allow reducing the amount of
2347 data stored on disk or forwarded by the log level.
2349 * TimerSlackNSec= can now be specified in system.conf for
2350 PID1. This allows system-wide power savings.
2352 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
2353 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
2358 * logind is now capable of (optionally) handling power and
2359 sleep keys as well as the lid switch.
2361 * journalctl now understands the syntax "journalctl
2362 /usr/bin/avahi-daemon" to get all log output of a specific
2365 * CapabilityBoundingSet= in system.conf now also influences
2366 the capability bound set of usermode helpers of the kernel.
2368 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
2369 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
2370 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
2371 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
2375 * Note that we skipped 139 releases here in order to set the
2376 new version to something that is greater than both udev's
2377 and systemd's most recent version number.
2379 * udev: all udev sources are merged into the systemd source tree now.
2380 All future udev development will happen in the systemd tree. It
2381 is still fully supported to use the udev daemon and tools without
2382 systemd running, like in initramfs or other init systems. Building
2383 udev though, will require the *build* of the systemd tree, but
2384 udev can be properly *run* without systemd.
2386 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
2387 should be used to create dead device nodes as workarounds for broken
2390 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
2391 no longer supported. udev_monitor_new_from_netlink() needs to be
2392 used to subscribe to events.
2394 * udev: when udevd is started by systemd, processes which are left
2395 behind by forking them off of udev rules, are unconditionally cleaned
2396 up and killed now after the event handling has finished. Services or
2397 daemons must be started as systemd services. Services can be
2398 pulled-in by udev to get started, but they can no longer be directly
2399 forked by udev rules.
2401 * udev: the daemon binary is called systemd-udevd now and installed
2402 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
2403 to adapt to that, create symlink, or rename the binary after building
2406 * libudev no longer provides these symbols:
2407 udev_monitor_from_socket()
2408 udev_queue_get_failed_list_entry()
2409 udev_get_{dev,sys,run}_path()
2410 The versions number was bumped and symbol versioning introduced.
2412 * systemd-loginctl and systemd-journalctl have been renamed
2413 to loginctl and journalctl to match systemctl.
2415 * The config files: /etc/systemd/systemd-logind.conf and
2416 /etc/systemd/systemd-journald.conf have been renamed to
2417 logind.conf and journald.conf. Package updates should rename
2418 the files to the new names on upgrade.
2420 * For almost all files the license is now LGPL2.1+, changed
2421 from the previous GPL2.0+. Exceptions are some minor stuff
2422 of udev (which will be changed to LGPL2.1 eventually, too),
2423 and the MIT licensed sd-daemon.[ch] library that is suitable
2424 to be used as drop-in files.
2426 * systemd and logind now handle system sleep states, in
2427 particular suspending and hibernating.
2429 * logind now implements a sleep/shutdown/idle inhibiting logic
2430 suitable for a variety of uses. Soonishly Lennart will blog
2431 about this in more detail.
2433 * var-run.mount and var-lock.mount are no longer provided
2434 (which prevously bind mounted these directories to their new
2435 places). Distributions which have not converted these
2436 directories to symlinks should consider stealing these files
2437 from git history and add them downstream.
2439 * We introduced the Documentation= field for units and added
2440 this to all our shipped units. This is useful to make it
2441 easier to explore the boot and the purpose of the various
2444 * All smaller setup units (such as
2445 systemd-vconsole-setup.service) now detect properly if they
2446 are run in a container and are skipped when
2447 appropriate. This guarantees an entirely noise-free boot in
2448 Linux container environments such as systemd-nspawn.
2450 * A framework for implementing offline system updates is now
2451 integrated, for details see:
2452 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
2454 * A new service type Type=idle is available now which helps us
2455 avoiding ugly interleaving of getty output and boot status
2458 * There's now a system-wide CapabilityBoundingSet= option to
2459 globally reduce the set of capabilities for the
2460 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
2461 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
2462 even CAP_NET_ADMIN system-wide for secure systems.
2464 * There are now system-wide DefaultLimitXXX= options to
2465 globally change the defaults of the various resource limits
2466 for all units started by PID 1.
2468 * Harald Hoyer's systemd test suite has been integrated into
2469 systemd which allows easy testing of systemd builds in qemu
2470 and nspawn. (This is really awesome! Ask us for details!)
2472 * The fstab parser is now implemented as generator, not inside
2475 * systemctl will now warn you if .mount units generated from
2476 /etc/fstab are out of date due to changes in fstab that
2477 haven't been read by systemd yet.
2479 * systemd is now suitable for usage in initrds. Dracut has
2480 already been updated to make use of this. With this in place
2481 initrds get a slight bit faster but primarily are much
2482 easier to introspect and debug since "systemctl status" in
2483 the host system can be used to introspect initrd services,
2484 and the journal from the initrd is kept around too.
2486 * systemd-delta has been added, a tool to explore differences
2487 between user/admin configuration and vendor defaults.
2489 * PrivateTmp= now affects both /tmp and /var/tmp.
2491 * Boot time status messages are now much prettier and feature
2492 proper english language. Booting up systemd has never been
2495 * Read-ahead pack files now include the inode number of all
2496 files to pre-cache. When the inode changes the pre-caching
2497 is not attempted. This should be nicer to deal with updated
2498 packages which might result in changes of read-ahead
2501 * We now temporaritly lower the kernel's read_ahead_kb variable
2502 when collecting read-ahead data to ensure the kernel's
2503 built-in read-ahead does not add noise to our measurements
2504 of necessary blocks to pre-cache.
2506 * There's now RequiresMountsFor= to add automatic dependencies
2507 for all mounts necessary for a specific file system path.
2509 * MountAuto= and SwapAuto= have been removed from
2510 system.conf. Mounting file systems at boot has to take place
2513 * nspawn now learned a new switch --uuid= to set the machine
2514 ID on the command line.
2516 * nspawn now learned the -b switch to automatically search
2519 * vt102 is now the default TERM for serial TTYs, upgraded from
2522 * systemd-logind now works on VT-less systems.
2524 * The build tree has been reorganized. The individual
2525 components now have directories of their own.
2527 * A new condition type ConditionPathIsReadWrite= is now available.
2529 * nspawn learned the new -C switch to create cgroups for the
2530 container in other hierarchies.
2532 * We now have support for hardware watchdogs, configurable in
2535 * The scheduled shutdown logic now has a public API.
2537 * We now mount /tmp as tmpfs by default, but this can be
2538 masked and /etc/fstab can override it.
2540 * Since udisks doesn't make use of /media anymore we are not
2541 mounting a tmpfs on it anymore.
2543 * journalctl gained a new --local switch to only interleave
2544 locally generated journal files.
2546 * We can now load the IMA policy at boot automatically.
2548 * The GTK tools have been split off into a systemd-ui.
2550 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
2551 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
2552 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
2553 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
2554 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
2555 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
2556 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
2557 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
2558 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
2563 * This is mostly a bugfix release
2565 * Support optional initialization of the machine ID from the
2566 KVM or container configured UUID.
2568 * Support immediate reboots with "systemctl reboot -ff"
2570 * Show /etc/os-release data in systemd-analyze output
2572 * Many bugfixes for the journal, including endianness fixes and
2573 ensuring that disk space enforcement works
2575 * sd-login.h is C++ comptaible again
2577 * Extend the /etc/os-release format on request of the Debian
2580 * We now refuse non-UTF8 strings used in various configuration
2581 and unit files. This is done to ensure we don't pass invalid
2582 data over D-Bus or expose it elsewhere.
2584 * Register Mimo USB Screens as suitable for automatic seat
2587 * Read SELinux client context from journal clients in a race
2590 * Reorder configuration file lookup order. /etc now always
2591 overrides /run in order to allow the administrator to always
2592 and unconditionally override vendor supplied or
2593 automatically generated data.
2595 * The various user visible bits of the journal now have man
2596 pages. We still lack man pages for the journal API calls
2599 * We now ship all man pages in HTML format again in the
2602 Contributions from: Dave Reisner, Dirk Eibach, Frederic
2603 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
2604 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
2609 * This is mostly a bugfix release
2611 * systems lacking /etc/os-release are no longer supported.
2613 * Various functionality updates to libsystemd-login.so
2615 * Track class of PAM logins to distuingish greeters from
2618 Contributions from: Kay Sievers, Lennart Poettering, Michael
2623 * This is an important bugfix release for v41.
2625 * Building man pages is now optional which should be useful
2626 for those building systemd from git but unwilling to install
2629 * Watchdog support for supervising services is now usable. In
2630 a future release support for hardware watchdogs
2631 (i.e. /dev/watchdog) will be added building on this.
2633 * Service start rate limiting is now configurable and can be
2634 turned off per service. When a start rate limit is hit a
2635 reboot can automatically be triggered.
2637 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
2639 Contributions from: Benjamin Franzke, Bill Nottingham,
2640 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
2641 Schmidt, Michał Górny, Piotr Drąg
2645 * The systemd binary is installed /usr/lib/systemd/systemd now;
2646 An existing /sbin/init symlink needs to be adapted with the
2649 * The code that loads kernel modules has been ported to invoke
2650 libkmod directly, instead of modprobe. This means we do not
2651 support systems with module-init-tools anymore.
2653 * Watchdog support is now already useful, but still not
2656 * A new kernel command line option systemd.setenv= is
2657 understood to set system wide environment variables
2658 dynamically at boot.
2660 * We now limit the set of capabilities of systemd-journald.
2662 * We now set SIGPIPE to ignore by default, since it only is
2663 useful in shell pipelines, and has little use in general
2664 code. This can be disabled with IgnoreSIPIPE=no in unit
2667 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
2668 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
2673 * This is mostly a bugfix release
2675 * We now expose the reason why a service failed in the
2676 "Result" D-Bus property.
2678 * Rudimentary service watchdog support (will be completed over
2679 the next few releases.)
2681 * When systemd forks off in order execute some service we will
2682 now immediately changes its argv[0] to reflect which process
2683 it will execute. This is useful to minimize the time window
2684 with a generic argv[0], which makes bootcharts more useful
2686 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
2687 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
2688 Mike Kazantsev, Ray Strode
2692 * This is mostly a test release, but incorporates many
2695 * New systemd-cgtop tool to show control groups by their
2698 * Linking against libacl for ACLs is optional again. If
2699 disabled, support tracking device access for active logins
2700 goes becomes unavailable, and so does access to the user
2701 journals by the respective users.
2703 * If a group "adm" exists, journal files are automatically
2704 owned by them, thus allow members of this group full access
2705 to the system journal as well as all user journals.
2707 * The journal now stores the SELinux context of the logging
2708 client for all entries.
2710 * Add C++ inclusion guards to all public headers
2712 * New output mode "cat" in the journal to print only text
2713 messages, without any meta data like date or time.
2715 * Include tiny X server wrapper as a temporary stop-gap to
2716 teach XOrg udev display enumeration. This is used by display
2717 managers such as gdm, and will go away as soon as XOrg
2718 learned native udev hotplugging for display devices.
2720 * Add new systemd-cat tool for executing arbitrary programs
2721 with STDERR/STDOUT connected to the journal. Can also act as
2722 BSD logger replacement, and does so by default.
2724 * Optionally store all locally generated coredumps in the
2725 journal along with meta data.
2727 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
2728 writing short strings to files (for usage for /sys), and for
2729 creating symlinks, character and block device nodes.
2731 * New unit file option ControlGroupPersistent= to make cgroups
2732 persistent, following the mechanisms outlined in
2733 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
2735 * Support multiple local RTCs in a sane way
2737 * No longer monopolize IO when replaying readahead data on
2738 rotating disks, since we might starve non-file-system IO to
2739 death, since fanotify() will not see accesses done by blkid,
2742 * Don't show kernel threads in systemd-cgls anymore, unless
2743 requested with new -k switch.
2745 Contributions from: Dan Horák, Kay Sievers, Lennart
2746 Poettering, Michal Schmidt
2750 * This is mostly a test release, but incorporates many
2753 * The git repository moved to:
2754 git://anongit.freedesktop.org/systemd/systemd
2755 ssh://git.freedesktop.org/git/systemd/systemd
2757 * First release with the journal
2758 http://0pointer.de/blog/projects/the-journal.html
2760 * The journal replaces both systemd-kmsg-syslogd and
2761 systemd-stdout-bridge.
2763 * New sd_pid_get_unit() API call in libsystemd-logind
2765 * Many systemadm clean-ups
2767 * Introduce remote-fs-pre.target which is ordered before all
2768 remote mounts and may be used to start services before all
2771 * Added Mageia support
2773 * Add bash completion for systemd-loginctl
2775 * Actively monitor PID file creation for daemons which exit in
2776 the parent process before having finished writing the PID
2777 file in the daemon process. Daemons which do this need to be
2778 fixed (i.e. PID file creation must have finished before the
2779 parent exits), but we now react a bit more gracefully to them.
2781 * Add colourful boot output, mimicking the well-known output
2782 of existing distributions.
2784 * New option PassCredentials= for socket units, for
2785 compatibility with a recent kernel ABI breakage.
2787 * /etc/rc.local is now hooked in via a generator binary, and
2788 thus will no longer act as synchronization point during
2791 * systemctl list-unit-files now supports --root=.
2793 * systemd-tmpfiles now understands two new commands: z, Z for
2794 relabelling files according to the SELinux database. This is
2795 useful to apply SELinux labels to specific files in /sys,
2798 * Output of SysV services is now forwarded to both the console
2799 and the journal by default, not only just the console.
2801 * New man pages for all APIs from libsystemd-login.
2803 * The build tree got reorganized and a the build system is a
2804 lot more modular allowing embedded setups to specifically
2805 select the components of systemd they are interested in.
2807 * Support for Linux systems lacking the kernel VT subsystem is
2810 * configure's --with-rootdir= got renamed to
2811 --with-rootprefix= to follow the naming used by udev and
2814 * Unless specified otherwise we'll now install to /usr instead
2815 of /usr/local by default.
2817 * Processes with '@' in argv[0][0] are now excluded from the
2818 final shut-down killing spree, following the logic explained
2820 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
2822 * All processes remaining in a service cgroup when we enter
2823 the START or START_PRE states are now killed with
2824 SIGKILL. That means it is no longer possible to spawn
2825 background processes from ExecStart= lines (which was never
2826 supported anyway, and bad style).
2828 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
2829 reloading of units together.
2831 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
2832 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
2833 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
2834 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
2835 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek