1 systemd System and Service Manager
3 CHANGES WITH 239 in spe:
5 * NETWORK INTERFACE DEVICE NAMING CHANGES: elogind-udevd's "net_id"
6 builtin may name network interfaces differently than in previous
7 versions. SR-IOV virtual functions and NPAR partitions with PCI
8 function numbers of 8 and above will be named more predictably, and
9 udev may generate names based on the PCI slot number in some cases
10 where it previously did not.
12 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
13 elogind.service. Since v235, IPAddressDeny=any has been set to
14 the unit. So, it is expected that the default behavior of elogind
15 is not changed. However, if distribution packagers or administrators
16 disabled or modified IPAddressDeny= setting by a drop-in config file,
17 then it may be necessary to update the file to re-enable AF_INET and
18 AF_INET6 to support network user name services, e.g. NIS.
20 * When the RestrictNamespaces= unit property is specified multiple times,
21 then the specified types are merged now. Previously, only the last
22 assignment was used. So, if distribution packagers or administrators
23 modified the setting by a drop-in config file, then it may be necessary
26 * When OnFailure= is used in combination with Restart= on a service
27 unit, then the specified units will no longer be triggered on
28 failures that result in restarting. Previously, the specified units
29 would be activated each time the unit failed, even when the unit was
30 going to be restarted automatically. This behaviour contradicted the
31 documentation. With this release the code is adjusted to match the
34 * elogind-tmpfiles will now print a notice whenever it encounters
35 tmpfiles.d/ lines referencing the /var/run/ directory. It will
36 recommend reworking them to use the /run/ directory instead (for
37 which /var/run/ is simply a symlinked compatibility alias). This way
38 elogind-tmpfiles can properly detect line conflicts and merge lines
39 referencing the same file by two paths, without having to access
42 * systemctl disable/unmask/preset/preset-all cannot be used with
43 --runtime. Previously this was allowed, but resulted in unintuitive
44 behaviour that wasn't useful. systemctl disable/unmask will now
45 undo both runtime and persistent enablement/masking, i.e. it will
46 remove any relevant symlinks both in /run and /etc.
48 * sd-boot acquired new loader configuration settings to optionally turn
49 off Windows and MacOS boot partition discovery as well as
50 reboot-into-firmware menu items. It is also able to pick a better
51 screen resolution for HiDPI systems, and now provides loader
52 configuration settings to change the resolution explicitly.
54 * The elogind-resolve tool has been renamed to resolvectl (it also
55 remains available under the old name, for compatibility), and its
56 interface is now verb-based, similar in style to the other <xyz>ctl
57 tools, such as systemctl or loginctl.
59 * The resolvectl/elogind-resolve tool also provides 'resolveconf'
60 compatibility. It may be symlinked under the 'resolveconf' name, in
61 which case it will take arguments and input compatible with the
62 Debian and FreeBSD resolvconf tool.
64 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
65 where the system initially suspends, and after a time-out resumes and
68 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
69 set the client will only send a DUID as client identifier.
71 * The nss-elogind glibc NSS module will now enumerate dynamic users and
72 groups in effect. Previously, it could resolve UIDs/GIDs to user
73 names/groups and vice versa, but did not support enumeration.
75 * journald's Compress= configuration setting now optionally accepts a
76 byte threshold value. All journal objects larger than this threshold
77 will be compressed, smaller ones will not. Previously this threshold
78 was not configurable and set to 512.
80 * A new system.conf setting NoNewPrivileges= is now available which may
81 be used to turn off acquisition of new privileges system-wide
82 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
83 for all its children). Note that turning this option on means setuid
84 binaries and file system capabilities lose their special powers.
85 While turning on this option is a big step towards a more secure
86 system, doing so is likely to break numerous pre-existing UNIX tools,
87 in particular su and sudo.
89 * A new service elogind-time-sync-wait.service has been added. If
90 enabled it will delay the time-sync.target unit at boot until time
91 synchronization has been received from the network. This
92 functionality is useful on systems lacking a local RTC or where it is
93 acceptable that the boot process shall be delayed by external network
96 * When hibernating, elogind will now inform the kernel of the image
97 write offset, on kernels new enough to support this. This means swap
98 files should work for hibernation now.
100 * When loading unit files, elogind will now look for drop-in unit files
101 extensions in additional places. Previously, for a unit file name
102 "foo-bar-baz.service" it would look for dropin files in
103 "foo-bar-baz.service.d/*.conf". Now, it will also look in
104 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
105 service name truncated after all inner dashes. This scheme allows
106 writing drop-ins easily that apply to a whole set of unit files at
107 once. It's particularly useful for mount and slice units (as their
108 naming is prefix based), but is also useful for service and other
109 units, for packages that install multiple unit files at once,
110 following a strict naming regime of beginning the unit file name with
111 the package's name. Two new specifiers are now supported in unit
112 files to match this: %j and %J are replaced by the part of the unit
113 name following the last dash.
115 * Unit files and other configuration files that support specifier
116 expansion now understand another two new specifiers: %T and %V will
117 resolve to /tmp and /var/tmp respectively, or whatever temporary
118 directory has been set for the calling user.
120 * The ExecStart= lines of unit files are no longer required to
121 reference absolute paths. If non-absolute paths are specified the
122 specified binary name is searched within the service manager's
123 built-in $PATH, which may be queried with 'elogind-path
124 search-binaries-default'. It's generally recommended to continue to
125 use absolute paths for all binaries specified in unit files.
127 * Units gained a new load state "bad-setting", which is used when a
128 unit file was loaded, but contained fatal errors which prevent it
129 from being started (for example, an ExecStart= path which references
130 a non-existent executable).
132 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
133 support alternative debuggers, for example lldb. The old name
134 continues to be available however, for compatibility reasons. Use the
135 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
136 to pick an alternative debugger instead of the default gdb.
138 * systemctl and the other tools will now output escape sequences that
139 generate proper clickable hyperlinks in various terminal emulators
140 where useful (for example, in the "systemctl status" output you can
141 now click on the unit file name to quickly open it in the
142 editor/viewer of your choice). Note that not all terminal emulators
143 support this functionality yet, but many do. Unfortunately, the
144 "less" pager doesn't support this yet, hence this functionality is
145 currently automatically turned off when a pager is started (which
146 happens quite often due to auto-paging). We hope to remove this
147 limitation as soon as "less" learns these escape sequences. This new
148 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
149 environment variable. For details on these escape sequences see:
150 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
152 * networkd's .network files now support a new IPv6MTUBytes= option for
153 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
154 option in the [Route] section to configure the MTU to use for
155 specific routes. It also gained support for configuration of the DHCP
156 "UserClass" option through the new UserClass= setting. It gained
157 three new options in the new [CAN] section for configuring CAN
158 networks. The MULTICAST and ALLMULTI interface flags may now be
159 controlled explicitly with the new Multicast= and AllMulticast=
162 * networkd will now automatically make use of the kernel's route
163 expiration feature, if it is available.
165 * udevd's .link files now support setting the number of receive and
166 transmit channels, using the RxChannels=, TxChannels=,
167 OtherChannels=, CombinedChannels= settings.
169 * Support for UDPSegmentationOffload= has been removed, given its
170 limited support in hardware, and waning software support.
172 * networkd's .netdev files now support creating "netdevsim" interfaces.
174 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
175 to query the unit belonging to a specific kernel control group.
177 * elogind-analyze gained a new verb "cat-config", which may be used to
178 dump the contents of any configuration file, with all its matching
179 drop-in files added in, and honouring the usual search and masking
180 logic applied to elogind configuration files. For example use
181 "elogind-analyze cat-config elogind/system.conf" to get the complete
182 system configuration file of elogind how it would be loaded by PID 1
183 itself. Similar to this, various tools such as elogind-tmpfiles or
184 elogind-sysusers, gained a new option "--cat-config", which does the
185 corresponding operation for their own configuration settings. For
186 example, "elogind-tmpfiles --cat-config" will now output the full
187 list of tmpfiles.d/ lines in place.
189 * timedatectl gained two new verbs "timesync-status" (to show the
190 current NTP synchronization state of elogind-timesyncd) and
191 "show-timesync" (to show bus properties of elogind-timesyncd).
193 * elogind-timesyncd gained a bus interface on which it exposes details
196 * elogind-nspawn gained a new --rlimit= switch for setting initial
197 resource limits for the container payload. There's a new switch
198 --hostname= to explicitly override the container's hostname. A new
199 --no-new-privileges= switch may be used to control the
200 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
201 --oom-score-adjust= switch controls the OOM scoring adjustment value
202 for the payload. The new --cpu-affinity= switch controls the CPU
203 affinity of the container payload. The new --resolv-conf= switch
204 allows more detailed control of /etc/resolv.conf handling of the
205 container. Similarly, the new --timezone= switch allows more detailed
206 control of /etc/localtime handling of the container.
208 * elogind-detect-virt gained a new --list switch, which will print a
209 list of all currently known VM and container environments.
211 * Support for "Portable Services" has been added, see
212 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
213 experimental, but this is expected to change soon. Reflecting this
214 experimental state, the "portablectl" binary is not installed into
215 /usr/bin yet. The binary has to be called with the full path
216 /usr/lib/elogind/portablectl instead.
218 * journalctl's and systemctl's -o switch now knows a new log output
219 mode "with-unit". The output it generates is very similar to the
220 regular "short" mode, but displays the unit name instead of the
221 syslog tag for each log line. Also, the date is shown with timezone
222 information. This mode is probably more useful than the classic
223 "short" output mode for most purposes, except where pixel-perfect
224 compatibility with classic /var/log/messages formatting is required.
226 * A new --dump-bus-properties switch has been added to the elogind
227 binary, which may be used to dump all supported D-Bus properties.
228 (Options which are still supported, but are deprecated, are *not*
231 * sd-bus gained a set of new calls:
232 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
233 enable/disable the "floating" state of a bus slot object,
234 i.e. whether the slot object pins the bus it is allocated for into
235 memory or if the bus slot object gets disconnected when the bus goes
236 away. sd_bus_open_with_description(),
237 sd_bus_open_user_with_description(),
238 sd_bus_open_system_with_description() may be used to allocate bus
239 objects and set their description string already during allocation.
241 * sd-event gained support for watching inotify events from the event
242 loop, in an efficient way, sharing inotify handles between multiple
243 users. For this a new function sd_event_add_inotify() has been added.
245 * sd-event and sd-bus gained support for calling special user-supplied
246 destructor functions for userdata pointers associated with
247 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
248 functions sd_bus_slot_set_destroy_callback,
249 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
250 sd_bus_track_get_destroy_callback,
251 sd_event_source_set_destroy_callback,
252 sd_event_source_get_destroy_callback have been added.
254 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
256 * PID 1 will now automatically reschedule .timer units whenever the
257 local timezone changes. (They previously got rescheduled
258 automatically when the system clock changed.)
260 * New documentation has been added to document cgroups delegation,
261 portable services and the various code quality tools we have set up:
263 https://github.com/systemd/systemd/blob/master/doc/CGROUP_DELEGATION.md
264 https://github.com/systemd/systemd/blob/master/doc/PORTABLE_SERVICES.md
265 https://github.com/systemd/systemd/blob/master/doc/CODE_QUALITY.md
267 * pam_elogind will now look for PAM userdata fields elogind.memory_max,
268 elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
269 earlier PAM modules. The data in these fields is used to initialize
270 the session scope's resource properties. Thus external PAM modules
271 may now configure per-session limits, for example sourced from
272 external user databases.
274 * socket units with Accept=yes will now maintain a "refused" counter in
275 addition to the existing "accepted" counter, counting connections
276 refused due to the enforced limits.
278 * The "elogind-path search-binaries-default" command may now be use to
279 query the default, built-in $PATH PID 1 will pass to the services it
282 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
283 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Antique, Arnaud
284 Rebillout, Brian J. Murrell, Bruno Vernay, Chris Lesiak, Christian
285 Brauner, Christian Hesse, Daniel Dao, Daniel Lin, Danylo Korostil,
286 Davide Cavalca, David Tardon, Dimitri John Ledkov, Dmitriy Geels,
287 Douglas Christman, Elia Geretto, emelenas, Evegeny Vereshchagin, Evgeny
288 Vereshchagin, Felipe Sateler, Feng Sun, Filipe Brandenburger, Franck
289 Bui, futpib, Giuseppe Scrivano, Guillem Jover, guixxx, Hans de Goede,
290 Henrique Dante de Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko,
291 Ivan Shapovalov, James Cowgill, Jan Janssen, Jan Synacek, Jared
292 Kazimir, João Paulo Rechi Vita, Joost Heitbrink, juergbi, Jui-Chi Ricky
293 Liang, Kai-Heng Feng, Karol Augustin, Krzysztof Nowicki, Lauri
294 Tirkkonen, Lennart Poettering, Leonard, Long Li, Luca Boccassi, Lucas
295 Werkmeister, Marcel Hoppe, Marc Kleine-Budde, Mario Limonciello, Martin
296 Jansa, Martin Wilck, Mathieu Malaterre, Matteo F. Vescovi, Michael
297 Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal Sekletar,
298 Mike Gilbert, Mikhail Kasimov, Milan, Milan Broz, mourikwa, Muhammet
299 Kara, Nicolas Boichat, Omer Katz, Paride Legovini, Paul Menzel, Paul
300 Milliken, Peter A. Bigot, Peter Hutterer, Peter Jones, Philip Sequeira,
301 Philip Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de
302 Araujo, Rosen Penev, rubensa, Ryan Gonzalez, Salvo 'LtWorf' Tomaselli,
303 Sebastian Reichel, Sergio Lindo Mansilla, Stefan Schweter, Stephen
304 Hemminger, Stuart Hayes, Susant Sahani, Sylvain Plantefève, Thomas
305 H. P. Andersen, Tobias Jungel, Tomasz Torcz, Vito Caputo, Will Dietz,
306 Will Thompson, xginn8, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
312 * The MemoryAccounting= unit property now defaults to on. After
313 discussions with the upstream control group maintainers we learnt
314 that the negative impact of cgroup memory accounting on current
315 kernels is finally relatively minimal, so that it should be safe to
316 enable this by default without affecting system performance. Besides
317 memory accounting only task accounting is turned on by default, all
318 other forms of resource accounting (CPU, IO, IP) remain off for now,
319 because it's not clear yet that their impact is small enough to move
320 from opt-in to opt-out. We recommend downstreams to leave memory
321 accounting on by default if kernel 4.14 or higher is primarily
322 used. On very resource constrained systems or when support for old
323 kernels is a necessity, -Dmemory-accounting-default=false can be used
324 to revert this change.
326 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
327 %udev_rules_update) and the journal catalog (%journal_catalog_update)
328 from the upgrade scriptlets of individual packages now do nothing.
329 Transfiletriggers have been added which will perform those updates
330 once at the end of the transaction.
332 Similar transfiletriggers have been added to execute any sysctl.d
333 and binfmt.d rules. Thus, it should be unnecessary to provide any
334 scriptlets to execute this configuration from package installation
337 * elogind-sysusers gained a mode where the configuration to execute is
338 specified on the command line, but this configuration is not executed
339 directly, but instead it is merged with the configuration on disk,
340 and the result is executed. This is useful for package installation
341 scripts which want to create the user before installing any files on
342 disk (in case some of those files are owned by that user), while
343 still allowing local admin overrides.
345 This functionality is exposed to rpm scriptlets through a new
346 %sysusers_create_package macro. Old %sysusers_create and
347 %sysusers_create_inline macros are deprecated.
349 A transfiletrigger for sysusers.d configuration is now installed,
350 which means that it should be unnecessary to call elogind-sysusers from
351 package installation scripts, unless the package installs any files
352 owned by those newly-created users, in which case
353 %sysusers_create_package should be used.
355 * Analogous change has been done for elogind-tmpfiles: it gained a mode
356 where the command-line configuration is merged with the configuration
357 on disk. This is exposed as the new %tmpfiles_create_package macro,
358 and %tmpfiles_create is deprecated. A transfiletrigger is installed
359 for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
360 from package installation scripts.
362 * sysusers.d configuration for a user may now also specify the group
363 number, in addition to the user number ("u username 123:456"), or
364 without the user number ("u username -:456").
366 * Configution items for elogind-sysusers can now be specified as
367 positional arguments when the new --inline switch is used.
369 * The login shell of users created through sysusers.d may now be
370 specified (previously, it was always /bin/sh for root and
371 /sbin/nologin for other users).
373 * elogind-analyze gained a new --global switch to look at global user
374 configuration. It also gained a unit-paths verb to list the unit load
375 paths that are compiled into elogind (which can be used with
376 --elogind, --user, or --global).
378 * udevadm trigger gained a new --settle/-w option to wait for any
379 triggered events to finish (but just those, and not any other events
380 which are triggered meanwhile).
382 * The action that elogind takes when the lid is closed and the
383 machine is connected to external power can now be configured using
384 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
385 was determined by HandleLidSwitch=, and, for backwards compatibility,
386 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
388 * journalctl will periodically call sd_journal_process() to make it
389 resilient against inotify queue overruns when journal files are
390 rotated very quickly.
392 * Two new functions in libelogind — sd_bus_get_n_queued_read and
393 sd_bus_get_n_queued_write — may be used to check the number of
394 pending bus messages.
396 * elogind gained a new
397 org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
398 which can be used to migrate foreign processes to scope and service
399 units. The primary user for this new API is elogind itself: the
400 elogind --user instance uses this call of the elogind --system
401 instance to migrate processes if it itself gets the request to
402 migrate processes and the kernel refuses this due to access
403 restrictions. Thanks to this "elogind-run --scope --user …" works
404 again in pure cgroups v2 environments when invoked from the user
407 * A new TemporaryFileSystem= setting can be used to mask out part of
408 the real file system tree with tmpfs mounts. This may be combined
409 with BindPaths= and BindReadOnlyPaths= to hide files or directories
410 not relevant to the unit, while still allowing some paths lower in
411 the tree to be accessed.
413 ProtectHome=tmpfs may now be used to hide user home and runtime
414 directories from units, in a way that is mostly equivalent to
415 "TemporaryFileSystem=/home /run/user /root".
417 * Non-service units are now started with KeyringMode=shared by default.
418 This means that mount and swapon and other mount tools have access
419 to keys in the main keyring.
421 * /sys/fs/bpf is now mounted automatically.
423 * QNX virtualization is now detected by elogind-detect-virt and may
424 be used in ConditionVirtualization=.
426 * IPAccounting= may now be enabled also for slice units.
428 * A new -Dsplit-bin= build configuration switch may be used to specify
429 whether bin and sbin directories are merged, or if they should be
430 included separately in $PATH and various listings of executable
431 directories. The build configuration scripts will try to autodetect
432 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
433 system, but distributions are encouraged to configure this
436 * A new -Dok-color= build configuration switch may be used to change
437 the colour of "OK" status messages.
439 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
440 PrivateNetwork=yes was buggy in previous versions of elogind. This
441 means that after the upgrade and daemon-reexec, any such units must
444 * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
445 will not exclude read-only files owned by root from cleanup.
447 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
448 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
449 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
450 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
451 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
452 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
453 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
454 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
455 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
456 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
457 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
458 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
459 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
460 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
461 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
462 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
468 * Some keyboards come with a zoom see-saw or rocker which until now got
469 mapped to the Linux "zoomin/out" keys in hwdb. However, these
470 keycodes are not recognized by any major desktop. They now produce
471 Up/Down key events so that they can be used for scrolling.
473 * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
474 slightly: previously, if an argument was specified for lines of this
475 type (i.e. the right-most column was set) this string was appended to
476 existing files each time elogind-tmpfiles was run. This behaviour was
477 different from what the documentation said, and not particularly
478 useful, as repeated elogind-tmpfiles invocations would not be
479 idempotent and grow such files without bounds. With this release
480 behaviour has been altered slightly, to match what the documentation
481 says: lines of this type only have an effect if the indicated files
482 don't exist yet, and only then the argument string is written to the
485 * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
486 elogind-tmpfiles behaviour: previously, read-only files owned by root
487 were always excluded from the file "aging" algorithm (i.e. the
488 automatic clean-up of directories like /tmp based on
489 atime/mtime/ctime). We intend to drop this restriction, and age files
490 by default even when owned by root and read-only. This behaviour was
491 inherited from older tools, but there have been requests to remove
492 it, and it's not obvious why this restriction was made in the first
493 place. Please speak up now, if you are aware of software that reqires
494 this behaviour, otherwise we'll remove the restriction in v238.
496 * A new environment variable $SYSTEMD_OFFLINE is now understood by
497 systemctl. It takes a boolean argument. If on, systemctl assumes it
498 operates on an "offline" OS tree, and will not attempt to talk to the
499 service manager. Previously, this mode was implicitly enabled if a
500 chroot() environment was detected, and this new environment variable
501 now provides explicit control.
503 * .path and .socket units may now be created transiently, too.
504 Previously only service, mount, automount and timer units were
505 supported as transient units. The elogind-run tool has been updated
506 to expose this new functionality, you may hence use it now to bind
507 arbitrary commands to path or socket activation on-the-fly from the
508 command line. Moreover, almost all properties are now exposed for the
509 unit types that already supported transient operation.
511 * The elogind-mount command gained support for a new --owner= parameter
512 which takes a user name, which is then resolved and included in uid=
513 and gid= mount options string of the file system to mount.
515 * A new unit condition ConditionControlGroupController= has been added
516 that checks whether a specific cgroup controller is available.
518 * Unit files, udev's .link files, and elogind-networkd's .netdev and
519 .network files all gained support for a new condition
520 ConditionKernelVersion= for checking against specific kernel
523 * In elogind-networkd, the [IPVLAN] section in .netdev files gained
524 support for configuring device flags in the Flags= setting. In the
525 same files, the [Tunnel] section gained support for configuring
526 AllowLocalRemote=. The [Route] section in .network files gained
527 support for configuring InitialCongestionWindow=,
528 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
529 understands RapidCommit=.
531 * elogind-networkd's DHCPv6 support gained support for Prefix
534 * sd-bus gained support for a new "watch-bind" feature. When this
535 feature is enabled, an sd_bus connection may be set up to connect to
536 an AF_UNIX socket in the file system as soon as it is created. This
537 functionality is useful for writing early-boot services that
538 automatically connect to the system bus as soon as it is started,
539 without ugly time-based polling. elogind-networkd and
540 elogind-resolved have been updated to make use of this
541 functionality. busctl exposes this functionality in a new
542 --watch-bind= command line switch.
544 * sd-bus will now optionally synthesize a local "Connected" signal as
545 soon as a D-Bus connection is set up fully. This message mirrors the
546 already existing "Disconnected" signal which is synthesized when the
547 connection is terminated. This signal is generally useful but
548 particularly handy in combination with the "watch-bind" feature
549 described above. Synthesizing of this message has to be requested
550 explicitly through the new API call sd_bus_set_connected_signal(). In
551 addition a new call sd_bus_is_ready() has been added that checks
552 whether a connection is fully set up (i.e. between the "Connected" and
553 "Disconnected" signals).
555 * sd-bus gained two new calls sd_bus_request_name_async() and
556 sd_bus_release_name_async() for asynchronously registering bus
557 names. Similar, there is now sd_bus_add_match_async() for installing
558 a signal match asynchronously. All of elogind's own services have
559 been updated to make use of these calls. Doing these operations
560 asynchronously has two benefits: it reduces the risk of deadlocks in
561 case of cyclic dependencies between bus services, and it speeds up
562 service initialization since synchronization points for bus
563 round-trips are removed.
565 * sd-bus gained two new calls sd_bus_match_signal() and
566 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
567 and sd_bus_add_match_async() but instead of taking a D-Bus match
568 string take match fields as normal function parameters.
570 * sd-bus gained two new calls sd_bus_set_sender() and
571 sd_bus_message_set_sender() for setting the sender name of outgoing
572 messages (either for all outgoing messages or for just one specific
573 one). These calls are only useful in direct connections as on
574 brokered connections the broker fills in the sender anyway,
575 overwriting whatever the client filled in.
577 * sd-event gained a new pseudo-handle that may be specified on all API
578 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
579 used this refers to the default event loop object of the calling
580 thread. Note however that this does not implicitly allocate one —
581 which has to be done prior by using sd_event_default(). Similarly
582 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
583 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
584 to the default bus of the specified type of the calling thread. Here
585 too this does not implicitly allocate bus connection objects, this
586 has to be done prior with sd_bus_default() and friends.
588 * sd-event gained a new call pair
589 sd_event_source_{get|set}_io_fd_own(). This may be used to request
590 automatic closure of the file descriptor an IO event source watches
591 when the event source is destroyed.
593 * elogind-networkd gained support for natively configuring WireGuard
596 * In previous versions elogind synthesized user records both for the
597 "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
598 internally. In order to simplify distribution-wide renames of the
599 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
600 new transitional flag file has been added: if
601 /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
602 user and group record within the elogind codebase is disabled.
604 * elogind-notify gained a new --uid= option for selecting the source
605 user/UID to use for notification messages sent to the service
608 * journalctl gained a new --grep= option to list only entries in which
609 the message matches a certain pattern. By default matching is case
610 insensitive if the pattern is lowercase, and case sensitive
611 otherwise. Option --case-sensitive=yes|no can be used to override
612 this an specify case sensitivity or case insensitivity.
614 * There's now a "elogind-analyze service-watchdogs" command for printing
615 the current state of the service runtime watchdog, and optionally
616 enabling or disabling the per-service watchdogs system-wide if given a
617 boolean argument (i.e. the concept you configure in WatchdogSec=), for
618 debugging purposes. There's also a kernel command line option
619 elogind.service_watchdogs= for controlling the same.
621 * Two new "log-level" and "log-target" options for elogind-analyze were
622 added that merge the now deprecated get-log-level, set-log-level and
623 get-log-target, set-log-target pairs. The deprecated options are still
624 understood for backwards compatibility. The two new options print the
625 current value when no arguments are given, and set them when a
626 level/target is given as an argument.
628 * sysusers.d's "u" lines now optionally accept both a UID and a GID
629 specification, separated by a ":" character, in order to create users
630 where UID and GID do not match.
632 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
633 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
634 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
635 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
636 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
637 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
638 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
639 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
640 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
641 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
642 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
643 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
644 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
645 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
646 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
647 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
648 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
649 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
650 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
651 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
658 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
659 in v235 has been extended to also set the dummy.ko module option
660 numdummies=0, preventing the kernel from automatically creating
661 dummy0. All dummy interfaces must now be explicitly created.
663 * Unknown '%' specifiers in configuration files are now rejected. This
664 applies to units and tmpfiles.d configuration. Any percent characters
665 that are followed by a letter or digit that are not supposed to be
666 interpreted as the beginning of a specifier should be escaped by
667 doubling ("%%"). (So "size=5%" is still accepted, as well as
668 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
669 valid specifiers today.)
671 * systemd-resolved now maintains a new dynamic
672 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
673 recommended to make /etc/resolv.conf a symlink to it. This file
674 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
675 includes dynamically acquired search domains, achieving more correct
676 DNS resolution by software that bypasses local DNS APIs such as NSS.
678 * The "uaccess" udev tag has been dropped from /dev/kvm and
679 /dev/dri/renderD*. These devices now have the 0666 permissions by
680 default (but this may be changed at build-time). /dev/dri/renderD*
681 will now be owned by the "render" group along with /dev/kfd.
683 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
684 systemd-journal-gatewayd.service and
685 systemd-journal-upload.service. This means "nss-systemd" must be
686 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
687 services are resolved properly.
689 * In /etc/fstab two new mount options are now understood:
690 x-systemd.makefs and x-systemd.growfs. The former has the effect that
691 the configured file system is formatted before it is mounted, the
692 latter that the file system is resized to the full block device size
693 after it is mounted (i.e. if the file system is smaller than the
694 partition it resides on, it's grown). This is similar to the fsck
695 logic in /etc/fstab, and pulls in systemd-makefs@.service and
696 systemd-growfs@.service as necessary, similar to
697 systemd-fsck@.service. Resizing is currently only supported on ext4
700 * In systemd-networkd, the IPv6 RA logic now optionally may announce
701 DNS server and domain information.
703 * Support for the LUKS2 on-disk format for encrypted partitions has
704 been added. This requires libcryptsetup2 during compilation and
707 * The systemd --user instance will now signal "readiness" when its
708 basic.target unit has been reached, instead of when the run queue ran
709 empty for the first time.
711 * Tmpfiles.d with user configuration are now also supported.
712 systemd-tmpfiles gained a new --user switch, and snippets placed in
713 ~/.config/user-tmpfiles.d/ and corresponding directories will be
714 executed by systemd-tmpfiles --user running in the new
715 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
716 running in the user session.
718 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
719 %S resolves to the top-level state directory (/var/lib for the system
720 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
721 top-level cache directory (/var/cache for the system instance,
722 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
723 logs directory (/var/log for the system instance,
724 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
725 existing %t specifier, that resolves to the top-level runtime
726 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
729 * journalctl learnt a new parameter --output-fields= for limiting the
730 set of journal fields to output in verbose and JSON output modes.
732 * systemd-timesyncd's configuration file gained a new option
733 RootDistanceMaxSec= for setting the maximum root distance of servers
734 it'll use, as well as the new options PollIntervalMinSec= and
735 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
737 * bootctl gained a new command "list" for listing all available boot
738 menu items on systems that follow the boot loader specification.
740 * systemctl gained a new --dry-run switch that shows what would be done
741 instead of doing it, and is currently supported by the shutdown and
744 * ConditionSecurity= can now detect the TOMOYO security module.
746 * Unit file [Install] sections are now also respected in unit drop-in
747 files. This is intended to be used by drop-ins under /usr/lib/.
749 * systemd-firstboot may now also set the initial keyboard mapping.
751 * Udev "changed" events for devices which are exposed as systemd
752 .device units are now propagated to units specified in
753 ReloadPropagatedFrom= as reload requests.
755 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
756 unit template name (i.e. a name in the form of 'foobar@.service',
757 without the instance component between the '@' and - the '.'), then
758 the escaped sysfs path of the device is automatically used as the
761 * SystemCallFilter= in unit files has been extended so that an "errno"
762 can be specified individually for each system call. Example:
763 SystemCallFilter=~uname:EILSEQ.
765 * The cgroup delegation logic has been substantially updated. Delegate=
766 now optionally takes a list of controllers (instead of a boolean, as
767 before), which lists the controllers to delegate at least.
769 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
771 * A new LogLevelMax= setting configures the maximum log level any
772 process of the service may log at (i.e. anything with a lesser
773 priority than what is specified is automatically dropped). A new
774 LogExtraFields= setting allows configuration of additional journal
775 fields to attach to all log records generated by any of the unit's
778 * New StandardInputData= and StandardInputText= settings along with the
779 new option StandardInput=data may be used to configure textual or
780 binary data that shall be passed to the executed service process via
781 standard input, encoded in-line in the unit file.
783 * StandardInput=, StandardOutput= and StandardError= may now be used to
784 connect stdin/stdout/stderr of executed processes directly with a
785 file or AF_UNIX socket in the file system, using the new "file:" option.
787 * A new unit file option CollectMode= has been added, that allows
788 tweaking the garbage collection logic for units. It may be used to
789 tell systemd to garbage collect units that have failed automatically
790 (normally it only GCs units that exited successfully). systemd-run
791 and systemd-mount expose this new functionality with a new -G option.
793 * "machinectl bind" may now be used to bind mount non-directories
794 (i.e. regularfiles, devices, fifos, sockets).
796 * systemd-analyze gained a new verb "calendar" for validating and
797 testing calendar time specifications to use for OnCalendar= in timer
798 units. Besides validating the expression it will calculate the next
799 time the specified expression would elapse.
801 * In addition to the pre-existing FailureAction= unit file setting
802 there's now SuccessAction=, for configuring a shutdown action to
803 execute when a unit completes successfully. This is useful in
804 particular inside containers that shall terminate after some workload
805 has been completed. Also, both options are now supported for all unit
806 types, not just services.
808 * networkds's IP rule support gained two new options
809 IncomingInterface= and OutgoingInterface= for configuring the incoming
810 and outgoing interfaces of configured rules. systemd-networkd also
811 gained support for "vxcan" network devices.
813 * networkd gained a new setting RequiredForOnline=, taking a
814 boolean. If set, systemd-wait-online will take it into consideration
815 when determining that the system is up, otherwise it will ignore the
816 interface for this purpose.
818 * The sd_notify() protocol gained support for a new operation: with
819 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
820 store again, ahead of POLLHUP or POLLERR when they are removed
823 * A new document doc/UIDS-GIDS.md has been added to the source tree,
824 that documents the UID/GID range and assignment assumptions and
825 requirements of systemd.
827 * The watchdog device PID 1 will ping may now be configured through the
828 WatchdogDevice= configuration file setting, or by setting the
829 systemd.watchdog_service= kernel commandline option.
831 * systemd-resolved's gained support for registering DNS-SD services on
832 the local network using MulticastDNS. Services may either be
833 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
834 the same dir below /run, /usr/lib), or through its D-Bus API.
836 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
837 extend the effective start, runtime, and stop time. The service must
838 continue to send EXTEND_TIMEOUT_USEC within the period specified to
839 prevent the service manager from making the service as timedout.
841 * elogind-resolved's DNSSEC support gained support for RFC 8080
842 (Ed25519 keys and signatures).
844 * The elogind-resolve command line tool gained a new set of options
845 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
846 --set-nta= and --revert to configure per-interface DNS configuration
847 dynamically during runtime. It's useful for pushing DNS information
848 into elogind-resolved from DNS hook scripts that various interface
849 managing software supports (such as pppd).
851 * elogind-nspawn gained a new --network-namespace-path= command line
852 option, which may be used to make a container join an existing
853 network namespace, by specifying a path to a "netns" file.
855 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
856 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
857 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
858 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
859 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
860 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
861 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
862 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
863 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
864 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
865 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
866 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
867 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
868 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
869 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
870 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
871 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
872 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
873 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
874 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
875 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
876 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
877 Jędrzejewski-Szmek, Zeal Jagannatha
883 * INCOMPATIBILITY: systemd-logind.service and other long-running
884 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
885 communication with the outside. This generally improves security of
886 the system, and is in almost all cases a safe and good choice, as
887 these services do not and should not provide any network-facing
888 functionality. However, systemd-logind uses the glibc NSS API to
889 query the user database. This creates problems on systems where NSS
890 is set up to directly consult network services for user database
891 lookups. In particular, this creates incompatibilities with the
892 "nss-nis" module, which attempts to directly contact the NIS/YP
893 network servers it is configured for, and will now consistently
894 fail. In such cases, it is possible to turn off IP sandboxing for
895 systemd-logind.service (set IPAddressDeny= in its [Service] section
896 to the empty string, via a .d/ unit file drop-in). Downstream
897 distributions might want to update their nss-nis packaging to include
898 such a drop-in snippet, accordingly, to hide this incompatibility
899 from the user. Another option is to make use of glibc's nscd service
900 to proxy such network requests through a privilege-separated, minimal
901 local caching daemon, or to switch to more modern technologies such
902 sssd, whose NSS hook-ups generally do not involve direct network
903 access. In general, we think it's definitely time to question the
904 implementation choices of nss-nis, i.e. whether it's a good idea
905 today to embed a network-facing loadable module into all local
906 processes that need to query the user database, including the most
907 trivial and benign ones, such as "ls". For more details about
908 IPAddressDeny= see below.
910 * A new modprobe.d drop-in is now shipped by default that sets the
911 bonding module option max_bonds=0. This overrides the kernel default,
912 to avoid conflicts and ambiguity as to whether or not bond0 should be
913 managed by systemd-networkd or not. This resolves multiple issues
914 with bond0 properties not being applied, when bond0 is configured
915 with systemd-networkd. Distributors may choose to not package this,
916 however in that case users will be prevented from correctly managing
917 bond0 interface using systemd-networkd.
919 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
920 which print the logging level and target of the system manager. They
921 complement the existing "set-log-level" and "set-log-target" verbs
922 used to change those values.
924 * journald.conf gained a new boolean setting ReadKMsg= which defaults
925 to on. If turned off kernel log messages will not be read by
926 systemd-journald or included in the logs. It also gained a new
927 setting LineMax= for configuring the maximum line length in
928 STDOUT/STDERR log streams. The new default for this value is 48K, up
929 from the previous hardcoded 2048.
931 * A new unit setting RuntimeDirectoryPreserve= has been added, which
932 allows more detailed control of what to do with a runtime directory
933 configured with RuntimeDirectory= (i.e. a directory below /run or
934 $XDG_RUNTIME_DIR) after a unit is stopped.
936 * The RuntimeDirectory= setting for units gained support for creating
937 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
938 one top-level directory.
940 * Units gained new options StateDirectory=, CacheDirectory=,
941 LogsDirectory= and ConfigurationDirectory= which are closely related
942 to RuntimeDirectory= but manage per-service directories below
943 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
944 possible to write unit files which when activated automatically gain
945 properly owned service specific directories in these locations, thus
946 making unit files self-contained and increasing compatibility with
947 stateless systems and factory reset where /etc or /var are
948 unpopulated at boot. Matching these new settings there's also
949 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
950 ConfigurationDirectoryMode= for configuring the access mode of these
951 directories. These settings are particularly useful in combination
952 with DynamicUser=yes as they provide secure, properly-owned,
953 writable, and stateful locations for storage, excluded from the
954 sandbox that such services live in otherwise.
956 * Automake support has been removed from this release. systemd is now
959 * systemd-journald will now aggressively cache client metadata during
960 runtime, speeding up log write performance under pressure. This comes
961 at a small price though: as much of the metadata is read
962 asynchronously from /proc/ (and isn't implicitly attached to log
963 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
964 metadata stored alongside a log entry might be slightly
965 out-of-date. Previously it could only be slightly newer than the log
966 message. The time window is small however, and given that the kernel
967 is unlikely to be improved anytime soon in this regard, this appears
970 * nss-myhostname/systemd-resolved will now by default synthesize an
971 A/AAAA resource record for the "_gateway" hostname, pointing to the
972 current default IP gateway. Previously it did that for the "gateway"
973 name, hampering adoption, as some distributions wanted to leave that
974 host name open for local use. The old behaviour may still be
975 requested at build time.
977 * systemd-networkd's [Address] section in .network files gained a new
978 Scope= setting for configuring the IP address scope. The [Network]
979 section gained a new boolean setting ConfigureWithoutCarrier= that
980 tells systemd-networkd to ignore link sensing when configuring the
981 device. The [DHCP] section gained a new Anonymize= boolean option for
982 turning on a number of options suggested in RFC 7844. A new
983 [RoutingPolicyRule] section has been added for configuring the IP
984 routing policy. The [Route] section has gained support for a new
985 Type= setting which permits configuring
986 blackhole/unreachable/prohibit routes.
988 * The [VRF] section in .netdev files gained a new Table= setting for
989 configuring the routing table to use. The [Tunnel] section gained a
990 new Independent= boolean field for configuring tunnels independent of
991 an underlying network interface. The [Bridge] section gained a new
992 GroupForwardMask= option for configuration of propagation of link
993 local frames between bridge ports.
995 * The WakeOnLan= setting in .link files gained support for a number of
996 new modes. A new TCP6SegmentationOffload= setting has been added for
997 configuring TCP/IPv6 hardware segmentation offload.
999 * The IPv6 RA sender implementation may now optionally send out RDNSS
1000 and RDNSSL records to supply DNS configuration to peers.
1002 * systemd-nspawn gained support for a new --system-call-filter= command
1003 line option for adding and removing entries in the default system
1004 call filter it applies. Moreover systemd-nspawn has been changed to
1005 implement a system call whitelist instead of a blacklist.
1007 * systemd-run gained support for a new --pipe command line option. If
1008 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1009 are directly passed on to the activated transient service
1010 executable. This allows invoking arbitrary processes as systemd
1011 services (for example to take benefit of dependency management,
1012 accounting management, resource management or log management that is
1013 done automatically for services) — while still allowing them to be
1014 integrated in a classic UNIX shell pipeline.
1016 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1017 using ReloadPropagationTo= is configured, a reload is now propagated
1018 to configured units. (Previously this was only done on explicitly
1019 requested reloads, using "systemctl reload" or an equivalent
1022 * For each service unit a restart counter is now kept: it is increased
1023 each time the service is restarted due to Restart=, and may be
1024 queried using "systemctl show -p NRestarts …".
1026 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1027 @signal and @timer have been added, for usage with SystemCallFilter=
1028 in unit files and the new --system-call-filter= command line option
1029 of systemd-nspawn (see above).
1031 * ExecStart= lines in unit files gained two new modifiers: when a
1032 command line is prefixed with "!" the command will be executed as
1033 configured, except for the credentials applied by
1034 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1035 "+", but does still apply namespacing options unlike "+". There's
1036 also "!!" now, which is mostly identical, but becomes a NOP on
1037 systems that support ambient capabilities. This is useful to write
1038 unit files that work with ambient capabilities where possible but
1039 automatically fall back to traditional privilege dropping mechanisms
1040 on systems where this is not supported.
1042 * ListenNetlink= settings in socket units now support RDMA netlink
1045 * A new unit file setting LockPersonality= has been added which permits
1046 locking down the chosen execution domain ("personality") of a service
1049 * A new special target "getty-pre.target" has been added, which is
1050 ordered before all text logins, and may be used to order services
1051 before textual logins acquire access to the console.
1053 * systemd will now attempt to load the virtio-rng.ko kernel module very
1054 early on if a VM environment supporting this is detected. This should
1055 improve entropy during early boot in virtualized environments.
1057 * A _netdev option is now supported in /etc/crypttab that operates in a
1058 similar way as the same option in /etc/fstab: it permits configuring
1059 encrypted devices that need to be ordered after the network is up.
1060 Following this logic, two new special targets
1061 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1062 added that are to cryptsetup.target what remote-fs.target and
1063 remote-fs-pre.target are to local-fs.target.
1065 * Service units gained a new UnsetEnvironment= setting which permits
1066 unsetting specific environment variables for services that are
1067 normally passed to it (for example in order to mask out locale
1068 settings for specific services that can't deal with it).
1070 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1071 traffic accounting (packet count as well as byte count) is done for
1072 the service, and shown as part of "systemctl status" or "systemd-run
1075 * Service units acquired two new options IPAddressAllow= and
1076 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1077 for configuring a simple IP access control list for all sockets of
1078 the unit. These options are available also on .slice and .socket
1079 units, permitting flexible access list configuration for individual
1080 services as well as groups of services (as defined by a slice unit),
1081 including system-wide. Note that IP ACLs configured this way are
1082 enforced on every single IPv4 and IPv6 socket created by any process
1083 of the service unit, and apply to ingress as well as egress traffic.
1085 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
1086 structured log message is generated each time the unit is stopped,
1087 containing information about the consumed resources of this
1090 * A new setting KeyringMode= has been added to unit files, which may be
1091 used to control how the kernel keyring is set up for executed
1094 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1095 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1096 behaviour (that is: these commands return immediately after the
1097 operation was enqueued instead of waiting for the operation to
1098 complete). Previously, "systemctl poweroff" and "systemctl reboot"
1099 were asynchronous on systems using systemd-logind (i.e. almost
1100 always, and like they were on sysvinit), and the other three commands
1101 were unconditionally synchronous. With this release this is cleaned
1102 up, and callers will see the same asynchronous behaviour on all
1103 systems for all five operations.
1105 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1108 * .timer units now accept calendar specifications in other timezones
1109 than UTC or the local timezone.
1111 * The tmpfiles snippet var.conf has been changed to create
1112 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1113 the "utmp" group already, and it appears to be generally understood
1114 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1115 databases. Previously this was implemented correctly for all these
1116 databases excepts btmp, which has been opened up like this now
1117 too. Note that while the other databases are world-readable
1118 (i.e. 0644), btmp is not and remains more restrictive.
1120 * The systemd-resolve tool gained a new --reset-server-features
1121 switch. When invoked like this systemd-resolved will forget
1122 everything it learnt about the features supported by the configured
1123 upstream DNS servers, and restarts the feature probing logic on the
1124 next resolver look-up for them at the highest feature level
1127 * The status dump systemd-resolved sends to the logs upon receiving
1128 SIGUSR1 now also includes information about all DNS servers it is
1129 configured to use, and the features levels it probed for them.
1131 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1132 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
1133 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1134 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1135 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1136 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1137 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1138 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1139 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1140 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1141 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1142 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1143 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1144 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1145 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1146 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1147 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1148 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1149 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1150 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1152 — Berlin, 2017-10-06
1156 * Meson is now supported as build system in addition to Automake. It is
1157 our plan to remove Automake in one of our next releases, so that
1158 Meson becomes our exclusive build system. Hence, please start using
1159 the Meson build system in your downstream packaging. There's plenty
1160 of documentation around how to use Meson, the extremely brief
1163 ./autogen.sh && ./configure && make && sudo make install
1167 meson build && ninja -C build && sudo ninja -C build install
1169 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1170 which permits configuring a timeout on the time a job is
1171 running. This is particularly useful for setting timeouts on jobs for
1174 * Unit files gained two new options ConditionUser= and ConditionGroup=
1175 for conditionalizing units based on the identity of the user/group
1176 running a systemd user instance.
1178 * systemd-networkd now understands a new FlowLabel= setting in the
1179 [VXLAN] section of .network files, as well as a Priority= in
1180 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1181 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1182 gained support for configuration of GENEVE links, and IPv6 address
1183 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1185 * .link files now understand a new Port= setting.
1187 * systemd-networkd's DHCP support gained support for DHCP option 119
1188 (domain search list).
1190 * systemd-networkd gained support for serving IPv6 address ranges using
1191 the Router Advertisement protocol. The new .network configuration
1192 section [IPv6Prefix] may be used to configure the ranges to
1193 serve. This is implemented based on a new, minimal, native server
1194 implementation of RA.
1196 * journalctl's --output= switch gained support for a new parameter
1197 "short-iso-precise" for a mode where timestamps are shown as precise
1200 * systemd-udevd's "net_id" builtin may now generate stable network
1201 interface names from IBM PowerVM VIO devices as well as ACPI platform
1204 * MulticastDNS support in systemd-resolved may now be explicitly
1205 enabled/disabled using the new MulticastDNS= configuration file
1208 * systemd-resolved may now optionally use libidn2 instead of the libidn
1209 for processing internationalized domain names. Support for libidn2
1210 should be considered experimental and should not be enabled by
1213 * "machinectl pull-tar" and related call may now do verification of
1214 downloaded images using SUSE-style .sha256 checksum files in addition
1215 to the already existing support for validating using Ubuntu-style
1218 * sd-bus gained support for a new sd_bus_message_appendv() call which
1219 is va_list equivalent of sd_bus_message_append().
1221 * sd-boot gained support for validating images using SHIM/MOK.
1223 * The SMACK code learnt support for "onlycap".
1225 * systemd-mount --umount is now much smarter in figuring out how to
1226 properly unmount a device given its mount or device path.
1228 * The code to call libnss_dns as a fallback from libnss_resolve when
1229 the communication with systemd-resolved fails was removed. This
1230 fallback was redundant and interfered with the [!UNAVAIL=return]
1231 suffix. See nss-resolve(8) for the recommended configuration.
1233 * systemd-logind may now be restarted without losing state. It stores
1234 the file descriptors for devices it manages in the system manager
1235 using the FDSTORE= mechanism. Please note that further changes in
1236 other components may be required to make use of this (for example
1237 Xorg has code to listen for stops of systemd-logind and terminate
1238 itself when logind is stopped or restarted, in order to avoid using
1239 stale file descriptors for graphical devices, which is now
1240 counterproductive and must be reverted in order for restarts of
1241 systemd-logind to be safe. See
1242 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1244 * All kernel install plugins are called with the environment variable
1245 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1246 /etc/machine-id. If the file is missing or empty, the variable is
1247 empty and BOOT_DIR_ABS is the path of a temporary directory which is
1248 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
1249 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1251 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
1252 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1253 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1254 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1255 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
1256 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1257 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1258 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1259 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
1260 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1261 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1262 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1263 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1264 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1265 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1266 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1267 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1268 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1269 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1270 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1271 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1272 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1273 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1274 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1275 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
1276 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1277 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
1278 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1279 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1282 — Berlin, 2017-07-12
1286 * The "hybrid" control group mode has been modified to improve
1287 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1288 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1289 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1290 cgroups-v1 hierarchy), the only externally visible change being that
1291 the cgroups-v2 hierarchy is also mounted, to
1292 /sys/fs/cgroup/unified. This should provide a large degree of
1293 compatibility with "legacy" cgroups-v1, while taking benefit of the
1294 better management capabilities of cgroups-v2.
1296 * The default control group setup mode may be selected both a boot-time
1297 via a set of kernel command line parameters (specifically:
1298 systemd.unified_cgroup_hierarchy= and
1299 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1300 default selected on the configure command line
1301 (--with-default-hierarchy=). The upstream default is "hybrid"
1302 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1303 this will change in a future systemd version to be "unified" (pure
1304 cgroups-v2 mode). The third option for the compile time option is
1305 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1306 distributions to default to "hybrid" mode for release distributions,
1307 starting with v233. We recommend "unified" for development
1308 distributions (specifically: distributions such as Fedora's rawhide)
1309 as that's where things are headed in the long run. Use "legacy" for
1310 greatest stability and compatibility only.
1312 * Note one current limitation of "unified" and "hybrid" control group
1313 setup modes: the kernel currently does not permit the systemd --user
1314 instance (i.e. unprivileged code) to migrate processes between two
1315 disconnected cgroup subtrees, even if both are managed and owned by
1316 the user. This effectively means "systemd-run --user --scope" doesn't
1317 work when invoked from outside of any "systemd --user" service or
1318 scope. Specifically, it is not supported from session scopes. We are
1319 working on fixing this in a future systemd version. (See #3388 for
1320 further details about this.)
1322 * DBus policy files are now installed into /usr rather than /etc. Make
1323 sure your system has dbus >= 1.9.18 running before upgrading to this
1324 version, or override the install path with --with-dbuspolicydir= .
1326 * All python scripts shipped with systemd (specifically: the various
1327 tests written in Python) now require Python 3.
1329 * systemd unit tests can now run standalone (without the source or
1330 build directories), and can be installed into /usr/lib/systemd/tests/
1331 with 'make install-tests'.
1333 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1334 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1337 * Support for the %c, %r, %R specifiers in unit files has been
1338 removed. Specifiers are not supposed to be dependent on configuration
1339 in the unit file itself (so that they resolve the same regardless
1340 where used in the unit files), but these specifiers were influenced
1341 by the Slice= option.
1343 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1344 all cases. If distributions want to use a different shell for this
1345 purpose (for example Fedora's /sbin/sushell) they need to specify
1346 this explicitly at configure time using --with-debug-shell=.
1348 * The confirmation spawn prompt has been reworked to offer the
1351 (c)ontinue, proceed without asking anymore
1352 (D)ump, show the state of the unit
1353 (f)ail, don't execute the command and pretend it failed
1355 (i)nfo, show a short summary of the unit
1356 (j)obs, show jobs that are in progress
1357 (s)kip, don't execute the command and pretend it succeeded
1358 (y)es, execute the command
1360 The 'n' choice for the confirmation spawn prompt has been removed,
1361 because its meaning was confusing.
1363 The prompt may now also be redirected to an alternative console by
1364 specifying the console as parameter to systemd.confirm_spawn=.
1366 * Services of Type=notify require a READY=1 notification to be sent
1367 during startup. If no such message is sent, the service now fails,
1368 even if the main process exited with a successful exit code.
1370 * Services that fail to start up correctly now always have their
1371 ExecStopPost= commands executed. Previously, they'd enter "failed"
1372 state directly, without executing these commands.
1374 * The option MulticastDNS= of network configuration files has acquired
1375 an actual implementation. With MulticastDNS=yes a host can resolve
1376 names of remote hosts and reply to mDNS A and AAAA requests.
1378 * When units are about to be started an additional check is now done to
1379 ensure that all dependencies of type BindsTo= (when used in
1380 combination with After=) have been started.
1382 * systemd-analyze gained a new verb "syscall-filter" which shows which
1383 system call groups are defined for the SystemCallFilter= unit file
1384 setting, and which system calls they contain.
1386 * A new system call filter group "@filesystem" has been added,
1387 consisting of various file system related system calls. Group
1388 "@reboot" has been added, covering reboot, kexec and shutdown related
1389 calls. Finally, group "@swap" has been added covering swap
1390 configuration related calls.
1392 * A new unit file option RestrictNamespaces= has been added that may be
1393 used to restrict access to the various process namespace types the
1394 Linux kernel provides. Specifically, it may be used to take away the
1395 right for a service unit to create additional file system, network,
1396 user, and other namespaces. This sandboxing option is particularly
1397 relevant due to the high amount of recently discovered namespacing
1398 related vulnerabilities in the kernel.
1400 * systemd-udev's .link files gained support for a new AutoNegotiation=
1401 setting for configuring Ethernet auto-negotiation.
1403 * systemd-networkd's .network files gained support for a new
1404 ListenPort= setting in the [DHCP] section to explicitly configure the
1405 UDP client port the DHCP client shall listen on.
1407 * .network files gained a new Unmanaged= boolean setting for explicitly
1408 excluding one or more interfaces from management by systemd-networkd.
1410 * The systemd-networkd ProxyARP= option has been renamed to
1411 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1412 renamed to ReduceARPProxy=. The old names continue to be available
1415 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1416 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1418 * systemd-networkd's bonding device support gained support for two new
1419 configuration options ActiveSlave= and PrimarySlave=.
1421 * The various options in the [Match] section of .network files gained
1422 support for negative matching.
1424 * New systemd-specific mount options are now understood in /etc/fstab:
1426 x-systemd.mount-timeout= may be used to configure the maximum
1427 permitted runtime of the mount command.
1429 x-systemd.device-bound may be set to bind a mount point to its
1430 backing device unit, in order to automatically remove a mount point
1431 if its backing device is unplugged. This option may also be
1432 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1433 on the block device, which is now automatically set for all CDROM
1434 drives, so that mounted CDs are automatically unmounted when they are
1435 removed from the drive.
1437 x-systemd.after= and x-systemd.before= may be used to explicitly
1438 order a mount after or before another unit or mount point.
1440 * Enqueued start jobs for device units are now automatically garbage
1441 collected if there are no jobs waiting for them anymore.
1443 * systemctl list-jobs gained two new switches: with --after, for every
1444 queued job the jobs it's waiting for are shown; with --before the
1445 jobs which it's blocking are shown.
1447 * systemd-nspawn gained support for ephemeral boots from disk images
1448 (or in other words: --ephemeral and --image= may now be
1449 combined). Moreover, ephemeral boots are now supported for normal
1450 directories, even if the backing file system is not btrfs. Of course,
1451 if the file system does not support file system snapshots or
1452 reflinks, the initial copy operation will be relatively expensive, but
1453 this should still be suitable for many use cases.
1455 * Calendar time specifications in .timer units now support
1456 specifications relative to the end of a month by using "~" instead of
1457 "-" as separator between month and day. For example, "*-02~03" means
1458 "the third last day in February". In addition a new syntax for
1459 repeated events has been added using the "/" character. For example,
1460 "9..17/2:00" means "every two hours from 9am to 5pm".
1462 * systemd-socket-proxyd gained a new parameter --connections-max= for
1463 configuring the maximum number of concurrent connections.
1465 * sd-id128 gained a new API for generating unique IDs for the host in a
1466 way that does not leak the machine ID. Specifically,
1467 sd_id128_get_machine_app_specific() derives an ID based on the
1468 machine ID a in well-defined, non-reversible, stable way. This is
1469 useful whenever an identifier for the host is needed but where the
1470 identifier shall not be useful to identify the system beyond the
1471 scope of the application itself. (Internally this uses HMAC-SHA256 as
1472 keyed hash function using the machine ID as input.)
1474 * NotifyAccess= gained a new supported value "exec". When set
1475 notifications are accepted from all processes systemd itself invoked,
1476 including all control processes.
1478 * .nspawn files gained support for defining overlay mounts using the
1479 Overlay= and OverlayReadOnly= options. Previously this functionality
1480 was only available on the systemd-nspawn command line.
1482 * systemd-nspawn's --bind= and --overlay= options gained support for
1483 bind/overlay mounts whose source lies within the container tree by
1484 prefixing the source path with "+".
1486 * systemd-nspawn's --bind= and --overlay= options gained support for
1487 automatically allocating a temporary source directory in /var/tmp
1488 that is removed when the container dies. Specifically, if the source
1489 directory is specified as empty string this mechanism is selected. An
1490 example usage is --overlay=+/var::/var, which creates an overlay
1491 mount based on the original /var contained in the image, overlayed
1492 with a temporary directory in the host's /var/tmp. This way changes
1493 to /var are automatically flushed when the container shuts down.
1495 * systemd-nspawn --image= option does now permit raw file system block
1496 devices (in addition to images containing partition tables, as
1499 * The disk image dissection logic in systemd-nspawn gained support for
1500 automatically setting up LUKS encrypted as well as Verity protected
1501 partitions. When a container is booted from an encrypted image the
1502 passphrase is queried at start-up time. When a container with Verity
1503 data is started, the root hash is search in a ".roothash" file
1504 accompanying the disk image (alternatively, pass the root hash via
1505 the new --root-hash= command line option).
1507 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1508 be used to dissect disk images the same way as systemd-nspawn does
1509 it, following the Bootable Partition Specification. It may even be
1510 used to mount disk images with complex partition setups (including
1511 LUKS and Verity partitions) to a local host directory, in order to
1512 inspect them. This tool is not considered public API (yet), and is
1513 thus not installed into /usr/bin. Please do not rely on its
1514 existence, since it might go away or be changed in later systemd
1517 * A new generator "systemd-verity-generator" has been added, similar in
1518 style to "systemd-cryptsetup-generator", permitting automatic setup of
1519 Verity root partitions when systemd boots up. In order to make use of
1520 this your partition setup should follow the Discoverable Partitions
1521 Specification, and the GPT partition ID of the root file system
1522 partition should be identical to the upper 128bit of the Verity root
1523 hash. The GPT partition ID of the Verity partition protecting it
1524 should be the lower 128bit of the Verity root hash. If the partition
1525 image follows this model it is sufficient to specify a single
1526 "roothash=" kernel command line argument to both configure which root
1527 image and verity partition to use as well as the root hash for
1528 it. Note that systemd-nspawn's Verity support follows the same
1529 semantics, meaning that disk images with proper Verity data in place
1530 may be booted in containers with systemd-nspawn as well as on
1531 physical systems via the verity generator. Also note that the "mkosi"
1532 tool available at https://github.com/systemd/mkosi has been updated
1533 to generate Verity protected disk images following this scheme. In
1534 fact, it has been updated to generate disk images that optionally
1535 implement a complete UEFI SecureBoot trust chain, involving a signed
1536 kernel and initrd image that incorporates such a root hash as well as
1537 a Verity-enabled root partition.
1539 * The hardware database (hwdb) udev supports has been updated to carry
1540 accelerometer quirks.
1542 * All system services are now run with a fresh kernel keyring set up
1543 for them. The invocation ID is stored by default in it, thus
1544 providing a safe, non-overridable way to determine the invocation
1547 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1548 options for bind mounting arbitrary paths in a service-specific
1549 way. When these options are used, arbitrary host or service files and
1550 directories may be mounted to arbitrary locations in the service's
1553 * Documentation has been added that lists all of systemd's low-level
1554 environment variables:
1556 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
1558 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1559 whether a specific socket file descriptor matches a specified socket
1562 * systemd-firstboot has been updated to check for the
1563 systemd.firstboot= kernel command line option. It accepts a boolean
1564 and when set to false the first boot questions are skipped.
1566 * systemd-fstab-generator has been updated to check for the
1567 systemd.volatile= kernel command line option, which either takes an
1568 optional boolean parameter or the special value "state". If used the
1569 system may be booted in a "volatile" boot mode. Specifically,
1570 "systemd.volatile" is used, the root directory will be mounted as
1571 tmpfs, and only /usr is mounted from the actual root file system. If
1572 "systemd.volatile=state" is used, the root directory will be mounted
1573 as usual, but /var is mounted as tmpfs. This concept provides similar
1574 functionality as systemd-nspawn's --volatile= option, but provides it
1575 on physical boots. Use this option for implementing stateless
1576 systems, or testing systems with all state and/or configuration reset
1577 to the defaults. (Note though that many distributions are not
1578 prepared to boot up without a populated /etc or /var, though.)
1580 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1581 partitions. Previously it only supported LUKS encrypted partitions
1582 for all other uses, except for the root partition itself.
1584 * Socket units gained support for listening on AF_VSOCK sockets for
1585 communication in virtualized QEMU environments.
1587 * The "configure" script gained a new option --with-fallback-hostname=
1588 for specifying the fallback hostname to use if none is configured in
1589 /etc/hostname. For example, by specifying
1590 --with-fallback-hostname=fedora it is possible to default to a
1591 hostname of "fedora" on pristine installations.
1593 * systemd-cgls gained support for a new --unit= switch for listing only
1594 the control groups of a specific unit. Similar --user-unit= has been
1595 added for listing only the control groups of a specific user unit.
1597 * systemd-mount gained a new --umount switch for unmounting a mount or
1598 automount point (and all mount/automount points below it).
1600 * systemd will now refuse full configuration reloads (via systemctl
1601 daemon-reload and related calls) unless at least 16MiB of free space
1602 are available in /run. This is a safety precaution in order to ensure
1603 that generators can safely operate after the reload completed.
1605 * A new unit file option RootImage= has been added, which has a similar
1606 effect as RootDirectory= but mounts the service's root directory from
1607 a disk image instead of plain directory. This logic reuses the same
1608 image dissection and mount logic that systemd-nspawn already uses,
1609 and hence supports any disk images systemd-nspawn supports, including
1610 those following the Discoverable Partition Specification, as well as
1611 Verity enabled images. This option enables systemd to run system
1612 services directly off disk images acting as resource bundles,
1613 possibly even including full integrity data.
1615 * A new MountAPIVFS= unit file option has been added, taking a boolean
1616 argument. If enabled /proc, /sys and /dev (collectively called the
1617 "API VFS") will be mounted for the service. This is only relevant if
1618 RootDirectory= or RootImage= is used for the service, as these mounts
1619 are of course in place in the host mount namespace anyway.
1621 * systemd-nspawn gained support for a new --pivot-root= switch. If
1622 specified the root directory within the container image is pivoted to
1623 the specified mount point, while the original root disk is moved to a
1624 different place. This option enables booting of ostree images
1625 directly with systemd-nspawn.
1627 * The systemd build scripts will no longer complain if the NTP server
1628 addresses are not changed from the defaults. Google now supports
1629 these NTP servers officially. We still recommend downstreams to
1630 properly register an NTP pool with the NTP pool project though.
1632 * coredumpctl gained a new "--reverse" option for printing the list
1633 of coredumps in reverse order.
1635 * coredumpctl will now show additional information about truncated and
1636 inaccessible coredumps, as well as coredumps that are still being
1637 processed. It also gained a new --quiet switch for suppressing
1638 additional informational message in its output.
1640 * coredumpctl gained support for only showing coredumps newer and/or
1641 older than specific timestamps, using the new --since= and --until=
1642 options, reminiscent of journalctl's options by the same name.
1644 * The systemd-coredump logic has been improved so that it may be reused
1645 to collect backtraces in non-compiled languages, for example in
1646 scripting languages such as Python.
1648 * machinectl will now show the UID shift of local containers, if user
1649 namespacing is enabled for them.
1651 * systemd will now optionally run "environment generator" binaries at
1652 configuration load time. They may be used to add environment
1653 variables to the environment block passed to services invoked. One
1654 user environment generator is shipped by default that sets up
1655 environment variables based on files dropped into /etc/environment.d
1656 and ~/.config/environment.d/.
1658 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1661 * hostnamed has been updated to report a new chassis type of
1662 "convertible" to cover "foldable" laptops that can both act as a
1663 tablet and as a laptop, such as various Lenovo Yoga devices.
1665 Contributions from: Adrián López, Alexander Galanin, Alexander
1666 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1667 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1668 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1669 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1670 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1671 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1672 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1673 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
1674 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1675 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1676 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1677 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1678 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1679 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1680 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1681 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1682 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1683 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1684 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1685 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1686 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1687 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1688 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1689 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1690 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1691 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1692 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1695 — Berlin, 2017-03-01
1699 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1700 RestrictAddressFamilies= enabled. These sandboxing options should
1701 generally be compatible with the various external udev call-out
1702 binaries we are aware of, however there may be exceptions, in
1703 particular when exotic languages for these call-outs are used. In
1704 this case, consider turning off these settings locally.
1706 * The new RemoveIPC= option can be used to remove IPC objects owned by
1707 the user or group of a service when that service exits.
1709 * The new ProtectKernelModules= option can be used to disable explicit
1710 load and unload operations of kernel modules by a service. In
1711 addition access to /usr/lib/modules is removed if this option is set.
1713 * ProtectSystem= option gained a new value "strict", which causes the
1714 whole file system tree with the exception of /dev, /proc, and /sys,
1715 to be remounted read-only for a service.
1717 * The new ProtectKernelTunables= option can be used to disable
1718 modification of configuration files in /sys and /proc by a service.
1719 Various directories and files are remounted read-only, so access is
1720 restricted even if the file permissions would allow it.
1722 * The new ProtectControlGroups= option can be used to disable write
1723 access by a service to /sys/fs/cgroup.
1725 * Various systemd services have been hardened with
1726 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1727 RestrictAddressFamilies=.
1729 * Support for dynamically creating users for the lifetime of a service
1730 has been added. If DynamicUser=yes is specified, user and group IDs
1731 will be allocated from the range 61184..65519 for the lifetime of the
1732 service. They can be resolved using the new nss-systemd.so NSS
1733 module. The module must be enabled in /etc/nsswitch.conf. Services
1734 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1735 any resources allocated by the service will be cleaned up when the
1736 service exits. They also have ProtectHome=read-only and
1737 ProtectSystem=strict enabled, so they are not able to make any
1738 permanent modifications to the system.
1740 * The nss-systemd module also always resolves root and nobody, making
1741 it possible to have no /etc/passwd or /etc/group files in minimal
1742 container or chroot environments.
1744 * Services may be started with their own user namespace using the new
1745 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1746 under which the service is running are mapped. All other users are
1749 * Support for the cgroup namespace has been added to systemd-nspawn. If
1750 supported by kernel, the container system started by systemd-nspawn
1751 will have its own view of the cgroup hierarchy. This new behaviour
1752 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1754 * The new MemorySwapMax= option can be used to limit the maximum swap
1755 usage under the unified cgroup hierarchy.
1757 * Support for the CPU controller in the unified cgroup hierarchy has
1758 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1759 options. This controller requires out-of-tree patches for the kernel
1760 and the support is provisional.
1762 * Mount and automount units may now be created transiently
1763 (i.e. dynamically at runtime via the bus API, instead of requiring
1764 unit files in the file system).
1766 * systemd-mount is a new tool which may mount file systems – much like
1767 mount(8), optionally pulling in additional dependencies through
1768 transient .mount and .automount units. For example, this tool
1769 automatically runs fsck on a backing block device before mounting,
1770 and allows the automount logic to be used dynamically from the
1771 command line for establishing mount points. This tool is particularly
1772 useful when dealing with removable media, as it will ensure fsck is
1773 run – if necessary – before the first access and that the file system
1774 is quickly unmounted after each access by utilizing the automount
1775 logic. This maximizes the chance that the file system on the
1776 removable media stays in a clean state, and if it isn't in a clean
1777 state is fixed automatically.
1779 * LazyUnmount=yes option for mount units has been added to expose the
1780 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1783 * /efi will be used as the mount point of the EFI boot partition, if
1784 the directory is present, and the mount point was not configured
1785 through other means (e.g. fstab). If /efi directory does not exist,
1786 /boot will be used as before. This makes it easier to automatically
1787 mount the EFI partition on systems where /boot is used for something
1790 * When operating on GPT disk images for containers, systemd-nspawn will
1791 now mount the ESP to /boot or /efi according to the same rules as PID
1792 1 running on a host. This allows tools like "bootctl" to operate
1793 correctly within such containers, in order to make container images
1794 bootable on physical systems.
1796 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
1798 * Two new user session targets have been added to support running
1799 graphical sessions under the systemd --user instance:
1800 graphical-session.target and graphical-session-pre.target. See
1801 systemd.special(7) for a description of how those targets should be
1804 * The vconsole initialization code has been significantly reworked to
1805 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
1806 support unicode keymaps. Font and keymap configuration will now be
1807 copied to all allocated virtual consoles.
1809 * FreeBSD's bhyve virtualization is now detected.
1811 * Information recorded in the journal for core dumps now includes the
1812 contents of /proc/mountinfo and the command line of the process at
1813 the top of the process hierarchy (which is usually the init process
1816 * systemd-journal-gatewayd learned the --directory= option to serve
1817 files from the specified location.
1819 * journalctl --root=… can be used to peruse the journal in the
1820 /var/log/ directories inside of a container tree. This is similar to
1821 the existing --machine= option, but does not require the container to
1824 * The hardware database has been extended to support
1825 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1828 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1829 specify the click rate for mice which include a horizontal wheel with
1830 a click rate that is different than the one for the vertical wheel.
1832 * systemd-run gained a new --wait option that makes service execution
1833 synchronous. (Specifically, the command will not return until the
1834 specified service binary exited.)
1836 * systemctl gained a new --wait option that causes the start command to
1837 wait until the units being started have terminated again.
1839 * A new journal output mode "short-full" has been added which displays
1840 timestamps with abbreviated English day names and adds a timezone
1841 suffix. Those timestamps include more information than the default
1842 "short" output mode, and can be passed directly to journalctl's
1843 --since= and --until= options.
1845 * /etc/resolv.conf will be bind-mounted into containers started by
1846 systemd-nspawn, if possible, so any changes to resolv.conf contents
1847 are automatically propagated to the container.
1849 * The number of instances for socket-activated services originating
1850 from a single IP address can be limited with
1851 MaxConnectionsPerSource=, extending the existing setting of
1854 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1857 * .netdev and .network configuration can now be extended through
1860 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1861 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1862 can be enabled and disabled using the new UDPSegmentationOffload=,
1863 TCPSegmentationOffload=, GenericSegmentationOffload=,
1864 GenericReceiveOffload=, LargeReceiveOffload= options in the
1865 [Link] section of .link files.
1867 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1868 Port VLAN ID can be configured for bridge devices using the new STP=,
1869 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1870 section of .netdev files.
1872 * The route table to which routes received over DHCP or RA should be
1873 added can be configured with the new RouteTable= option in the [DHCP]
1874 and [IPv6AcceptRA] sections of .network files.
1876 * The Address Resolution Protocol can be disabled on links managed by
1877 systemd-networkd using the ARP=no setting in the [Link] section of
1880 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1881 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1882 encode information about the result and exit codes of the current
1883 service runtime cycle.
1885 * systemd-sysctl will now configure kernel parameters in the order
1886 they occur in the configuration files. This matches what sysctl
1887 has been traditionally doing.
1889 * kernel-install "plugins" that are executed to perform various
1890 tasks after a new kernel is added and before an old one is removed
1891 can now return a special value to terminate the procedure and
1892 prevent any later plugins from running.
1894 * Journald's SplitMode=login setting has been deprecated. It has been
1895 removed from documentation, and its use is discouraged. In a future
1896 release it will be completely removed, and made equivalent to current
1897 default of SplitMode=uid.
1899 * Storage=both option setting in /etc/systemd/coredump.conf has been
1900 removed. With fast LZ4 compression storing the core dump twice is not
1903 * The --share-system systemd-nspawn option has been replaced with an
1904 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1905 this functionality is discouraged. In addition the variables
1906 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1907 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1908 individual namespaces.
1910 * "machinectl list" now shows the IP address of running containers in
1911 the output, as well as OS release information.
1913 * "loginctl list" now shows the TTY of each session in the output.
1915 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1916 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1917 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1918 tracking objects in a "recursive" mode, where a single client can be
1919 counted multiple times, if it takes multiple references.
1921 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
1922 sd_bus_get_exit_on_disconnect(). They may be used to make a
1923 process using sd-bus automatically exit if the bus connection is
1926 * Bus clients of the service manager may now "pin" loaded units into
1927 memory, by taking an explicit reference on them. This is useful to
1928 ensure the client can retrieve runtime data about the service even
1929 after the service completed execution. Taking such a reference is
1930 available only for privileged clients and should be helpful to watch
1931 running services in a race-free manner, and in particular collect
1932 information about exit statuses and results.
1934 * The nss-resolve module has been changed to strictly return UNAVAIL
1935 when communication via D-Bus with resolved failed, and NOTFOUND when
1936 a lookup completed but was negative. This means it is now possible to
1937 neatly configure fallbacks using nsswitch.conf result checking
1938 expressions. Taking benefit of this, the new recommended
1939 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
1941 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
1943 * A new setting CtrlAltDelBurstAction= has been added to
1944 /etc/systemd/system.conf which may be used to configure the precise
1945 behaviour if the user on the console presses Ctrl-Alt-Del more often
1946 than 7 times in 2s. Previously this would unconditionally result in
1947 an expedited, immediate reboot. With this new setting the precise
1948 operation may be configured in more detail, and also turned off
1951 * In .netdev files two new settings RemoteChecksumTx= and
1952 RemoteChecksumRx= are now understood that permit configuring the
1953 remote checksumming logic for VXLAN networks.
1955 * The service manager learnt a new "invocation ID" concept for invoked
1956 services. Each runtime cycle of a service will get a new invocation
1957 ID (a 128bit random UUID) assigned that identifies the current
1958 run of the service uniquely and globally. A new invocation ID
1959 is generated each time a service starts up. The journal will store
1960 the invocation ID of a service along with any logged messages, thus
1961 making the invocation ID useful for matching the online runtime of a
1962 service with the offline log data it generated in a safe way without
1963 relying on synchronized timestamps. In many ways this new service
1964 invocation ID concept is similar to the kernel's boot ID concept that
1965 uniquely and globally identifies the runtime of each boot. The
1966 invocation ID of a service is passed to the service itself via an
1967 environment variable ($INVOCATION_ID). A new bus call
1968 GetUnitByInvocationID() has been added that is similar to GetUnit()
1969 but instead of retrieving the bus path for a unit by its name
1970 retrieves it by its invocation ID. The returned path is valid only as
1971 long as the passed invocation ID is current.
1973 * systemd-resolved gained a new "DNSStubListener" setting in
1974 resolved.conf. It either takes a boolean value or the special values
1975 "udp" and "tcp", and configures whether to enable the stub DNS
1976 listener on 127.0.0.53:53.
1978 * IP addresses configured via networkd may now carry additional
1979 configuration settings supported by the kernel. New options include:
1980 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
1981 PrefixRoute=, AutoJoin=.
1983 * The PAM configuration fragment file for "user@.service" shipped with
1984 systemd (i.e. the --user instance of systemd) has been stripped to
1985 the minimum necessary to make the system boot. Previously, it
1986 contained Fedora-specific stanzas that did not apply to other
1987 distributions. It is expected that downstream distributions add
1988 additional configuration lines, matching their needs to this file,
1989 using it only as rough template of what systemd itself needs. Note
1990 that this reduced fragment does not even include an invocation of
1991 pam_limits which most distributions probably want to add, even though
1992 systemd itself does not need it. (There's also the new build time
1993 option --with-pamconfdir=no to disable installation of the PAM
1996 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
1997 capability is now also dropped from its set (in addition to
1998 CAP_SYS_MKNOD as before).
2000 * In service unit files it is now possible to connect a specific named
2001 file descriptor with stdin/stdout/stdout of an executed service. The
2002 name may be specified in matching .socket units using the
2003 FileDescriptorName= setting.
2005 * A number of journal settings may now be configured on the kernel
2006 command line. Specifically, the following options are now understood:
2007 systemd.journald.max_level_console=,
2008 systemd.journald.max_level_store=,
2009 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2010 systemd.journald.max_level_wall=.
2012 * "systemctl is-enabled --full" will now show by which symlinks a unit
2013 file is enabled in the unit dependency tree.
2015 * Support for VeraCrypt encrypted partitions has been added to the
2016 "cryptsetup" logic and /etc/crypttab.
2018 * systemd-detect-virt gained support for a new --private-users switch
2019 that checks whether the invoking processes are running inside a user
2020 namespace. Similar, a new special value "private-users" for the
2021 existing ConditionVirtualization= setting has been added, permitting
2022 skipping of specific units in user namespace environments.
2024 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2025 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2026 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2027 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2028 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2029 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2030 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2031 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2032 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2033 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2034 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2035 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2036 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2037 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2038 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2039 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2040 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2041 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2042 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2043 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2044 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2045 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2046 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2047 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2048 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2049 Jędrzejewski-Szmek, Zeal Jagannatha
2051 — Santa Fe, 2016-11-03
2055 * In service units the various ExecXYZ= settings have been extended
2056 with an additional special character as first argument of the
2057 assigned value: if the character '+' is used the specified command
2058 line it will be run with full privileges, regardless of User=,
2059 Group=, CapabilityBoundingSet= and similar options. The effect is
2060 similar to the existing PermissionsStartOnly= option, but allows
2061 configuration of this concept for each executed command line
2064 * Services may now alter the service watchdog timeout at runtime by
2065 sending a WATCHDOG_USEC= message via sd_notify().
2067 * MemoryLimit= and related unit settings now optionally take percentage
2068 specifications. The percentage is taken relative to the amount of
2069 physical memory in the system (or in case of containers, the assigned
2070 amount of memory). This allows scaling service resources neatly with
2071 the amount of RAM available on the system. Similarly, systemd-logind's
2072 RuntimeDirectorySize= option now also optionally takes percentage
2075 * In similar fashion TasksMax= takes percentage values now, too. The
2076 value is taken relative to the configured maximum number of processes
2077 on the system. The per-service task maximum has been changed to 15%
2078 using this functionality. (Effectively this is an increase of 512 →
2079 4915 for service units, given the kernel's default pid_max setting.)
2081 * Calendar time specifications in .timer units now understand a ".."
2082 syntax for time ranges. Example: "4..7:10" may now be used for
2083 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2086 * The InaccessableDirectories=, ReadOnlyDirectories= and
2087 ReadWriteDirectories= unit file settings have been renamed to
2088 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2089 applied to all kinds of file nodes, and not just directories, with
2090 the exception of symlinks. Specifically these settings may now be
2091 used on block and character device nodes, UNIX sockets and FIFOS as
2092 well as regular files. The old names of these settings remain
2093 available for compatibility.
2095 * systemd will now log about all service processes it kills forcibly
2096 (using SIGKILL) because they remained after the clean shutdown phase
2097 of the service completed. This should help identifying services that
2098 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2099 systemd-logind's configuration a similar log message is generated for
2100 processes killed at the end of each session due to this setting.
2102 * systemd will now set the $JOURNAL_STREAM environment variable for all
2103 services whose stdout/stderr are connected to the Journal (which
2104 effectively means by default: all services). The variable contains
2105 the device and inode number of the file descriptor used for
2106 stdout/stderr. This may be used by invoked programs to detect whether
2107 their stdout/stderr is connected to the Journal, in which case they
2108 can switch over to direct Journal communication, thus being able to
2109 pass extended, structured metadata along with their log messages. As
2110 one example, this is now used by glib's logging primitives.
2112 * When using systemd's default tmp.mount unit for /tmp, the mount point
2113 will now be established with the "nosuid" and "nodev" options. This
2114 avoids privilege escalation attacks that put traps and exploits into
2115 /tmp. However, this might cause problems if you e. g. put container
2116 images or overlays into /tmp; if you need this, override tmp.mount's
2117 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2120 * systemd now supports the "memory" cgroup controller also on
2123 * The systemd-cgtop tool now optionally takes a control group path as
2124 command line argument. If specified, the control group list shown is
2125 limited to subgroups of that group.
2127 * The SystemCallFilter= unit file setting gained support for
2128 pre-defined, named system call filter sets. For example
2129 SystemCallFilter=@clock is now an effective way to make all clock
2130 changing-related system calls unavailable to a service. A number of
2131 similar pre-defined groups are defined. Writing system call filters
2132 for system services is simplified substantially with this new
2133 concept. Accordingly, all of systemd's own, long-running services now
2134 enable system call filtering based on this, by default.
2136 * A new service setting MemoryDenyWriteExecute= has been added, taking
2137 a boolean value. If turned on, a service may no longer create memory
2138 mappings that are writable and executable at the same time. This
2139 enhances security for services where this is enabled as it becomes
2140 harder to dynamically write and then execute memory in exploited
2141 service processes. This option has been enabled for all of systemd's
2142 own long-running services.
2144 * A new RestrictRealtime= service setting has been added, taking a
2145 boolean argument. If set the service's processes may no longer
2146 acquire realtime scheduling. This improves security as realtime
2147 scheduling may otherwise be used to easily freeze the system.
2149 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2150 value. This may be used for requesting that the system manager inside
2151 of the container reports start-up completion to nspawn which then
2152 propagates this notification further to the service manager
2153 supervising nspawn itself. A related option NotifyReady= in .nspawn
2154 files has been added too. This functionality allows ordering of the
2155 start-up of multiple containers using the usual systemd ordering
2158 * machinectl gained a new command "stop" that is an alias for
2161 * systemd-resolved gained support for contacting DNS servers on
2162 link-local IPv6 addresses.
2164 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2165 its caches. A method call for requesting the same operation has been
2166 added to the bus API too, and is made available via "systemd-resolve
2169 * systemd-resolve gained a new --status switch. If passed a brief
2170 summary of the used DNS configuration with per-interface information
2173 * resolved.conf gained a new Cache= boolean option, defaulting to
2174 on. If turned off local DNS caching is disabled. This comes with a
2175 performance penalty in particular when DNSSEC is enabled. Note that
2176 resolved disables its internal caching implicitly anyway, when the
2177 configured DNS server is on a host-local IP address such as ::1 or
2178 127.0.0.1, thus automatically avoiding double local caching.
2180 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2181 for DNS requests. This improves compatibility with local programs
2182 that do not use the libc NSS or systemd-resolved's bus APIs for name
2183 resolution. This minimal DNS service is only available to local
2184 programs and does not implement the full DNS protocol, but enough to
2185 cover local DNS clients. A new, static resolv.conf file, listing just
2186 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2187 now recommended to make /etc/resolv.conf a symlink to this file in
2188 order to route all DNS lookups to systemd-resolved, regardless if
2189 done via NSS, the bus API or raw DNS packets. Note that this local
2190 DNS service is not as fully featured as the libc NSS or
2191 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2192 used to deliver link-local address information (as this implies
2193 sending a local interface index along), LLMNR/mDNS support via this
2194 interface is severely restricted. It is thus strongly recommended for
2195 all applications to use the libc NSS API or native systemd-resolved
2198 * systemd-networkd's bridge support learned a new setting
2199 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2200 in .network files has been added for configuring VLAN bridging in
2201 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2203 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2204 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2205 now be acquired without relying on DHCPv6. Two new options
2206 UseDomains= and UseDNS= have been added to configure this behaviour.
2208 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2209 renamed IPv6AcceptRA=, without altering its behaviour. The old
2210 setting name remains available for compatibility reasons.
2212 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2213 Key=, InputKey= and OutputKey=.
2215 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2216 interface configuration.
2218 * "systemctl edit" may now be used to create new unit files by
2219 specifying the --force switch.
2221 * sd-event gained a new function sd_event_get_iteration() for
2222 requesting the current iteration counter of the event loop. It starts
2223 at zero and is increased by one with each event loop iteration.
2225 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2226 file. It can be used in lieu of %systemd_requires in packages which
2227 don't use any systemd functionality and are intended to be installed
2228 in minimal containers without systemd present. This macro provides
2229 ordering dependencies to ensure that if the package is installed in
2230 the same rpm transaction as systemd, systemd will be installed before
2231 the scriptlets for the package are executed, allowing unit presets
2234 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2235 been added to simplify packaging of generators.
2237 * The os-release file gained VERSION_CODENAME field for the
2238 distribution nickname (e.g. VERSION_CODENAME=woody).
2240 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2241 can be set to disable parsing of metadata and the creation
2242 of persistent symlinks for that device.
2244 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2245 to make them available to logged-in users has been reverted.
2247 * Much of the common code of the various systemd components is now
2248 built into an internal shared library libsystemd-shared-231.so
2249 (incorporating the systemd version number in the name, to be updated
2250 with future releases) that the components link to. This should
2251 decrease systemd footprint both in memory during runtime and on
2252 disk. Note that the shared library is not for public use, and is
2253 neither API nor ABI stable, but is likely to change with every new
2254 released update. Packagers need to make sure that binaries
2255 linking to libsystemd-shared.so are updated in step with the
2258 * Configuration for "mkosi" is now part of the systemd
2259 repository. mkosi is a tool to easily build legacy-free OS images,
2260 and is available on github: https://github.com/systemd/mkosi. If
2261 "mkosi" is invoked in the build tree a new raw OS image is generated
2262 incorporating the systemd sources currently being worked on and a
2263 clean, fresh distribution installation. The generated OS image may be
2264 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
2265 UEFI PC. This functionality is particularly useful to easily test
2266 local changes made to systemd in a pristine, defined environment. See
2267 doc/HACKING for details.
2269 * configure learned the --with-support-url= option to specify the
2270 distribution's bugtracker.
2272 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2273 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2274 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2275 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2276 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2277 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2278 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2279 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2280 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2281 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2282 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2283 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2284 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2285 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2286 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2287 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
2288 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2289 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
2290 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
2292 — Berlin, 2016-07-25
2296 * DNSSEC is now turned on by default in systemd-resolved (in
2297 "allow-downgrade" mode), but may be turned off during compile time by
2298 passing "--with-default-dnssec=no" to "configure" (and of course,
2299 during runtime with DNSSEC= in resolved.conf). We recommend
2300 downstreams to leave this on at least during development cycles and
2301 report any issues with the DNSSEC logic upstream. We are very
2302 interested in collecting feedback about the DNSSEC validator and its
2303 limitations in the wild. Note however, that DNSSEC support is
2304 probably nothing downstreams should turn on in stable distros just
2305 yet, as it might create incompatibilities with a few DNS servers and
2306 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2307 automatically whenever we detect such incompatible setups, but there
2308 might be systems we do not cover yet. Hence: please help us testing
2309 the DNSSEC code, leave this on where you can, report back, but then
2310 again don't consider turning this on in your stable, LTS or
2311 production release just yet. (Note that you have to enable
2312 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2313 and its DNSSEC mode for host name resolution from local
2316 * systemd-resolve conveniently resolves DANE records with the --tlsa
2317 option and OPENPGPKEY records with the --openpgp option. It also
2318 supports dumping raw DNS record data via the new --raw= switch.
2320 * systemd-logind will now by default terminate user processes that are
2321 part of the user session scope unit (session-XX.scope) when the user
2322 logs out. This behavior is controlled by the KillUserProcesses=
2323 setting in logind.conf, and the previous default of "no" is now
2324 changed to "yes". This means that user sessions will be properly
2325 cleaned up after, but additional steps are necessary to allow
2326 intentionally long-running processes to survive logout.
2328 While the user is logged in at least once, user@.service is running,
2329 and any service that should survive the end of any individual login
2330 session can be started at a user service or scope using systemd-run.
2331 systemd-run(1) man page has been extended with an example which shows
2332 how to run screen in a scope unit underneath user@.service. The same
2333 command works for tmux.
2335 After the user logs out of all sessions, user@.service will be
2336 terminated too, by default, unless the user has "lingering" enabled.
2337 To effectively allow users to run long-term tasks even if they are
2338 logged out, lingering must be enabled for them. See loginctl(1) for
2339 details. The default polkit policy was modified to allow users to
2340 set lingering for themselves without authentication.
2342 Previous defaults can be restored at compile time by the
2343 --without-kill-user-processes option to "configure".
2345 * systemd-logind gained new configuration settings SessionsMax= and
2346 InhibitorsMax=, both with a default of 8192. It will not register new
2347 user sessions or inhibitors above this limit.
2349 * systemd-logind will now reload configuration on SIGHUP.
2351 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
2352 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
2353 enable. Also, support for the "io" cgroup controller in the unified
2354 hierarchy has been added, so that the "memory", "pids" and "io" are
2355 now the controllers that are supported on the unified hierarchy.
2357 WARNING: it is not possible to use previous systemd versions with
2358 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2359 is necessary to also update systemd in the initramfs if using the
2360 unified hierarchy. An updated SELinux policy is also required.
2362 * LLDP support has been extended, and both passive (receive-only) and
2363 active (sender) modes are supported. Passive mode ("routers-only") is
2364 enabled by default in systemd-networkd. Active LLDP mode is enabled
2365 by default for containers on the internal network. The "networkctl
2366 lldp" command may be used to list information gathered. "networkctl
2367 status" will also show basic LLDP information on connected peers now.
2369 * The IAID and DUID unique identifier sent in DHCP requests may now be
2370 configured for the system and each .network file managed by
2371 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2373 * systemd-networkd gained support for configuring proxy ARP support for
2374 each interface, via the ProxyArp= setting in .network files. It also
2375 gained support for configuring the multicast querier feature of
2376 bridge devices, via the new MulticastQuerier= setting in .netdev
2377 files. Similarly, snooping on the IGMP traffic can be controlled
2378 via the new setting MulticastSnooping=.
2380 A new setting PreferredLifetime= has been added for addresses
2381 configured in .network file to configure the lifetime intended for an
2384 The systemd-networkd DHCP server gained the option EmitRouter=, which
2385 defaults to yes, to configure whether the DHCP Option 3 (Router)
2388 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
2389 systemd-socket-activate and installed into /usr/bin. It is now fully
2392 * systemd-journald now uses separate threads to flush changes to disk
2393 when closing journal files, thus reducing impact of slow disk I/O on
2394 logging performance.
2396 * The sd-journal API gained two new calls
2397 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2398 can be used to open journal files using file descriptors instead of
2399 file or directory paths. sd_journal_open_container() has been
2400 deprecated, sd_journal_open_directory_fd() should be used instead
2401 with the flag SD_JOURNAL_OS_ROOT.
2403 * journalctl learned a new output mode "-o short-unix" that outputs log
2404 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2405 UTC). It also gained support for a new --no-hostname setting to
2406 suppress the hostname column in the family of "short" output modes.
2408 * systemd-ask-password now optionally skips printing of the password to
2409 stdout with --no-output which can be useful in scripts.
2411 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2412 (devices tagged with ID_MAKER_TOOL) are now tagged with
2413 "uaccess" and are available to logged in users.
2415 * The DeviceAllow= unit setting now supports specifiers (with "%").
2417 * "systemctl show" gained a new --value switch, which allows print a
2418 only the contents of a specific unit property, without also printing
2419 the property's name. Similar support was added to "show*" verbs
2420 of loginctl and machinectl that output "key=value" lists.
2422 * A new unit type "generated" was added for files dynamically generated
2423 by generator tools. Similarly, a new unit type "transient" is used
2424 for unit files created using the runtime API. "systemctl enable" will
2425 refuse to operate on such files.
2427 * A new command "systemctl revert" has been added that may be used to
2428 revert to the vendor version of a unit file, in case local changes
2429 have been made by adding drop-ins or overriding the unit file.
2431 * "machinectl clean" gained a new verb to automatically remove all or
2432 just hidden container images.
2434 * systemd-tmpfiles gained support for a new line type "e" for emptying
2435 directories, if they exist, without creating them if they don't.
2437 * systemd-nspawn gained support for automatically patching the UID/GIDs
2438 of the owners and the ACLs of all files and directories in a
2439 container tree to match the UID/GID user namespacing range selected
2440 for the container invocation. This mode is enabled via the new
2441 --private-users-chown switch. It also gained support for
2442 automatically choosing a free, previously unused UID/GID range when
2443 starting a container, via the new --private-users=pick setting (which
2444 implies --private-users-chown). Together, these options for the first
2445 time make user namespacing for nspawn containers fully automatic and
2446 thus deployable. The systemd-nspawn@.service template unit file has
2447 been changed to use this functionality by default.
2449 * systemd-nspawn gained a new --network-zone= switch, that allows
2450 creating ad-hoc virtual Ethernet links between multiple containers,
2451 that only exist as long as at least one container referencing them is
2452 running. This allows easy connecting of multiple containers with a
2453 common link that implements an Ethernet broadcast domain. Each of
2454 these network "zones" may be named relatively freely by the user, and
2455 may be referenced by any number of containers, but each container may
2456 only reference one of these "zones". On the lower level, this is
2457 implemented by an automatically managed bridge network interface for
2458 each zone, that is created when the first container referencing its
2459 zone is created and removed when the last one referencing its zone
2462 * The default start timeout may now be configured on the kernel command
2463 line via systemd.default_timeout_start_sec=. It was already
2464 configurable via the DefaultTimeoutStartSec= option in
2465 /etc/systemd/system.conf.
2467 * Socket units gained a new TriggerLimitIntervalSec= and
2468 TriggerLimitBurst= setting to configure a limit on the activation
2469 rate of the socket unit.
2471 * The LimitNICE= setting now optionally takes normal UNIX nice values
2472 in addition to the raw integer limit value. If the specified
2473 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2474 value is understood as UNIX nice value. If not prefixed like this it
2475 is understood as raw RLIMIT_NICE limit.
2477 * Note that the effect of the PrivateDevices= unit file setting changed
2478 slightly with this release: the per-device /dev file system will be
2479 mounted read-only from this version on, and will have "noexec"
2480 set. This (minor) change of behavior might cause some (exceptional)
2481 legacy software to break, when PrivateDevices=yes is set for its
2482 service. Please leave PrivateDevices= off if you run into problems
2485 * systemd-bootchart has been split out to a separate repository:
2486 https://github.com/systemd/systemd-bootchart
2488 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2489 merged into the kernel in its current form.
2491 * The compatibility libraries libsystemd-daemon.so,
2492 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2493 which have been deprecated since systemd-209 have been removed along
2494 with the corresponding pkg-config files. All symbols provided by
2495 those libraries are provided by libsystemd.so.
2497 * The Capabilities= unit file setting has been removed (it is ignored
2498 for backwards compatibility). AmbientCapabilities= and
2499 CapabilityBoundingSet= should be used instead.
2501 * A new special target has been added, initrd-root-device.target,
2502 which creates a synchronization point for dependencies of the root
2503 device in early userspace. Initramfs builders must ensure that this
2504 target is now included in early userspace.
2506 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2507 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2508 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2509 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2510 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2511 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2512 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2513 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
2514 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2515 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
2516 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2517 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2518 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
2519 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2520 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2521 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
2522 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2523 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2524 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2525 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2526 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2527 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
2528 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2529 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2530 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2533 — Fairfax, 2016-05-21
2537 * The systemd-resolved DNS resolver service has gained a substantial
2538 set of new features, most prominently it may now act as a DNSSEC
2539 validating stub resolver. DNSSEC mode is currently turned off by
2540 default, but is expected to be turned on by default in one of the
2541 next releases. For now, we invite everybody to test the DNSSEC logic
2542 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2543 service also gained a full set of D-Bus interfaces, including calls
2544 to configure DNS and DNSSEC settings per link (for use by external
2545 network management software). systemd-resolved and systemd-networkd
2546 now distinguish between "search" and "routing" domains. The former
2547 are used to qualify single-label names, the latter are used purely
2548 for routing lookups within certain domains to specific links.
2549 resolved now also synthesizes RRs for all entries from /etc/hosts.
2551 * The systemd-resolve tool (which is a client utility for
2552 systemd-resolved) has been improved considerably and is now fully
2553 supported and documented. Hence it has moved from /usr/lib/systemd to
2556 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2559 * The coredump collection logic has been reworked: when a coredump is
2560 collected it is now written to disk, compressed and processed
2561 (including stacktrace extraction) from a new instantiated service
2562 systemd-coredump@.service, instead of directly from the
2563 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2564 processing large coredumps can take up a substantial amount of
2565 resources and time, and this previously happened entirely outside of
2566 systemd's service supervision. With the new logic the core_pattern
2567 hook only does minimal metadata collection before passing off control
2568 to the new instantiated service, which is configured with a time
2569 limit, a nice level and other settings to minimize negative impact on
2570 the rest of the system. Also note that the new logic will honour the
2571 RLIMIT_CORE setting of the crashed process, which now allows users
2572 and processes to turn off coredumping for their processes by setting
2575 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2576 and all forked processes by default. Previously, PID 1 would leave
2577 the setting at "0" for all processes, as set by the kernel. Note that
2578 the resource limit traditionally has no effect on the generated
2579 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2580 logic is used. Since the limit is now honoured (see above) its
2581 default has been changed so that the coredumping logic is enabled by
2582 default for all processes, while allowing specific opt-out.
2584 * When the stacktrace is extracted from processes of system users, this
2585 is now done as "systemd-coredump" user, in order to sandbox this
2586 potentially security sensitive parsing operation. (Note that when
2587 processing coredumps of normal users this is done under the user ID
2588 of process that crashed, as before.) Packagers should take notice
2589 that it is now necessary to create the "systemd-coredump" system user
2590 and group at package installation time.
2592 * The systemd-activate socket activation testing tool gained support
2593 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2594 and --seqpacket switches. It also has been extended to support both
2595 new-style and inetd-style file descriptor passing. Use the new
2596 --inetd switch to request inetd-style file descriptor passing.
2598 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2599 variable, which takes a boolean value. If set to false, ANSI color
2600 output is disabled in the tools even when run on a terminal that
2603 * The VXLAN support in networkd now supports two new settings
2604 DestinationPort= and PortRange=.
2606 * A new systemd.machine_id= kernel command line switch has been added,
2607 that may be used to set the machine ID in /etc/machine-id if it is
2608 not initialized yet. This command line option has no effect if the
2609 file is already initialized.
2611 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2612 specified command line as PID 2 rather than PID 1 in the
2613 container. In this mode PID 1 is a minimal stub init process that
2614 implements the special POSIX and Linux semantics of PID 1 regarding
2615 signal and child process management. Note that this stub init process
2616 is implemented in nspawn itself and requires no support from the
2617 container image. This new logic is useful to support running
2618 arbitrary commands in the container, as normal processes are
2619 generally not prepared to run as PID 1.
2621 * systemd-nspawn gained a new --chdir= switch for setting the current
2622 working directory for the process started in the container.
2624 * "journalctl /dev/sda" will now output all kernel log messages for
2625 specified device from the current boot, in addition to all devices
2626 that are parents of it. This should make log output about devices
2627 pretty useful, as long as kernel drivers attach enough metadata to
2628 the log messages. (The usual SATA drivers do.)
2630 * The sd-journal API gained two new calls
2631 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2632 that report whether log data from /run or /var has been found.
2634 * journalctl gained a new switch "--fields" that prints all journal
2635 record field names currently in use in the journal. This is backed
2636 by two new sd-journal API calls sd_journal_enumerate_fields() and
2637 sd_journal_restart_fields().
2639 * Most configurable timeouts in systemd now expect an argument of
2640 "infinity" to turn them off, instead of "0" as before. The semantics
2641 from now on is that a timeout of "0" means "now", and "infinity"
2642 means "never". To maintain backwards compatibility, "0" continues to
2643 turn off previously existing timeout settings.
2645 * "systemctl reload-or-try-restart" has been renamed to "systemctl
2646 try-reload-or-restart" to clarify what it actually does: the "try"
2647 logic applies to both reloading and restarting, not just restarting.
2648 The old name continues to be accepted for compatibility.
2650 * On boot-up, when PID 1 detects that the system clock is behind the
2651 release date of the systemd version in use, the clock is now set
2652 to the latter. Previously, this was already done in timesyncd, in order
2653 to avoid running with clocks set to the various clock epochs such as
2654 1902, 1938 or 1970. With this change the logic is now done in PID 1
2655 in addition to timesyncd during early boot-up, so that it is enforced
2656 before the first process is spawned by systemd. Note that the logic
2657 in timesyncd remains, as it is more comprehensive and ensures
2658 clock monotonicity by maintaining a persistent timestamp file in
2659 /var. Since /var is generally not available in earliest boot or the
2660 initrd, this part of the logic remains in timesyncd, and is not done
2663 * Support for tweaking details in net_cls.class_id through the
2664 NetClass= configuration directive has been removed, as the kernel
2665 people have decided to deprecate that controller in cgroup v2.
2666 Userspace tools such as nftables are moving over to setting rules
2667 that are specific to the full cgroup path of a task, which obsoletes
2668 these controllers anyway. The NetClass= directive is kept around for
2669 legacy compatibility reasons. For a more in-depth description of the
2670 kernel change, please refer to the respective upstream commit:
2672 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2674 * A new service setting RuntimeMaxSec= has been added that may be used
2675 to specify a maximum runtime for a service. If the timeout is hit, the
2676 service is terminated and put into a failure state.
2678 * A new service setting AmbientCapabilities= has been added. It allows
2679 configuration of additional Linux process capabilities that are
2680 passed to the activated processes. This is only available on very
2683 * The process resource limit settings in service units may now be used
2684 to configure hard and soft limits individually.
2686 * The various libsystemd APIs such as sd-bus or sd-event now publicly
2687 expose support for gcc's __attribute__((cleanup())) C extension.
2688 Specifically, for many object destructor functions alternative
2689 versions have been added that have names suffixed with "p" and take a
2690 pointer to a pointer to the object to destroy, instead of just a
2691 pointer to the object itself. This is useful because these destructor
2692 functions may be used directly as parameters to the cleanup
2693 construct. Internally, systemd has been a heavy user of this GCC
2694 extension for a long time, and with this change similar support is
2695 now available to consumers of the library outside of systemd. Note
2696 that by using this extension in your sources compatibility with old
2697 and strictly ANSI compatible C compilers is lost. However, all gcc or
2698 LLVM versions of recent years support this extension.
2700 * Timer units gained support for a new setting RandomizedDelaySec= that
2701 allows configuring some additional randomized delay to the configured
2702 time. This is useful to spread out timer events to avoid load peaks in
2703 clusters or larger setups.
2705 * Calendar time specifications now support sub-second accuracy.
2707 * Socket units now support listening on SCTP and UDP-lite protocol
2710 * The sd-event API now comes with a full set of man pages.
2712 * Older versions of systemd contained experimental support for
2713 compressing journal files and coredumps with the LZ4 compressor that
2714 was not compatible with the lz4 binary (due to API limitations of the
2715 lz4 library). This support has been removed; only support for files
2716 compatible with the lz4 binary remains. This LZ4 logic is now
2717 officially supported and no longer considered experimental.
2719 * The dkr image import logic has been removed again from importd. dkr's
2720 micro-services focus doesn't fit into the machine image focus of
2721 importd, and quickly got out of date with the upstream dkr API.
2723 * Creation of the /run/lock/lockdev/ directory was dropped from
2724 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
2725 been available for many years. If you still need this, you need to
2726 create your own tmpfiles.d config file with:
2728 d /run/lock/lockdev 0775 root lock -
2730 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2731 and RebootArgument= have been moved from the [Service] section of
2732 unit files to [Unit], and they are now supported on all unit types,
2733 not just service units. Of course, elogind will continue to
2734 understand these settings also at the old location, in order to
2735 maintain compatibility.
2737 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2738 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2739 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2740 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2741 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2742 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2743 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2744 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2745 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2746 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2747 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2748 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2749 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2750 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2751 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2752 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2753 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2754 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2755 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2757 — Berlin, 2016-02-11
2761 * A number of properties previously only settable in unit
2762 files are now also available as properties to set when
2763 creating transient units programmatically via the bus, as it
2764 is exposed with systemd-run's --property=
2765 setting. Specifically, these are: SyslogIdentifier=,
2766 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2767 EnvironmentFile=, ReadWriteDirectories=,
2768 ReadOnlyDirectories=, InaccessibleDirectories=,
2769 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2771 * When creating transient services via the bus API it is now
2772 possible to pass in a set of file descriptors to use as
2773 STDIN/STDOUT/STDERR for the invoked process.
2775 * Slice units may now be created transiently via the bus APIs,
2776 similar to the way service and scope units may already be
2777 created transiently.
2779 * Wherever systemd expects a calendar timestamp specification
2780 (like in journalctl's --since= and --until= switches) UTC
2781 timestamps are now supported. Timestamps suffixed with "UTC"
2782 are now considered to be in Universal Time Coordinated
2783 instead of the local timezone. Also, timestamps may now
2784 optionally be specified with sub-second accuracy. Both of
2785 these additions also apply to recurring calendar event
2786 specification, such as OnCalendar= in timer units.
2788 * journalctl gained a new "--sync" switch that asks the
2789 journal daemon to write all so far unwritten log messages to
2790 disk and sync the files, before returning.
2792 * systemd-tmpfiles learned two new line types "q" and "Q" that
2793 operate like "v", but also set up a basic btrfs quota
2794 hierarchy when used on a btrfs file system with quota
2797 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2798 instead of a subvolume (even on a btrfs file system) if the
2799 root directory is a plain directory, and not a
2800 subvolume. This should simplify things with certain chroot()
2801 environments which are not aware of the concept of btrfs
2804 * systemd-detect-virt gained a new --chroot switch to detect
2805 whether execution takes place in a chroot() environment.
2807 * CPUAffinity= now takes CPU index ranges in addition to
2810 * The various memory-related resource limit settings (such as
2811 LimitAS=) now understand the usual K, M, G, ... suffixes to
2812 the base of 1024 (IEC). Similar, the time-related resource
2813 limit settings understand the usual min, h, day, ...
2816 * There's a new system.conf setting DefaultTasksMax= to
2817 control the default TasksMax= setting for services and
2818 scopes running on the system. (TasksMax= is the primary
2819 setting that exposes the "pids" cgroup controller on systemd
2820 and was introduced in the previous systemd release.) The
2821 setting now defaults to 512, which means services that are
2822 not explicitly configured otherwise will only be able to
2823 create 512 processes or threads at maximum, from this
2824 version on. Note that this means that thread- or
2825 process-heavy services might need to be reconfigured to set
2826 TasksMax= to a higher value. It is sufficient to set
2827 TasksMax= in these specific unit files to a higher value, or
2828 even "infinity". Similar, there's now a logind.conf setting
2829 UserTasksMax= that defaults to 4096 and limits the total
2830 number of processes or tasks each user may own
2831 concurrently. nspawn containers also have the TasksMax=
2832 value set by default now, to 8192. Note that all of this
2833 only has an effect if the "pids" cgroup controller is
2834 enabled in the kernel. The general benefit of these changes
2835 should be a more robust and safer system, that provides a
2836 certain amount of per-service fork() bomb protection.
2838 * systemd-nspawn gained the new --network-veth-extra= switch
2839 to define additional and arbitrarily-named virtual Ethernet
2840 links between the host and the container.
2842 * A new service execution setting PassEnvironment= has been
2843 added that allows importing select environment variables
2844 from PID1's environment block into the environment block of
2847 * Timer units gained support for a new RemainAfterElapse=
2848 setting which takes a boolean argument. It defaults to on,
2849 exposing behaviour unchanged to previous releases. If set to
2850 off, timer units are unloaded after they elapsed if they
2851 cannot elapse again. This is particularly useful for
2852 transient timer units, which shall not stay around longer
2853 than until they first elapse.
2855 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
2856 default now (the kernel default is 16). This is beneficial
2857 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
2858 allows substantially larger numbers of queued
2859 datagrams. This should increase the capability of systemd to
2860 parallelize boot-up, as logging and sd_notify() are unlikely
2861 to stall execution anymore. If you need to change the value
2862 from the new defaults, use the usual sysctl.d/ snippets.
2864 * The compression framing format used by the journal or
2865 coredump processing has changed to be in line with what the
2866 official LZ4 tools generate. LZ4 compression support in
2867 systemd was considered unsupported previously, as the format
2868 was not compatible with the normal tools. With this release
2869 this has changed now, and it is hence safe for downstream
2870 distributions to turn it on. While not compressing as well
2871 as the XZ, LZ4 is substantially faster, which makes
2872 it a good default choice for the compression logic in the
2873 journal and in coredump handling.
2875 * Any reference to /etc/mtab has been dropped from
2876 systemd. The file has been obsolete since a while, but
2877 systemd refused to work on systems where it was incorrectly
2878 set up (it should be a symlink or non-existent). Please make
2879 sure to update to util-linux 2.27.1 or newer in conjunction
2880 with this systemd release, which also drops any reference to
2881 /etc/mtab. If you maintain a distribution make sure that no
2882 software you package still references it, as this is a
2883 likely source of bugs. There's also a glibc bug pending,
2884 asking for removal of any reference to this obsolete file:
2886 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
2888 Note that only util-linux versions built with
2889 --enable-libmount-force-mountinfo are supported.
2891 * Support for the ".snapshot" unit type has been removed. This
2892 feature turned out to be little useful and little used, and
2893 has now been removed from the core and from systemctl.
2895 * The dependency types RequiresOverridable= and
2896 RequisiteOverridable= have been removed from systemd. They
2897 have been used only very sparingly to our knowledge and
2898 other options that provide a similar effect (such as
2899 systemctl --mode=ignore-dependencies) are much more useful
2900 and commonly used. Moreover, they were only half-way
2901 implemented as the option to control behaviour regarding
2902 these dependencies was never added to systemctl. By removing
2903 these dependency types the execution engine becomes a bit
2904 simpler. Unit files that use these dependencies should be
2905 changed to use the non-Overridable dependency types
2906 instead. In fact, when parsing unit files with these
2907 options, that's what systemd will automatically convert them
2908 too, but it will also warn, asking users to fix the unit
2909 files accordingly. Removal of these dependency types should
2910 only affect a negligible number of unit files in the wild.
2912 * Behaviour of networkd's IPForward= option changed
2913 (again). It will no longer maintain a per-interface setting,
2914 but propagate one way from interfaces where this is enabled
2915 to the global kernel setting. The global setting will be
2916 enabled when requested by a network that is set up, but
2917 never be disabled again. This change was made to make sure
2918 IPv4 and IPv6 behaviour regarding packet forwarding is
2919 similar (as the Linux IPv6 stack does not support
2920 per-interface control of this setting) and to minimize
2923 * In unit files the behaviour of %u, %U, %h, %s has
2924 changed. These specifiers will now unconditionally resolve
2925 to the various user database fields of the user that the
2926 systemd instance is running as, instead of the user
2927 configured in the specific unit via User=. Note that this
2928 effectively doesn't change much, as resolving of these
2929 specifiers was already turned off in the --system instance
2930 of systemd, as we cannot do NSS lookups from PID 1. In the
2931 --user instance of systemd these specifiers where correctly
2932 resolved, but hardly made any sense, since the user instance
2933 lacks privileges to do user switches anyway, and User= is
2934 hence useless. Moreover, even in the --user instance of
2935 systemd behaviour was awkward as it would only take settings
2936 from User= assignment placed before the specifier into
2937 account. In order to unify and simplify the logic around
2938 this the specifiers will now always resolve to the
2939 credentials of the user invoking the manager (which in case
2940 of PID 1 is the root user).
2942 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
2943 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
2944 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
2945 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
2946 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
2947 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
2948 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
2949 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
2950 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
2951 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
2952 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
2953 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
2954 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
2955 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
2958 — Berlin, 2015-11-18
2962 * systemd now depends on util-linux v2.27. More specifically,
2963 the newly added mount monitor feature in libmount now
2964 replaces systemd's former own implementation.
2966 * libmount mandates /etc/mtab not to be regular file, and
2967 systemd now enforces this condition at early boot.
2968 /etc/mtab has been deprecated and warned about for a very
2969 long time, so systems running systemd should already have
2970 stopped having this file around as anything else than a
2971 symlink to /proc/self/mounts.
2973 * Support for the "pids" cgroup controller has been added. It
2974 allows accounting the number of tasks in a cgroup and
2975 enforcing limits on it. This adds two new setting
2976 TasksAccounting= and TasksMax= to each unit, as well as a
2977 global option DefaultTasksAccounting=.
2979 * Support for the "net_cls" cgroup controller has been added.
2980 It allows assigning a net class ID to each task in the
2981 cgroup, which can then be used in firewall rules and traffic
2982 shaping configurations. Note that the kernel netfilter net
2983 class code does not currently work reliably for ingress
2984 packets on unestablished sockets.
2986 This adds a new config directive called NetClass= to CGroup
2987 enabled units. Allowed values are positive numbers for fixed
2988 assignments and "auto" for picking a free value
2991 * 'systemctl is-system-running' now returns 'offline' if the
2992 system is not booted with systemd. This command can now be
2993 used as a substitute for 'systemd-notify --booted'.
2995 * Watchdog timeouts have been increased to 3 minutes for all
2996 in-tree service files. Apparently, disk IO issues are more
2997 frequent than we hoped, and user reported >1 minute waiting
3000 * 'machine-id-commit' functionality has been merged into
3001 'machine-id-setup --commit'. The separate binary has been
3004 * The WorkingDirectory= directive in unit files may now be set
3005 to the special value '~'. In this case, the working
3006 directory is set to the home directory of the user
3007 configured in User=.
3009 * "machinectl shell" will now open the shell in the home
3010 directory of the selected user by default.
3012 * The CrashChVT= configuration file setting is renamed to
3013 CrashChangeVT=, following our usual logic of not
3014 abbreviating unnecessarily. The old directive is still
3015 supported for compat reasons. Also, this directive now takes
3016 an integer value between 1 and 63, or a boolean value. The
3017 formerly supported '-1' value for disabling stays around for
3020 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3021 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3022 RootDirectory= properties can now be set for transient
3025 * The systemd-analyze tool gained a new "set-log-target" verb
3026 to change the logging target the system manager logs to
3027 dynamically during runtime. This is similar to how
3028 "systemd-analyze set-log-level" already changes the log
3031 * In nspawn /sys is now mounted as tmpfs, with only a selected
3032 set of subdirectories mounted in from the real sysfs. This
3033 enhances security slightly, and is useful for ensuring user
3034 namespaces work correctly.
3036 * Support for USB FunctionFS activation has been added. This
3037 allows implementation of USB gadget services that are
3038 activated as soon as they are requested, so that they don't
3039 have to run continuously, similar to classic socket
3042 * The "systemctl exit" command now optionally takes an
3043 additional parameter that sets the exit code to return from
3044 the systemd manager when exiting. This is only relevant when
3045 running the systemd user instance, or when running the
3046 system instance in a container.
3048 * sd-bus gained the new API calls sd_bus_path_encode_many()
3049 and sd_bus_path_decode_many() that allow easy encoding and
3050 decoding of multiple identifier strings inside a D-Bus
3051 object path. Another new call sd_bus_default_flush_close()
3052 has been added to flush and close per-thread default
3055 * systemd-cgtop gained support for a -M/--machine= switch to
3056 show the control groups within a certain container only.
3058 * "systemctl kill" gained support for an optional --fail
3059 switch. If specified the requested operation will fail of no
3060 processes have been killed, because the unit had no
3061 processes attached, or similar.
3063 * A new systemd.crash_reboot=1 kernel command line option has
3064 been added that triggers a reboot after crashing. This can
3065 also be set through CrashReboot= in systemd.conf.
3067 * The RuntimeDirectory= setting now understands unit
3068 specifiers like %i or %f.
3070 * A new (still internal) library API sd-ipv4acd has been added,
3071 that implements address conflict detection for IPv4. It's
3072 based on code from sd-ipv4ll, and will be useful for
3073 detecting DHCP address conflicts.
3075 * File descriptors passed during socket activation may now be
3076 named. A new API sd_listen_fds_with_names() is added to
3077 access the names. The default names may be overridden,
3078 either in the .socket file using the FileDescriptorName=
3079 parameter, or by passing FDNAME= when storing the file
3080 descriptors using sd_notify().
3082 * systemd-networkd gained support for:
3084 - Setting the IPv6 Router Advertisement settings via
3085 IPv6AcceptRouterAdvertisements= in .network files.
3087 - Configuring the HelloTimeSec=, MaxAgeSec= and
3088 ForwardDelaySec= bridge parameters in .netdev files.
3090 - Configuring PreferredSource= for static routes in
3093 * The "ask-password" framework used to query for LUKS harddisk
3094 passwords or SSL passwords during boot gained support for
3095 caching passwords in the kernel keyring, if it is
3096 available. This makes sure that the user only has to type in
3097 a passphrase once if there are multiple objects to unlock
3098 with the same one. Previously, such password caching was
3099 available only when Plymouth was used; this moves the
3100 caching logic into the systemd codebase itself. The
3101 "systemd-ask-password" utility gained a new --keyname=
3102 switch to control which kernel keyring key to use for
3103 caching a password in. This functionality is also useful for
3104 enabling display managers such as gdm to automatically
3105 unlock the user's GNOME keyring if its passphrase, the
3106 user's password and the harddisk password are the same, if
3107 gdm-autologin is used.
3109 * When downloading tar or raw images using "machinectl
3110 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3111 file is now also downloaded, if it is available and stored
3112 next to the image file.
3114 * Units of type ".socket" gained a new boolean setting
3115 Writable= which is only useful in conjunction with
3116 ListenSpecial=. If true, enables opening the specified
3117 special file in O_RDWR mode rather than O_RDONLY mode.
3119 * systemd-rfkill has been reworked to become a singleton
3120 service that is activated through /dev/rfkill on each rfkill
3121 state change and saves the settings to disk. This way,
3122 systemd-rfkill is now compatible with devices that exist
3123 only intermittendly, and even restores state if the previous
3124 system shutdown was abrupt rather than clean.
3126 * The journal daemon gained support for vacuuming old journal
3127 files controlled by the number of files that shall remain,
3128 in addition to the already existing control by size and by
3129 date. This is useful as journal interleaving performance
3130 degrades with too many separate journal files, and allows
3131 putting an effective limit on them. The new setting defaults
3132 to 100, but this may be changed by setting SystemMaxFiles=
3133 and RuntimeMaxFiles= in journald.conf. Also, the
3134 "journalctl" tool gained the new --vacuum-files= switch to
3135 manually vacuum journal files to leave only the specified
3136 number of files in place.
3138 * udev will now create /dev/disk/by-path links for ATA devices
3139 on kernels where that is supported.
3141 * Galician, Serbian, Turkish and Korean translations were added.
3143 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3144 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3145 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3146 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3147 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3148 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3149 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3150 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3151 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3152 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3153 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3154 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3155 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3156 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3157 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3158 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3159 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3160 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3162 — Berlin, 2015-10-07
3166 * The DHCP implementation of systemd-networkd gained a set of
3169 - The DHCP server now supports emitting DNS and NTP
3170 information. It may be enabled and configured via
3171 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3172 and NTP information is enabled, but no servers are
3173 configured, the corresponding uplink information (if there
3174 is any) is propagated.
3176 - Server and client now support transmission and reception
3177 of timezone information. It can be configured via the
3178 newly introduced network options UseTimezone=,
3179 EmitTimezone=, and Timezone=. Transmission of timezone
3180 information is enabled between host and containers by
3181 default now: the container will change its local timezone
3182 to what the host has set.
3184 - Lease timeouts can now be configured via
3185 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3187 - The DHCP server improved on the stability of
3188 leases. Clients are more likely to get the same lease
3189 information back, even if the server loses state.
3191 - The DHCP server supports two new configuration options to
3192 control the lease address pool metrics, PoolOffset= and
3195 * The encapsulation limit of tunnels in systemd-networkd may
3196 now be configured via 'EncapsulationLimit='. It allows
3197 modifying the maximum additional levels of encapsulation
3198 that are permitted to be prepended to a packet.
3200 * systemd now supports the concept of user buses replacing
3201 session buses, if used with dbus-1.10 (and enabled via dbus
3202 --enable-user-session). It previously only supported this on
3203 kdbus-enabled systems, and this release expands this to
3204 'dbus-daemon' systems.
3206 * systemd-networkd now supports predictable interface names
3209 * systemd now optionally supports the new Linux kernel
3210 "unified" control group hierarchy. If enabled via the kernel
3211 command-line option 'systemd.unified_cgroup_hierarchy=1',
3212 systemd will try to mount the unified cgroup hierarchy
3213 directly on /sys/fs/cgroup. If not enabled, or not
3214 available, systemd will fall back to the legacy cgroup
3215 hierarchy setup, as before. Host system and containers can
3216 mix and match legacy and unified hierarchies as they
3217 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
3218 environment variable to individually select the hierarchy to
3219 use for executed containers. By default, nspawn will use the
3220 unified hierarchy for the containers if the host uses the
3221 unified hierarchy, and the legacy hierarchy otherwise.
3222 Please note that at this point the unified hierarchy is an
3223 experimental kernel feature and is likely to change in one
3224 of the next kernel releases. Therefore, it should not be
3225 enabled by default in downstream distributions yet. The
3226 minimum required kernel version for the unified hierarchy to
3227 work is 4.2. Note that when the unified hierarchy is used
3228 for the first time delegated access to controllers is
3229 safe. Because of this systemd-nspawn containers will get
3230 access to controllers now, as will systemd user
3231 sessions. This means containers and user sessions may now
3232 manage their own resources, partitioning up what the system
3235 * A new special scope unit "init.scope" has been introduced
3236 that encapsulates PID 1 of the system. It may be used to
3237 determine resource usage and enforce resource limits on PID
3238 1 itself. PID 1 hence moved out of the root of the control
3241 * The cgtop tool gained support for filtering out kernel
3242 threads when counting tasks in a control group. Also, the
3243 count of processes is now recursively summed up by
3244 default. Two options -k and --recursive= have been added to
3245 revert to old behaviour. The tool has also been updated to
3246 work correctly in containers now.
3248 * systemd-nspawn's --bind= and --bind-ro= options have been
3249 extended to allow creation of non-recursive bind mounts.
3251 * libsystemd gained two new calls sd_pid_get_cgroup() and
3252 sd_peer_get_cgroup() which return the control group path of
3253 a process or peer of a connected AF_UNIX socket. This
3254 function call is particularly useful when implementing
3255 delegated subtrees support in the control group hierarchy.
3257 * The "sd-event" event loop API of libsystemd now supports
3258 correct dequeuing of real-time signals, without losing
3261 * When systemd requests a PolicyKit decision when managing
3262 units it will now add additional fields to the request,
3263 including unit name and desired operation. This enables more
3264 powerful PolicyKit policies, that make decisions depending
3265 on these parameters.
3267 * nspawn learnt support for .nspawn settings files, that may
3268 accompany the image files or directories of containers, and
3269 may contain additional settings for the container. This is
3270 an alternative to configuring container parameters via the
3271 nspawn command line.
3273 Contributions from: Cristian Rodríguez, Daniel Mack, David
3274 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3275 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3276 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3277 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3278 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3279 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
3280 Andersen, Tom Gundersen, Torstein Husebø
3282 — Berlin, 2015-09-08
3286 * machinectl gained a new verb 'shell' which opens a fresh
3287 shell on the target container or the host. It is similar to
3288 the existing 'login' command of machinectl, but spawns the
3289 shell directly without prompting for username or
3290 password. The pseudo machine '.host' now refers to the local
3291 host and is used by default. Hence, 'machinectl shell' can
3292 be used as replacement for 'su -' which spawns a session as
3293 a fresh systemd unit in a way that is fully isolated from
3294 the originating session.
3296 * systemd-networkd learned to cope with private-zone DHCP
3297 options and allows other programs to query the values.
3299 * SELinux access control when enabling/disabling units is no
3300 longer enforced with this release. The previous
3301 implementation was incorrect, and a new corrected
3302 implementation is not yet available. As unit file operations
3303 are still protected via PolicyKit and D-Bus policy this is
3304 not a security problem. Yet, distributions which care about
3305 optimal SELinux support should probably not stabilize on
3308 * sd-bus gained support for matches of type "arg0has=", that
3309 test for membership of strings in string arrays sent in bus
3312 * systemd-resolved now dumps the contents of its DNS and LLMNR
3313 caches to the logs on reception of the SIGUSR1 signal. This
3314 is useful to debug DNS behaviour.
3316 * The coredumpctl tool gained a new --directory= option to
3317 operate on journal files in a specific directory.
3319 * "systemctl reboot" and related commands gained a new
3320 "--message=" option which may be used to set a free-text
3321 wall message when shutting down or rebooting the
3322 system. This message is also logged, which is useful for
3323 figuring out the reason for a reboot or shutdown a
3326 * The "systemd-resolve-host" tool's -i switch now takes
3327 network interface numbers as alternative to interface names.
3329 * A new unit file setting for services has been introduced:
3330 UtmpMode= allows configuration of how precisely systemd
3331 handles utmp and wtmp entries for the service if this is
3332 enabled. This allows writing services that appear similar to
3333 user sessions in the output of the "w", "who", "last" and
3336 * systemd-resolved will now locally synthesize DNS resource
3337 records for the "localhost" and "gateway" domains as well as
3338 the local hostname. This should ensure that clients querying
3339 RRs via resolved will get similar results as those going via
3340 NSS, if nss-myhostname is enabled.
3342 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3343 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3344 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3345 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3346 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3347 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3348 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3349 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3350 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3351 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3352 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3353 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3355 — Berlin, 2015-08-27
3359 * The systemd-efi-boot-generator functionality was merged into
3360 systemd-gpt-auto-generator.
3362 * systemd-networkd now supports Group Policy for vxlan
3363 devices. It can be enabled via the new boolean configuration
3364 option called 'GroupPolicyExtension='.
3366 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3367 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3368 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3370 — Berlin, 2015-07-31
3374 * The python-systemd code has been removed from the systemd repository.
3375 A new repository has been created which accommodates the code from
3376 now on, and we kindly ask distributions to create a separate package
3377 for this: https://github.com/systemd/python-systemd
3379 * The systemd daemon will now reload its main configuration
3380 (/etc/systemd/system.conf) on daemon-reload.
3382 * sd-dhcp now exposes vendor specific extensions via
3383 sd_dhcp_lease_get_vendor_specific().
3385 * systemd-networkd gained a number of new configuration options.
3387 - A new boolean configuration option for TAP devices called
3388 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
3389 device, thus allowing to send and receive GSO packets.
3391 - A new tunnel configuration option called 'CopyDSCP='.
3392 If enabled, the DSCP field of ip6 tunnels is copied into the
3393 decapsulated packet.
3395 - A set of boolean bridge configuration options were added.
3396 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3397 and 'UnicastFlood=' are now parsed by networkd and applied to the
3398 respective bridge link device via the respective IFLA_BRPORT_*
3401 - A new string configuration option to override the hostname sent
3402 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3403 is true, networkd will use the configured hostname instead of the
3404 system hostname when sending DHCP requests.
3406 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3407 networkd will configure the IPv6 flow-label of the tunnel device
3408 according to RFC2460.
3410 - The 'macvtap' virtual network devices are now supported, similar to
3411 the already supported 'macvlan' devices.
3413 * systemd-resolved now implements RFC5452 to improve resilience against
3414 cache poisoning. Additionally, source port randomization is enabled
3415 by default to further protect against DNS spoofing attacks.
3417 * nss-mymachines now supports translating UIDs and GIDs of running
3418 containers with user-namespaces enabled. If a container 'foo'
3419 translates a host uid 'UID' to the container uid 'TUID', then
3420 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3421 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3422 mapped as 'vg-foo-TGID'.
3424 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
3425 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3426 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3427 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3428 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3429 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3430 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3431 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3432 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3433 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3435 — Berlin, 2015-07-29
3439 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3440 There are no known issues with current sysfs, and udev does not need
3441 or should be used to work around such bugs.
3443 * udev does no longer enable USB HID power management. Several reports
3444 indicate, that some devices cannot handle that setting.
3446 * The udev accelerometer helper was removed. The functionality
3447 is now fully included in iio-sensor-proxy. But this means,
3448 older iio-sensor-proxy versions will no longer provide
3449 accelerometer/orientation data with this systemd version.
3450 Please upgrade iio-sensor-proxy to version 1.0.
3452 * networkd gained a new configuration option IPv6PrivacyExtensions=
3453 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3454 for Stateless Address") on selected networks.
3456 * For the sake of fewer build-time dependencies and less code in the
3457 main repository, the python bindings are about to be removed in the
3458 next release. A new repository has been created which accommodates
3459 the code from now on, and we kindly ask distributions to create a
3460 separate package for this. The removal will take place in v223.
3462 https://github.com/systemd/python-systemd
3464 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3465 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3466 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3467 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
3468 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3469 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
3470 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3471 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
3472 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3473 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3475 — Berlin, 2015-07-07
3479 * The sd-bus.h and sd-event.h APIs have now been declared
3480 stable and have been added to the official interface of
3481 libsystemd.so. sd-bus implements an alternative D-Bus client
3482 library, that is relatively easy to use, very efficient and
3483 supports both classic D-Bus as well as kdbus as transport
3484 backend. sd-event is a generic event loop abstraction that
3485 is built around Linux epoll, but adds features such as event
3486 prioritization or efficient timer handling. Both APIs are good
3487 choices for C programs looking for a bus and/or event loop
3488 implementation that is minimal and does not have to be
3489 portable to other kernels.
3491 * kdbus support is no longer compile-time optional. It is now
3492 always built-in. However, it can still be disabled at
3493 runtime using the kdbus=0 kernel command line setting, and
3494 that setting may be changed to default to off, by specifying
3495 --disable-kdbus at build-time. Note though that the kernel
3496 command line setting has no effect if the kdbus.ko kernel
3497 module is not installed, in which case kdbus is (obviously)
3498 also disabled. We encourage all downstream distributions to
3499 begin testing kdbus by adding it to the kernel images in the
3500 development distributions, and leaving kdbus support in
3503 * The minimal required util-linux version has been bumped to
3506 * Support for chkconfig (--enable-chkconfig) was removed in
3507 favor of calling an abstraction tool
3508 /lib/systemd/systemd-sysv-install. This needs to be
3509 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3510 in README for details.
3512 * If there's a systemd unit and a SysV init script for the
3513 same service name, and the user executes "systemctl enable"
3514 for it (or a related call), then this will now enable both
3515 (or execute the related operation on both), not just the
3518 * The libudev API documentation has been converted from gtkdoc
3521 * gudev has been removed from the systemd tree, it is now an
3524 * The systemd-cgtop tool learnt a new --raw switch to generate
3525 "raw" (machine parsable) output.
3527 * networkd's IPForwarding= .network file setting learnt the
3528 new setting "kernel", which ensures that networkd does not
3529 change the IP forwarding sysctl from the default kernel
3532 * The systemd-logind bus API now exposes a new boolean
3533 property "Docked" that reports whether logind considers the
3534 system "docked", i.e. connected to a docking station or not.
3536 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3537 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3538 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3539 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3540 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3541 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3542 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3543 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3544 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3545 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3546 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
3547 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3548 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3549 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3550 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3551 Fink, Zbigniew Jędrzejewski-Szmek
3553 — Berlin, 2015-06-19
3557 * The gudev library has been extracted into a separate repository
3558 available at: https://git.gnome.org/browse/libgudev/
3559 It is now managed as part of the Gnome project. Distributions
3560 are recommended to pass --disable-gudev to systemd and use
3561 gudev from the Gnome project instead. gudev is still included
3562 in systemd, for now. It will be removed soon, though. Please
3563 also see the announcement-thread on systemd-devel:
3564 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
3566 * systemd now exposes a CPUUsageNSec= property for each
3567 service unit on the bus, that contains the overall consumed
3568 CPU time of a service (the sum of what each process of the
3569 service consumed). This value is only available if
3570 CPUAccounting= is turned on for a service, and is then shown
3571 in the "systemctl status" output.
3573 * Support for configuring alternative mappings of the old SysV
3574 runlevels to systemd targets has been removed. They are now
3575 hardcoded in a way that runlevels 2, 3, 4 all map to
3576 multi-user.target and 5 to graphical.target (which
3577 previously was already the default behaviour).
3579 * The auto-mounter logic gained support for mount point
3580 expiry, using a new TimeoutIdleSec= setting in .automount
3581 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3583 * The EFI System Partition (ESP) as mounted to /boot by
3584 systemd-efi-boot-generator will now be unmounted
3585 automatically after 2 minutes of not being used. This should
3586 minimize the risk of ESP corruptions.
3588 * New /etc/fstab options x-systemd.requires= and
3589 x-systemd.requires-mounts-for= are now supported to express
3590 additional dependencies for mounts. This is useful for
3591 journalling file systems that support external journal
3592 devices or overlay file systems that require underlying file
3593 systems to be mounted.
3595 * systemd does not support direct live-upgrades (via systemctl
3596 daemon-reexec) from versions older than v44 anymore. As no
3597 distribution we are aware of shipped such old versions in a
3598 stable release this should not be problematic.
3600 * When systemd forks off a new per-connection service instance
3601 it will now set the $REMOTE_ADDR environment variable to the
3602 remote IP address, and $REMOTE_PORT environment variable to
3603 the remote IP port. This behaviour is similar to the
3604 corresponding environment variables defined by CGI.
3606 * systemd-networkd gained support for uplink failure
3607 detection. The BindCarrier= option allows binding interface
3608 configuration dynamically to the link sense of other
3609 interfaces. This is useful to achieve behaviour like in
3612 * systemd-networkd gained support for configuring the DHCP
3613 client identifier to use when requesting leases.
3615 * systemd-networkd now has a per-network UseNTP= option to
3616 configure whether NTP server information acquired via DHCP
3617 is passed on to services like systemd-timesyncd.
3619 * systemd-networkd gained support for vti6 tunnels.
3621 * Note that systemd-networkd manages the sysctl variable
3622 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3623 it is configured for since v219. The variable controls IP
3624 forwarding, and is a per-interface alternative to the global
3625 /proc/sys/net/ipv[46]/ip_forward. This setting is
3626 configurable in the IPForward= option, which defaults to
3627 "no". This means if networkd is used for an interface it is
3628 no longer sufficient to set the global sysctl option to turn
3629 on IP forwarding! Instead, the .network file option
3630 IPForward= needs to be turned on! Note that the
3631 implementation of this behaviour was broken in v219 and has
3634 * Many bonding and vxlan options are now configurable in
3637 * systemd-nspawn gained a new --property= setting to set unit
3638 properties for the container scope. This is useful for
3639 setting resource parameters (e.g. "CPUShares=500") on
3640 containers started from the command line.
3642 * systemd-nspawn gained a new --private-users= switch to make
3643 use of user namespacing available on recent Linux kernels.
3645 * systemd-nspawn may now be called as part of a shell pipeline
3646 in which case the pipes used for stdin and stdout are passed
3647 directly to the process invoked in the container, without
3648 indirection via a pseudo tty.
3650 * systemd-nspawn gained a new switch to control the UNIX
3651 signal to use when killing the init process of the container
3654 * systemd-nspawn gained a new --overlay= switch for mounting
3655 overlay file systems into the container using the new kernel
3658 * When a container image is imported via systemd-importd and
3659 the host file system is not btrfs, a loopback block device
3660 file is created in /var/lib/machines.raw with a btrfs file
3661 system inside. It is then mounted to /var/lib/machines to
3662 enable btrfs features for container management. The loopback
3663 file and btrfs file system is grown as needed when container
3664 images are imported via systemd-importd.
3666 * systemd-machined/systemd-importd gained support for btrfs
3667 quota, to enforce container disk space limits on disk. This
3668 is exposed in "machinectl set-limit".
3670 * systemd-importd now can import containers from local .tar,
3671 .raw and .qcow2 images, and export them to .tar and .raw. It
3672 can also import dkr v2 images now from the network (on top
3675 * systemd-importd gained support for verifying downloaded
3676 images with gpg2 (previously only gpg1 was supported).
3678 * systemd-machined, systemd-logind, systemd: most bus calls
3679 are now accessible to unprivileged processes via
3680 PolicyKit. Also, systemd-logind will now allow users to kill
3681 their own sessions without further privileges or
3684 * systemd-shutdownd has been removed. This service was
3685 previously responsible for implementing scheduled shutdowns
3686 as exposed in /usr/bin/shutdown's time parameter. This
3687 functionality has now been moved into systemd-logind and is
3688 accessible via a bus interface.
3690 * "systemctl reboot" gained a new switch --firmware-setup that
3691 can be used to reboot into the EFI firmware setup, if that
3692 is available. systemd-logind now exposes an API on the bus
3693 to trigger such reboots, in case graphical desktop UIs want
3694 to cover this functionality.
3696 * "systemctl enable", "systemctl disable" and "systemctl mask"
3697 now support a new "--now" switch. If specified the units
3698 that are enabled will also be started, and the ones
3699 disabled/masked also stopped.
3701 * The Gummiboot EFI boot loader tool has been merged into
3702 systemd, and renamed to "systemd-boot". The bootctl tool has been
3703 updated to support systemd-boot.
3705 * An EFI kernel stub has been added that may be used to create
3706 kernel EFI binaries that contain not only the actual kernel,
3707 but also an initrd, boot splash, command line and OS release
3708 information. This combined binary can then be signed as a
3709 single image, so that the firmware can verify it all in one
3710 step. systemd-boot has special support for EFI binaries created
3711 like this and can extract OS release information from them
3712 and show them in the boot menu. This functionality is useful
3713 to implement cryptographically verified boot schemes.
3715 * Optional support has been added to systemd-fsck to pass
3716 fsck's progress report to an AF_UNIX socket in the file
3719 * udev will no longer create device symlinks for all block
3720 devices by default. A blacklist for excluding special block
3721 devices from this logic has been turned into a whitelist
3722 that requires picking block devices explicitly that require
3725 * A new (currently still internal) API sd-device.h has been
3726 added to libsystemd. This modernized API is supposed to
3727 replace libudev eventually. In fact, already much of libudev
3728 is now just a wrapper around sd-device.h.
3730 * A new hwdb database for storing metadata about pointing
3731 stick devices has been added.
3733 * systemd-tmpfiles gained support for setting file attributes
3734 similar to the "chattr" tool with new 'h' and 'H' lines.
3736 * systemd-journald will no longer unconditionally set the
3737 btrfs NOCOW flag on new journal files. This is instead done
3738 with tmpfiles snippet using the new 'h' line type. This
3739 allows easy disabling of this logic, by masking the
3740 journal-nocow.conf tmpfiles file.
3742 * systemd-journald will now translate audit message types to
3743 human readable identifiers when writing them to the
3744 journal. This should improve readability of audit messages.
3746 * The LUKS logic gained support for the offset= and skip=
3747 options in /etc/crypttab, as previously implemented by
3750 * /usr/lib/os-release gained a new optional field VARIANT= for
3751 distributions that support multiple variants (such as a
3752 desktop edition, a server edition, ...)
3754 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3755 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3756 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3757 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3758 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3759 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3760 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3761 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3762 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3763 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3764 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3765 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3766 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3767 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3768 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3769 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3770 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3771 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3772 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3773 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3774 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3775 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3776 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3777 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3778 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3779 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3780 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3782 — Berlin, 2015-05-22
3786 * Introduce a new API "sd-hwdb.h" for querying the hardware
3787 metadata database. With this minimal interface one can query
3788 and enumerate the udev hwdb, decoupled from the old libudev
3789 library. libudev's interface for this is now only a wrapper
3790 around sd-hwdb. A new tool systemd-hwdb has been added to
3791 interface with and update the database.
3793 * When any of systemd's tools copies files (for example due to
3794 tmpfiles' C lines) a btrfs reflink will attempted first,
3795 before bytewise copying is done.
3797 * systemd-nspawn gained a new --ephemeral switch. When
3798 specified a btrfs snapshot is taken of the container's root
3799 directory, and immediately removed when the container
3800 terminates again. Thus, a container can be started whose
3801 changes never alter the container's root directory, and are
3802 lost on container termination. This switch can also be used
3803 for starting a container off the root file system of the
3804 host without affecting the host OS. This switch is only
3805 available on btrfs file systems.
3807 * systemd-nspawn gained a new --template= switch. It takes the
3808 path to a container tree to use as template for the tree
3809 specified via --directory=, should that directory be
3810 missing. This allows instantiating containers dynamically,
3811 on first run. This switch is only available on btrfs file
3814 * When a .mount unit refers to a mount point on which multiple
3815 mounts are stacked, and the .mount unit is stopped all of
3816 the stacked mount points will now be unmounted until no
3817 mount point remains.
3819 * systemd now has an explicit notion of supported and
3820 unsupported unit types. Jobs enqueued for unsupported unit
3821 types will now fail with an "unsupported" error code. More
3822 specifically .swap, .automount and .device units are not
3823 supported in containers, .busname units are not supported on
3824 non-kdbus systems. .swap and .automount are also not
3825 supported if their respective kernel compile time options
3828 * machinectl gained support for two new "copy-from" and
3829 "copy-to" commands for copying files from a running
3830 container to the host or vice versa.
3832 * machinectl gained support for a new "bind" command to bind
3833 mount host directories into local containers. This is
3834 currently only supported for nspawn containers.
3836 * networkd gained support for configuring bridge forwarding
3837 database entries (fdb) from .network files.
3839 * A new tiny daemon "systemd-importd" has been added that can
3840 download container images in tar, raw, qcow2 or dkr formats,
3841 and make them available locally in /var/lib/machines, so
3842 that they can run as nspawn containers. The daemon can GPG
3843 verify the downloads (not supported for dkr, since it has no
3844 provisions for verifying downloads). It will transparently
3845 decompress bz2, xz, gzip compressed downloads if necessary,
3846 and restore sparse files on disk. The daemon uses privilege
3847 separation to ensure the actual download logic runs with
3848 fewer privileges than the daemon itself. machinectl has
3849 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3850 make the functionality of importd available to the
3851 user. With this in place the Fedora and Ubuntu "Cloud"
3852 images can be downloaded and booted as containers unmodified
3853 (the Fedora images lack the appropriate GPG signature files
3854 currently, so they cannot be verified, but this will change
3855 soon, hopefully). Note that downloading images is currently
3856 only fully supported on btrfs.
3858 * machinectl is now able to list container images found in
3859 /var/lib/machines, along with some metadata about sizes of
3860 disk and similar. If the directory is located on btrfs and
3861 quota is enabled, this includes quota display. A new command
3862 "image-status" has been added that shows additional
3863 information about images.
3865 * machinectl is now able to clone container images
3866 efficiently, if the underlying file system (btrfs) supports
3867 it, with the new "machinectl clone" command. It also
3868 gained commands for renaming and removing images, as well as
3869 marking them read-only or read-write (supported also on
3870 legacy file systems).
3872 * networkd gained support for collecting LLDP network
3873 announcements, from hardware that supports this. This is
3874 shown in networkctl output.
3876 * systemd-run gained support for a new -t (--pty) switch for
3877 invoking a binary on a pty whose input and output is
3878 connected to the invoking terminal. This allows executing
3879 processes as system services while interactively
3880 communicating with them via the terminal. Most interestingly
3881 this is supported across container boundaries. Invoking
3882 "systemd-run -t /bin/bash" is an alternative to running a
3883 full login session, the difference being that the former
3884 will not register a session, nor go through the PAM session
3887 * tmpfiles gained support for a new "v" line type for creating
3888 btrfs subvolumes. If the underlying file system is a legacy
3889 file system, this automatically degrades to creating a
3890 normal directory. Among others /var/lib/machines is now
3891 created like this at boot, should it be missing.
3893 * The directory /var/lib/containers/ has been deprecated and
3894 been replaced by /var/lib/machines. The term "machines" has
3895 been used in the systemd context as generic term for both
3896 VMs and containers, and hence appears more appropriate for
3897 this, as the directory can also contain raw images bootable
3900 * systemd-nspawn when invoked with -M but without --directory=
3901 or --image= is now capable of searching for the container
3902 root directory, subvolume or disk image automatically, in
3903 /var/lib/machines. systemd-nspawn@.service has been updated
3904 to make use of this, thus allowing it to be used for raw
3907 * A new machines.target unit has been introduced that is
3908 supposed to group all containers/VMs invoked as services on
3909 the system. systemd-nspawn@.service has been updated to
3910 integrate with that.
3912 * machinectl gained a new "start" command, for invoking a
3913 container as a service. "machinectl start foo" is mostly
3914 equivalent to "systemctl start systemd-nspawn@foo.service",
3915 but handles escaping in a nicer way.
3917 * systemd-nspawn will now mount most of the cgroupfs tree
3918 read-only into each container, with the exception of the
3919 container's own subtree in the name=systemd hierarchy.
3921 * journald now sets the special FS_NOCOW file flag for its
3922 journal files. This should improve performance on btrfs, by
3923 avoiding heavy fragmentation when journald's write-pattern
3924 is used on COW file systems. It degrades btrfs' data
3925 integrity guarantees for the files to the same levels as for
3926 ext3/ext4 however. This should be OK though as journald does
3927 its own data integrity checks and all its objects are
3928 checksummed on disk. Also, journald should handle btrfs disk
3929 full events a lot more gracefully now, by processing SIGBUS
3930 errors, and not relying on fallocate() anymore.
3932 * When journald detects that journal files it is writing to
3933 have been deleted it will immediately start new journal
3936 * systemd now provides a way to store file descriptors
3937 per-service in PID 1. This is useful for daemons to ensure
3938 that fds they require are not lost during a daemon
3939 restart. The fds are passed to the daemon on the next
3940 invocation in the same way socket activation fds are
3941 passed. This is now used by journald to ensure that the
3942 various sockets connected to all the system's stdout/stderr
3943 are not lost when journald is restarted. File descriptors
3944 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
3945 an extension to sd_notify(). Note that a limit is enforced
3946 on the number of fds a service can store in PID 1, and it
3947 defaults to 0, so that no fds may be stored, unless this is
3948 explicitly turned on.
3950 * The default TERM variable to use for units connected to a
3951 terminal, when no other value is explicitly is set is now
3952 vt220 rather than vt102. This should be fairly safe still,
3953 but allows PgUp/PgDn work.
3955 * The /etc/crypttab option header= as known from Debian is now
3958 * "loginctl user-status" and "loginctl session-status" will
3959 now show the last 10 lines of log messages of the
3960 user/session following the status output. Similar,
3961 "machinectl status" will show the last 10 log lines
3962 associated with a virtual machine or container
3963 service. (Note that this is usually not the log messages
3964 done in the VM/container itself, but simply what the
3965 container manager logs. For nspawn this includes all console
3968 * "loginctl session-status" without further argument will now
3969 show the status of the session of the caller. Similar,
3970 "lock-session", "unlock-session", "activate",
3971 "enable-linger", "disable-linger" may now be called without
3972 session/user parameter in which case they apply to the
3973 caller's session/user.
3975 * An X11 session scriptlet is now shipped that uploads
3976 $DISPLAY and $XAUTHORITY into the environment of the systemd
3977 --user daemon if a session begins. This should improve
3978 compatibility with X11 enabled applications run as systemd
3981 * Generators are now subject to masking via /etc and /run, the
3982 same way as unit files.
3984 * networkd .network files gained support for configuring
3985 per-link IPv4/IPv6 packet forwarding as well as IPv4
3986 masquerading. This is by default turned on for veth links to
3987 containers, as registered by systemd-nspawn. This means that
3988 nspawn containers run with --network-veth will now get
3989 automatic routed access to the host's networks without any
3990 further configuration or setup, as long as networkd runs on
3993 * systemd-nspawn gained the --port= (-p) switch to expose TCP
3994 or UDP posts of a container on the host. With this in place
3995 it is possible to run containers with private veth links
3996 (--network-veth), and have their functionality exposed on
3997 the host as if their services were running directly on the
4000 * systemd-nspawn's --network-veth switch now gained a short
4001 version "-n", since with the changes above it is now truly
4002 useful out-of-the-box. The systemd-nspawn@.service has been
4003 updated to make use of it too by default.
4005 * systemd-nspawn will now maintain a per-image R/W lock, to
4006 ensure that the same image is not started more than once
4007 writable. (It's OK to run an image multiple times
4008 simultaneously in read-only mode.)
4010 * systemd-nspawn's --image= option is now capable of
4011 dissecting and booting MBR and GPT disk images that contain
4012 only a single active Linux partition. Previously it
4013 supported only GPT disk images with proper GPT type
4014 IDs. This allows running cloud images from major
4015 distributions directly with systemd-nspawn, without
4018 * In addition to collecting mouse dpi data in the udev
4019 hardware database, there's now support for collecting angle
4020 information for mouse scroll wheels. The database is
4021 supposed to guarantee similar scrolling behavior on mice
4022 that it knows about. There's also support for collecting
4023 information about Touchpad types.
4025 * udev's input_id built-in will now also collect touch screen
4026 dimension data and attach it to probed devices.
4028 * /etc/os-release gained support for a Distribution Privacy
4031 * networkd gained support for creating "ipvlan", "gretap",
4032 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4034 * systemd-tmpfiles gained support for "a" lines for setting
4037 * systemd-nspawn will now mount /tmp in the container to
4038 tmpfs, automatically.
4040 * systemd now exposes the memory.usage_in_bytes cgroup
4041 attribute and shows it for each service in the "systemctl
4042 status" output, if available.
4044 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4045 immediate reboot is triggered. This useful if shutdown is
4046 hung and is unable to complete, to expedite the
4047 operation. Note that this kind of reboot will still unmount
4048 all file systems, and hence should not result in fsck being
4051 * A .device unit for an optical block device will now be
4052 considered active only when a medium is in the drive. Also,
4053 mount units are now bound to their backing devices thus
4054 triggering automatic unmounting when devices become
4055 unavailable. With this in place systemd will now
4056 automatically unmount left-over mounts when a CD-ROM is
4057 ejected or an USB stick is yanked from the system.
4059 * networkd-wait-online now has support for waiting for
4060 specific interfaces only (with globbing), and for giving up
4061 after a configurable timeout.
4063 * networkd now exits when idle. It will be automatically
4064 restarted as soon as interfaces show up, are removed or
4065 change state. networkd will stay around as long as there is
4066 at least one DHCP state machine or similar around, that keep
4069 * networkd may now configure IPv6 link-local addressing in
4070 addition to IPv4 link-local addressing.
4072 * The IPv6 "token" for use in SLAAC may now be configured for
4073 each .network interface in networkd.
4075 * Routes configured with networkd may now be assigned a scope
4078 * networkd's [Match] sections now support globbing and lists
4079 of multiple space-separated matches per item.
4081 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4082 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4083 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4084 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4085 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4086 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4087 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4088 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4089 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4090 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4091 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4092 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4093 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4094 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4095 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
4096 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4097 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4098 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4099 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4100 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4101 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4102 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
4103 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4104 Hoffmann, Zbigniew Jędrzejewski-Szmek
4106 — Berlin, 2015-02-16
4110 * When querying unit file enablement status (for example via
4111 "systemctl is-enabled"), a new state "indirect" is now known
4112 which indicates that a unit might not be enabled itself, but
4113 another unit listed in its Also= setting might be.
4115 * Similar to the various existing ConditionXYZ= settings for
4116 units, there are now matching AssertXYZ= settings. While
4117 failing conditions cause a unit to be skipped, but its job
4118 to succeed, failing assertions declared like this will cause
4119 a unit start operation and its job to fail.
4121 * hostnamed now knows a new chassis type "embedded".
4123 * systemctl gained a new "edit" command. When used on a unit
4124 file, this allows extending unit files with .d/ drop-in
4125 configuration snippets or editing the full file (after
4126 copying it from /usr/lib to /etc). This will invoke the
4127 user's editor (as configured with $EDITOR), and reload the
4128 modified configuration after editing.
4130 * "systemctl status" now shows the suggested enablement state
4131 for a unit, as declared in the (usually vendor-supplied)
4132 system preset files.
4134 * nss-myhostname will now resolve the single-label host name
4135 "gateway" to the locally configured default IP routing
4136 gateways, ordered by their metrics. This assigns a stable
4137 name to the used gateways, regardless which ones are
4138 currently configured. Note that the name will only be
4139 resolved after all other name sources (if nss-myhostname is
4140 configured properly) and should hence not negatively impact
4141 systems that use the single-label host name "gateway" in
4144 * systemd-inhibit now allows filtering by mode when listing
4147 * Scope and service units gained a new "Delegate" boolean
4148 property, which, when set, allows processes running inside the
4149 unit to further partition resources. This is primarily
4150 useful for systemd user instances as well as container
4153 * journald will now pick up audit messages directly from
4154 the kernel, and log them like any other log message. The
4155 audit fields are split up and fully indexed. This means that
4156 journalctl in many ways is now a (nicer!) alternative to
4157 ausearch, the traditional audit client. Note that this
4158 implements only a minimal audit client. If you want the
4159 special audit modes like reboot-on-log-overflow, please use
4160 the traditional auditd instead, which can be used in
4161 parallel to journald.
4163 * The ConditionSecurity= unit file option now understands the
4164 special string "audit" to check whether auditing is
4167 * journalctl gained two new commands --vacuum-size= and
4168 --vacuum-time= to delete old journal files until the
4169 remaining ones take up no more than the specified size on disk,
4170 or are not older than the specified time.
4172 * A new, native PPPoE library has been added to sd-network,
4173 systemd's library of light-weight networking protocols. This
4174 library will be used in a future version of networkd to
4175 enable PPPoE communication without an external pppd daemon.
4177 * The busctl tool now understands a new "capture" verb that
4178 works similar to "monitor", but writes a packet capture
4179 trace to STDOUT that can be redirected to a file which is
4180 compatible with libcap's capture file format. This can then
4181 be loaded in Wireshark and similar tools to inspect bus
4184 * The busctl tool now understands a new "tree" verb that shows
4185 the object trees of a specific service on the bus, or of all
4188 * The busctl tool now understands a new "introspect" verb that
4189 shows all interfaces and members of objects on the bus,
4190 including their signature and values. This is particularly
4191 useful to get more information about bus objects shown by
4192 the new "busctl tree" command.
4194 * The busctl tool now understands new verbs "call",
4195 "set-property" and "get-property" for invoking bus method
4196 calls, setting and getting bus object properties in a
4199 * busctl gained a new --augment-creds= argument that controls
4200 whether the tool shall augment credential information it
4201 gets from the bus with data from /proc, in a possibly
4204 * nspawn's --link-journal= switch gained two new values
4205 "try-guest" and "try-host" that work like "guest" and
4206 "host", but do not fail if the host has no persistent
4207 journalling enabled. -j is now equivalent to
4208 --link-journal=try-guest.
4210 * macvlan network devices created by nspawn will now have
4211 stable MAC addresses.
4213 * A new SmackProcessLabel= unit setting has been added, which
4214 controls the SMACK security label processes forked off by
4215 the respective unit shall use.
4217 * If compiled with --enable-xkbcommon, systemd-localed will
4218 verify x11 keymap settings by compiling the given keymap. It
4219 will spew out warnings if the compilation fails. This
4220 requires libxkbcommon to be installed.
4222 * When a coredump is collected, a larger number of metadata
4223 fields is now collected and included in the journal records
4224 created for it. More specifically, control group membership,
4225 environment variables, memory maps, working directory,
4226 chroot directory, /proc/$PID/status, and a list of open file
4227 descriptors is now stored in the log entry.
4229 * The udev hwdb now contains DPI information for mice. For
4232 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4234 * All systemd programs that read standalone configuration
4235 files in /etc now also support a corresponding series of
4236 .conf.d configuration directories in /etc/, /run/,
4237 /usr/local/lib/, /usr/lib/, and (if configured with
4238 --enable-split-usr) /lib/. In particular, the following
4239 configuration files now have corresponding configuration
4240 directories: system.conf user.conf, logind.conf,
4241 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4242 resolved.conf, timesyncd.conf, journal-remote.conf, and
4243 journal-upload.conf. Note that distributions should use the
4244 configuration directories in /usr/lib/; the directories in
4245 /etc/ are reserved for the system administrator.
4247 * systemd-rfkill will no longer take the rfkill device name
4248 into account when storing rfkill state on disk, as the name
4249 might be dynamically assigned and not stable. Instead, the
4250 ID_PATH udev variable combined with the rfkill type (wlan,
4251 bluetooth, ...) is used.
4253 * A new service systemd-machine-id-commit.service has been
4254 added. When used on systems where /etc is read-only during
4255 boot, and /etc/machine-id is not initialized (but an empty
4256 file), this service will copy the temporary machine ID
4257 created as replacement into /etc after the system is fully
4258 booted up. This is useful for systems that are freshly
4259 installed with a non-initialized machine ID, but should get
4260 a fixed machine ID for subsequent boots.
4262 * networkd's .netdev files now provide a large set of
4263 configuration parameters for VXLAN devices. Similarly, the
4264 bridge port cost parameter is now configurable in .network
4265 files. There's also new support for configuring IP source
4266 routing. networkd .link files gained support for a new
4267 OriginalName= match that is useful to match against the
4268 original interface name the kernel assigned. .network files
4269 may include MTU= and MACAddress= fields for altering the MTU
4270 and MAC address while being connected to a specific network
4273 * The LUKS logic gained supported for configuring
4274 UUID-specific key files. There's also new support for naming
4275 LUKS device from the kernel command line, using the new
4276 luks.name= argument.
4278 * Timer units may now be transiently created via the bus API
4279 (this was previously already available for scope and service
4280 units). In addition it is now possible to create multiple
4281 transient units at the same time with a single bus call. The
4282 "systemd-run" tool has been updated to make use of this for
4283 running commands on a specified time, in at(1)-style.
4285 * tmpfiles gained support for "t" lines, for assigning
4286 extended attributes to files. Among other uses this may be
4287 used to assign SMACK labels to files.
4289 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4290 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4291 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4292 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4293 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4294 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4295 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4296 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4297 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4298 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4299 Michael Biebl, Michael Chapman, Michael Marineau, Michal
4300 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4301 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4302 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4303 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4304 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4305 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4306 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
4308 — Berlin, 2014-12-10
4312 * journalctl gained the new options -t/--identifier= to match
4313 on the syslog identifier (aka "tag"), as well as --utc to
4314 show log timestamps in the UTC timezone. journalctl now also
4315 accepts -n/--lines=all to disable line capping in a pager.
4317 * journalctl gained a new switch, --flush, that synchronously
4318 flushes logs from /run/log/journal to /var/log/journal if
4319 persistent storage is enabled. systemd-journal-flush.service
4320 now waits until the operation is complete.
4322 * Services can notify the manager before they start a reload
4323 (by sending RELOADING=1) or shutdown (by sending
4324 STOPPING=1). This allows the manager to track and show the
4325 internal state of daemons and closes a race condition when
4326 the process is still running but has closed its D-Bus
4329 * Services with Type=oneshot do not have to have any ExecStart
4332 * User units are now loaded also from
4333 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4334 /run/systemd/user directory that was already previously
4335 supported, but is under the control of the user.
4337 * Job timeouts (i.e. time-outs on the time a job that is
4338 queued stays in the run queue) can now optionally result in
4339 immediate reboot or power-off actions (JobTimeoutAction= and
4340 JobTimeoutRebootArgument=). This is useful on ".target"
4341 units, to limit the maximum time a target remains
4342 undispatched in the run queue, and to trigger an emergency
4343 operation in such a case. This is now used by default to
4344 turn off the system if boot-up (as defined by everything in
4345 basic.target) hangs and does not complete for at least
4346 15min. Also, if power-off or reboot hang for at least 30min
4347 an immediate power-off/reboot operation is triggered. This
4348 functionality is particularly useful to increase reliability
4349 on embedded devices, but also on laptops which might
4350 accidentally get powered on when carried in a backpack and
4351 whose boot stays stuck in a hard disk encryption passphrase
4354 * systemd-logind can be configured to also handle lid switch
4355 events even when the machine is docked or multiple displays
4356 are attached (HandleLidSwitchDocked= option).
4358 * A helper binary and a service have been added which can be
4359 used to resume from hibernation in the initramfs. A
4360 generator will parse the resume= option on the kernel
4361 command line to trigger resume.
4363 * A user console daemon systemd-consoled has been
4364 added. Currently, it is a preview, and will so far open a
4365 single terminal on each session of the user marked as
4366 Desktop=systemd-console.
4368 * Route metrics can be specified for DHCP routes added by
4371 * The SELinux context of socket-activated services can be set
4372 from the information provided by the networking stack
4373 (SELinuxContextFromNet= option).
4375 * Userspace firmware loading support has been removed and
4376 the minimum supported kernel version is thus bumped to 3.7.
4378 * Timeout for udev workers has been increased from 1 to 3
4379 minutes, but a warning will be printed after 1 minute to
4380 help diagnose kernel modules that take a long time to load.
4382 * Udev rules can now remove tags on devices with TAG-="foobar".
4384 * systemd's readahead implementation has been removed. In many
4385 circumstances it didn't give expected benefits even for
4386 rotational disk drives and was becoming less relevant in the
4387 age of SSDs. As none of the developers has been using
4388 rotating media anymore, and nobody stepped up to actively
4389 maintain this component of systemd it has now been removed.
4391 * Swap units can use Options= to specify discard options.
4392 Discard options specified for swaps in /etc/fstab are now
4395 * Docker containers are now detected as a separate type of
4398 * The Password Agent protocol gained support for queries where
4399 the user input is shown, useful e.g. for user names.
4400 systemd-ask-password gained a new --echo option to turn that
4403 * The default sysctl.d/ snippets will now set:
4405 net.core.default_qdisc = fq_codel
4407 This selects Fair Queuing Controlled Delay as the default
4408 queuing discipline for network interfaces. fq_codel helps
4409 fight the network bufferbloat problem. It is believed to be
4410 a good default with no tuning required for most workloads.
4411 Downstream distributions may override this choice. On 10Gbit
4412 servers that do not do forwarding, "fq" may perform better.
4413 Systems without a good clocksource should use "pfifo_fast".
4415 * If kdbus is enabled during build a new option BusPolicy= is
4416 available for service units, that allows locking all service
4417 processes into a stricter bus policy, in order to limit
4418 access to various bus services, or even hide most of them
4419 from the service's view entirely.
4421 * networkctl will now show the .network and .link file
4422 networkd has applied to a specific interface.
4424 * sd-login gained a new API call sd_session_get_desktop() to
4425 query which desktop environment has been selected for a
4428 * UNIX utmp support is now compile-time optional to support
4429 legacy-free systems.
4431 * systemctl gained two new commands "add-wants" and
4432 "add-requires" for pulling in units from specific targets
4435 * If the word "rescue" is specified on the kernel command line
4436 the system will now boot into rescue mode (aka
4437 rescue.target), which was previously available only by
4438 specifying "1" or "systemd.unit=rescue.target" on the kernel
4439 command line. This new kernel command line option nicely
4440 mirrors the already existing "emergency" kernel command line
4443 * New kernel command line options mount.usr=, mount.usrflags=,
4444 mount.usrfstype= have been added that match root=, rootflags=,
4445 rootfstype= but allow mounting a specific file system to
4448 * The $NOTIFY_SOCKET is now also passed to control processes of
4449 services, not only the main process.
4451 * This version reenables support for fsck's -l switch. This
4452 means at least version v2.25 of util-linux is required for
4453 operation, otherwise dead-locks on device nodes may
4454 occur. Again: you need to update util-linux to at least
4455 v2.25 when updating systemd to v217.
4457 * The "multi-seat-x" tool has been removed from systemd, as
4458 its functionality has been integrated into X servers 1.16,
4459 and the tool is hence redundant. It is recommended to update
4460 display managers invoking this tool to simply invoke X
4461 directly from now on, again.
4463 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4464 message flag has been added for all of systemd's PolicyKit
4465 authenticated method calls has been added. In particular
4466 this now allows optional interactive authorization via
4467 PolicyKit for many of PID1's privileged operations such as
4468 unit file enabling and disabling.
4470 * "udevadm hwdb --update" learnt a new switch "--usr" for
4471 placing the rebuilt hardware database in /usr instead of
4472 /etc. When used only hardware database entries stored in
4473 /usr will be used, and any user database entries in /etc are
4474 ignored. This functionality is useful for vendors to ship a
4475 pre-built database on systems where local configuration is
4476 unnecessary or unlikely.
4478 * Calendar time specifications in .timer units now also
4479 understand the strings "semi-annually", "quarterly" and
4480 "minutely" as shortcuts (in addition to the preexisting
4481 "anually", "hourly", ...).
4483 * systemd-tmpfiles will now correctly create files in /dev
4484 at boot which are marked for creation only at boot. It is
4485 recommended to always create static device nodes with 'c!'
4486 and 'b!', so that they are created only at boot and not
4487 overwritten at runtime.
4489 * When the watchdog logic is used for a service (WatchdogSec=)
4490 and the watchdog timeout is hit the service will now be
4491 terminated with SIGABRT (instead of just SIGTERM), in order
4492 to make sure a proper coredump and backtrace is
4493 generated. This ensures that hanging services will result in
4494 similar coredump/backtrace behaviour as services that hit a
4497 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4498 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4499 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4500 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4501 Herrmann, David Sommerseth, David Strauss, Emil Renner
4502 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4503 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4504 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4505 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4506 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4507 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4508 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4509 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4510 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4511 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4512 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4513 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4514 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4515 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4516 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4517 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
4520 — Berlin, 2014-10-28
4524 * timedated no longer reads NTP implementation unit names from
4525 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
4526 implementations should add a
4528 Conflicts=systemd-timesyncd.service
4530 to their unit files to take over and replace systemd's NTP
4531 default functionality.
4533 * systemd-sysusers gained a new line type "r" for configuring
4534 which UID/GID ranges to allocate system users/groups
4535 from. Lines of type "u" may now add an additional column
4536 that specifies the home directory for the system user to be
4537 created. Also, systemd-sysusers may now optionally read user
4538 information from STDIN instead of a file. This is useful for
4539 invoking it from RPM preinst scriptlets that need to create
4540 users before the first RPM file is installed since these
4541 files might need to be owned by them. A new
4542 %sysusers_create_inline RPM macro has been introduced to do
4543 just that. systemd-sysusers now updates the shadow files as
4544 well as the user/group databases, which should enhance
4545 compatibility with certain tools like grpck.
4547 * A number of bus APIs of PID 1 now optionally consult
4548 PolicyKit to permit access for otherwise unprivileged
4549 clients under certain conditions. Note that this currently
4550 doesn't support interactive authentication yet, but this is
4551 expected to be added eventually, too.
4553 * /etc/machine-info now has new fields for configuring the
4554 deployment environment of the machine, as well as the
4555 location of the machine. hostnamectl has been updated with
4556 new command to update these fields.
4558 * systemd-timesyncd has been updated to automatically acquire
4559 NTP server information from systemd-networkd, which might
4560 have been discovered via DHCP.
4562 * systemd-resolved now includes a caching DNS stub resolver
4563 and a complete LLMNR name resolution implementation. A new
4564 NSS module "nss-resolve" has been added which can be used
4565 instead of glibc's own "nss-dns" to resolve hostnames via
4566 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4567 be resolved via systemd-resolved D-Bus APIs. In contrast to
4568 the glibc internal resolver systemd-resolved is aware of
4569 multi-homed system, and keeps DNS server and caches separate
4570 and per-interface. Queries are sent simultaneously on all
4571 interfaces that have DNS servers configured, in order to
4572 properly handle VPNs and local LANs which might resolve
4573 separate sets of domain names. systemd-resolved may acquire
4574 DNS server information from systemd-networkd automatically,
4575 which in turn might have discovered them via DHCP. A tool
4576 "systemd-resolve-host" has been added that may be used to
4577 query the DNS logic in resolved. systemd-resolved implements
4578 IDNA and automatically uses IDNA or UTF-8 encoding depending
4579 on whether classic DNS or LLMNR is used as transport. In the
4580 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4581 implementation to systemd-resolved.
4583 * A new NSS module nss-mymachines has been added, that
4584 automatically resolves the names of all local registered
4585 containers to their respective IP addresses.
4587 * A new client tool "networkctl" for systemd-networkd has been
4588 added. It currently is entirely passive and will query
4589 networking configuration from udev, rtnetlink and networkd,
4590 and present it to the user in a very friendly
4591 way. Eventually, we hope to extend it to become a full
4592 control utility for networkd.
4594 * .socket units gained a new DeferAcceptSec= setting that
4595 controls the kernels' TCP_DEFER_ACCEPT sockopt for
4596 TCP. Similarly, support for controlling TCP keep-alive
4597 settings has been added (KeepAliveTimeSec=,
4598 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4599 turning off Nagle's algorithm on TCP has been added
4602 * logind learned a new session type "web", for use in projects
4603 like Cockpit which register web clients as PAM sessions.
4605 * timer units with at least one OnCalendar= setting will now
4606 be started only after time-sync.target has been
4607 reached. This way they will not elapse before the system
4608 clock has been corrected by a local NTP client or
4609 similar. This is particular useful on RTC-less embedded
4610 machines, that come up with an invalid system clock.
4612 * systemd-nspawn's --network-veth= switch should now result in
4613 stable MAC addresses for both the outer and the inner side
4616 * systemd-nspawn gained a new --volatile= switch for running
4617 container instances with /etc or /var unpopulated.
4619 * The kdbus client code has been updated to use the new Linux
4620 3.17 memfd subsystem instead of the old kdbus-specific one.
4622 * systemd-networkd's DHCP client and server now support
4623 FORCERENEW. There are also new configuration options to
4624 configure the vendor client identifier and broadcast mode
4627 * systemd will no longer inform the kernel about the current
4628 timezone, as this is necessarily incorrect and racy as the
4629 kernel has no understanding of DST and similar
4630 concepts. This hence means FAT timestamps will be always
4631 considered UTC, similar to what Android is already
4632 doing. Also, when the RTC is configured to the local time
4633 (rather than UTC) systemd will never synchronize back to it,
4634 as this might confuse Windows at a later boot.
4636 * systemd-analyze gained a new command "verify" for offline
4637 validation of unit files.
4639 * systemd-networkd gained support for a couple of additional
4640 settings for bonding networking setups. Also, the metric for
4641 statically configured routes may now be configured. For
4642 network interfaces where this is appropriate the peer IP
4643 address may now be configured.
4645 * systemd-networkd's DHCP client will no longer request
4646 broadcasting by default, as this tripped up some networks.
4647 For hardware where broadcast is required the feature should
4648 be switched back on using RequestBroadcast=yes.
4650 * systemd-networkd will now set up IPv4LL addresses (when
4651 enabled) even if DHCP is configured successfully.
4653 * udev will now default to respect network device names given
4654 by the kernel when the kernel indicates that these are
4655 predictable. This behavior can be tweaked by changing
4656 NamePolicy= in the relevant .link file.
4658 * A new library systemd-terminal has been added that
4659 implements full TTY stream parsing and rendering. This
4660 library is supposed to be used later on for implementing a
4661 full userspace VT subsystem, replacing the current kernel
4664 * A new tool systemd-journal-upload has been added to push
4665 journal data to a remote system running
4666 systemd-journal-remote.
4668 * journald will no longer forward all local data to another
4669 running syslog daemon. This change has been made because
4670 rsyslog (which appears to be the most commonly used syslog
4671 implementation these days) no longer makes use of this, and
4672 instead pulls the data out of the journal on its own. Since
4673 forwarding the messages to a non-existent syslog server is
4674 more expensive than we assumed we have now turned this
4675 off. If you run a syslog server that is not a recent rsyslog
4676 version, you have to turn this option on again
4677 (ForwardToSyslog= in journald.conf).
4679 * journald now optionally supports the LZ4 compressor for
4680 larger journal fields. This compressor should perform much
4681 better than XZ which was the previous default.
4683 * machinectl now shows the IP addresses of local containers,
4684 if it knows them, plus the interface name of the container.
4686 * A new tool "systemd-escape" has been added that makes it
4687 easy to escape strings to build unit names and similar.
4689 * sd_notify() messages may now include a new ERRNO= field
4690 which is parsed and collected by systemd and shown among the
4691 "systemctl status" output for a service.
4693 * A new component "systemd-firstboot" has been added that
4694 queries the most basic systemd information (timezone,
4695 hostname, root password) interactively on first
4696 boot. Alternatively it may also be used to provision these
4697 things offline on OS images installed into directories.
4699 * The default sysctl.d/ snippets will now set
4701 net.ipv4.conf.default.promote_secondaries=1
4703 This has the benefit of no flushing secondary IP addresses
4704 when primary addresses are removed.
4706 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4707 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4708 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4709 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4710 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4711 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4712 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4713 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4714 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4715 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4716 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4717 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4718 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4719 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4720 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4722 — Berlin, 2014-08-19
4726 * A new tool systemd-sysusers has been added. This tool
4727 creates system users and groups in /etc/passwd and
4728 /etc/group, based on static declarative system user/group
4729 definitions in /usr/lib/sysusers.d/. This is useful to
4730 enable factory resets and volatile systems that boot up with
4731 an empty /etc directory, and thus need system users and
4732 groups created during early boot. systemd now also ships
4733 with two default sysusers.d/ files for the most basic
4734 users and groups systemd and the core operating system
4737 * A new tmpfiles snippet has been added that rebuilds the
4738 essential files in /etc on boot, should they be missing.
4740 * A directive for ensuring automatic clean-up of
4741 /var/cache/man/ has been removed from the default
4742 configuration. This line should now be shipped by the man
4743 implementation. The necessary change has been made to the
4744 man-db implementation. Note that you need to update your man
4745 implementation to one that ships this line, otherwise no
4746 automatic clean-up of /var/cache/man will take place.
4748 * A new condition ConditionNeedsUpdate= has been added that
4749 may conditionalize services to only run when /etc or /var
4750 are "older" than the vendor operating system resources in
4751 /usr. This is useful for reconstructing or updating /etc
4752 after an offline update of /usr or a factory reset, on the
4753 next reboot. Services that want to run once after such an
4754 update or reset should use this condition and order
4755 themselves before the new systemd-update-done.service, which
4756 will mark the two directories as fully updated. A number of
4757 service files have been added making use of this, to rebuild
4758 the udev hardware database, the journald message catalog and
4759 dynamic loader cache (ldconfig). The systemd-sysusers tool
4760 described above also makes use of this now. With this in
4761 place it is now possible to start up a minimal operating
4762 system with /etc empty cleanly. For more information on the
4763 concepts involved see this recent blog story:
4765 http://0pointer.de/blog/projects/stateless.html
4767 * A new system group "input" has been introduced, and all
4768 input device nodes get this group assigned. This is useful
4769 for system-level software to get access to input devices. It
4770 complements what is already done for "audio" and "video".
4772 * systemd-networkd learnt minimal DHCPv4 server support in
4773 addition to the existing DHCPv4 client support. It also
4774 learnt DHCPv6 client and IPv6 Router Solicitation client
4775 support. The DHCPv4 client gained support for static routes
4776 passed in from the server. Note that the [DHCPv4] section
4777 known in older systemd-networkd versions has been renamed to
4778 [DHCP] and is now also used by the DHCPv6 client. Existing
4779 .network files using settings of this section should be
4780 updated, though compatibility is maintained. Optionally, the
4781 client hostname may now be sent to the DHCP server.
4783 * networkd gained support for vxlan virtual networks as well
4784 as tun/tap and dummy devices.
4786 * networkd gained support for automatic allocation of address
4787 ranges for interfaces from a system-wide pool of
4788 addresses. This is useful for dynamically managing a large
4789 number of interfaces with a single network configuration
4790 file. In particular this is useful to easily assign
4791 appropriate IP addresses to the veth links of a large number
4792 of nspawn instances.
4794 * RPM macros for processing sysusers, sysctl and binfmt
4795 drop-in snippets at package installation time have been
4798 * The /etc/os-release file should now be placed in
4799 /usr/lib/os-release. The old location is automatically
4800 created as symlink. /usr/lib is the more appropriate
4801 location of this file, since it shall actually describe the
4802 vendor operating system shipped in /usr, and not the
4803 configuration stored in /etc.
4805 * .mount units gained a new boolean SloppyOptions= setting
4806 that maps to mount(8)'s -s option which enables permissive
4807 parsing of unknown mount options.
4809 * tmpfiles learnt a new "L+" directive which creates a symlink
4810 but (unlike "L") deletes a pre-existing file first, should
4811 it already exist and not already be the correct
4812 symlink. Similarly, "b+", "c+" and "p+" directives have been
4813 added as well, which create block and character devices, as
4814 well as fifos in the filesystem, possibly removing any
4815 pre-existing files of different types.
4817 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4818 'argument' field (which so far specified the source to
4819 symlink/copy the files from) is now optional. If omitted the
4820 same file os copied from /usr/share/factory/ suffixed by the
4821 full destination path. This is useful for populating /etc
4822 with essential files, by copying them from vendor defaults
4823 shipped in /usr/share/factory/etc.
4825 * A new command "systemctl preset-all" has been added that
4826 applies the service preset settings to all installed unit
4827 files. A new switch --preset-mode= has been added that
4828 controls whether only enable or only disable operations
4831 * A new command "systemctl is-system-running" has been added
4832 that allows checking the overall state of the system, for
4833 example whether it is fully up and running.
4835 * When the system boots up with an empty /etc, the equivalent
4836 to "systemctl preset-all" is executed during early boot, to
4837 make sure all default services are enabled after a factory
4840 * systemd now contains a minimal preset file that enables the
4841 most basic services systemd ships by default.
4843 * Unit files' [Install] section gained a new DefaultInstance=
4844 field for defining the default instance to create if a
4845 template unit is enabled with no instance specified.
4847 * A new passive target cryptsetup-pre.target has been added
4848 that may be used by services that need to make they run and
4849 finish before the first LUKS cryptographic device is set up.
4851 * The /dev/loop-control and /dev/btrfs-control device nodes
4852 are now owned by the "disk" group by default, opening up
4853 access to this group.
4855 * systemd-coredump will now automatically generate a
4856 stack trace of all core dumps taking place on the system,
4857 based on elfutils' libdw library. This stack trace is logged
4860 * systemd-coredump may now optionally store coredumps directly
4861 on disk (in /var/lib/systemd/coredump, possibly compressed),
4862 instead of storing them unconditionally in the journal. This
4863 mode is the new default. A new configuration file
4864 /etc/systemd/coredump.conf has been added to configure this
4865 and other parameters of systemd-coredump.
4867 * coredumpctl gained a new "info" verb to show details about a
4868 specific coredump. A new switch "-1" has also been added
4869 that makes sure to only show information about the most
4870 recent entry instead of all entries. Also, as the tool is
4871 generally useful now the "systemd-" prefix of the binary
4872 name has been removed. Distributions that want to maintain
4873 compatibility with the old name should add a symlink from
4874 the old name to the new name.
4876 * journald's SplitMode= now defaults to "uid". This makes sure
4877 that unprivileged users can access their own coredumps with
4878 coredumpctl without restrictions.
4880 * New kernel command line options "systemd.wants=" (for
4881 pulling an additional unit during boot), "systemd.mask="
4882 (for masking a specific unit for the boot), and
4883 "systemd.debug-shell" (for enabling the debug shell on tty9)
4884 have been added. This is implemented in the new generator
4885 "systemd-debug-generator".
4887 * systemd-nspawn will now by default filter a couple of
4888 syscalls for containers, among them those required for
4889 kernel module loading, direct x86 IO port access, swap
4890 management, and kexec. Most importantly though
4891 open_by_handle_at() is now prohibited for containers,
4892 closing a hole similar to a recently discussed vulnerability
4893 in docker regarding access to files on file hierarchies the
4894 container should normally not have access to. Note that, for
4895 nspawn, we generally make no security claims anyway (and
4896 this is explicitly documented in the man page), so this is
4897 just a fix for one of the most obvious problems.
4899 * A new man page file-hierarchy(7) has been added that
4900 contains a minimized, modernized version of the file system
4901 layout systemd expects, similar in style to the FHS
4902 specification or hier(5). A new tool systemd-path(1) has
4903 been added to query many of these paths for the local
4906 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4907 longer done. Since the directory now has a per-user size
4908 limit, and is cleaned on logout this appears unnecessary,
4909 in particular since this now brings the lifecycle of this
4910 directory closer in line with how IPC objects are handled.
4912 * systemd.pc now exports a number of additional directories,
4913 including $libdir (which is useful to identify the library
4914 path for the primary architecture of the system), and a
4915 couple of drop-in directories.
4917 * udev's predictable network interface names now use the dev_port
4918 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4919 distinguish between ports of the same PCI function. dev_id should
4920 only be used for ports using the same HW address, hence the need
4923 * machined has been updated to export the OS version of a
4924 container (read from /etc/os-release and
4925 /usr/lib/os-release) on the bus. This is now shown in
4926 "machinectl status" for a machine.
4928 * A new service setting RestartForceExitStatus= has been
4929 added. If configured to a set of exit signals or process
4930 return values, the service will be restarted when the main
4931 daemon process exits with any of them, regardless of the
4934 * systemctl's -H switch for connecting to remote systemd
4935 machines has been extended so that it may be used to
4936 directly connect to a specific container on the
4937 host. "systemctl -H root@foobar:waldi" will now connect as
4938 user "root" to host "foobar", and then proceed directly to
4939 the container named "waldi". Note that currently you have to
4940 authenticate as user "root" for this to work, as entering
4941 containers is a privileged operation.
4943 Contributions from: Andreas Henriksson, Benjamin Steinwender,
4944 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
4945 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
4946 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
4947 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
4948 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
4949 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
4950 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
4951 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
4952 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
4953 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
4954 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
4956 — Berlin, 2014-07-03
4960 * As an experimental feature, udev now tries to lock the
4961 disk device node (flock(LOCK_SH|LOCK_NB)) while it
4962 executes events for the disk or any of its partitions.
4963 Applications like partitioning programs can lock the
4964 disk device node (flock(LOCK_EX)) and claim temporary
4965 device ownership that way; udev will entirely skip all event
4966 handling for this disk and its partitions. If the disk
4967 was opened for writing, the close will trigger a partition
4968 table rescan in udev's "watch" facility, and if needed
4969 synthesize "change" events for the disk and all its partitions.
4970 This is now unconditionally enabled, and if it turns out to
4971 cause major problems, we might turn it on only for specific
4972 devices, or might need to disable it entirely. Device Mapper
4973 devices are excluded from this logic.
4975 * We temporarily dropped the "-l" switch for fsck invocations,
4976 since they collide with the flock() logic above. util-linux
4977 upstream has been changed already to avoid this conflict,
4978 and we will readd "-l" as soon as util-linux with this
4979 change has been released.
4981 * The dependency on libattr has been removed. Since a long
4982 time, the extended attribute calls have moved to glibc, and
4983 libattr is thus unnecessary.
4985 * Virtualization detection works without privileges now. This
4986 means the systemd-detect-virt binary no longer requires
4987 CAP_SYS_PTRACE file capabilities, and our daemons can run
4988 with fewer privileges.
4990 * systemd-networkd now runs under its own "systemd-network"
4991 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
4992 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
4993 loses the ability to write to files owned by root this way.
4995 * Similarly, systemd-resolved now runs under its own
4996 "systemd-resolve" user with no capabilities remaining.
4998 * Similarly, systemd-bus-proxyd now runs under its own
4999 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5001 * systemd-networkd gained support for setting up "veth"
5002 virtual Ethernet devices for container connectivity, as well
5003 as GRE and VTI tunnels.
5005 * systemd-networkd will no longer automatically attempt to
5006 manually load kernel modules necessary for certain tunnel
5007 transports. Instead, it is assumed the kernel loads them
5008 automatically when required. This only works correctly on
5009 very new kernels. On older kernels, please consider adding
5010 the kernel modules to /etc/modules-load.d/ as a work-around.
5012 * The resolv.conf file systemd-resolved generates has been
5013 moved to /run/systemd/resolve/. If you have a symlink from
5014 /etc/resolv.conf, it might be necessary to correct it.
5016 * Two new service settings, ProtectHome= and ProtectSystem=,
5017 have been added. When enabled, they will make the user data
5018 (such as /home) inaccessible or read-only and the system
5019 (such as /usr) read-only, for specific services. This allows
5020 very light-weight per-service sandboxing to avoid
5021 modifications of user data or system files from
5022 services. These two new switches have been enabled for all
5023 of systemd's long-running services, where appropriate.
5025 * Socket units gained new SocketUser= and SocketGroup=
5026 settings to set the owner user and group of AF_UNIX sockets
5027 and FIFOs in the file system.
5029 * Socket units gained a new RemoveOnStop= setting. If enabled,
5030 all FIFOS and sockets in the file system will be removed
5031 when the specific socket unit is stopped.
5033 * Socket units gained a new Symlinks= setting. It takes a list
5034 of symlinks to create to file system sockets or FIFOs
5035 created by the specific Unix sockets. This is useful to
5036 manage symlinks to socket nodes with the same life-cycle as
5039 * The /dev/log socket and /dev/initctl FIFO have been moved to
5040 /run, and have been replaced by symlinks. This allows
5041 connecting to these facilities even if PrivateDevices=yes is
5042 used for a service (which makes /dev/log itself unavailable,
5043 but /run is left). This also has the benefit of ensuring
5044 that /dev only contains device nodes, directories and
5045 symlinks, and nothing else.
5047 * sd-daemon gained two new calls sd_pid_notify() and
5048 sd_pid_notifyf(). They are similar to sd_notify() and
5049 sd_notifyf(), but allow overriding of the source PID of
5050 notification messages if permissions permit this. This is
5051 useful to send notify messages on behalf of a different
5052 process (for example, the parent process). The
5053 systemd-notify tool has been updated to make use of this
5054 when sending messages (so that notification messages now
5055 originate from the shell script invoking systemd-notify and
5056 not the systemd-notify process itself. This should minimize
5057 a race where systemd fails to associate notification
5058 messages to services when the originating process already
5061 * A new "on-abnormal" setting for Restart= has been added. If
5062 set, it will result in automatic restarts on all "abnormal"
5063 reasons for a process to exit, which includes unclean
5064 signals, core dumps, timeouts and watchdog timeouts, but
5065 does not include clean and unclean exit codes or clean
5066 signals. Restart=on-abnormal is an alternative for
5067 Restart=on-failure for services that shall be able to
5068 terminate and avoid restarts on certain errors, by
5069 indicating so with an unclean exit code. Restart=on-failure
5070 or Restart=on-abnormal is now the recommended setting for
5071 all long-running services.
5073 * If the InaccessibleDirectories= service setting points to a
5074 mount point (or if there are any submounts contained within
5075 it), it is now attempted to completely unmount it, to make
5076 the file systems truly unavailable for the respective
5079 * The ReadOnlyDirectories= service setting and
5080 systemd-nspawn's --read-only parameter are now recursively
5081 applied to all submounts, too.
5083 * Mount units may now be created transiently via the bus APIs.
5085 * The support for SysV and LSB init scripts has been removed
5086 from the systemd daemon itself. Instead, it is now
5087 implemented as a generator that creates native systemd units
5088 from these scripts when needed. This enables us to remove a
5089 substantial amount of legacy code from PID 1, following the
5090 fact that many distributions only ship a very small number
5091 of LSB/SysV init scripts nowadays.
5093 * Privileged Xen (dom0) domains are not considered
5094 virtualization anymore by the virtualization detection
5095 logic. After all, they generally have unrestricted access to
5096 the hardware and usually are used to manage the unprivileged
5099 * systemd-tmpfiles gained a new "C" line type, for copying
5100 files or entire directories.
5102 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
5103 lines. So far, they have been non-globbing versions of the
5104 latter, and have thus been redundant. In future, it is
5105 recommended to only use "z". "m" has hence been removed
5106 from the documentation, even though it stays supported.
5108 * A tmpfiles snippet to recreate the most basic structure in
5109 /var has been added. This is enough to create the /var/run →
5110 /run symlink and create a couple of structural
5111 directories. This allows systems to boot up with an empty or
5112 volatile /var. Of course, while with this change, the core OS
5113 now is capable with dealing with a volatile /var, not all
5114 user services are ready for it. However, we hope that sooner
5115 or later, many service daemons will be changed upstream so
5116 that they are able to automatically create their necessary
5117 directories in /var at boot, should they be missing. This is
5118 the first step to allow state-less systems that only require
5119 the vendor image for /usr to boot.
5121 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5122 empty tmpfs instance to a specific directory. This is
5123 particularly useful for making use of the automatic
5124 reconstruction of /var (see above), by passing --tmpfs=/var.
5126 * Access modes specified in tmpfiles snippets may now be
5127 prefixed with "~", which indicates that they shall be masked
5128 by whether the existing file or directory is currently
5129 writable, readable or executable at all. Also, if specified,
5130 the sgid/suid/sticky bits will be masked for all
5133 * A new passive target unit "network-pre.target" has been
5134 added which is useful for services that shall run before any
5135 network is configured, for example firewall scripts.
5137 * The "floppy" group that previously owned the /dev/fd*
5138 devices is no longer used. The "disk" group is now used
5139 instead. Distributions should probably deprecate usage of
5142 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5143 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5144 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5145 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5146 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5147 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5150 — Berlin, 2014-06-11
5154 * A new "systemd-timesyncd" daemon has been added for
5155 synchronizing the system clock across the network. It
5156 implements an SNTP client. In contrast to NTP
5157 implementations such as chrony or the NTP reference server,
5158 this only implements a client side, and does not bother with
5159 the full NTP complexity, focusing only on querying time from
5160 one remote server and synchronizing the local clock to
5161 it. Unless you intend to serve NTP to networked clients or
5162 want to connect to local hardware clocks, this simple NTP
5163 client should be more than appropriate for most
5164 installations. The daemon runs with minimal privileges, and
5165 has been hooked up with networkd to only operate when
5166 network connectivity is available. The daemon saves the
5167 current clock to disk every time a new NTP sync has been
5168 acquired, and uses this to possibly correct the system clock
5169 early at bootup, in order to accommodate for systems that
5170 lack an RTC such as the Raspberry Pi and embedded devices,
5171 and to make sure that time monotonically progresses on these
5172 systems, even if it is not always correct. To make use of
5173 this daemon, a new system user and group "systemd-timesync"
5174 needs to be created on installation of systemd.
5176 * The queue "seqnum" interface of libudev has been disabled, as
5177 it was generally incompatible with device namespacing as
5178 sequence numbers of devices go "missing" if the devices are
5179 part of a different namespace.
5181 * "systemctl list-timers" and "systemctl list-sockets" gained
5182 a --recursive switch for showing units of these types also
5183 for all local containers, similar in style to the already
5184 supported --recursive switch for "systemctl list-units".
5186 * A new RebootArgument= setting has been added for service
5187 units, which may be used to specify a kernel reboot argument
5188 to use when triggering reboots with StartLimitAction=.
5190 * A new FailureAction= setting has been added for service
5191 units which may be used to specify an operation to trigger
5192 when a service fails. This works similarly to
5193 StartLimitAction=, but unlike it, controls what is done
5194 immediately rather than only after several attempts to
5195 restart the service in question.
5197 * hostnamed got updated to also expose the kernel name,
5198 release, and version on the bus. This is useful for
5199 executing commands like hostnamectl with the -H switch.
5200 systemd-analyze makes use of this to properly display
5201 details when running non-locally.
5203 * The bootchart tool can now show cgroup information in the
5204 graphs it generates.
5206 * The CFS CPU quota cgroup attribute is now exposed for
5207 services. The new CPUQuota= switch has been added for this
5208 which takes a percentage value. Setting this will have the
5209 result that a service may never get more CPU time than the
5210 specified percentage, even if the machine is otherwise idle.
5212 * systemd-networkd learned IPIP and SIT tunnel support.
5214 * LSB init scripts exposing a dependency on $network will now
5215 get a dependency on network-online.target rather than simply
5216 network.target. This should bring LSB handling closer to
5217 what it was on SysV systems.
5219 * A new fsck.repair= kernel option has been added to control
5220 how fsck shall deal with unclean file systems at boot.
5222 * The (.ini) configuration file parser will now silently
5223 ignore sections whose name begins with "X-". This may be
5224 used to maintain application-specific extension sections in unit
5227 * machined gained a new API to query the IP addresses of
5228 registered containers. "machinectl status" has been updated
5229 to show these addresses in its output.
5231 * A new call sd_uid_get_display() has been added to the
5232 sd-login APIs for querying the "primary" session of a
5233 user. The "primary" session of the user is elected from the
5234 user's sessions and generally a graphical session is
5235 preferred over a text one.
5237 * A minimal systemd-resolved daemon has been added. It
5238 currently simply acts as a companion to systemd-networkd and
5239 manages resolv.conf based on per-interface DNS
5240 configuration, possibly supplied via DHCP. In the long run
5241 we hope to extend this into a local DNSSEC enabled DNS and
5244 * The systemd-networkd-wait-online tool is now enabled by
5245 default. It will delay network-online.target until a network
5246 connection has been configured. The tool primarily integrates
5247 with networkd, but will also make a best effort to make sense
5248 of network configuration performed in some other way.
5250 * Two new service options StartupCPUShares= and
5251 StartupBlockIOWeight= have been added that work similarly to
5252 CPUShares= and BlockIOWeight= however only apply during
5253 system startup. This is useful to prioritize certain services
5254 differently during bootup than during normal runtime.
5256 * hostnamed has been changed to prefer the statically
5257 configured hostname in /etc/hostname (unless set to
5258 'localhost' or empty) over any dynamic one supplied by
5259 dhcp. With this change, the rules for picking the hostname
5260 match more closely the rules of other configuration settings
5261 where the local administrator's configuration in /etc always
5262 overrides any other settings.
5264 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
5265 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5266 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5267 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5268 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5269 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5270 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5271 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5272 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
5273 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5274 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5275 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5276 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5277 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5278 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5279 Lindskog, WaLyong Cho, Will Woods, Zbigniew
5282 — Beijing, 2014-05-28
5286 * When restoring the screen brightness at boot, stay away from
5287 the darkest setting or from the lowest 5% of the available
5288 range, depending on which is the larger value of both. This
5289 should effectively protect the user from rebooting into a
5290 black screen, should the brightness have been set to minimum
5293 * sd-login gained a new sd_machine_get_class() call to
5294 determine the class ("vm" or "container") of a machine
5295 registered with machined.
5297 * sd-login gained new calls
5298 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5299 to query the identity of the peer of a local AF_UNIX
5300 connection. They operate similarly to their sd_pid_get_xyz()
5303 * PID 1 will now maintain a system-wide system state engine
5304 with the states "starting", "running", "degraded",
5305 "maintenance", "stopping". These states are bound to system
5306 startup, normal runtime, runtime with at least one failed
5307 service, rescue/emergency mode and system shutdown. This
5308 state is shown in the "systemctl status" output when no unit
5309 name is passed. It is useful to determine system state, in
5310 particularly when doing so for many systems or containers at
5313 * A new command "list-machines" has been added to "systemctl"
5314 that lists all local OS containers and shows their system
5315 state (see above), if systemd runs inside of them.
5317 * systemctl gained a new "-r" switch to recursively enumerate
5318 units on all local containers, when used with the
5319 "list-unit" command (which is the default one that is
5320 executed when no parameters are specified).
5322 * The GPT automatic partition discovery logic will now honour
5323 two GPT partition flags: one may be set on a partition to
5324 cause it to be mounted read-only, and the other may be set
5325 on a partition to ignore it during automatic discovery.
5327 * Two new GPT type UUIDs have been added for automatic root
5328 partition discovery, for 32-bit and 64-bit ARM. This is not
5329 particularly useful for discovering the root directory on
5330 these architectures during bare-metal boots (since UEFI is
5331 not common there), but still very useful to allow booting of
5332 ARM disk images in nspawn with the -i option.
5334 * MAC addresses of interfaces created with nspawn's
5335 --network-interface= switch will now be generated from the
5336 machine name, and thus be stable between multiple invocations
5339 * logind will now automatically remove all IPC objects owned
5340 by a user if she or he fully logs out. This makes sure that
5341 users who are logged out cannot continue to consume IPC
5342 resources. This covers SysV memory, semaphores and message
5343 queues as well as POSIX shared memory and message
5344 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5345 limits. With this functionality, that is corrected. This may
5346 be turned off by using the RemoveIPC= switch of logind.conf.
5348 * The systemd-machine-id-setup and tmpfiles tools gained a
5349 --root= switch to operate on a specific root directory,
5352 * journald can now forward logged messages to the TTYs of all
5353 logged in users ("wall"). This is the default for all
5354 emergency messages now.
5356 * A new tool systemd-journal-remote has been added to stream
5357 journal log messages across the network.
5359 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5360 controller trees are mounted into it. Note that the
5361 directories mounted beneath it are not read-only. This is a
5362 security measure and is particularly useful because glibc
5363 actually includes a search logic to pick any tmpfs it can
5364 find to implement shm_open() if /dev/shm is not available
5365 (which it might very well be in namespaced setups).
5367 * machinectl gained a new "poweroff" command to cleanly power
5368 down a local OS container.
5370 * The PrivateDevices= unit file setting will now also drop the
5371 CAP_MKNOD capability from the capability bound set, and
5372 imply DevicePolicy=closed.
5374 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5375 comprehensively on all long-running systemd services where
5376 this is appropriate.
5378 * systemd-udevd will now run in a disassociated mount
5379 namespace. To mount directories from udev rules, make sure to
5380 pull in mount units via SYSTEMD_WANTS properties.
5382 * The kdbus support gained support for uploading policy into
5383 the kernel. sd-bus gained support for creating "monitoring"
5384 connections that can eavesdrop into all bus communication
5385 for debugging purposes.
5387 * Timestamps may now be specified in seconds since the UNIX
5388 epoch Jan 1st, 1970 by specifying "@" followed by the value
5391 * Native tcpwrap support in systemd has been removed. tcpwrap
5392 is old code, not really maintained anymore and has serious
5393 shortcomings, and better options such as firewalls
5394 exist. For setups that require tcpwrap usage, please
5395 consider invoking your socket-activated service via tcpd,
5396 like on traditional inetd.
5398 * A new system.conf configuration option
5399 DefaultTimerAccuracySec= has been added that controls the
5400 default AccuracySec= setting of .timer units.
5402 * Timer units gained a new WakeSystem= switch. If enabled,
5403 timers configured this way will cause the system to resume
5404 from system suspend (if the system supports that, which most
5407 * Timer units gained a new Persistent= switch. If enabled,
5408 timers configured this way will save to disk when they have
5409 been last triggered. This information is then used on next
5410 reboot to possible execute overdue timer events, that
5411 could not take place because the system was powered off.
5412 This enables simple anacron-like behaviour for timer units.
5414 * systemctl's "list-timers" will now also list the time a
5415 timer unit was last triggered in addition to the next time
5416 it will be triggered.
5418 * systemd-networkd will now assign predictable IPv4LL
5419 addresses to its local interfaces.
5421 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5422 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5423 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5424 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5425 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5426 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5427 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5428 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5431 — Berlin, 2014-03-25
5435 * A new unit file setting RestrictAddressFamilies= has been
5436 added to restrict which socket address families unit
5437 processes gain access to. This takes address family names
5438 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5439 attack surface of services via exotic protocol stacks. This
5440 is built on seccomp system call filters.
5442 * Two new unit file settings RuntimeDirectory= and
5443 RuntimeDirectoryMode= have been added that may be used to
5444 manage a per-daemon runtime directories below /run. This is
5445 an alternative for setting up directory permissions with
5446 tmpfiles snippets, and has the advantage that the runtime
5447 directory's lifetime is bound to the daemon runtime and that
5448 the daemon starts up with an empty directory each time. This
5449 is particularly useful when writing services that drop
5450 privileges using the User= or Group= setting.
5452 * The DeviceAllow= unit setting now supports globbing for
5453 matching against device group names.
5455 * The systemd configuration file system.conf gained new
5456 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5457 DefaultMemoryAccounting= to globally turn on/off accounting
5458 for specific resources (cgroups) for all units. These
5459 settings may still be overridden individually in each unit
5462 * systemd-gpt-auto-generator is now able to discover /srv and
5463 root partitions in addition to /home and swap partitions. It
5464 also supports LUKS-encrypted partitions now. With this in
5465 place, automatic discovery of partitions to mount following
5466 the Discoverable Partitions Specification
5467 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
5468 is now a lot more complete. This allows booting without
5469 /etc/fstab and without root= on the kernel command line on
5470 systems prepared appropriately.
5472 * systemd-nspawn gained a new --image= switch which allows
5473 booting up disk images and Linux installations on any block
5474 device that follow the Discoverable Partitions Specification
5475 (see above). This means that installations made with
5476 appropriately updated installers may now be started and
5477 deployed using container managers, completely
5478 unmodified. (We hope that libvirt-lxc will add support for
5479 this feature soon, too.)
5481 * systemd-nspawn gained a new --network-macvlan= setting to
5482 set up a private macvlan interface for the
5483 container. Similarly, systemd-networkd gained a new
5484 Kind=macvlan setting in .netdev files.
5486 * systemd-networkd now supports configuring local addresses
5489 * A new tool systemd-network-wait-online has been added to
5490 synchronously wait for network connectivity using
5493 * The sd-bus.h bus API gained a new sd_bus_track object for
5494 tracking the life-cycle of bus peers. Note that sd-bus.h is
5495 still not a public API though (unless you specify
5496 --enable-kdbus on the configure command line, which however
5497 voids your warranty and you get no API stability guarantee).
5499 * The $XDG_RUNTIME_DIR runtime directories for each user are
5500 now individual tmpfs instances, which has the benefit of
5501 introducing separate pools for each user, with individual
5502 size limits, and thus making sure that unprivileged clients
5503 can no longer negatively impact the system or other users by
5504 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5505 RuntimeDirectorySize= has been introduced that allows
5506 controlling the default size limit for all users. It
5507 defaults to 10% of the available physical memory. This is no
5508 replacement for quotas on tmpfs though (which the kernel
5509 still does not support), as /dev/shm and /tmp are still
5510 shared resources used by both the system and unprivileged
5513 * logind will now automatically turn off automatic suspending
5514 on laptop lid close when more than one display is
5515 connected. This was previously expected to be implemented
5516 individually in desktop environments (such as GNOME),
5517 however has been added to logind now, in order to fix a
5518 boot-time race where a desktop environment might not have
5519 been started yet and thus not been able to take an inhibitor
5520 lock at the time where logind already suspends the system
5521 due to a closed lid.
5523 * logind will now wait at least 30s after each system
5524 suspend/resume cycle, and 3min after system boot before
5525 suspending the system due to a closed laptop lid. This
5526 should give USB docking stations and similar enough time to
5527 be probed and configured after system resume and boot in
5528 order to then act as suspend blocker.
5530 * systemd-run gained a new --property= setting which allows
5531 initialization of resource control properties (and others)
5532 for the created scope or service unit. Example: "systemd-run
5533 --property=BlockIOWeight=10 updatedb" may be used to run
5534 updatedb at a low block IO scheduling weight.
5536 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5537 now also work in --scope mode.
5539 * When systemd is compiled with kdbus support, basic support
5540 for enforced policies is now in place. (Note that enabling
5541 kdbus still voids your warranty and no API compatibility
5544 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5545 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5546 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5547 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5548 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5549 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5550 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5551 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5552 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5553 Zbigniew Jędrzejewski-Szmek
5555 — Berlin, 2014-03-12
5559 * systemd will now relabel /dev after loading the SMACK policy
5560 according to SMACK rules.
5562 * A new unit file option AppArmorProfile= has been added to
5563 set the AppArmor profile for the processes of a unit.
5565 * A new condition check ConditionArchitecture= has been added
5566 to conditionalize units based on the system architecture, as
5567 reported by uname()'s "machine" field.
5569 * systemd-networkd now supports matching on the system
5570 virtualization, architecture, kernel command line, host name
5573 * logind is now a lot more aggressive when suspending the
5574 machine due to a closed laptop lid. Instead of acting only
5575 on the lid close action, it will continuously watch the lid
5576 status and act on it. This is useful for laptops where the
5577 power button is on the outside of the chassis so that it can
5578 be reached without opening the lid (such as the Lenovo
5579 Yoga). On those machines, logind will now immediately
5580 re-suspend the machine if the power button has been
5581 accidentally pressed while the laptop was suspended and in a
5582 backpack or similar.
5584 * logind will now watch SW_DOCK switches and inhibit reaction
5585 to the lid switch if it is pressed. This means that logind
5586 will not suspend the machine anymore if the lid is closed
5587 and the system is docked, if the laptop supports SW_DOCK
5588 notifications via the input layer. Note that ACPI docking
5589 stations do not generate this currently. Also note that this
5590 logic is usually not fully sufficient and Desktop
5591 Environments should take a lid switch inhibitor lock when an
5592 external display is connected, as systemd will not watch
5595 * nspawn will now make use of the devices cgroup controller by
5596 default, and only permit creation of and access to the usual
5597 API device nodes like /dev/null or /dev/random, as well as
5598 access to (but not creation of) the pty devices.
5600 * We will now ship a default .network file for
5601 systemd-networkd that automatically configures DHCP for
5602 network interfaces created by nspawn's --network-veth or
5603 --network-bridge= switches.
5605 * systemd will now understand the usual M, K, G, T suffixes
5606 according to SI conventions (i.e. to the base 1000) when
5607 referring to throughput and hardware metrics. It will stay
5608 with IEC conventions (i.e. to the base 1024) for software
5609 metrics, according to what is customary according to
5610 Wikipedia. We explicitly document which base applies for
5611 each configuration option.
5613 * The DeviceAllow= setting in unit files now supports a syntax
5614 to whitelist an entire group of devices node majors at once,
5615 based on the /proc/devices listing. For example, with the
5616 string "char-pts", it is now possible to whitelist all
5617 current and future pseudo-TTYs at once.
5619 * sd-event learned a new "post" event source. Event sources of
5620 this type are triggered by the dispatching of any event
5621 source of a type that is not "post". This is useful for
5622 implementing clean-up and check event sources that are
5623 triggered by other work being done in the program.
5625 * systemd-networkd is no longer statically enabled, but uses
5626 the usual [Install] sections so that it can be
5627 enabled/disabled using systemctl. It still is enabled by
5630 * When creating a veth interface pair with systemd-nspawn, the
5631 host side will now be prefixed with "vb-" if
5632 --network-bridge= is used, and with "ve-" if --network-veth
5633 is used. This way, it is easy to distinguish these cases on
5634 the host, for example to apply different configuration to
5635 them with systemd-networkd.
5637 * The compatibility libraries for libsystemd-journal.so,
5638 libsystem-id128.so, libsystemd-login.so and
5639 libsystemd-daemon.so do not make use of IFUNC
5640 anymore. Instead, we now build libsystemd.so multiple times
5641 under these alternative names. This means that the footprint
5642 is drastically increased, but given that these are
5643 transitional compatibility libraries, this should not matter
5644 much. This change has been made necessary to support the ARM
5645 platform for these compatibility libraries, as the ARM
5646 toolchain is not really at the same level as the toolchain
5647 for other architectures like x86 and does not support
5648 IFUNC. Please make sure to use --enable-compat-libs only
5649 during a transitional period!
5651 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
5652 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5653 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5654 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5655 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5656 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5657 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5658 Zbigniew Jędrzejewski-Szmek
5660 — Berlin, 2014-02-24
5664 * A new component "systemd-networkd" has been added that can
5665 be used to configure local network interfaces statically or
5666 via DHCP. It is capable of bringing up bridges, VLANs, and
5667 bonding. Currently, no hook-ups for interactive network
5668 configuration are provided. Use this for your initrd,
5669 container, embedded, or server setup if you need a simple,
5670 yet powerful, network configuration solution. This
5671 configuration subsystem is quite nifty, as it allows wildcard
5672 hotplug matching in interfaces. For example, with a single
5673 configuration snippet, you can configure that all Ethernet
5674 interfaces showing up are automatically added to a bridge,
5675 or similar. It supports link-sensing and more.
5677 * A new tool "systemd-socket-proxyd" has been added which can
5678 act as a bidirectional proxy for TCP sockets. This is
5679 useful for adding socket activation support to services that
5680 do not actually support socket activation, including virtual
5681 machines and the like.
5683 * Add a new tool to save/restore rfkill state on
5686 * Save/restore state of keyboard backlights in addition to
5687 display backlights on shutdown/boot.
5689 * udev learned a new SECLABEL{} construct to label device
5690 nodes with a specific security label when they appear. For
5691 now, only SECLABEL{selinux} is supported, but the syntax is
5692 prepared for additional security frameworks.
5694 * udev gained a new scheme to configure link-level attributes
5695 from files in /etc/systemd/network/*.link. These files can
5696 match against MAC address, device path, driver name and type,
5697 and will apply attributes like the naming policy, link speed,
5698 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
5699 address assignment policy (randomized, ...).
5701 * The configuration of network interface naming rules for
5702 "permanent interface names" has changed: a new NamePolicy=
5703 setting in the [Link] section of .link files determines the
5704 priority of possible naming schemes (onboard, slot, MAC,
5705 path). The default value of this setting is determined by
5706 /usr/lib/net/links/99-default.link. Old
5707 80-net-name-slot.rules udev configuration file has been
5708 removed, so local configuration overriding this file should
5709 be adapted to override 99-default.link instead.
5711 * When the User= switch is used in a unit file, also
5712 initialize $SHELL= based on the user database entry.
5714 * systemd no longer depends on libdbus. All communication is
5715 now done with sd-bus, systemd's low-level bus library
5718 * kdbus support has been added to PID 1 itself. When kdbus is
5719 enabled, this causes PID 1 to set up the system bus and
5720 enable support for a new ".busname" unit type that
5721 encapsulates bus name activation on kdbus. It works a little
5722 bit like ".socket" units, except for bus names. A new
5723 generator has been added that converts classic dbus1 service
5724 activation files automatically into native systemd .busname
5727 * sd-bus: add a light-weight vtable implementation that allows
5728 defining objects on the bus with a simple static const
5729 vtable array of its methods, signals and properties.
5731 * systemd will not generate or install static dbus
5732 introspection data anymore to /usr/share/dbus-1/interfaces,
5733 as the precise format of these files is unclear, and
5734 nothing makes use of it.
5736 * A proxy daemon is now provided to proxy clients connecting
5737 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5738 compatibility with classic D-Bus.
5740 * A bus driver implementation has been added that supports the
5741 classic D-Bus bus driver calls on kdbus, also for
5742 compatibility purposes.
5744 * A new API "sd-event.h" has been added that implements a
5745 minimal event loop API built around epoll. It provides a
5746 couple of features that direct epoll usage is lacking:
5747 prioritization of events, scales to large numbers of timer
5748 events, per-event timer slack (accuracy), system-wide
5749 coalescing of timer events, exit handlers, watchdog
5750 supervision support using systemd's sd_notify() API, child
5753 * A new API "sd-rntl.h" has been added that provides an API
5754 around the route netlink interface of the kernel, similar in
5755 style to "sd-bus.h".
5757 * A new API "sd-dhcp-client.h" has been added that provides a
5758 small DHCPv4 client-side implementation. This is used by
5761 * There is a new kernel command line option
5762 "systemd.restore_state=0|1". When set to "0", none of the
5763 systemd tools will restore saved runtime state to hardware
5764 devices. More specifically, the rfkill and backlight states
5767 * The FsckPassNo= compatibility option in mount/service units
5768 has been removed. The fstab generator will now add the
5769 necessary dependencies automatically, and does not require
5770 PID1's support for that anymore.
5772 * journalctl gained a new switch, --list-boots, that lists
5773 recent boots with their times and boot IDs.
5775 * The various tools like systemctl, loginctl, timedatectl,
5776 busctl, systemd-run, ... have gained a new switch "-M" to
5777 connect to a specific, local OS container (as direct
5778 connection, without requiring SSH). This works on any
5779 container that is registered with machined, such as those
5780 created by libvirt-lxc or nspawn.
5782 * systemd-run and systemd-analyze also gained support for "-H"
5783 to connect to remote hosts via SSH. This is particularly
5784 useful for systemd-run because it enables queuing of jobs
5785 onto remote systems.
5787 * machinectl gained a new command "login" to open a getty
5788 login in any local container. This works with any container
5789 that is registered with machined (such as those created by
5790 libvirt-lxc or nspawn), and which runs systemd inside.
5792 * machinectl gained a new "reboot" command that may be used to
5793 trigger a reboot on a specific container that is registered
5794 with machined. This works on any container that runs an init
5795 system of some kind.
5797 * systemctl gained a new "list-timers" command to print a nice
5798 listing of installed timer units with the times they elapse
5801 * Alternative reboot() parameters may now be specified on the
5802 "systemctl reboot" command line and are passed to the
5803 reboot() system call.
5805 * systemctl gained a new --job-mode= switch to configure the
5806 mode to queue a job with. This is a more generic version of
5807 --fail, --irreversible, and --ignore-dependencies, which are
5808 still available but not advertised anymore.
5810 * /etc/systemd/system.conf gained new settings to configure
5811 various default timeouts of units, as well as the default
5812 start limit interval and burst. These may still be overridden
5815 * PID1 will now export on the bus profile data of the security
5816 policy upload process (such as the SELinux policy upload to
5819 * journald: when forwarding logs to the console, include
5820 timestamps (following the setting in
5821 /sys/module/printk/parameters/time).
5823 * OnCalendar= in timer units now understands the special
5824 strings "yearly" and "annually". (Both are equivalent)
5826 * The accuracy of timer units is now configurable with the new
5827 AccuracySec= setting. It defaults to 1min.
5829 * A new dependency type JoinsNamespaceOf= has been added that
5830 allows running two services within the same /tmp and network
5831 namespace, if PrivateNetwork= or PrivateTmp= are used.
5833 * A new command "cat" has been added to systemctl. It outputs
5834 the original unit file of a unit, and concatenates the
5835 contents of additional "drop-in" unit file snippets, so that
5836 the full configuration is shown.
5838 * systemctl now supports globbing on the various "list-xyz"
5839 commands, like "list-units" or "list-sockets", as well as on
5840 those commands which take multiple unit names.
5842 * journalctl's --unit= switch gained support for globbing.
5844 * All systemd daemons now make use of the watchdog logic so
5845 that systemd automatically notices when they hang.
5847 * If the $container_ttys environment variable is set,
5848 getty-generator will automatically spawn a getty for each
5849 listed tty. This is useful for container managers to request
5850 login gettys to be spawned on as many ttys as needed.
5852 * %h, %s, %U specifier support is not available anymore when
5853 used in unit files for PID 1. This is because NSS calls are
5854 not safe from PID 1. They stay available for --user
5855 instances of systemd, and as special case for the root user.
5857 * loginctl gained a new "--no-legend" switch to turn off output
5860 * The "sd-login.h" API gained three new calls:
5861 sd_session_is_remote(), sd_session_get_remote_user(),
5862 sd_session_get_remote_host() to query information about
5865 * The udev hardware database now also carries vendor/product
5866 information of SDIO devices.
5868 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5869 determine whether watchdog notifications are requested by
5872 * Socket-activated per-connection services now include a
5873 short description of the connection parameters in the
5876 * tmpfiles gained a new "--boot" option. When this is not used,
5877 only lines where the command character is not suffixed with
5878 "!" are executed. When this option is specified, those
5879 options are executed too. This partitions tmpfiles
5880 directives into those that can be safely executed at any
5881 time, and those which should be run only at boot (for
5882 example, a line that creates /run/nologin).
5884 * A new API "sd-resolve.h" has been added which provides a simple
5885 asynchronous wrapper around glibc NSS host name resolution
5886 calls, such as getaddrinfo(). In contrast to glibc's
5887 getaddrinfo_a(), it does not use signals. In contrast to most
5888 other asynchronous name resolution libraries, this one does
5889 not reimplement DNS, but reuses NSS, so that alternate
5890 host name resolution systems continue to work, such as mDNS,
5891 LDAP, etc. This API is based on libasyncns, but it has been
5892 cleaned up for inclusion in systemd.
5894 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5895 "sd-daemon.h" are no longer found in individual libraries
5896 libsystemd-journal.so, libsystemd-login.so,
5897 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
5898 merged them into a single library, libsystemd.so, which
5899 provides all symbols. The reason for this is cyclic
5900 dependencies, as these libraries tend to use each other's
5901 symbols. So far, we have managed to workaround that by linking
5902 a copy of a good part of our code into each of these
5903 libraries again and again, which, however, makes certain
5904 things hard to do, like sharing static variables. Also, it
5905 substantially increases footprint. With this change, there
5906 is only one library for the basic APIs systemd
5907 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5908 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5909 library as well, however are subject to the --enable-kdbus
5910 switch (see below). Note that "sd-dhcp-client.h" is not part
5911 of this library (this is because it only consumes, never
5912 provides, services of/to other APIs). To make the transition
5913 easy from the separate libraries to the unified one, we
5914 provide the --enable-compat-libs compile-time switch which
5915 will generate stub libraries that are compatible with the
5916 old ones but redirect all calls to the new one.
5918 * All of the kdbus logic and the new APIs "sd-bus.h",
5919 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
5920 and "sd-utf8.h" are compile-time optional via the
5921 "--enable-kdbus" switch, and they are not compiled in by
5922 default. To make use of kdbus, you have to explicitly enable
5923 the switch. Note however, that neither the kernel nor the
5924 userspace API for all of this is considered stable yet. We
5925 want to maintain the freedom to still change the APIs for
5926 now. By specifying this build-time switch, you acknowledge
5927 that you are aware of the instability of the current
5930 * Also, note that while kdbus is pretty much complete,
5931 it lacks one thing: proper policy support. This means you
5932 can build a fully working system with all features; however,
5933 it will be highly insecure. Policy support will be added in
5934 one of the next releases, at the same time that we will
5935 declare the APIs stable.
5937 * When the kernel command line argument "kdbus" is specified,
5938 systemd will automatically load the kdbus.ko kernel module. At
5939 this stage of development, it is only useful for testing kdbus
5940 and should not be used in production. Note: if "--enable-kdbus"
5941 is specified, and the kdbus.ko kernel module is available, and
5942 "kdbus" is added to the kernel command line, the entire system
5943 runs with kdbus instead of dbus-daemon, with the above mentioned
5944 problem of missing the system policy enforcement. Also a future
5945 version of kdbus.ko or a newer systemd will not be compatible with
5946 each other, and will unlikely be able to boot the machine if only
5947 one of them is updated.
5949 * systemctl gained a new "import-environment" command which
5950 uploads the caller's environment (or parts thereof) into the
5951 service manager so that it is inherited by services started
5952 by the manager. This is useful to upload variables like
5953 $DISPLAY into the user service manager.
5955 * A new PrivateDevices= switch has been added to service units
5956 which allows running a service with a namespaced /dev
5957 directory that does not contain any device nodes for
5958 physical devices. More specifically, it only includes devices
5959 such as /dev/null, /dev/urandom, and /dev/zero which are API
5962 * logind has been extended to support behaviour like VT
5963 switching on seats that do not support a VT. This makes
5964 multi-session available on seats that are not the first seat
5965 (seat0), and on systems where kernel support for VTs has
5966 been disabled at compile-time.
5968 * If a process holds a delay lock for system sleep or shutdown
5969 and fails to release it in time, we will now log its
5970 identity. This makes it easier to identify processes that
5971 cause slow suspends or power-offs.
5973 * When parsing /etc/crypttab, support for a new key-slot=
5974 option as supported by Debian is added. It allows indicating
5975 which LUKS slot to use on disk, speeding up key loading.
5977 * The sd_journal_sendv() API call has been checked and
5978 officially declared to be async-signal-safe so that it may
5979 be invoked from signal handlers for logging purposes.
5981 * Boot-time status output is now enabled automatically after a
5982 short timeout if boot does not progress, in order to give
5983 the user an indication what she or he is waiting for.
5985 * The boot-time output has been improved to show how much time
5986 remains until jobs expire.
5988 * The KillMode= switch in service units gained a new possible
5989 value "mixed". If set, and the unit is shut down, then the
5990 initial SIGTERM signal is sent only to the main daemon
5991 process, while the following SIGKILL signal is sent to
5992 all remaining processes of the service.
5994 * When a scope unit is registered, a new property "Controller"
5995 may be set. If set to a valid bus name, systemd will send a
5996 RequestStop() signal to this name when it would like to shut
5997 down the scope. This may be used to hook manager logic into
5998 the shutdown logic of scope units. Also, scope units may now
5999 be put in a special "abandoned" state, in which case the
6000 manager process which created them takes no further
6001 responsibilities for it.
6003 * When reading unit files, systemd will now verify
6004 the access mode of these files, and warn about certain
6005 suspicious combinations. This has been added to make it
6006 easier to track down packaging bugs where unit files are
6007 marked executable or world-writable.
6009 * systemd-nspawn gained a new "--setenv=" switch to set
6010 container-wide environment variables. The similar option in
6011 systemd-activate was renamed from "--environment=" to
6012 "--setenv=" for consistency.
6014 * systemd-nspawn has been updated to create a new kdbus domain
6015 for each container that is invoked, thus allowing each
6016 container to have its own set of system and user buses,
6017 independent of the host.
6019 * systemd-nspawn gained a new --drop-capability= switch to run
6020 the container with less capabilities than the default. Both
6021 --drop-capability= and --capability= now take the special
6022 string "all" for dropping or keeping all capabilities.
6024 * systemd-nspawn gained new switches for executing containers
6025 with specific SELinux labels set.
6027 * systemd-nspawn gained a new --quiet switch to not generate
6028 any additional output but the container's own console
6031 * systemd-nspawn gained a new --share-system switch to run a
6032 container without PID namespacing enabled.
6034 * systemd-nspawn gained a new --register= switch to control
6035 whether the container is registered with systemd-machined or
6036 not. This is useful for containers that do not run full
6037 OS images, but only specific apps.
6039 * systemd-nspawn gained a new --keep-unit which may be used
6040 when invoked as the only program from a service unit, and
6041 results in registration of the unit service itself in
6042 systemd-machined, instead of a newly opened scope unit.
6044 * systemd-nspawn gained a new --network-interface= switch for
6045 moving arbitrary interfaces to the container. The new
6046 --network-veth switch creates a virtual Ethernet connection
6047 between host and container. The new --network-bridge=
6048 switch then allows assigning the host side of this virtual
6049 Ethernet connection to a bridge device.
6051 * systemd-nspawn gained a new --personality= switch for
6052 setting the kernel personality for the container. This is
6053 useful when running a 32-bit container on a 64-bit host. A
6054 similar option Personality= is now also available for service
6057 * logind will now also track a "Desktop" identifier for each
6058 session which encodes the desktop environment of it. This is
6059 useful for desktop environments that want to identify
6060 multiple running sessions of itself easily.
6062 * A new SELinuxContext= setting for service units has been
6063 added that allows setting a specific SELinux execution
6064 context for a service.
6066 * Most systemd client tools will now honour $SYSTEMD_LESS for
6067 settings of the "less" pager. By default, these tools will
6068 override $LESS to allow certain operations to work, such as
6069 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6070 influence this logic.
6072 * systemd's "seccomp" hook-up has been changed to make use of
6073 the libseccomp library instead of using its own
6074 implementation. This has benefits for portability among
6077 * For usage together with SystemCallFilter=, a new
6078 SystemCallErrorNumber= setting has been introduced that
6079 allows configuration of a system error number to be returned
6080 on filtered system calls, instead of immediately killing the
6081 process. Also, SystemCallArchitectures= has been added to
6082 limit access to system calls of a particular architecture
6083 (in order to turn off support for unused secondary
6084 architectures). There is also a global
6085 SystemCallArchitectures= setting in system.conf now to turn
6086 off support for non-native system calls system-wide.
6088 * systemd requires a kernel with a working name_to_handle_at(),
6089 please see the kernel config requirements in the README file.
6091 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6092 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6093 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6094 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6095 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6096 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6097 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6098 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6099 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6100 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6101 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6102 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6103 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6104 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6105 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6106 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6107 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6108 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6109 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6110 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6111 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6112 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6113 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6114 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6116 — Berlin, 2014-02-20
6120 * logind has gained support for facilitating privileged input
6121 and drm device access for unprivileged clients. This work is
6122 useful to allow Wayland display servers (and similar
6123 programs, such as kmscon) to run under the user's ID and
6124 access input and drm devices which are normally
6125 protected. When this is used (and the kernel is new enough)
6126 logind will "mute" IO on the file descriptors passed to
6127 Wayland as long as it is in the background and "unmute" it
6128 if it returns into the foreground. This allows secure
6129 session switching without allowing background sessions to
6130 eavesdrop on input and display data. This also introduces
6131 session switching support if VT support is turned off in the
6132 kernel, and on seats that are not seat0.
6134 * A new kernel command line option luks.options= is understood
6135 now which allows specifying LUKS options for usage for LUKS
6136 encrypted partitions specified with luks.uuid=.
6138 * tmpfiles.d(5) snippets may now use specifier expansion in
6139 path names. More specifically %m, %b, %H, %v, are now
6140 replaced by the local machine id, boot id, hostname, and
6141 kernel version number.
6143 * A new tmpfiles.d(5) command "m" has been introduced which
6144 may be used to change the owner/group/access mode of a file
6145 or directory if it exists, but do nothing if it does not.
6147 * This release removes high-level support for the
6148 MemorySoftLimit= cgroup setting. The underlying kernel
6149 cgroup attribute memory.soft_limit= is currently badly
6150 designed and likely to be removed from the kernel API in its
6151 current form, hence we should not expose it for now.
6153 * The memory.use_hierarchy cgroup attribute is now enabled for
6154 all cgroups systemd creates in the memory cgroup
6155 hierarchy. This option is likely to be come the built-in
6156 default in the kernel anyway, and the non-hierarchical mode
6157 never made much sense in the intrinsically hierarchical
6160 * A new field _SYSTEMD_SLICE= is logged along with all journal
6161 messages containing the slice a message was generated
6162 from. This is useful to allow easy per-customer filtering of
6163 logs among other things.
6165 * systemd-journald will no longer adjust the group of journal
6166 files it creates to the "systemd-journal" group. Instead we
6167 rely on the journal directory to be owned by the
6168 "systemd-journal" group, and its setgid bit set, so that the
6169 kernel file system layer will automatically enforce that
6170 journal files inherit this group assignment. The reason for
6171 this change is that we cannot allow NSS look-ups from
6172 journald which would be necessary to resolve
6173 "systemd-journal" to a numeric GID, because this might
6174 create deadlocks if NSS involves synchronous queries to
6175 other daemons (such as nscd, or sssd) which in turn are
6176 logging clients of journald and might block on it, which
6177 would then dead lock. A tmpfiles.d(5) snippet included in
6178 systemd will make sure the setgid bit and group are
6179 properly set on the journal directory if it exists on every
6180 boot. However, we recommend adjusting it manually after
6181 upgrades too (or from RPM scriptlets), so that the change is
6182 not delayed until next reboot.
6184 * Backlight and random seed files in /var/lib/ have moved into
6185 the /var/lib/systemd/ directory, in order to centralize all
6186 systemd generated files in one directory.
6188 * Boot time performance measurements (as displayed by
6189 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6190 performance information if that's available to determine how
6191 much time BIOS and boot loader initialization required. With
6192 a sufficiently new BIOS you hence no longer need to boot
6193 with Gummiboot to get access to such information.
6195 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6196 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6197 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6198 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6199 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6200 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6201 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6203 — Berlin, 2013-10-02
6207 * The Restart= option for services now understands a new
6208 on-watchdog setting, which will restart the service
6209 automatically if the service stops sending out watchdog keep
6210 alive messages (as configured with WatchdogSec=).
6212 * The getty generator (which is responsible for bringing up a
6213 getty on configured serial consoles) will no longer only
6214 start a getty on the primary kernel console but on all
6215 others, too. This makes the order in which console= is
6216 specified on the kernel command line less important.
6218 * libsystemd-logind gained a new sd_session_get_vt() call to
6219 retrieve the VT number of a session.
6221 * If the option "tries=0" is set for an entry of /etc/crypttab
6222 its passphrase is queried indefinitely instead of any
6223 maximum number of tries.
6225 * If a service with a configure PID file terminates its PID
6226 file will now be removed automatically if it still exists
6227 afterwards. This should put an end to stale PID files.
6229 * systemd-run will now also take relative binary path names
6230 for execution and no longer insists on absolute paths.
6232 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6233 paths that are optionally prefixed with "-" to indicate that
6234 it should not be considered a failure if they do not exist.
6236 * journalctl -o (and similar commands) now understands a new
6237 output mode "short-precise", it is similar to "short" but
6238 shows timestamps with usec accuracy.
6240 * The option "discard" (as known from Debian) is now
6241 synonymous to "allow-discards" in /etc/crypttab. In fact,
6242 "discard" is preferred now (since it is easier to remember
6245 * Some licensing clean-ups were made, so that more code is now
6246 LGPL-2.1 licensed than before.
6248 * A minimal tool to save/restore the display backlight
6249 brightness across reboots has been added. It will store the
6250 backlight setting as late as possible at shutdown, and
6251 restore it as early as possible during reboot.
6253 * A logic to automatically discover and enable home and swap
6254 partitions on GPT disks has been added. With this in place
6255 /etc/fstab becomes optional for many setups as systemd can
6256 discover certain partitions located on the root disk
6257 automatically. Home partitions are recognized under their
6258 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6259 partitions are recognized under their GPT type ID
6260 0657fd6da4ab43c484e50933c84b4f4f.
6262 * systemd will no longer pass any environment from the kernel
6263 or initrd to system services. If you want to set an
6264 environment for all services, do so via the kernel command
6265 line systemd.setenv= assignment.
6267 * The systemd-sysctl tool no longer natively reads the file
6268 /etc/sysctl.conf. If desired, the file should be symlinked
6269 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6270 legacy support by a symlink rather than built-in code, it
6271 also makes the otherwise hidden order of application of the
6272 different files visible. (Note that this partly reverts to a
6273 pre-198 application order of sysctl knobs!)
6275 * The "systemctl set-log-level" and "systemctl dump" commands
6276 have been moved to systemd-analyze.
6278 * systemd-run learned the new --remain-after-exit switch,
6279 which causes the scope unit not to be cleaned up
6280 automatically after the process terminated.
6282 * tmpfiles learned a new --exclude-prefix= switch to exclude
6283 certain paths from operation.
6285 * journald will now automatically flush all messages to disk
6286 as soon as a message at the log level CRIT, ALERT or EMERG
6289 Contributions from: Andrew Cook, Brandon Philips, Christian
6290 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6291 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6292 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6293 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6294 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6295 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6296 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6297 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6298 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6299 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6300 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6301 William Giokas, Zbigniew Jędrzejewski-Szmek
6303 — Berlin, 2013-09-13
6307 * The documentation has been updated to cover the various new
6308 concepts introduced with 205.
6310 * Unit files now understand the new %v specifier which
6311 resolves to the kernel version string as returned by "uname
6314 * systemctl now supports filtering the unit list output by
6315 load state, active state and sub state, using the new
6318 * "systemctl status" will now show the results of the
6319 condition checks (like ConditionPathExists= and similar) of
6320 the last start attempts of the unit. They are also logged to
6323 * "journalctl -b" may now be used to look for boot output of a
6324 specific boot. Try "journalctl -b -1" for the previous boot,
6325 but the syntax is substantially more powerful.
6327 * "journalctl --show-cursor" has been added which prints the
6328 cursor string the last shown log line. This may then be used
6329 with the new "journalctl --after-cursor=" switch to continue
6330 browsing logs from that point on.
6332 * "journalctl --force" may now be used to force regeneration
6335 * Creation of "dead" device nodes has been moved from udev
6336 into kmod and tmpfiles. Previously, udev would read the kmod
6337 databases to pre-generate dead device nodes based on meta
6338 information contained in kernel modules, so that these would
6339 be auto-loaded on access rather then at boot. As this
6340 does not really have much to do with the exposing actual
6341 kernel devices to userspace this has always been slightly
6342 alien in the udev codebase. Following the new scheme kmod
6343 will now generate a runtime snippet for tmpfiles from the
6344 module meta information and it now is tmpfiles' job to the
6345 create the nodes. This also allows overriding access and
6346 other parameters for the nodes using the usual tmpfiles
6347 facilities. As side effect this allows us to remove the
6348 CAP_SYS_MKNOD capability bit from udevd entirely.
6350 * logind's device ACLs may now be applied to these "dead"
6351 devices nodes too, thus finally allowing managed access to
6352 devices such as /dev/snd/sequencer without loading the
6353 backing module right-away.
6355 * A new RPM macro has been added that may be used to apply
6356 tmpfiles configuration during package installation.
6358 * systemd-detect-virt and ConditionVirtualization= now can
6359 detect User-Mode-Linux machines (UML).
6361 * journald will now implicitly log the effective capabilities
6362 set of processes in the message metadata.
6364 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6366 * The initrd interface has been simplified (more specifically,
6367 support for passing performance data via environment
6368 variables and fsck results via files in /run has been
6369 removed). These features were non-essential, and are
6370 nowadays available in a much nicer way by having systemd in
6371 the initrd serialize its state and have the hosts systemd
6372 deserialize it again.
6374 * The udev "keymap" data files and tools to apply keyboard
6375 specific mappings of scan to key codes, and force-release
6376 scan code lists have been entirely replaced by a udev
6377 "keyboard" builtin and a hwdb data file.
6379 * systemd will now honour the kernel's "quiet" command line
6380 argument also during late shutdown, resulting in a
6381 completely silent shutdown when used.
6383 * There's now an option to control the SO_REUSEPORT socket
6384 option in .socket units.
6386 * Instance units will now automatically get a per-template
6387 subslice of system.slice unless something else is explicitly
6388 configured. For example, instances of sshd@.service will now
6389 implicitly be placed in system-sshd.slice rather than
6390 system.slice as before.
6392 * Test coverage support may now be enabled at build time.
6394 Contributions from: Dave Reisner, Frederic Crozat, Harald
6395 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6396 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6397 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6398 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6399 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6400 Giokas, Zbigniew Jędrzejewski-Szmek
6402 — Berlin, 2013-07-23
6406 * Two new unit types have been introduced:
6408 Scope units are very similar to service units, however, are
6409 created out of pre-existing processes — instead of PID 1
6410 forking off the processes. By using scope units it is
6411 possible for system services and applications to group their
6412 own child processes (worker processes) in a powerful way
6413 which then maybe used to organize them, or kill them
6414 together, or apply resource limits on them.
6416 Slice units may be used to partition system resources in an
6417 hierarchical fashion and then assign other units to them. By
6418 default there are now three slices: system.slice (for all
6419 system services), user.slice (for all user sessions),
6420 machine.slice (for VMs and containers).
6422 Slices and scopes have been introduced primarily in
6423 context of the work to move cgroup handling to a
6424 single-writer scheme, where only PID 1
6425 creates/removes/manages cgroups.
6427 * There's a new concept of "transient" units. In contrast to
6428 normal units these units are created via an API at runtime,
6429 not from configuration from disk. More specifically this
6430 means it is now possible to run arbitrary programs as
6431 independent services, with all execution parameters passed
6432 in via bus APIs rather than read from disk. Transient units
6433 make systemd substantially more dynamic then it ever was,
6434 and useful as a general batch manager.
6436 * logind has been updated to make use of scope and slice units
6437 for managing user sessions. As a user logs in he will get
6438 his own private slice unit, to which all sessions are added
6439 as scope units. We also added support for automatically
6440 adding an instance of user@.service for the user into the
6441 slice. Effectively logind will no longer create cgroup
6442 hierarchies on its own now, it will defer entirely to PID 1
6443 for this by means of scope, service and slice units. Since
6444 user sessions this way become entities managed by PID 1
6445 the output of "systemctl" is now a lot more comprehensive.
6447 * A new mini-daemon "systemd-machined" has been added which
6448 may be used by virtualization managers to register local
6449 VMs/containers. nspawn has been updated accordingly, and
6450 libvirt will be updated shortly. machined will collect a bit
6451 of meta information about the VMs/containers, and assign
6452 them their own scope unit (see above). The collected
6453 meta-data is then made available via the "machinectl" tool,
6454 and exposed in "ps" and similar tools. machined/machinectl
6455 is compile-time optional.
6457 * As discussed earlier, the low-level cgroup configuration
6458 options ControlGroup=, ControlGroupModify=,
6459 ControlGroupPersistent=, ControlGroupAttribute= have been
6460 removed. Please use high-level attribute settings instead as
6461 well as slice units.
6463 * A new bus call SetUnitProperties() has been added to alter
6464 various runtime parameters of a unit. This is primarily
6465 useful to alter cgroup parameters dynamically in a nice way,
6466 but will be extended later on to make more properties
6467 modifiable at runtime. systemctl gained a new set-properties
6468 command that wraps this call.
6470 * A new tool "systemd-run" has been added which can be used to
6471 run arbitrary command lines as transient services or scopes,
6472 while configuring a number of settings via the command
6473 line. This tool is currently very basic, however already
6474 very useful. We plan to extend this tool to even allow
6475 queuing of execution jobs with time triggers from the
6476 command line, similar in fashion to "at".
6478 * nspawn will now inform the user explicitly that kernels with
6479 audit enabled break containers, and suggest the user to turn
6482 * Support for detecting the IMA and AppArmor security
6483 frameworks with ConditionSecurity= has been added.
6485 * journalctl gained a new "-k" switch for showing only kernel
6486 messages, mimicking dmesg output; in addition to "--user"
6487 and "--system" switches for showing only user's own logs
6490 * systemd-delta can now show information about drop-in
6491 snippets extending unit files.
6493 * libsystemd-bus has been substantially updated but is still
6494 not available as public API.
6496 * systemd will now look for the "debug" argument on the kernel
6497 command line and enable debug logging, similar to what
6498 "systemd.log_level=debug" already did before.
6500 * "systemctl set-default", "systemctl get-default" has been
6501 added to configure the default.target symlink, which
6502 controls what to boot into by default.
6504 * "systemctl set-log-level" has been added as a convenient
6505 way to raise and lower systemd logging threshold.
6507 * "systemd-analyze plot" will now show the time the various
6508 generators needed for execution, as well as information
6509 about the unit file loading.
6511 * libsystemd-journal gained a new sd_journal_open_files() call
6512 for opening specific journal files. journactl also gained a
6513 new switch to expose this new functionality. Previously we
6514 only supported opening all files from a directory, or all
6515 files from the system, as opening individual files only is
6516 racy due to journal file rotation.
6518 * systemd gained the new DefaultEnvironment= setting in
6519 /etc/systemd/system.conf to set environment variables for
6522 * If a privileged process logs a journal message with the
6523 OBJECT_PID= field set, then journald will automatically
6524 augment this with additional OBJECT_UID=, OBJECT_GID=,
6525 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6526 system services want to log events about specific client
6527 processes. journactl/systemctl has been updated to make use
6528 of this information if all log messages regarding a specific
6531 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6532 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6533 Reisner, David Coppa, David King, David Strauss, Eelco
6534 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6535 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6536 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6537 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6538 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6539 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6540 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6541 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6542 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6543 Łukasz Stelmach, 장동준
6547 * The Python bindings gained some minimal support for the APIs
6548 exposed by libsystemd-logind.
6550 * ConditionSecurity= gained support for detecting SMACK. Since
6551 this condition already supports SELinux and AppArmor we only
6552 miss IMA for this. Patches welcome!
6554 Contributions from: Karol Lewandowski, Lennart Poettering,
6555 Zbigniew Jędrzejewski-Szmek
6559 * systemd-nspawn will now create /etc/resolv.conf if
6560 necessary, before bind-mounting the host's file onto it.
6562 * systemd-nspawn will now store meta information about a
6563 container on the container's cgroup as extended attribute
6564 fields, including the root directory.
6566 * The cgroup hierarchy has been reworked in many ways. All
6567 objects any of the components systemd creates in the cgroup
6568 tree are now suffixed. More specifically, user sessions are
6569 now placed in cgroups suffixed with ".session", users in
6570 cgroups suffixed with ".user", and nspawn containers in
6571 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6572 names are now escaped in a simple scheme to avoid collision
6573 of userspace object names with kernel filenames. This work
6574 is preparation for making these objects relocatable in the
6575 cgroup tree, in order to allow easy resource partitioning of
6576 these objects without causing naming conflicts.
6578 * systemctl list-dependencies gained the new switches
6579 --plain, --reverse, --after and --before.
6581 * systemd-inhibit now shows the process name of processes that
6582 have taken an inhibitor lock.
6584 * nss-myhostname will now also resolve "localhost"
6585 implicitly. This makes /etc/hosts an optional file and
6586 nicely handles that on IPv6 ::1 maps to both "localhost" and
6589 * libsystemd-logind.so gained a new call
6590 sd_get_machine_names() to enumerate running containers and
6591 VMs (currently only supported by very new libvirt and
6592 nspawn). sd_login_monitor can now be used to watch
6593 VMs/containers coming and going.
6595 * .include is not allowed recursively anymore, and only in
6596 unit files. Usually it is better to use drop-in snippets in
6597 .d/*.conf anyway, as introduced with systemd 198.
6599 * systemd-analyze gained a new "critical-chain" command that
6600 determines the slowest chain of units run during system
6601 boot-up. It is very useful for tracking down where
6602 optimizing boot time is the most beneficial.
6604 * systemd will no longer allow manipulating service paths in
6605 the name=systemd:/system cgroup tree using ControlGroup= in
6606 units. (But is still fine with it in all other dirs.)
6608 * There's a new systemd-nspawn@.service service file that may
6609 be used to easily run nspawn containers as system
6610 services. With the container's root directory in
6611 /var/lib/container/foobar it is now sufficient to run
6612 "systemctl start systemd-nspawn@foobar.service" to boot it.
6614 * systemd-cgls gained a new parameter "--machine" to list only
6615 the processes within a certain container.
6617 * ConditionSecurity= now can check for "apparmor". We still
6618 are lacking checks for SMACK and IMA for this condition
6619 check though. Patches welcome!
6621 * A new configuration file /etc/systemd/sleep.conf has been
6622 added that may be used to configure which kernel operation
6623 systemd is supposed to execute when "suspend", "hibernate"
6624 or "hybrid-sleep" is requested. This makes the new kernel
6625 "freeze" state accessible to the user.
6627 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6628 the passed argument if applicable.
6630 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6631 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6632 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6633 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6634 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6635 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6636 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6641 * The output of 'systemctl list-jobs' got some polishing. The
6642 '--type=' argument may now be passed more than once. A new
6643 command 'systemctl list-sockets' has been added which shows
6644 a list of kernel sockets systemd is listening on with the
6645 socket units they belong to, plus the units these socket
6648 * The experimental libsystemd-bus library got substantial
6649 updates to work in conjunction with the (also experimental)
6650 kdbus kernel project. It works well enough to exchange
6651 messages with some sophistication. Note that kdbus is not
6652 ready yet, and the library is mostly an elaborate test case
6653 for now, and not installable.
6655 * systemd gained a new unit 'systemd-static-nodes.service'
6656 that generates static device nodes earlier during boot, and
6657 can run in conjunction with udev.
6659 * libsystemd-login gained a new call sd_pid_get_user_unit()
6660 to retrieve the user systemd unit a process is running
6661 in. This is useful for systems where systemd is used as
6664 * systemd-nspawn now places all containers in the new /machine
6665 top-level cgroup directory in the name=systemd
6666 hierarchy. libvirt will soon do the same, so that we get a
6667 uniform separation of /system, /user and /machine for system
6668 services, user processes and containers/virtual
6669 machines. This new cgroup hierarchy is also useful to stick
6670 stable names to specific container instances, which can be
6671 recognized later this way (this name may be controlled
6672 via systemd-nspawn's new -M switch). libsystemd-login also
6673 gained a new call sd_pid_get_machine_name() to retrieve the
6674 name of the container/VM a specific process belongs to.
6676 * bootchart can now store its data in the journal.
6678 * libsystemd-journal gained a new call
6679 sd_journal_add_conjunction() for AND expressions to the
6680 matching logic. This can be used to express more complex
6681 logical expressions.
6683 * journactl can now take multiple --unit= and --user-unit=
6686 * The cryptsetup logic now understands the "luks.key=" kernel
6687 command line switch for specifying a file to read the
6688 decryption key from. Also, if a configured key file is not
6689 found the tool will now automatically fall back to prompting
6692 * Python systemd.journal module was updated to wrap recently
6693 added functions from libsystemd-journal. The interface was
6694 changed to bring the low level interface in s.j._Reader
6695 closer to the C API, and the high level interface in
6696 s.j.Reader was updated to wrap and convert all data about
6699 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6700 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6701 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6702 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6703 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6704 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6708 * journalctl --update-catalog now understands a new --root=
6709 option to operate on catalogs found in a different root
6712 * During shutdown after systemd has terminated all running
6713 services a final killing loop kills all remaining left-over
6714 processes. We will now print the name of these processes
6715 when we send SIGKILL to them, since this usually indicates a
6718 * If /etc/crypttab refers to password files stored on
6719 configured mount points automatic dependencies will now be
6720 generated to ensure the specific mount is established first
6721 before the key file is attempted to be read.
6723 * 'systemctl status' will now show information about the
6724 network sockets a socket unit is listening on.
6726 * 'systemctl status' will also shown information about any
6727 drop-in configuration file for units. (Drop-In configuration
6728 files in this context are files such as
6729 /etc/systemd/systemd/foobar.service.d/*.conf)
6731 * systemd-cgtop now optionally shows summed up CPU times of
6732 cgroups. Press '%' while running cgtop to switch between
6733 percentage and absolute mode. This is useful to determine
6734 which cgroups use up the most CPU time over the entire
6735 runtime of the system. systemd-cgtop has also been updated
6736 to be 'pipeable' for processing with further shell tools.
6738 * 'hostnamectl set-hostname' will now allow setting of FQDN
6741 * The formatting and parsing of time span values has been
6742 changed. The parser now understands fractional expressions
6743 such as "5.5h". The formatter will now output fractional
6744 expressions for all time spans under 1min, i.e. "5.123456s"
6745 rather than "5s 123ms 456us". For time spans under 1s
6746 millisecond values are shown, for those under 1ms
6747 microsecond values are shown. This should greatly improve
6748 all time-related output of systemd.
6750 * libsystemd-login and libsystemd-journal gained new
6751 functions for querying the poll() events mask and poll()
6752 timeout value for integration into arbitrary event
6755 * localectl gained the ability to list available X11 keymaps
6756 (models, layouts, variants, options).
6758 * 'systemd-analyze dot' gained the ability to filter for
6759 specific units via shell-style globs, to create smaller,
6760 more useful graphs. I.e. it is now possible to create simple
6761 graphs of all the dependencies between only target units, or
6762 of all units that Avahi has dependencies with.
6764 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6765 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6766 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6767 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6768 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6769 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6770 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6774 * The boot-time readahead implementation for rotating media
6775 will now read the read-ahead data in multiple passes which
6776 consist of all read requests made in equidistant time
6777 intervals. This means instead of strictly reading read-ahead
6778 data in its physical order on disk we now try to find a
6779 middle ground between physical and access time order.
6781 * /etc/os-release files gained a new BUILD_ID= field for usage
6782 on operating systems that provide continuous builds of OS
6785 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6786 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6787 William Douglas, Zbigniew Jędrzejewski-Szmek
6791 * systemd-python gained an API exposing libsystemd-daemon.
6793 * The SMACK setup logic gained support for uploading CIPSO
6796 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6797 ReadOnlyDirectories= and InaccessibleDirectories= has
6798 changed. The private /tmp and /var/tmp directories are now
6799 shared by all processes of a service (which means
6800 ExecStartPre= may now leave data in /tmp that ExecStart= of
6801 the same service can still access). When a service is
6802 stopped its temporary directories are immediately deleted
6803 (normal clean-up with tmpfiles is still done in addition to
6806 * By default, systemd will now set a couple of sysctl
6807 variables in the kernel: the safe sysrq options are turned
6808 on, IP route verification is turned on, and source routing
6809 disabled. The recently added hardlink and softlink
6810 protection of the kernel is turned on. These settings should
6811 be reasonably safe, and good defaults for all new systems.
6813 * The predictable network naming logic may now be turned off
6814 with a new kernel command line switch: net.ifnames=0.
6816 * A new libsystemd-bus module has been added that implements a
6817 pretty complete D-Bus client library. For details see:
6819 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
6821 * journald will now explicitly flush the journal files to disk
6822 at the latest 5min after each write. The file will then also
6823 be marked offline until the next write. This should increase
6824 reliability in case of a crash. The synchronization delay
6825 can be configured via SyncIntervalSec= in journald.conf.
6827 * There's a new remote-fs-setup.target unit that can be used
6828 to pull in specific services when at least one remote file
6829 system is to be mounted.
6831 * There are new targets timers.target and paths.target as
6832 canonical targets to pull user timer and path units in
6833 from. This complements sockets.target with a similar
6834 purpose for socket units.
6836 * libudev gained a new call udev_device_set_attribute_value()
6837 to set sysfs attributes of a device.
6839 * The udev daemon now sets the default number of worker
6840 processes executed in parallel based on the number of available
6841 CPUs instead of the amount of available RAM. This is supposed
6842 to provide a more reliable default and limit a too aggressive
6843 parallelism for setups with 1000s of devices connected.
6845 Contributions from: Auke Kok, Colin Walters, Cristian
6846 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6847 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6848 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6849 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6850 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6851 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6852 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6853 Zbigniew Jędrzejewski-Szmek
6857 * Configuration of unit files may now be extended via drop-in
6858 files without having to edit/override the unit files
6859 themselves. More specifically, if the administrator wants to
6860 change one value for a service file foobar.service he can
6861 now do so by dropping in a configuration snippet into
6862 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
6863 will load all these snippets and apply them on top of the
6864 main unit configuration file, possibly extending or
6865 overriding its settings. Using these drop-in snippets is
6866 generally nicer than the two earlier options for changing
6867 unit files locally: copying the files from
6868 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6869 them there; or creating a new file in /etc/systemd/system/
6870 that incorporates the original one via ".include". Drop-in
6871 snippets into these .d/ directories can be placed in any
6872 directory systemd looks for units in, and the usual
6873 overriding semantics between /usr/lib, /etc and /run apply
6876 * Most unit file settings which take lists of items can now be
6877 reset by assigning the empty string to them. For example,
6878 normally, settings such as Environment=FOO=BAR append a new
6879 environment variable assignment to the environment block,
6880 each time they are used. By assigning Environment= the empty
6881 string the environment block can be reset to empty. This is
6882 particularly useful with the .d/*.conf drop-in snippets
6883 mentioned above, since this adds the ability to reset list
6884 settings from vendor unit files via these drop-ins.
6886 * systemctl gained a new "list-dependencies" command for
6887 listing the dependencies of a unit recursively.
6889 * Inhibitors are now honored and listed by "systemctl
6890 suspend", "systemctl poweroff" (and similar) too, not only
6891 GNOME. These commands will also list active sessions by
6894 * Resource limits (as exposed by the various control group
6895 controllers) can now be controlled dynamically at runtime
6896 for all units. More specifically, you can now use a command
6897 like "systemctl set-cgroup-attr foobar.service cpu.shares
6898 2000" to alter the CPU shares a specific service gets. These
6899 settings are stored persistently on disk, and thus allow the
6900 administrator to easily adjust the resource usage of
6901 services with a few simple commands. This dynamic resource
6902 management logic is also available to other programs via the
6903 bus. Almost any kernel cgroup attribute and controller is
6906 * systemd-vconsole-setup will now copy all font settings to
6907 all allocated VTs, where it previously applied them only to
6910 * libsystemd-login gained the new sd_session_get_tty() API
6913 * This release drops support for a few legacy or
6914 distribution-specific LSB facility names when parsing init
6915 scripts: $x-display-manager, $mail-transfer-agent,
6916 $mail-transport-agent, $mail-transfer-agent, $smtp,
6917 $null. Also, the mail-transfer-agent.target unit backing
6918 this has been removed. Distributions which want to retain
6919 compatibility with this should carry the burden for
6920 supporting this themselves and patch support for these back
6921 in, if they really need to. Also, the facilities $syslog and
6922 $local_fs are now ignored, since systemd does not support
6923 early-boot LSB init scripts anymore, and these facilities
6924 are implied anyway for normal services. syslog.target has
6927 * There are new bus calls on PID1's Manager object for
6928 cancelling jobs, and removing snapshot units. Previously,
6929 both calls were only available on the Job and Snapshot
6932 * systemd-journal-gatewayd gained SSL support.
6934 * The various "environment" files, such as /etc/locale.conf
6935 now support continuation lines with a backslash ("\") as
6936 last character in the line, similarly in style (but different)
6937 to how this is supported in shells.
6939 * For normal user processes the _SYSTEMD_USER_UNIT= field is
6940 now implicitly appended to every log entry logged. systemctl
6941 has been updated to filter by this field when operating on a
6942 user systemd instance.
6944 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
6945 CAP_AUDIT_CONTROL capabilities to the capabilities set for
6946 the container. This makes it easier to boot unmodified
6947 Fedora systems in a container, which however still requires
6948 audit=0 to be passed on the kernel command line. Auditing in
6949 kernel and userspace is unfortunately still too broken in
6950 context of containers, hence we recommend compiling it out
6951 of the kernel or using audit=0. Hopefully this will be fixed
6952 one day for good in the kernel.
6954 * nspawn gained the new --bind= and --bind-ro= parameters to
6955 bind mount specific directories from the host into the
6958 * nspawn will now mount its own devpts file system instance
6959 into the container, in order not to leak pty devices from
6960 the host into the container.
6962 * systemd will now read the firmware boot time performance
6963 information from the EFI variables, if the used boot loader
6964 supports this, and takes it into account for boot performance
6965 analysis via "systemd-analyze". This is currently supported
6966 only in conjunction with Gummiboot, but could be supported
6967 by other boot loaders too. For details see:
6969 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
6971 * A new generator has been added that automatically mounts the
6972 EFI System Partition (ESP) to /boot, if that directory
6973 exists, is empty, and no other file system has been
6974 configured to be mounted there.
6976 * logind will now send out PrepareForSleep(false) out
6977 unconditionally, after coming back from suspend. This may be
6978 used by applications as asynchronous notification for
6979 system resume events.
6981 * "systemctl unlock-sessions" has been added, that allows
6982 unlocking the screens of all user sessions at once, similar
6983 to how "systemctl lock-sessions" already locked all users
6984 sessions. This is backed by a new D-Bus call UnlockSessions().
6986 * "loginctl seat-status" will now show the master device of a
6987 seat. (i.e. the device of a seat that needs to be around for
6988 the seat to be considered available, usually the graphics
6991 * tmpfiles gained a new "X" line type, that allows
6992 configuration of files and directories (with wildcards) that
6993 shall be excluded from automatic cleanup ("aging").
6995 * udev default rules set the device node permissions now only
6996 at "add" events, and do not change them any longer with a
6997 later "change" event.
6999 * The log messages for lid events and power/sleep keypresses
7000 now carry a message ID.
7002 * We now have a substantially larger unit test suite, but this
7003 continues to be work in progress.
7005 * udevadm hwdb gained a new --root= parameter to change the
7006 root directory to operate relative to.
7008 * logind will now issue a background sync() request to the kernel
7009 early at shutdown, so that dirty buffers are flushed to disk early
7010 instead of at the last moment, in order to optimize shutdown
7013 * A new bootctl tool has been added that is an interface for
7014 certain boot loader operations. This is currently a preview
7015 and is likely to be extended into a small mechanism daemon
7016 like timedated, localed, hostnamed, and can be used by
7017 graphical UIs to enumerate available boot options, and
7018 request boot into firmware operations.
7020 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7021 the rest of the package. It also has been updated to work
7022 correctly in initrds.
7024 * Policykit previously has been runtime optional, and is now
7025 also compile time optional via a configure switch.
7027 * systemd-analyze has been reimplemented in C. Also "systemctl
7028 dot" has moved into systemd-analyze.
7030 * "systemctl status" with no further parameters will now print
7031 the status of all active or failed units.
7033 * Operations such as "systemctl start" can now be executed
7034 with a new mode "--irreversible" which may be used to queue
7035 operations that cannot accidentally be reversed by a later
7036 job queuing. This is by default used to make shutdown
7037 requests more robust.
7039 * The Python API of systemd now gained a new module for
7040 reading journal files.
7042 * A new tool kernel-install has been added that can install
7043 kernel images according to the Boot Loader Specification:
7045 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7047 * Boot time console output has been improved to provide
7048 animated boot time output for hanging jobs.
7050 * A new tool systemd-activate has been added which can be used
7051 to test socket activation with, directly from the command
7052 line. This should make it much easier to test and debug
7053 socket activation in daemons.
7055 * journalctl gained a new "--reverse" (or -r) option to show
7056 journal output in reverse order (i.e. newest line first).
7058 * journalctl gained a new "--pager-end" (or -e) option to jump
7059 to immediately jump to the end of the journal in the
7060 pager. This is only supported in conjunction with "less".
7062 * journalctl gained a new "--user-unit=" option, that works
7063 similarly to "--unit=" but filters for user units rather than
7066 * A number of unit files to ease adoption of systemd in
7067 initrds has been added. This moves some minimal logic from
7068 the various initrd implementations into systemd proper.
7070 * The journal files are now owned by a new group
7071 "systemd-journal", which exists specifically to allow access
7072 to the journal, and nothing else. Previously, we used the
7073 "adm" group for that, which however possibly covers more
7074 than just journal/log file access. This new group is now
7075 already used by systemd-journal-gatewayd to ensure this
7076 daemon gets access to the journal files and as little else
7077 as possible. Note that "make install" will also set FS ACLs
7078 up for /var/log/journal to give "adm" and "wheel" read
7079 access to it, in addition to "systemd-journal" which owns
7080 the journal files. We recommend that packaging scripts also
7081 add read access to "adm" + "wheel" to /var/log/journal, and
7082 all existing/future journal files. To normal users and
7083 administrators little changes, however packagers need to
7084 ensure to create the "systemd-journal" system group at
7085 package installation time.
7087 * The systemd-journal-gatewayd now runs as unprivileged user
7088 systemd-journal-gateway:systemd-journal-gateway. Packaging
7089 scripts need to create these system user/group at
7092 * timedated now exposes a new boolean property CanNTP that
7093 indicates whether a local NTP service is available or not.
7095 * systemd-detect-virt will now also detect xen PVs
7097 * The pstore file system is now mounted by default, if it is
7100 * In addition to the SELinux and IMA policies we will now also
7101 load SMACK policies at early boot.
7103 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7104 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7105 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7106 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7107 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7108 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7109 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7110 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7111 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7112 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7113 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7114 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7115 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7116 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7120 * Timer units now support calendar time events in addition to
7121 monotonic time events. That means you can now trigger a unit
7122 based on a calendar time specification such as "Thu,Fri
7123 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7124 or fifth day of any month of the year 2013, given that it is
7125 a thursday or friday. This brings timer event support
7126 considerably closer to cron's capabilities. For details on
7127 the supported calendar time specification language see
7130 * udev now supports a number of different naming policies for
7131 network interfaces for predictable names, and a combination
7132 of these policies is now the default. Please see this wiki
7133 document for details:
7135 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
7137 * Auke Kok's bootchart implementation has been added to the
7138 systemd tree. It is an optional component that can graph the
7139 boot in quite some detail. It is one of the best bootchart
7140 implementations around and minimal in its code and
7143 * nss-myhostname has been integrated into the systemd source
7144 tree. nss-myhostname guarantees that the local hostname
7145 always stays resolvable via NSS. It has been a weak
7146 requirement of systemd-hostnamed since a long time, and
7147 since its code is actually trivial we decided to just
7148 include it in systemd's source tree. It can be turned off
7149 with a configure switch.
7151 * The read-ahead logic is now capable of properly detecting
7152 whether a btrfs file system is on SSD or rotating media, in
7153 order to optimize the read-ahead scheme. Previously, it was
7154 only capable of detecting this on traditional file systems
7157 * In udev, additional device properties are now read from the
7158 IAB in addition to the OUI database. Also, Bluetooth company
7159 identities are attached to the devices as well.
7161 * In service files %U may be used as specifier that is
7162 replaced by the configured user name of the service.
7164 * nspawn may now be invoked without a controlling TTY. This
7165 makes it suitable for invocation as its own service. This
7166 may be used to set up a simple containerized server system
7167 using only core OS tools.
7169 * systemd and nspawn can now accept socket file descriptors
7170 when they are started for socket activation. This enables
7171 implementation of socket activated nspawn
7172 containers. i.e. think about autospawning an entire OS image
7173 when the first SSH or HTTP connection is received. We expect
7174 that similar functionality will also be added to libvirt-lxc
7177 * journalctl will now suppress ANSI color codes when
7178 presenting log data.
7180 * systemctl will no longer show control group information for
7181 a unit if the control group is empty anyway.
7183 * logind can now automatically suspend/hibernate/shutdown the
7186 * /etc/machine-info and hostnamed now also expose the chassis
7187 type of the system. This can be used to determine whether
7188 the local system is a laptop, desktop, handset or
7189 tablet. This information may either be configured by the
7190 user/vendor or is automatically determined from ACPI and DMI
7191 information if possible.
7193 * A number of PolicyKit actions are now bound together with
7194 "imply" rules. This should simplify creating UIs because
7195 many actions will now authenticate similar ones as well.
7197 * Unit files learnt a new condition ConditionACPower= which
7198 may be used to conditionalize a unit depending on whether an
7199 AC power source is connected or not, of whether the system
7200 is running on battery power.
7202 * systemctl gained a new "is-failed" verb that may be used in
7203 shell scripts and suchlike to check whether a specific unit
7204 is in the "failed" state.
7206 * The EnvironmentFile= setting in unit files now supports file
7207 globbing, and can hence be used to easily read a number of
7208 environment files at once.
7210 * systemd will no longer detect and recognize specific
7211 distributions. All distribution-specific #ifdeffery has been
7212 removed, systemd is now fully generic and
7213 distribution-agnostic. Effectively, not too much is lost as
7214 a lot of the code is still accessible via explicit configure
7215 switches. However, support for some distribution specific
7216 legacy configuration file formats has been dropped. We
7217 recommend distributions to simply adopt the configuration
7218 files everybody else uses now and convert the old
7219 configuration from packaging scripts. Most distributions
7220 already did that. If that's not possible or desirable,
7221 distributions are welcome to forward port the specific
7222 pieces of code locally from the git history.
7224 * When logging a message about a unit systemd will now always
7225 log the unit name in the message meta data.
7227 * localectl will now also discover system locale data that is
7228 not stored in locale archives, but directly unpacked.
7230 * logind will no longer unconditionally use framebuffer
7231 devices as seat masters, i.e. as devices that are required
7232 to be existing before a seat is considered preset. Instead,
7233 it will now look for all devices that are tagged as
7234 "seat-master" in udev. By default, framebuffer devices will
7235 be marked as such, but depending on local systems, other
7236 devices might be marked as well. This may be used to
7237 integrate graphics cards using closed source drivers (such
7238 as NVidia ones) more nicely into logind. Note however, that
7239 we recommend using the open source NVidia drivers instead,
7240 and no udev rules for the closed-source drivers will be
7241 shipped from us upstream.
7243 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7244 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7245 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7246 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7247 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7248 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7249 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7250 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7251 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7252 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7253 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7254 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7259 * udev gained support for loading additional device properties
7260 from an indexed database that is keyed by vendor/product IDs
7261 and similar device identifiers. For the beginning this
7262 "hwdb" is populated with data from the well-known PCI and
7263 USB database, but also includes PNP, ACPI and OID data. In
7264 the longer run this indexed database shall grow into
7265 becoming the one central database for non-essential
7266 userspace device metadata. Previously, data from the PCI/USB
7267 database was only attached to select devices, since the
7268 lookup was a relatively expensive operation due to O(n) time
7269 complexity (with n being the number of entries in the
7270 database). Since this is now O(1), we decided to add in this
7271 data for all devices where this is available, by
7272 default. Note that the indexed database needs to be rebuilt
7273 when new data files are installed. To achieve this you need
7274 to update your packaging scripts to invoke "udevadm hwdb
7275 --update" after installation of hwdb data files. For
7276 RPM-based distributions we introduced the new
7277 %udev_hwdb_update macro for this purpose.
7279 * The Journal gained support for the "Message Catalog", an
7280 indexed database to link up additional information with
7281 journal entries. For further details please check:
7283 https://www.freedesktop.org/wiki/Software/systemd/catalog
7285 The indexed message catalog database also needs to be
7286 rebuilt after installation of message catalog files. Use
7287 "journalctl --update-catalog" for this. For RPM-based
7288 distributions we introduced the %journal_catalog_update
7289 macro for this purpose.
7291 * The Python Journal bindings gained support for the standard
7292 Python logging framework.
7294 * The Journal API gained new functions for checking whether
7295 the underlying file system of a journal file is capable of
7296 properly reporting file change notifications, or whether
7297 applications that want to reflect journal changes "live"
7298 need to recheck journal files continuously in appropriate
7301 * It is now possible to set the "age" field for tmpfiles
7302 entries to 0, indicating that files matching this entry
7303 shall always be removed when the directories are cleaned up.
7305 * coredumpctl gained a new "gdb" verb which invokes gdb
7306 right-away on the selected coredump.
7308 * There's now support for "hybrid sleep" on kernels that
7309 support this, in addition to "suspend" and "hibernate". Use
7310 "systemctl hybrid-sleep" to make use of this.
7312 * logind's HandleSuspendKey= setting (and related settings)
7313 now gained support for a new "lock" setting to simply
7314 request the screen lock on all local sessions, instead of
7315 actually executing a suspend or hibernation.
7317 * systemd will now mount the EFI variables file system by
7320 * Socket units now gained support for configuration of the
7321 SMACK security label.
7323 * timedatectl will now output the time of the last and next
7324 daylight saving change.
7326 * We dropped support for various legacy and distro-specific
7327 concepts, such as insserv, early-boot SysV services
7328 (i.e. those for non-standard runlevels such as 'b' or 'S')
7329 or ArchLinux /etc/rc.conf support. We recommend the
7330 distributions who still need support this to either continue
7331 to maintain the necessary patches downstream, or find a
7332 different solution. (Talk to us if you have questions!)
7334 * Various systemd components will now bypass PolicyKit checks
7335 for root and otherwise handle properly if PolicyKit is not
7336 found to be around. This should fix most issues for
7337 PolicyKit-less systems. Quite frankly this should have been
7338 this way since day one. It is absolutely our intention to
7339 make systemd work fine on PolicyKit-less systems, and we
7340 consider it a bug if something does not work as it should if
7341 PolicyKit is not around.
7343 * For embedded systems it is now possible to build udev and
7344 systemd without blkid and/or kmod support.
7346 * "systemctl switch-root" is now capable of switching root
7347 more than once. I.e. in addition to transitions from the
7348 initrd to the host OS it is now possible to transition to
7349 further OS images from the host. This is useful to implement
7350 offline updating tools.
7352 * Various other additions have been made to the RPM macros
7353 shipped with systemd. Use %udev_rules_update() after
7354 installing new udev rules files. %_udevhwdbdir,
7355 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7356 %_sysctldir are now available which resolve to the right
7357 directories for packages to place various data files in.
7359 * journalctl gained the new --full switch (in addition to
7360 --all, to disable ellipsation for long messages.
7362 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7363 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7364 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7365 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7366 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7367 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7368 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7369 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7370 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7374 * journalctl gained new --since= and --until= switches to
7375 filter by time. It also now supports nice filtering for
7376 units via --unit=/-u.
7378 * Type=oneshot services may use ExecReload= and do the
7381 * The journal daemon now supports time-based rotation and
7382 vacuuming, in addition to the usual disk-space based
7385 * The journal will now index the available field values for
7386 each field name. This enables clients to show pretty drop
7387 downs of available match values when filtering. The bash
7388 completion of journalctl has been updated
7389 accordingly. journalctl gained a new switch -F to list all
7390 values a certain field takes in the journal database.
7392 * More service events are now written as structured messages
7393 to the journal, and made recognizable via message IDs.
7395 * The timedated, localed and hostnamed mini-services which
7396 previously only provided support for changing time, locale
7397 and hostname settings from graphical DEs such as GNOME now
7398 also have a minimal (but very useful) text-based client
7399 utility each. This is probably the nicest way to changing
7400 these settings from the command line now, especially since
7401 it lists available options and is fully integrated with bash
7404 * There's now a new tool "systemd-coredumpctl" to list and
7405 extract coredumps from the journal.
7407 * We now install a README each in /var/log/ and
7408 /etc/rc.d/init.d explaining where the system logs and init
7409 scripts went. This hopefully should help folks who go to
7410 that dirs and look into the otherwise now empty void and
7411 scratch their heads.
7413 * When user-services are invoked (by systemd --user) the
7414 $MANAGERPID env var is set to the PID of systemd.
7416 * SIGRTMIN+24 when sent to a --user instance will now result
7417 in immediate termination of systemd.
7419 * gatewayd received numerous feature additions such as a
7420 "follow" mode, for live syncing and filtering.
7422 * browse.html now allows filtering and showing detailed
7423 information on specific entries. Keyboard navigation and
7424 mouse screen support has been added.
7426 * gatewayd/journalctl now supports HTML5/JSON
7427 Server-Sent-Events as output.
7429 * The SysV init script compatibility logic will now
7430 heuristically determine whether a script supports the
7431 "reload" verb, and only then make this available as
7434 * "systemctl status --follow" has been removed, use "journalctl
7437 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7438 have been removed since they are hardly useful to be
7441 * And I'd like to take the opportunity to specifically mention
7442 Zbigniew for his great contributions. Zbigniew, you rock!
7444 Contributions from: Andrew Eikum, Christian Hesse, Colin
7445 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
7446 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7447 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7448 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7449 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7450 Jędrzejewski-Szmek, Сковорода Никита Андреевич
7454 * If /etc/vconsole.conf is non-existent or empty we will no
7455 longer load any console font or key map at boot by
7456 default. Instead the kernel defaults will be left
7457 intact. This is definitely the right thing to do, as no
7458 configuration should mean no configuration, and hard-coding
7459 font names that are different on all archs is probably a bad
7460 idea. Also, the kernel default key map and font should be
7461 good enough for most cases anyway, and mostly identical to
7462 the userspace fonts/key maps we previously overloaded them
7463 with. If distributions want to continue to default to a
7464 non-kernel font or key map they should ship a default
7465 /etc/vconsole.conf with the appropriate contents.
7467 Contributions from: Colin Walters, Daniel J Walsh, Dave
7468 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7469 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7473 * journalctl gained a new --cursor= switch to show entries
7474 starting from the specified location in the journal.
7476 * We now enforce a size limit on journal entry fields exported
7477 with "-o json" in journalctl. Fields larger than 4K will be
7478 assigned null. This can be turned off with --all.
7480 * An (optional) journal gateway daemon is now available as
7481 "systemd-journal-gatewayd.service". This service provides
7482 access to the journal via HTTP and JSON. This functionality
7483 will be used to implement live log synchronization in both
7484 pull and push modes, but has various other users too, such
7485 as easy log access for debugging of embedded devices. Right
7486 now it is already useful to retrieve the journal via HTTP:
7488 # systemctl start systemd-journal-gatewayd.service
7489 # wget http://localhost:19531/entries
7491 This will download the journal contents in a
7492 /var/log/messages compatible format. The same as JSON:
7494 # curl -H"Accept: application/json" http://localhost:19531/entries
7496 This service is also accessible via a web browser where a
7497 single static HTML5 app is served that uses the JSON logic
7498 to enable the user to do some basic browsing of the
7499 journal. This will be extended later on. Here's an example
7500 screenshot of this app in its current state:
7502 http://0pointer.de/public/journal-gatewayd
7504 Contributions from: Kay Sievers, Lennart Poettering, Robert
7505 Milasan, Tom Gundersen
7509 * The bash completion logic is now available for journalctl
7512 * We do not mount the "cpuset" controller anymore together with
7513 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7514 started if no parameters are assigned to it. "cpuset" hence
7515 broke code that assumed it could create "cpu" groups and
7518 * journalctl -f will now subscribe to terminal size changes,
7519 and line break accordingly.
7521 Contributions from: Dave Reisner, Kay Sievers, Lennart
7522 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
7526 * nspawn will now create a symlink /etc/localtime in the
7527 container environment, copying the host's timezone
7528 setting. Previously this has been done via a bind mount, but
7529 since symlinks cannot be bind mounted this has now been
7530 changed to create/update the appropriate symlink.
7532 * journalctl -n's line number argument is now optional, and
7533 will default to 10 if omitted.
7535 * journald will now log the maximum size the journal files may
7536 take up on disk. This is particularly useful if the default
7537 built-in logic of determining this parameter from the file
7538 system size is used. Use "systemctl status
7539 systemd-journald.service" to see this information.
7541 * The multi-seat X wrapper tool has been stripped down. As X
7542 is now capable of enumerating graphics devices via udev in a
7543 seat-aware way the wrapper is not strictly necessary
7544 anymore. A stripped down temporary stop-gap is still shipped
7545 until the upstream display managers have been updated to
7546 fully support the new X logic. Expect this wrapper to be
7547 removed entirely in one of the next releases.
7549 * HandleSleepKey= in logind.conf has been split up into
7550 HandleSuspendKey= and HandleHibernateKey=. The old setting
7551 is not available anymore. X11 and the kernel are
7552 distinguishing between these keys and we should too. This
7553 also means the inhibition lock for these keys has been split
7556 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7557 Poettering, Lukas Nykryn, Václav Pavlín
7561 * Whenever a unit changes state we will now log this to the
7562 journal and show along the unit's own log output in
7565 * ConditionPathIsMountPoint= can now properly detect bind
7566 mount points too. (Previously, a bind mount of one file
7567 system to another place in the same file system could not be
7568 detected as mount, since they shared struct stat's st_dev
7571 * We will now mount the cgroup controllers cpu, cpuacct,
7572 cpuset and the controllers net_cls, net_prio together by
7575 * nspawn containers will now have a virtualized boot
7576 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7577 over with a randomized ID at container initialization). This
7578 has the effect of making "journalctl -b" do the right thing
7581 * The JSON output journal serialization has been updated not
7582 to generate "endless" list objects anymore, but rather one
7583 JSON object per line. This is more in line how most JSON
7584 parsers expect JSON objects. The new output mode
7585 "json-pretty" has been added to provide similar output, but
7586 neatly aligned for readability by humans.
7588 * We dropped all explicit sync() invocations in the shutdown
7589 code. The kernel does this implicitly anyway in the kernel
7590 reboot() syscall. halt(8)'s -n option is now a compatibility
7593 * We now support virtualized reboot() in containers, as
7594 supported by newer kernels. We will fall back to exit() if
7595 CAP_SYS_REBOOT is not available to the container. Also,
7596 nspawn makes use of this now and will actually reboot the
7597 container if the containerized OS asks for that.
7599 * journalctl will only show local log output by default
7600 now. Use --merge (-m) to show remote log output, too.
7602 * libsystemd-journal gained the new sd_journal_get_usage()
7603 call to determine the current disk usage of all journal
7604 files. This is exposed in the new "journalctl --disk-usage"
7607 * journald gained a new configuration setting SplitMode= in
7608 journald.conf which may be used to control how user journals
7609 are split off. See journald.conf(5) for details.
7611 * A new condition type ConditionFileNotEmpty= has been added.
7613 * tmpfiles' "w" lines now support file globbing, to write
7614 multiple files at once.
7616 * We added Python bindings for the journal submission
7617 APIs. More Python APIs for a number of selected APIs will
7618 likely follow. Note that we intend to add native bindings
7619 only for the Python language, as we consider it common
7620 enough to deserve bindings shipped within systemd. There are
7621 various projects outside of systemd that provide bindings
7622 for languages such as PHP or Lua.
7624 * Many conditions will now resolve specifiers such as %i. In
7625 addition, PathChanged= and related directives of .path units
7626 now support specifiers as well.
7628 * There's now a new RPM macro definition for the system preset
7631 * journald will now warn if it ca not forward a message to the
7632 syslog daemon because its socket is full.
7634 * timedated will no longer write or process /etc/timezone,
7635 except on Debian. As we do not support late mounted /usr
7636 anymore /etc/localtime always being a symlink is now safe,
7637 and hence the information in /etc/timezone is not necessary
7640 * logind will now always reserve one VT for a text getty (VT6
7641 by default). Previously if more than 6 X sessions where
7642 started they took up all the VTs with auto-spawned gettys,
7643 so that no text gettys were available anymore.
7645 * udev will now automatically inform the btrfs kernel logic
7646 about btrfs RAID components showing up. This should make
7647 simple hotplug based btrfs RAID assembly work.
7649 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7650 (but not for its children which will stay at the kernel
7651 default). This should allow setups with a lot more listening
7654 * systemd will now always pass the configured timezone to the
7655 kernel at boot. timedated will do the same when the timezone
7658 * logind's inhibition logic has been updated. By default,
7659 logind will now handle the lid switch, the power and sleep
7660 keys all the time, even in graphical sessions. If DEs want
7661 to handle these events on their own they should take the new
7662 handle-power-key, handle-sleep-key and handle-lid-switch
7663 inhibitors during their runtime. A simple way to achieve
7664 that is to invoke the DE wrapped in an invocation of:
7666 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7668 * Access to unit operations is now checked via SELinux taking
7669 the unit file label and client process label into account.
7671 * systemd will now notify the administrator in the journal
7672 when he over-mounts a non-empty directory.
7674 * There are new specifiers that are resolved in unit files,
7675 for the host name (%H), the machine ID (%m) and the boot ID
7678 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
7679 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7680 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7681 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7682 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7683 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7684 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7688 * Support for reading structured kernel messages from
7689 /dev/kmsg has now been added and is enabled by default.
7691 * Support for reading kernel messages from /proc/kmsg has now
7692 been removed. If you want kernel messages in the journal
7693 make sure to run a recent kernel (>= 3.5) that supports
7694 reading structured messages from /dev/kmsg (see
7695 above). /proc/kmsg is now exclusive property of classic
7696 syslog daemons again.
7698 * The libudev API gained the new
7699 udev_device_new_from_device_id() call.
7701 * The logic for file system namespace (ReadOnlyDirectory=,
7702 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7703 require pivot_root() anymore. This means fewer temporary
7704 directories are created below /tmp for this feature.
7706 * nspawn containers will now see and receive all submounts
7707 made on the host OS below the root file system of the
7710 * Forward Secure Sealing is now supported for Journal files,
7711 which provide cryptographical sealing of journal files so
7712 that attackers cannot alter log history anymore without this
7713 being detectable. Lennart will soon post a blog story about
7714 this explaining it in more detail.
7716 * There are two new service settings RestartPreventExitStatus=
7717 and SuccessExitStatus= which allow configuration of exit
7718 status (exit code or signal) which will be excepted from the
7719 restart logic, resp. consider successful.
7721 * journalctl gained the new --verify switch that can be used
7722 to check the integrity of the structure of journal files and
7723 (if Forward Secure Sealing is enabled) the contents of
7726 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7727 and similar symlinks pre-created. This makes running shells
7728 as container init process a lot more fun.
7730 * The fstab support can now handle PARTUUID= and PARTLABEL=
7733 * A new ConditionHost= condition has been added to match
7734 against the hostname (with globs) and machine ID. This is
7735 useful for clusters where a single OS image is used to
7736 provision a large number of hosts which shall run slightly
7737 different sets of services.
7739 * Services which hit the restart limit will now be placed in a
7742 Contributions from: Bertram Poettering, Dave Reisner, Huang
7743 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7744 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7748 * When running in --user mode systemd will now become a
7749 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7750 tree a lot more organized.
7752 * A new PartOf= unit dependency type has been introduced that
7753 may be used to group services in a natural way.
7755 * "systemctl enable" may now be used to enable instances of
7758 * journalctl now prints error log levels in red, and
7759 warning/notice log levels in bright white. It also supports
7760 filtering by log level now.
7762 * cgtop gained a new -n switch (similar to top), to configure
7763 the maximum number of iterations to run for. It also gained
7764 -b, to run in batch mode (accepting no input).
7766 * The suffix ".service" may now be omitted on most systemctl
7767 command lines involving service unit names.
7769 * There's a new bus call in logind to lock all sessions, as
7770 well as a loginctl verb for it "lock-sessions".
7772 * libsystemd-logind.so gained a new call sd_journal_perror()
7773 that works similar to libc perror() but logs to the journal
7774 and encodes structured information about the error number.
7776 * /etc/crypttab entries now understand the new keyfile-size=
7779 * shutdown(8) now can send a (configurable) wall message when
7780 a shutdown is cancelled.
7782 * The mount propagation mode for the root file system will now
7783 default to "shared", which is useful to make containers work
7784 nicely out-of-the-box so that they receive new mounts from
7785 the host. This can be undone locally by running "mount
7786 --make-rprivate /" if needed.
7788 * The prefdm.service file has been removed. Distributions
7789 should maintain this unit downstream if they intend to keep
7790 it around. However, we recommend writing normal unit files
7791 for display managers instead.
7793 * Since systemd is a crucial part of the OS we will now
7794 default to a number of compiler switches that improve
7795 security (hardening) such as read-only relocations, stack
7796 protection, and suchlike.
7798 * The TimeoutSec= setting for services is now split into
7799 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7800 of individual time outs for the start and the stop phase of
7803 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7804 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7805 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7806 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7807 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7808 Gundersen, Zbigniew Jędrzejewski-Szmek
7812 * The journal and id128 C APIs are now fully documented as man
7815 * Extra safety checks have been added when transitioning from
7816 the initial RAM disk to the main system to avoid accidental
7819 * /etc/crypttab entries now understand the new keyfile-offset=
7822 * systemctl -t can now be used to filter by unit load state.
7824 * The journal C API gained the new sd_journal_wait() call to
7825 make writing synchronous journal clients easier.
7827 * journalctl gained the new -D switch to show journals from a
7830 * journalctl now displays a special marker between log
7831 messages of two different boots.
7833 * The journal is now explicitly flushed to /var via a service
7834 systemd-journal-flush.service, rather than implicitly simply
7835 by seeing /var/log/journal to be writable.
7837 * journalctl (and the journal C APIs) can now match for much
7838 more complex expressions, with alternatives and
7841 * When transitioning from the initial RAM disk to the main
7842 system we will now kill all processes in a killing spree to
7843 ensure no processes stay around by accident.
7845 * Three new specifiers may be used in unit files: %u, %h, %s
7846 resolve to the user name, user home directory resp. user
7847 shell. This is useful for running systemd user instances.
7849 * We now automatically rotate journal files if their data
7850 object hash table gets a fill level > 75%. We also size the
7851 hash table based on the configured maximum file size. This
7852 together should lower hash collisions drastically and thus
7853 speed things up a bit.
7855 * journalctl gained the new "--header" switch to introspect
7856 header data of journal files.
7858 * A new setting SystemCallFilters= has been added to services
7859 which may be used to apply blacklists or whitelists to
7860 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7862 * nspawn gained a new --link-journal= switch (and quicker: -j)
7863 to link the container journal with the host. This makes it
7864 very easy to centralize log viewing on the host for all
7865 guests while still keeping the journal files separated.
7867 * Many bugfixes and optimizations
7869 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7870 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7871 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7876 * Several tools now understand kernel command line arguments,
7877 which are only read when run in an initial RAM disk. They
7878 usually follow closely their normal counterparts, but are
7881 * There's a new tool to analyze the readahead files that are
7882 automatically generated at boot. Use:
7884 /usr/lib/systemd/systemd-readahead analyze /.readahead
7886 * We now provide an early debug shell on tty9 if this enabled. Use:
7888 systemctl enable debug-shell.service
7890 * All plymouth related units have been moved into the Plymouth
7891 package. Please make sure to upgrade your Plymouth version
7894 * systemd-tmpfiles now supports getting passed the basename of
7895 a configuration file only, in which case it will look for it
7896 in all appropriate directories automatically.
7898 * udevadm info now takes a /dev or /sys path as argument, and
7899 does the right thing. Example:
7901 udevadm info /dev/sda
7902 udevadm info /sys/class/block/sda
7904 * systemctl now prints a warning if a unit is stopped but a
7905 unit that might trigger it continues to run. Example: a
7906 service is stopped but the socket that activates it is left
7909 * "systemctl status" will now mention if the log output was
7910 shortened due to rotation since a service has been started.
7912 * The journal API now exposes functions to determine the
7913 "cutoff" times due to rotation.
7915 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7916 immediately flushing of runtime logs to /var if possible,
7917 resp. for triggering immediate rotation of the journal
7920 * It is now considered an error if a service is attempted to
7921 be stopped that is not loaded.
7923 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7925 * systemd-analyze now supports Python 3
7927 * tmpfiles now supports cleaning up directories via aging
7928 where the first level dirs are always kept around but
7929 directories beneath it automatically aged. This is enabled
7930 by prefixing the age field with '~'.
7932 * Seat objects now expose CanGraphical, CanTTY properties
7933 which is required to deal with very fast bootups where the
7934 display manager might be running before the graphics drivers
7935 completed initialization.
7937 * Seat objects now expose a State property.
7939 * We now include RPM macros for service enabling/disabling
7940 based on the preset logic. We recommend RPM based
7941 distributions to make use of these macros if possible. This
7942 makes it simpler to reuse RPM spec files across
7945 * We now make sure that the collected systemd unit name is
7946 always valid when services log to the journal via
7949 * There's a new man page kernel-command-line(7) detailing all
7950 command line options we understand.
7952 * The fstab generator may now be disabled at boot by passing
7953 fstab=0 on the kernel command line.
7955 * A new kernel command line option modules-load= is now understood
7956 to load a specific kernel module statically, early at boot.
7958 * Unit names specified on the systemctl command line are now
7959 automatically escaped as needed. Also, if file system or
7960 device paths are specified they are automatically turned
7961 into the appropriate mount or device unit names. Example:
7963 systemctl status /home
7964 systemctl status /dev/sda
7966 * The SysVConsole= configuration option has been removed from
7967 system.conf parsing.
7969 * The SysV search path is no longer exported on the D-Bus
7972 * The Names= option has been removed from unit file parsing.
7974 * There's a new man page bootup(7) detailing the boot process.
7976 * Every unit and every generator we ship with systemd now
7977 comes with full documentation. The self-explanatory boot is
7980 * A couple of services gained "systemd-" prefixes in their
7981 name if they wrap systemd code, rather than only external
7982 code. Among them fsck@.service which is now
7983 systemd-fsck@.service.
7985 * The HaveWatchdog property has been removed from the D-Bus
7988 * systemd.confirm_spawn= on the kernel command line should now
7991 * There's a new man page crypttab(5) which details all options
7992 we actually understand.
7994 * systemd-nspawn gained a new --capability= switch to pass
7995 additional capabilities to the container.
7997 * timedated will now read known NTP implementation unit names
7998 from /usr/lib/systemd/ntp-units.d/*.list,
7999 systemd-timedated-ntp.target has been removed.
8001 * journalctl gained a new switch "-b" that lists log data of
8002 the current boot only.
8004 * The notify socket is in the abstract namespace again, in
8005 order to support daemons which chroot() at start-up.
8007 * There is a new Storage= configuration option for journald
8008 which allows configuration of where log data should go. This
8009 also provides a way to disable journal logging entirely, so
8010 that data collected is only forwarded to the console, the
8011 kernel log buffer or another syslog implementation.
8013 * Many bugfixes and optimizations
8015 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8016 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8017 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8018 Shawn Landden, Tom Gundersen
8022 * "systemctl help <unit>" now shows the man page if one is
8025 * Several new man pages have been added.
8027 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8028 MaxLevelConsole= can now be specified in
8029 journald.conf. These options allow reducing the amount of
8030 data stored on disk or forwarded by the log level.
8032 * TimerSlackNSec= can now be specified in system.conf for
8033 PID1. This allows system-wide power savings.
8035 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8036 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8041 * logind is now capable of (optionally) handling power and
8042 sleep keys as well as the lid switch.
8044 * journalctl now understands the syntax "journalctl
8045 /usr/bin/avahi-daemon" to get all log output of a specific
8048 * CapabilityBoundingSet= in system.conf now also influences
8049 the capability bound set of usermode helpers of the kernel.
8051 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8052 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8053 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8054 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8058 * Note that we skipped 139 releases here in order to set the
8059 new version to something that is greater than both udev's
8060 and systemd's most recent version number.
8062 * udev: all udev sources are merged into the systemd source tree now.
8063 All future udev development will happen in the systemd tree. It
8064 is still fully supported to use the udev daemon and tools without
8065 systemd running, like in initramfs or other init systems. Building
8066 udev though, will require the *build* of the systemd tree, but
8067 udev can be properly *run* without systemd.
8069 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8070 should be used to create dead device nodes as workarounds for broken
8073 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8074 no longer supported. udev_monitor_new_from_netlink() needs to be
8075 used to subscribe to events.
8077 * udev: when udevd is started by systemd, processes which are left
8078 behind by forking them off of udev rules, are unconditionally cleaned
8079 up and killed now after the event handling has finished. Services or
8080 daemons must be started as systemd services. Services can be
8081 pulled-in by udev to get started, but they can no longer be directly
8082 forked by udev rules.
8084 * udev: the daemon binary is called systemd-udevd now and installed
8085 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8086 to adapt to that, create symlink, or rename the binary after building
8089 * libudev no longer provides these symbols:
8090 udev_monitor_from_socket()
8091 udev_queue_get_failed_list_entry()
8092 udev_get_{dev,sys,run}_path()
8093 The versions number was bumped and symbol versioning introduced.
8095 * systemd-loginctl and systemd-journalctl have been renamed
8096 to loginctl and journalctl to match systemctl.
8098 * The config files: /etc/systemd/systemd-logind.conf and
8099 /etc/systemd/systemd-journald.conf have been renamed to
8100 logind.conf and journald.conf. Package updates should rename
8101 the files to the new names on upgrade.
8103 * For almost all files the license is now LGPL2.1+, changed
8104 from the previous GPL2.0+. Exceptions are some minor stuff
8105 of udev (which will be changed to LGPL2.1 eventually, too),
8106 and the MIT licensed sd-daemon.[ch] library that is suitable
8107 to be used as drop-in files.
8109 * systemd and logind now handle system sleep states, in
8110 particular suspending and hibernating.
8112 * logind now implements a sleep/shutdown/idle inhibiting logic
8113 suitable for a variety of uses. Soonishly Lennart will blog
8114 about this in more detail.
8116 * var-run.mount and var-lock.mount are no longer provided
8117 (which previously bind mounted these directories to their new
8118 places). Distributions which have not converted these
8119 directories to symlinks should consider stealing these files
8120 from git history and add them downstream.
8122 * We introduced the Documentation= field for units and added
8123 this to all our shipped units. This is useful to make it
8124 easier to explore the boot and the purpose of the various
8127 * All smaller setup units (such as
8128 systemd-vconsole-setup.service) now detect properly if they
8129 are run in a container and are skipped when
8130 appropriate. This guarantees an entirely noise-free boot in
8131 Linux container environments such as systemd-nspawn.
8133 * A framework for implementing offline system updates is now
8134 integrated, for details see:
8135 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
8137 * A new service type Type=idle is available now which helps us
8138 avoiding ugly interleaving of getty output and boot status
8141 * There's now a system-wide CapabilityBoundingSet= option to
8142 globally reduce the set of capabilities for the
8143 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8144 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8145 even CAP_NET_ADMIN system-wide for secure systems.
8147 * There are now system-wide DefaultLimitXXX= options to
8148 globally change the defaults of the various resource limits
8149 for all units started by PID 1.
8151 * Harald Hoyer's systemd test suite has been integrated into
8152 systemd which allows easy testing of systemd builds in qemu
8153 and nspawn. (This is really awesome! Ask us for details!)
8155 * The fstab parser is now implemented as generator, not inside
8158 * systemctl will now warn you if .mount units generated from
8159 /etc/fstab are out of date due to changes in fstab that
8160 have not been read by systemd yet.
8162 * systemd is now suitable for usage in initrds. Dracut has
8163 already been updated to make use of this. With this in place
8164 initrds get a slight bit faster but primarily are much
8165 easier to introspect and debug since "systemctl status" in
8166 the host system can be used to introspect initrd services,
8167 and the journal from the initrd is kept around too.
8169 * systemd-delta has been added, a tool to explore differences
8170 between user/admin configuration and vendor defaults.
8172 * PrivateTmp= now affects both /tmp and /var/tmp.
8174 * Boot time status messages are now much prettier and feature
8175 proper english language. Booting up systemd has never been
8178 * Read-ahead pack files now include the inode number of all
8179 files to pre-cache. When the inode changes the pre-caching
8180 is not attempted. This should be nicer to deal with updated
8181 packages which might result in changes of read-ahead
8184 * We now temporaritly lower the kernel's read_ahead_kb variable
8185 when collecting read-ahead data to ensure the kernel's
8186 built-in read-ahead does not add noise to our measurements
8187 of necessary blocks to pre-cache.
8189 * There's now RequiresMountsFor= to add automatic dependencies
8190 for all mounts necessary for a specific file system path.
8192 * MountAuto= and SwapAuto= have been removed from
8193 system.conf. Mounting file systems at boot has to take place
8196 * nspawn now learned a new switch --uuid= to set the machine
8197 ID on the command line.
8199 * nspawn now learned the -b switch to automatically search
8202 * vt102 is now the default TERM for serial TTYs, upgraded from
8205 * systemd-logind now works on VT-less systems.
8207 * The build tree has been reorganized. The individual
8208 components now have directories of their own.
8210 * A new condition type ConditionPathIsReadWrite= is now available.
8212 * nspawn learned the new -C switch to create cgroups for the
8213 container in other hierarchies.
8215 * We now have support for hardware watchdogs, configurable in
8218 * The scheduled shutdown logic now has a public API.
8220 * We now mount /tmp as tmpfs by default, but this can be
8221 masked and /etc/fstab can override it.
8223 * Since udisks does not make use of /media anymore we are not
8224 mounting a tmpfs on it anymore.
8226 * journalctl gained a new --local switch to only interleave
8227 locally generated journal files.
8229 * We can now load the IMA policy at boot automatically.
8231 * The GTK tools have been split off into a systemd-ui.
8233 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8234 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8235 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8236 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8237 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8238 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8239 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8240 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8241 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8246 * This is mostly a bugfix release
8248 * Support optional initialization of the machine ID from the
8249 KVM or container configured UUID.
8251 * Support immediate reboots with "systemctl reboot -ff"
8253 * Show /etc/os-release data in systemd-analyze output
8255 * Many bugfixes for the journal, including endianness fixes and
8256 ensuring that disk space enforcement works
8258 * sd-login.h is C++ compatible again
8260 * Extend the /etc/os-release format on request of the Debian
8263 * We now refuse non-UTF8 strings used in various configuration
8264 and unit files. This is done to ensure we do not pass invalid
8265 data over D-Bus or expose it elsewhere.
8267 * Register Mimo USB Screens as suitable for automatic seat
8270 * Read SELinux client context from journal clients in a race
8273 * Reorder configuration file lookup order. /etc now always
8274 overrides /run in order to allow the administrator to always
8275 and unconditionally override vendor-supplied or
8276 automatically generated data.
8278 * The various user visible bits of the journal now have man
8279 pages. We still lack man pages for the journal API calls
8282 * We now ship all man pages in HTML format again in the
8285 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8286 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8287 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8292 * This is mostly a bugfix release
8294 * systems lacking /etc/os-release are no longer supported.
8296 * Various functionality updates to libsystemd-login.so
8298 * Track class of PAM logins to distinguish greeters from
8301 Contributions from: Kay Sievers, Lennart Poettering, Michael
8306 * This is an important bugfix release for v41.
8308 * Building man pages is now optional which should be useful
8309 for those building systemd from git but unwilling to install
8312 * Watchdog support for supervising services is now usable. In
8313 a future release support for hardware watchdogs
8314 (i.e. /dev/watchdog) will be added building on this.
8316 * Service start rate limiting is now configurable and can be
8317 turned off per service. When a start rate limit is hit a
8318 reboot can automatically be triggered.
8320 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8322 Contributions from: Benjamin Franzke, Bill Nottingham,
8323 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8324 Schmidt, Michał Górny, Piotr Drąg
8328 * The systemd binary is installed /usr/lib/systemd/systemd now;
8329 An existing /sbin/init symlink needs to be adapted with the
8332 * The code that loads kernel modules has been ported to invoke
8333 libkmod directly, instead of modprobe. This means we do not
8334 support systems with module-init-tools anymore.
8336 * Watchdog support is now already useful, but still not
8339 * A new kernel command line option systemd.setenv= is
8340 understood to set system wide environment variables
8341 dynamically at boot.
8343 * We now limit the set of capabilities of systemd-journald.
8345 * We now set SIGPIPE to ignore by default, since it only is
8346 useful in shell pipelines, and has little use in general
8347 code. This can be disabled with IgnoreSIPIPE=no in unit
8350 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8351 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8356 * This is mostly a bugfix release
8358 * We now expose the reason why a service failed in the
8359 "Result" D-Bus property.
8361 * Rudimentary service watchdog support (will be completed over
8362 the next few releases.)
8364 * When systemd forks off in order execute some service we will
8365 now immediately changes its argv[0] to reflect which process
8366 it will execute. This is useful to minimize the time window
8367 with a generic argv[0], which makes bootcharts more useful
8369 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8370 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8371 Mike Kazantsev, Ray Strode
8375 * This is mostly a test release, but incorporates many
8378 * New systemd-cgtop tool to show control groups by their
8381 * Linking against libacl for ACLs is optional again. If
8382 disabled, support tracking device access for active logins
8383 goes becomes unavailable, and so does access to the user
8384 journals by the respective users.
8386 * If a group "adm" exists, journal files are automatically
8387 owned by them, thus allow members of this group full access
8388 to the system journal as well as all user journals.
8390 * The journal now stores the SELinux context of the logging
8391 client for all entries.
8393 * Add C++ inclusion guards to all public headers
8395 * New output mode "cat" in the journal to print only text
8396 messages, without any meta data like date or time.
8398 * Include tiny X server wrapper as a temporary stop-gap to
8399 teach XOrg udev display enumeration. This is used by display
8400 managers such as gdm, and will go away as soon as XOrg
8401 learned native udev hotplugging for display devices.
8403 * Add new systemd-cat tool for executing arbitrary programs
8404 with STDERR/STDOUT connected to the journal. Can also act as
8405 BSD logger replacement, and does so by default.
8407 * Optionally store all locally generated coredumps in the
8408 journal along with meta data.
8410 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8411 writing short strings to files (for usage for /sys), and for
8412 creating symlinks, character and block device nodes.
8414 * New unit file option ControlGroupPersistent= to make cgroups
8415 persistent, following the mechanisms outlined in
8416 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
8418 * Support multiple local RTCs in a sane way
8420 * No longer monopolize IO when replaying readahead data on
8421 rotating disks, since we might starve non-file-system IO to
8422 death, since fanotify() will not see accesses done by blkid,
8425 * Do not show kernel threads in systemd-cgls anymore, unless
8426 requested with new -k switch.
8428 Contributions from: Dan Horák, Kay Sievers, Lennart
8429 Poettering, Michal Schmidt
8433 * This is mostly a test release, but incorporates many
8436 * The git repository moved to:
8437 git://anongit.freedesktop.org/systemd/systemd
8438 ssh://git.freedesktop.org/git/systemd/systemd
8440 * First release with the journal
8441 http://0pointer.de/blog/projects/the-journal.html
8443 * The journal replaces both systemd-kmsg-syslogd and
8444 systemd-stdout-bridge.
8446 * New sd_pid_get_unit() API call in libsystemd-logind
8448 * Many systemadm clean-ups
8450 * Introduce remote-fs-pre.target which is ordered before all
8451 remote mounts and may be used to start services before all
8454 * Added Mageia support
8456 * Add bash completion for systemd-loginctl
8458 * Actively monitor PID file creation for daemons which exit in
8459 the parent process before having finished writing the PID
8460 file in the daemon process. Daemons which do this need to be
8461 fixed (i.e. PID file creation must have finished before the
8462 parent exits), but we now react a bit more gracefully to them.
8464 * Add colourful boot output, mimicking the well-known output
8465 of existing distributions.
8467 * New option PassCredentials= for socket units, for
8468 compatibility with a recent kernel ABI breakage.
8470 * /etc/rc.local is now hooked in via a generator binary, and
8471 thus will no longer act as synchronization point during
8474 * systemctl list-unit-files now supports --root=.
8476 * systemd-tmpfiles now understands two new commands: z, Z for
8477 relabelling files according to the SELinux database. This is
8478 useful to apply SELinux labels to specific files in /sys,
8481 * Output of SysV services is now forwarded to both the console
8482 and the journal by default, not only just the console.
8484 * New man pages for all APIs from libsystemd-login.
8486 * The build tree got reorganized and the build system is a
8487 lot more modular allowing embedded setups to specifically
8488 select the components of systemd they are interested in.
8490 * Support for Linux systems lacking the kernel VT subsystem is
8493 * configure's --with-rootdir= got renamed to
8494 --with-rootprefix= to follow the naming used by udev and
8497 * Unless specified otherwise we will now install to /usr instead
8498 of /usr/local by default.
8500 * Processes with '@' in argv[0][0] are now excluded from the
8501 final shut-down killing spree, following the logic explained
8503 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
8505 * All processes remaining in a service cgroup when we enter
8506 the START or START_PRE states are now killed with
8507 SIGKILL. That means it is no longer possible to spawn
8508 background processes from ExecStart= lines (which was never
8509 supported anyway, and bad style).
8511 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8512 reloading of units together.
8514 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
8515 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8516 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8517 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8518 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob