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 will name network interfaces differently than in previous
7 versions for virtual network interfaces created with SR-IOV and NPAR
8 and for devices where the PCI network controller device does not have
9 a slot number associated.
11 SR-IOV virtual devices are now named based on the name of the parent
12 interface, with a suffix of "v<N>", where <N> is the virtual device
13 number. Previously those virtual devices were named as if completely
16 The ninth and later NPAR virtual devices will be named following the
17 scheme used for the first eight NPAR partitions. Previously those
18 devices were not renamed and the kernel default (eth<n>) was used.
20 "net_id" will also generate names for PCI devices where the PCI
21 network controller device does not have an associated slot number
22 itself, but one of its parents does. Previously those devices were
23 not renamed and the kernel default (eth<n>) was used.
25 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
26 elogind.service. Since v235, IPAddressDeny=any has been set to
27 the unit. So, it is expected that the default behavior of
28 elogind is not changed. However, if distribution packagers or
29 administrators disabled or modified IPAddressDeny= setting by a
30 drop-in config file, then it may be necessary to update the file to
31 re-enable AF_INET and AF_INET6 to support network user name services,
34 * When the RestrictNamespaces= unit property is specified multiple
35 times, then the specified types are merged now. Previously, only the
36 last assignment was used. So, if distribution packagers or
37 administrators modified the setting by a drop-in config file, then it
38 may be necessary to update the file.
40 * When OnFailure= is used in combination with Restart= on a service
41 unit, then the specified units will no longer be triggered on
42 failures that result in restarting. Previously, the specified units
43 would be activated each time the unit failed, even when the unit was
44 going to be restarted automatically. This behaviour contradicted the
45 documentation. With this release the code is adjusted to match the
48 * elogind-tmpfiles will now print a notice whenever it encounters
49 tmpfiles.d/ lines referencing the /var/run/ directory. It will
50 recommend reworking them to use the /run/ directory instead (for
51 which /var/run/ is simply a symlinked compatibility alias). This way
52 elogind-tmpfiles can properly detect line conflicts and merge lines
53 referencing the same file by two paths, without having to access
56 * systemctl disable/unmask/preset/preset-all cannot be used with
57 --runtime. Previously this was allowed, but resulted in unintuitive
58 behaviour that wasn't useful. systemctl disable/unmask will now undo
59 both runtime and persistent enablement/masking, i.e. it will remove
60 any relevant symlinks both in /run and /etc.
62 * Note that all long-running system services shipped with elogind will
63 now default to a system call whitelist (rather than a blacklist, as
64 before). In particular, elogind-udevd will now enforce one too. For
65 most cases this should be safe, however downstream distributions
66 which disabled sandboxing of elogind-udevd (specifically the
67 MountFlags= setting), might want to disable this security feature
68 too, as the default whitelisting will prohibit all mount, swap,
69 reboot and clock changing operations from udev rules.
71 * sd-boot acquired new loader configuration settings to optionally turn
72 off Windows and MacOS boot partition discovery as well as
73 reboot-into-firmware menu items. It is also able to pick a better
74 screen resolution for HiDPI systems, and now provides loader
75 configuration settings to change the resolution explicitly.
77 * The elogind-resolve tool has been renamed to resolvectl (it also
78 * elogind-resolved now supports DNS-over-TLS. It's still
79 turned off by default, use DNSOverTLS=opportunistic to turn it on in
80 resolved.conf. We intend to make this the default as soon as couple
81 of additional techniques for optimizing the initial latency caused by
82 establishing a TLS/TCP connection are implemented.
84 * elogind-resolved.service and elogind-networkd.service now set
85 DynamicUser=yes. The users elogind-resolve and elogind-network are
86 not created by elogind-sysusers.
88 remains available under the old name, for compatibility), and its
89 interface is now verb-based, similar in style to the other <xyz>ctl
90 tools, such as systemctl or loginctl.
92 * The resolvectl/elogind-resolve tool also provides 'resolvconf'
93 compatibility. It may be symlinked under the 'resolvconf' name, in
94 which case it will take arguments and input compatible with the
95 Debian and FreeBSD resolvconf tool.
97 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
98 where the system initially suspends, and after a time-out resumes and
101 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
102 set the client will only send a DUID as client identifier.
104 * The nss-elogind glibc NSS module will now enumerate dynamic users and
105 groups in effect. Previously, it could resolve UIDs/GIDs to user
106 names/groups and vice versa, but did not support enumeration.
108 * journald's Compress= configuration setting now optionally accepts a
109 byte threshold value. All journal objects larger than this threshold
110 will be compressed, smaller ones will not. Previously this threshold
111 was not configurable and set to 512.
113 * A new system.conf setting NoNewPrivileges= is now available which may
114 be used to turn off acquisition of new privileges system-wide
115 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
116 for all its children). Note that turning this option on means setuid
117 binaries and file system capabilities lose their special powers.
118 While turning on this option is a big step towards a more secure
119 system, doing so is likely to break numerous pre-existing UNIX tools,
120 in particular su and sudo.
122 * A new service elogind-time-sync-wait.service has been added. If
123 enabled it will delay the time-sync.target unit at boot until time
124 synchronization has been received from the network. This
125 functionality is useful on systems lacking a local RTC or where it is
126 acceptable that the boot process shall be delayed by external network
129 * When hibernating, elogind will now inform the kernel of the image
130 write offset, on kernels new enough to support this. This means swap
131 files should work for hibernation now.
133 * When loading unit files, elogind will now look for drop-in unit files
134 extensions in additional places. Previously, for a unit file name
135 "foo-bar-baz.service" it would look for dropin files in
136 "foo-bar-baz.service.d/*.conf". Now, it will also look in
137 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
138 service name truncated after all inner dashes. This scheme allows
139 writing drop-ins easily that apply to a whole set of unit files at
140 once. It's particularly useful for mount and slice units (as their
141 naming is prefix based), but is also useful for service and other
142 units, for packages that install multiple unit files at once,
143 following a strict naming regime of beginning the unit file name with
144 the package's name. Two new specifiers are now supported in unit
145 files to match this: %j and %J are replaced by the part of the unit
146 name following the last dash.
148 * Unit files and other configuration files that support specifier
149 expansion now understand another two new specifiers: %T and %V will
150 resolve to /tmp and /var/tmp respectively, or whatever temporary
151 directory has been set for the calling user.
153 * The ExecStart= lines of unit files are no longer required to
154 reference absolute paths. If non-absolute paths are specified the
155 specified binary name is searched within the service manager's
156 built-in $PATH, which may be queried with 'elogind-path
157 search-binaries-default'. It's generally recommended to continue to
158 use absolute paths for all binaries specified in unit files.
160 * Units gained a new load state "bad-setting", which is used when a
161 unit file was loaded, but contained fatal errors which prevent it
162 from being started (for example, a service unit has been defined
163 lacking both ExecStart= and ExecStop= lines).
165 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
166 support alternative debuggers, for example lldb. The old name
167 continues to be available however, for compatibility reasons. Use the
168 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
169 to pick an alternative debugger instead of the default gdb.
171 * systemctl and the other tools will now output escape sequences that
172 generate proper clickable hyperlinks in various terminal emulators
173 where useful (for example, in the "systemctl status" output you can
174 now click on the unit file name to quickly open it in the
175 editor/viewer of your choice). Note that not all terminal emulators
176 support this functionality yet, but many do. Unfortunately, the
177 "less" pager doesn't support this yet, hence this functionality is
178 currently automatically turned off when a pager is started (which
179 happens quite often due to auto-paging). We hope to remove this
180 limitation as soon as "less" learns these escape sequences. This new
181 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
182 environment variable. For details on these escape sequences see:
183 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
185 * networkd's .network files now support a new IPv6MTUBytes= option for
186 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
187 option in the [Route] section to configure the MTU to use for
188 specific routes. It also gained support for configuration of the DHCP
189 "UserClass" option through the new UserClass= setting. It gained
190 three new options in the new [CAN] section for configuring CAN
191 networks. The MULTICAST and ALLMULTI interface flags may now be
192 controlled explicitly with the new Multicast= and AllMulticast=
195 * networkd will now automatically make use of the kernel's route
196 expiration feature, if it is available.
198 * udevd's .link files now support setting the number of receive and
199 transmit channels, using the RxChannels=, TxChannels=,
200 OtherChannels=, CombinedChannels= settings.
202 * Support for UDPSegmentationOffload= has been removed, given its
203 limited support in hardware, and waning software support.
205 * networkd's .netdev files now support creating "netdevsim" interfaces.
207 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
208 to query the unit belonging to a specific kernel control group.
210 * elogind-analyze gained a new verb "cat-config", which may be used to
211 dump the contents of any configuration file, with all its matching
212 drop-in files added in, and honouring the usual search and masking
213 logic applied to elogind configuration files. For example use
214 "elogind-analyze cat-config elogind/system.conf" to get the complete
215 system configuration file of elogind how it would be loaded by PID 1
216 itself. Similar to this, various tools such as elogind-tmpfiles or
217 elogind-sysusers, gained a new option "--cat-config", which does the
218 corresponding operation for their own configuration settings. For
219 example, "elogind-tmpfiles --cat-config" will now output the full
220 list of tmpfiles.d/ lines in place.
222 * timedatectl gained two new verbs "timesync-status" (to show the
223 current NTP synchronization state of elogind-timesyncd) and
224 "show-timesync" (to show bus properties of elogind-timesyncd).
226 * elogind-timesyncd gained a bus interface on which it exposes details
229 * elogind-nspawn gained a new --rlimit= switch for setting initial
230 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
231 understood by elogind-timedated. It takes a colon-separated list of
232 unit names of NTP client services. The list is used by
233 "timedatectl set-ntp".
235 resource limits for the container payload. There's a new switch
236 --hostname= to explicitly override the container's hostname. A new
237 --no-new-privileges= switch may be used to control the
238 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
239 --oom-score-adjust= switch controls the OOM scoring adjustment value
240 for the payload. The new --cpu-affinity= switch controls the CPU
241 affinity of the container payload. The new --resolv-conf= switch
242 allows more detailed control of /etc/resolv.conf handling of the
243 container. Similarly, the new --timezone= switch allows more detailed
244 control of /etc/localtime handling of the container.
246 * elogind-detect-virt gained a new --list switch, which will print a
247 list of all currently known VM and container environments.
249 * Support for "Portable Services" has been added, see
250 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
251 experimental, but this is expected to change soon. Reflecting this
252 experimental state, the "portablectl" binary is not installed into
253 /usr/bin yet. The binary has to be called with the full path
254 /usr/lib/elogind/portablectl instead.
256 * journalctl's and systemctl's -o switch now knows a new log output
257 mode "with-unit". The output it generates is very similar to the
258 regular "short" mode, but displays the unit name instead of the
259 syslog tag for each log line. Also, the date is shown with timezone
260 information. This mode is probably more useful than the classic
261 "short" output mode for most purposes, except where pixel-perfect
262 compatibility with classic /var/log/messages formatting is required.
264 * A new --dump-bus-properties switch has been added to the elogind
265 binary, which may be used to dump all supported D-Bus properties.
266 (Options which are still supported, but are deprecated, are *not*
269 * sd-bus gained a set of new calls:
270 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
271 enable/disable the "floating" state of a bus slot object,
272 i.e. whether the slot object pins the bus it is allocated for into
273 memory or if the bus slot object gets disconnected when the bus goes
274 away. sd_bus_open_with_description(),
275 sd_bus_open_user_with_description(),
276 sd_bus_open_system_with_description() may be used to allocate bus
277 objects and set their description string already during allocation.
279 * sd-event gained support for watching inotify events from the event
280 loop, in an efficient way, sharing inotify handles between multiple
281 users. For this a new function sd_event_add_inotify() has been added.
283 * sd-event and sd-bus gained support for calling special user-supplied
284 destructor functions for userdata pointers associated with
285 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
286 functions sd_bus_slot_set_destroy_callback,
287 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
288 sd_bus_track_get_destroy_callback,
289 sd_event_source_set_destroy_callback,
290 sd_event_source_get_destroy_callback have been added.
292 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
294 * PID 1 will now automatically reschedule .timer units whenever the
295 local timezone changes. (They previously got rescheduled
296 automatically when the system clock changed.)
298 * New documentation has been added to document cgroups delegation,
299 portable services and the various code quality tools we have set up:
301 https://github.com/systemd/systemd/blob/master/doc/CGROUP_DELEGATION.md
302 https://github.com/systemd/systemd/blob/master/doc/PORTABLE_SERVICES.md
303 https://github.com/systemd/systemd/blob/master/doc/CODE_QUALITY.md
305 * pam_elogind will now look for PAM userdata fields elogind.memory_max,
306 elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
307 earlier PAM modules. The data in these fields is used to initialize
308 the session scope's resource properties. Thus external PAM modules
309 may now configure per-session limits, for example sourced from
310 external user databases.
312 * socket units with Accept=yes will now maintain a "refused" counter in
313 addition to the existing "accepted" counter, counting connections
314 refused due to the enforced limits.
316 * The "elogind-path search-binaries-default" command may now be use to
317 query the default, built-in $PATH PID 1 will pass to the services it
320 * A new unit file setting PrivateMounts= has been added. It's a boolean
321 option. If enabled the unit's processes are invoked in their own file
322 system namespace. Note that this behaviour is also implied if any
323 other file system namespacing options (such as PrivateTmp=,
324 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
325 primarily useful for services that do not use any of the other file
326 system namespacing options. One such service is elogind-udevd.service
327 wher this is now used by default.
329 * A new unit "system-update-pre.target" is added, which defines an
330 optional synchronization point for offline system updates, as
331 implemented by the pre-existing "system-update.target" unit. It
332 allows ordering services before the service that executes the actual
333 update process in a generic way.
335 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
336 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
337 J. Murrell, Bruno Vernay, Chris Lesiak, Christian Brauner, Christian
338 Hesse, Daniel Dao, Daniel Lin, Danylo Korostil, Davide Cavalca, David
339 Tardon, Dimitri John Ledkov, Dmitriy Geels, Douglas Christman, Elia
340 Geretto, emelenas, Emil Velikov, Evgeny Vereshchagin, Felipe Sateler,
341 Feng Sun, Filipe Brandenburger, Franck Bui, futpib, Giuseppe Scrivano,
342 Guillem Jover, guixxx, Hans de Goede, Henrique Dante de Almeida, Hiram
343 van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov, Iwan Timmer,
344 James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir, João Paulo
345 Rechi Vita, Joost Heitbrink, Jui-Chi Ricky Liang, Jürg Billeter,
346 Kai-Heng Feng, Karol Augustin, Krzysztof Nowicki, Lauri Tirkkonen,
347 Lennart Poettering, Leonard König, Long Li, Luca Boccassi, Lucas
348 Werkmeister, Marcel Hoppe, Marc Kleine-Budde, Mario Limonciello, Martin
349 Jansa, Martin Wilck, Mathieu Malaterre, Matteo F. Vescovi, Matthew
350 McGinn, Matthias-Christian Ott, Michael Biebl, Michael Olbrich, Michael
351 Prokop, Michal Koutný, Michal Sekletar, Mike Gilbert, Mikhail Kasimov,
352 Milan Broz, Milan Pässler, Muhammet Kara, Nicolas Boichat, Omer Katz,
353 Paride Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter
354 A. Bigot, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
355 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
356 Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez, Salvo Tomaselli,
357 Sebastian Reichel, Sergio Lindo Mansilla, Stefan Schweter, Stephen
358 Hemminger, Stuart Hayes, Susant Sahani, Sylvain Plantefève, Thomas
359 H. P. Andersen, Tobias Jungel, Tomasz Torcz, Vito Caputo, Will Dietz,
360 Will Thompson, Wim van Mourik, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
366 * The MemoryAccounting= unit property now defaults to on. After
367 discussions with the upstream control group maintainers we learnt
368 that the negative impact of cgroup memory accounting on current
369 kernels is finally relatively minimal, so that it should be safe to
370 enable this by default without affecting system performance. Besides
371 memory accounting only task accounting is turned on by default, all
372 other forms of resource accounting (CPU, IO, IP) remain off for now,
373 because it's not clear yet that their impact is small enough to move
374 from opt-in to opt-out. We recommend downstreams to leave memory
375 accounting on by default if kernel 4.14 or higher is primarily
376 used. On very resource constrained systems or when support for old
377 kernels is a necessity, -Dmemory-accounting-default=false can be used
378 to revert this change.
380 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
381 %udev_rules_update) and the journal catalog (%journal_catalog_update)
382 from the upgrade scriptlets of individual packages now do nothing.
383 Transfiletriggers have been added which will perform those updates
384 once at the end of the transaction.
386 Similar transfiletriggers have been added to execute any sysctl.d
387 and binfmt.d rules. Thus, it should be unnecessary to provide any
388 scriptlets to execute this configuration from package installation
391 * elogind-sysusers gained a mode where the configuration to execute is
392 specified on the command line, but this configuration is not executed
393 directly, but instead it is merged with the configuration on disk,
394 and the result is executed. This is useful for package installation
395 scripts which want to create the user before installing any files on
396 disk (in case some of those files are owned by that user), while
397 still allowing local admin overrides.
399 This functionality is exposed to rpm scriptlets through a new
400 %sysusers_create_package macro. Old %sysusers_create and
401 %sysusers_create_inline macros are deprecated.
403 A transfiletrigger for sysusers.d configuration is now installed,
404 which means that it should be unnecessary to call elogind-sysusers from
405 package installation scripts, unless the package installs any files
406 owned by those newly-created users, in which case
407 %sysusers_create_package should be used.
409 * Analogous change has been done for elogind-tmpfiles: it gained a mode
410 where the command-line configuration is merged with the configuration
411 on disk. This is exposed as the new %tmpfiles_create_package macro,
412 and %tmpfiles_create is deprecated. A transfiletrigger is installed
413 for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
414 from package installation scripts.
416 * sysusers.d configuration for a user may now also specify the group
417 number, in addition to the user number ("u username 123:456"), or
418 without the user number ("u username -:456").
420 * Configution items for elogind-sysusers can now be specified as
421 positional arguments when the new --inline switch is used.
423 * The login shell of users created through sysusers.d may now be
424 specified (previously, it was always /bin/sh for root and
425 /sbin/nologin for other users).
427 * elogind-analyze gained a new --global switch to look at global user
428 configuration. It also gained a unit-paths verb to list the unit load
429 paths that are compiled into elogind (which can be used with
430 --elogind, --user, or --global).
432 * udevadm trigger gained a new --settle/-w option to wait for any
433 triggered events to finish (but just those, and not any other events
434 which are triggered meanwhile).
436 * The action that elogind takes when the lid is closed and the
437 machine is connected to external power can now be configured using
438 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
439 was determined by HandleLidSwitch=, and, for backwards compatibility,
440 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
442 * journalctl will periodically call sd_journal_process() to make it
443 resilient against inotify queue overruns when journal files are
444 rotated very quickly.
446 * Two new functions in libelogind — sd_bus_get_n_queued_read and
447 sd_bus_get_n_queued_write — may be used to check the number of
448 pending bus messages.
450 * elogind gained a new
451 org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
452 which can be used to migrate foreign processes to scope and service
453 units. The primary user for this new API is elogind itself: the
454 elogind --user instance uses this call of the elogind --system
455 instance to migrate processes if it itself gets the request to
456 migrate processes and the kernel refuses this due to access
457 restrictions. Thanks to this "elogind-run --scope --user …" works
458 again in pure cgroups v2 environments when invoked from the user
461 * A new TemporaryFileSystem= setting can be used to mask out part of
462 the real file system tree with tmpfs mounts. This may be combined
463 with BindPaths= and BindReadOnlyPaths= to hide files or directories
464 not relevant to the unit, while still allowing some paths lower in
465 the tree to be accessed.
467 ProtectHome=tmpfs may now be used to hide user home and runtime
468 directories from units, in a way that is mostly equivalent to
469 "TemporaryFileSystem=/home /run/user /root".
471 * Non-service units are now started with KeyringMode=shared by default.
472 This means that mount and swapon and other mount tools have access
473 to keys in the main keyring.
475 * /sys/fs/bpf is now mounted automatically.
477 * QNX virtualization is now detected by elogind-detect-virt and may
478 be used in ConditionVirtualization=.
480 * IPAccounting= may now be enabled also for slice units.
482 * A new -Dsplit-bin= build configuration switch may be used to specify
483 whether bin and sbin directories are merged, or if they should be
484 included separately in $PATH and various listings of executable
485 directories. The build configuration scripts will try to autodetect
486 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
487 system, but distributions are encouraged to configure this
490 * A new -Dok-color= build configuration switch may be used to change
491 the colour of "OK" status messages.
493 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
494 PrivateNetwork=yes was buggy in previous versions of elogind. This
495 means that after the upgrade and daemon-reexec, any such units must
498 * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
499 will not exclude read-only files owned by root from cleanup.
501 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
502 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
503 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
504 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
505 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
506 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
507 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
508 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
509 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
510 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
511 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
512 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
513 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
514 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
515 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
516 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
522 * Some keyboards come with a zoom see-saw or rocker which until now got
523 mapped to the Linux "zoomin/out" keys in hwdb. However, these
524 keycodes are not recognized by any major desktop. They now produce
525 Up/Down key events so that they can be used for scrolling.
527 * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
528 slightly: previously, if an argument was specified for lines of this
529 type (i.e. the right-most column was set) this string was appended to
530 existing files each time elogind-tmpfiles was run. This behaviour was
531 different from what the documentation said, and not particularly
532 useful, as repeated elogind-tmpfiles invocations would not be
533 idempotent and grow such files without bounds. With this release
534 behaviour has been altered slightly, to match what the documentation
535 says: lines of this type only have an effect if the indicated files
536 don't exist yet, and only then the argument string is written to the
539 * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
540 elogind-tmpfiles behaviour: previously, read-only files owned by root
541 were always excluded from the file "aging" algorithm (i.e. the
542 automatic clean-up of directories like /tmp based on
543 atime/mtime/ctime). We intend to drop this restriction, and age files
544 by default even when owned by root and read-only. This behaviour was
545 inherited from older tools, but there have been requests to remove
546 it, and it's not obvious why this restriction was made in the first
547 place. Please speak up now, if you are aware of software that reqires
548 this behaviour, otherwise we'll remove the restriction in v238.
550 * A new environment variable $SYSTEMD_OFFLINE is now understood by
551 systemctl. It takes a boolean argument. If on, systemctl assumes it
552 operates on an "offline" OS tree, and will not attempt to talk to the
553 service manager. Previously, this mode was implicitly enabled if a
554 chroot() environment was detected, and this new environment variable
555 now provides explicit control.
557 * .path and .socket units may now be created transiently, too.
558 Previously only service, mount, automount and timer units were
559 supported as transient units. The elogind-run tool has been updated
560 to expose this new functionality, you may hence use it now to bind
561 arbitrary commands to path or socket activation on-the-fly from the
562 command line. Moreover, almost all properties are now exposed for the
563 unit types that already supported transient operation.
565 * The elogind-mount command gained support for a new --owner= parameter
566 which takes a user name, which is then resolved and included in uid=
567 and gid= mount options string of the file system to mount.
569 * A new unit condition ConditionControlGroupController= has been added
570 that checks whether a specific cgroup controller is available.
572 * Unit files, udev's .link files, and elogind-networkd's .netdev and
573 .network files all gained support for a new condition
574 ConditionKernelVersion= for checking against specific kernel
577 * In elogind-networkd, the [IPVLAN] section in .netdev files gained
578 support for configuring device flags in the Flags= setting. In the
579 same files, the [Tunnel] section gained support for configuring
580 AllowLocalRemote=. The [Route] section in .network files gained
581 support for configuring InitialCongestionWindow=,
582 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
583 understands RapidCommit=.
585 * elogind-networkd's DHCPv6 support gained support for Prefix
588 * sd-bus gained support for a new "watch-bind" feature. When this
589 feature is enabled, an sd_bus connection may be set up to connect to
590 an AF_UNIX socket in the file system as soon as it is created. This
591 functionality is useful for writing early-boot services that
592 automatically connect to the system bus as soon as it is started,
593 without ugly time-based polling. elogind-networkd and
594 elogind-resolved have been updated to make use of this
595 functionality. busctl exposes this functionality in a new
596 --watch-bind= command line switch.
598 * sd-bus will now optionally synthesize a local "Connected" signal as
599 soon as a D-Bus connection is set up fully. This message mirrors the
600 already existing "Disconnected" signal which is synthesized when the
601 connection is terminated. This signal is generally useful but
602 particularly handy in combination with the "watch-bind" feature
603 described above. Synthesizing of this message has to be requested
604 explicitly through the new API call sd_bus_set_connected_signal(). In
605 addition a new call sd_bus_is_ready() has been added that checks
606 whether a connection is fully set up (i.e. between the "Connected" and
607 "Disconnected" signals).
609 * sd-bus gained two new calls sd_bus_request_name_async() and
610 sd_bus_release_name_async() for asynchronously registering bus
611 names. Similar, there is now sd_bus_add_match_async() for installing
612 a signal match asynchronously. All of elogind's own services have
613 been updated to make use of these calls. Doing these operations
614 asynchronously has two benefits: it reduces the risk of deadlocks in
615 case of cyclic dependencies between bus services, and it speeds up
616 service initialization since synchronization points for bus
617 round-trips are removed.
619 * sd-bus gained two new calls sd_bus_match_signal() and
620 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
621 and sd_bus_add_match_async() but instead of taking a D-Bus match
622 string take match fields as normal function parameters.
624 * sd-bus gained two new calls sd_bus_set_sender() and
625 sd_bus_message_set_sender() for setting the sender name of outgoing
626 messages (either for all outgoing messages or for just one specific
627 one). These calls are only useful in direct connections as on
628 brokered connections the broker fills in the sender anyway,
629 overwriting whatever the client filled in.
631 * sd-event gained a new pseudo-handle that may be specified on all API
632 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
633 used this refers to the default event loop object of the calling
634 thread. Note however that this does not implicitly allocate one —
635 which has to be done prior by using sd_event_default(). Similarly
636 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
637 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
638 to the default bus of the specified type of the calling thread. Here
639 too this does not implicitly allocate bus connection objects, this
640 has to be done prior with sd_bus_default() and friends.
642 * sd-event gained a new call pair
643 sd_event_source_{get|set}_io_fd_own(). This may be used to request
644 automatic closure of the file descriptor an IO event source watches
645 when the event source is destroyed.
647 * elogind-networkd gained support for natively configuring WireGuard
650 * In previous versions elogind synthesized user records both for the
651 "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
652 internally. In order to simplify distribution-wide renames of the
653 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
654 new transitional flag file has been added: if
655 /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
656 user and group record within the elogind codebase is disabled.
658 * elogind-notify gained a new --uid= option for selecting the source
659 user/UID to use for notification messages sent to the service
662 * journalctl gained a new --grep= option to list only entries in which
663 the message matches a certain pattern. By default matching is case
664 insensitive if the pattern is lowercase, and case sensitive
665 otherwise. Option --case-sensitive=yes|no can be used to override
666 this an specify case sensitivity or case insensitivity.
668 * There's now a "elogind-analyze service-watchdogs" command for printing
669 the current state of the service runtime watchdog, and optionally
670 enabling or disabling the per-service watchdogs system-wide if given a
671 boolean argument (i.e. the concept you configure in WatchdogSec=), for
672 debugging purposes. There's also a kernel command line option
673 elogind.service_watchdogs= for controlling the same.
675 * Two new "log-level" and "log-target" options for elogind-analyze were
676 added that merge the now deprecated get-log-level, set-log-level and
677 get-log-target, set-log-target pairs. The deprecated options are still
678 understood for backwards compatibility. The two new options print the
679 current value when no arguments are given, and set them when a
680 level/target is given as an argument.
682 * sysusers.d's "u" lines now optionally accept both a UID and a GID
683 specification, separated by a ":" character, in order to create users
684 where UID and GID do not match.
686 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
687 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
688 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
689 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
690 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
691 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
692 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
693 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
694 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
695 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
696 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
697 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
698 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
699 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
700 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
701 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
702 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
703 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
704 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
705 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
712 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
713 in v235 has been extended to also set the dummy.ko module option
714 numdummies=0, preventing the kernel from automatically creating
715 dummy0. All dummy interfaces must now be explicitly created.
717 * Unknown '%' specifiers in configuration files are now rejected. This
718 applies to units and tmpfiles.d configuration. Any percent characters
719 that are followed by a letter or digit that are not supposed to be
720 interpreted as the beginning of a specifier should be escaped by
721 doubling ("%%"). (So "size=5%" is still accepted, as well as
722 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
723 valid specifiers today.)
725 * systemd-resolved now maintains a new dynamic
726 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
727 recommended to make /etc/resolv.conf a symlink to it. This file
728 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
729 includes dynamically acquired search domains, achieving more correct
730 DNS resolution by software that bypasses local DNS APIs such as NSS.
732 * The "uaccess" udev tag has been dropped from /dev/kvm and
733 /dev/dri/renderD*. These devices now have the 0666 permissions by
734 default (but this may be changed at build-time). /dev/dri/renderD*
735 will now be owned by the "render" group along with /dev/kfd.
737 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
738 systemd-journal-gatewayd.service and
739 systemd-journal-upload.service. This means "nss-systemd" must be
740 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
741 services are resolved properly.
743 * In /etc/fstab two new mount options are now understood:
744 x-systemd.makefs and x-systemd.growfs. The former has the effect that
745 the configured file system is formatted before it is mounted, the
746 latter that the file system is resized to the full block device size
747 after it is mounted (i.e. if the file system is smaller than the
748 partition it resides on, it's grown). This is similar to the fsck
749 logic in /etc/fstab, and pulls in systemd-makefs@.service and
750 systemd-growfs@.service as necessary, similar to
751 systemd-fsck@.service. Resizing is currently only supported on ext4
754 * In systemd-networkd, the IPv6 RA logic now optionally may announce
755 DNS server and domain information.
757 * Support for the LUKS2 on-disk format for encrypted partitions has
758 been added. This requires libcryptsetup2 during compilation and
761 * The systemd --user instance will now signal "readiness" when its
762 basic.target unit has been reached, instead of when the run queue ran
763 empty for the first time.
765 * Tmpfiles.d with user configuration are now also supported.
766 systemd-tmpfiles gained a new --user switch, and snippets placed in
767 ~/.config/user-tmpfiles.d/ and corresponding directories will be
768 executed by systemd-tmpfiles --user running in the new
769 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
770 running in the user session.
772 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
773 %S resolves to the top-level state directory (/var/lib for the system
774 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
775 top-level cache directory (/var/cache for the system instance,
776 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
777 logs directory (/var/log for the system instance,
778 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
779 existing %t specifier, that resolves to the top-level runtime
780 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
783 * journalctl learnt a new parameter --output-fields= for limiting the
784 set of journal fields to output in verbose and JSON output modes.
786 * systemd-timesyncd's configuration file gained a new option
787 RootDistanceMaxSec= for setting the maximum root distance of servers
788 it'll use, as well as the new options PollIntervalMinSec= and
789 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
791 * bootctl gained a new command "list" for listing all available boot
792 menu items on systems that follow the boot loader specification.
794 * systemctl gained a new --dry-run switch that shows what would be done
795 instead of doing it, and is currently supported by the shutdown and
798 * ConditionSecurity= can now detect the TOMOYO security module.
800 * Unit file [Install] sections are now also respected in unit drop-in
801 files. This is intended to be used by drop-ins under /usr/lib/.
803 * systemd-firstboot may now also set the initial keyboard mapping.
805 * Udev "changed" events for devices which are exposed as systemd
806 .device units are now propagated to units specified in
807 ReloadPropagatedFrom= as reload requests.
809 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
810 unit template name (i.e. a name in the form of 'foobar@.service',
811 without the instance component between the '@' and - the '.'), then
812 the escaped sysfs path of the device is automatically used as the
815 * SystemCallFilter= in unit files has been extended so that an "errno"
816 can be specified individually for each system call. Example:
817 SystemCallFilter=~uname:EILSEQ.
819 * The cgroup delegation logic has been substantially updated. Delegate=
820 now optionally takes a list of controllers (instead of a boolean, as
821 before), which lists the controllers to delegate at least.
823 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
825 * A new LogLevelMax= setting configures the maximum log level any
826 process of the service may log at (i.e. anything with a lesser
827 priority than what is specified is automatically dropped). A new
828 LogExtraFields= setting allows configuration of additional journal
829 fields to attach to all log records generated by any of the unit's
832 * New StandardInputData= and StandardInputText= settings along with the
833 new option StandardInput=data may be used to configure textual or
834 binary data that shall be passed to the executed service process via
835 standard input, encoded in-line in the unit file.
837 * StandardInput=, StandardOutput= and StandardError= may now be used to
838 connect stdin/stdout/stderr of executed processes directly with a
839 file or AF_UNIX socket in the file system, using the new "file:" option.
841 * A new unit file option CollectMode= has been added, that allows
842 tweaking the garbage collection logic for units. It may be used to
843 tell systemd to garbage collect units that have failed automatically
844 (normally it only GCs units that exited successfully). systemd-run
845 and systemd-mount expose this new functionality with a new -G option.
847 * "machinectl bind" may now be used to bind mount non-directories
848 (i.e. regularfiles, devices, fifos, sockets).
850 * systemd-analyze gained a new verb "calendar" for validating and
851 testing calendar time specifications to use for OnCalendar= in timer
852 units. Besides validating the expression it will calculate the next
853 time the specified expression would elapse.
855 * In addition to the pre-existing FailureAction= unit file setting
856 there's now SuccessAction=, for configuring a shutdown action to
857 execute when a unit completes successfully. This is useful in
858 particular inside containers that shall terminate after some workload
859 has been completed. Also, both options are now supported for all unit
860 types, not just services.
862 * networkds's IP rule support gained two new options
863 IncomingInterface= and OutgoingInterface= for configuring the incoming
864 and outgoing interfaces of configured rules. systemd-networkd also
865 gained support for "vxcan" network devices.
867 * networkd gained a new setting RequiredForOnline=, taking a
868 boolean. If set, systemd-wait-online will take it into consideration
869 when determining that the system is up, otherwise it will ignore the
870 interface for this purpose.
872 * The sd_notify() protocol gained support for a new operation: with
873 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
874 store again, ahead of POLLHUP or POLLERR when they are removed
877 * A new document doc/UIDS-GIDS.md has been added to the source tree,
878 that documents the UID/GID range and assignment assumptions and
879 requirements of systemd.
881 * The watchdog device PID 1 will ping may now be configured through the
882 WatchdogDevice= configuration file setting, or by setting the
883 systemd.watchdog_service= kernel commandline option.
885 * systemd-resolved's gained support for registering DNS-SD services on
886 the local network using MulticastDNS. Services may either be
887 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
888 the same dir below /run, /usr/lib), or through its D-Bus API.
890 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
891 extend the effective start, runtime, and stop time. The service must
892 continue to send EXTEND_TIMEOUT_USEC within the period specified to
893 prevent the service manager from making the service as timedout.
895 * elogind-resolved's DNSSEC support gained support for RFC 8080
896 (Ed25519 keys and signatures).
898 * The elogind-resolve command line tool gained a new set of options
899 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
900 --set-nta= and --revert to configure per-interface DNS configuration
901 dynamically during runtime. It's useful for pushing DNS information
902 into elogind-resolved from DNS hook scripts that various interface
903 managing software supports (such as pppd).
905 * elogind-nspawn gained a new --network-namespace-path= command line
906 option, which may be used to make a container join an existing
907 network namespace, by specifying a path to a "netns" file.
909 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
910 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
911 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
912 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
913 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
914 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
915 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
916 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
917 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
918 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
919 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
920 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
921 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
922 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
923 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
924 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
925 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
926 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
927 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
928 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
929 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
930 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
931 Jędrzejewski-Szmek, Zeal Jagannatha
937 * INCOMPATIBILITY: systemd-logind.service and other long-running
938 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
939 communication with the outside. This generally improves security of
940 the system, and is in almost all cases a safe and good choice, as
941 these services do not and should not provide any network-facing
942 functionality. However, systemd-logind uses the glibc NSS API to
943 query the user database. This creates problems on systems where NSS
944 is set up to directly consult network services for user database
945 lookups. In particular, this creates incompatibilities with the
946 "nss-nis" module, which attempts to directly contact the NIS/YP
947 network servers it is configured for, and will now consistently
948 fail. In such cases, it is possible to turn off IP sandboxing for
949 systemd-logind.service (set IPAddressDeny= in its [Service] section
950 to the empty string, via a .d/ unit file drop-in). Downstream
951 distributions might want to update their nss-nis packaging to include
952 such a drop-in snippet, accordingly, to hide this incompatibility
953 from the user. Another option is to make use of glibc's nscd service
954 to proxy such network requests through a privilege-separated, minimal
955 local caching daemon, or to switch to more modern technologies such
956 sssd, whose NSS hook-ups generally do not involve direct network
957 access. In general, we think it's definitely time to question the
958 implementation choices of nss-nis, i.e. whether it's a good idea
959 today to embed a network-facing loadable module into all local
960 processes that need to query the user database, including the most
961 trivial and benign ones, such as "ls". For more details about
962 IPAddressDeny= see below.
964 * A new modprobe.d drop-in is now shipped by default that sets the
965 bonding module option max_bonds=0. This overrides the kernel default,
966 to avoid conflicts and ambiguity as to whether or not bond0 should be
967 managed by systemd-networkd or not. This resolves multiple issues
968 with bond0 properties not being applied, when bond0 is configured
969 with systemd-networkd. Distributors may choose to not package this,
970 however in that case users will be prevented from correctly managing
971 bond0 interface using systemd-networkd.
973 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
974 which print the logging level and target of the system manager. They
975 complement the existing "set-log-level" and "set-log-target" verbs
976 used to change those values.
978 * journald.conf gained a new boolean setting ReadKMsg= which defaults
979 to on. If turned off kernel log messages will not be read by
980 systemd-journald or included in the logs. It also gained a new
981 setting LineMax= for configuring the maximum line length in
982 STDOUT/STDERR log streams. The new default for this value is 48K, up
983 from the previous hardcoded 2048.
985 * A new unit setting RuntimeDirectoryPreserve= has been added, which
986 allows more detailed control of what to do with a runtime directory
987 configured with RuntimeDirectory= (i.e. a directory below /run or
988 $XDG_RUNTIME_DIR) after a unit is stopped.
990 * The RuntimeDirectory= setting for units gained support for creating
991 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
992 one top-level directory.
994 * Units gained new options StateDirectory=, CacheDirectory=,
995 LogsDirectory= and ConfigurationDirectory= which are closely related
996 to RuntimeDirectory= but manage per-service directories below
997 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
998 possible to write unit files which when activated automatically gain
999 properly owned service specific directories in these locations, thus
1000 making unit files self-contained and increasing compatibility with
1001 stateless systems and factory reset where /etc or /var are
1002 unpopulated at boot. Matching these new settings there's also
1003 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1004 ConfigurationDirectoryMode= for configuring the access mode of these
1005 directories. These settings are particularly useful in combination
1006 with DynamicUser=yes as they provide secure, properly-owned,
1007 writable, and stateful locations for storage, excluded from the
1008 sandbox that such services live in otherwise.
1010 * Automake support has been removed from this release. systemd is now
1013 * systemd-journald will now aggressively cache client metadata during
1014 runtime, speeding up log write performance under pressure. This comes
1015 at a small price though: as much of the metadata is read
1016 asynchronously from /proc/ (and isn't implicitly attached to log
1017 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1018 metadata stored alongside a log entry might be slightly
1019 out-of-date. Previously it could only be slightly newer than the log
1020 message. The time window is small however, and given that the kernel
1021 is unlikely to be improved anytime soon in this regard, this appears
1024 * nss-myhostname/systemd-resolved will now by default synthesize an
1025 A/AAAA resource record for the "_gateway" hostname, pointing to the
1026 current default IP gateway. Previously it did that for the "gateway"
1027 name, hampering adoption, as some distributions wanted to leave that
1028 host name open for local use. The old behaviour may still be
1029 requested at build time.
1031 * systemd-networkd's [Address] section in .network files gained a new
1032 Scope= setting for configuring the IP address scope. The [Network]
1033 section gained a new boolean setting ConfigureWithoutCarrier= that
1034 tells systemd-networkd to ignore link sensing when configuring the
1035 device. The [DHCP] section gained a new Anonymize= boolean option for
1036 turning on a number of options suggested in RFC 7844. A new
1037 [RoutingPolicyRule] section has been added for configuring the IP
1038 routing policy. The [Route] section has gained support for a new
1039 Type= setting which permits configuring
1040 blackhole/unreachable/prohibit routes.
1042 * The [VRF] section in .netdev files gained a new Table= setting for
1043 configuring the routing table to use. The [Tunnel] section gained a
1044 new Independent= boolean field for configuring tunnels independent of
1045 an underlying network interface. The [Bridge] section gained a new
1046 GroupForwardMask= option for configuration of propagation of link
1047 local frames between bridge ports.
1049 * The WakeOnLan= setting in .link files gained support for a number of
1050 new modes. A new TCP6SegmentationOffload= setting has been added for
1051 configuring TCP/IPv6 hardware segmentation offload.
1053 * The IPv6 RA sender implementation may now optionally send out RDNSS
1054 and RDNSSL records to supply DNS configuration to peers.
1056 * systemd-nspawn gained support for a new --system-call-filter= command
1057 line option for adding and removing entries in the default system
1058 call filter it applies. Moreover systemd-nspawn has been changed to
1059 implement a system call whitelist instead of a blacklist.
1061 * systemd-run gained support for a new --pipe command line option. If
1062 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1063 are directly passed on to the activated transient service
1064 executable. This allows invoking arbitrary processes as systemd
1065 services (for example to take benefit of dependency management,
1066 accounting management, resource management or log management that is
1067 done automatically for services) — while still allowing them to be
1068 integrated in a classic UNIX shell pipeline.
1070 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1071 using ReloadPropagationTo= is configured, a reload is now propagated
1072 to configured units. (Previously this was only done on explicitly
1073 requested reloads, using "systemctl reload" or an equivalent
1076 * For each service unit a restart counter is now kept: it is increased
1077 each time the service is restarted due to Restart=, and may be
1078 queried using "systemctl show -p NRestarts …".
1080 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1081 @signal and @timer have been added, for usage with SystemCallFilter=
1082 in unit files and the new --system-call-filter= command line option
1083 of systemd-nspawn (see above).
1085 * ExecStart= lines in unit files gained two new modifiers: when a
1086 command line is prefixed with "!" the command will be executed as
1087 configured, except for the credentials applied by
1088 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1089 "+", but does still apply namespacing options unlike "+". There's
1090 also "!!" now, which is mostly identical, but becomes a NOP on
1091 systems that support ambient capabilities. This is useful to write
1092 unit files that work with ambient capabilities where possible but
1093 automatically fall back to traditional privilege dropping mechanisms
1094 on systems where this is not supported.
1096 * ListenNetlink= settings in socket units now support RDMA netlink
1099 * A new unit file setting LockPersonality= has been added which permits
1100 locking down the chosen execution domain ("personality") of a service
1103 * A new special target "getty-pre.target" has been added, which is
1104 ordered before all text logins, and may be used to order services
1105 before textual logins acquire access to the console.
1107 * systemd will now attempt to load the virtio-rng.ko kernel module very
1108 early on if a VM environment supporting this is detected. This should
1109 improve entropy during early boot in virtualized environments.
1111 * A _netdev option is now supported in /etc/crypttab that operates in a
1112 similar way as the same option in /etc/fstab: it permits configuring
1113 encrypted devices that need to be ordered after the network is up.
1114 Following this logic, two new special targets
1115 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1116 added that are to cryptsetup.target what remote-fs.target and
1117 remote-fs-pre.target are to local-fs.target.
1119 * Service units gained a new UnsetEnvironment= setting which permits
1120 unsetting specific environment variables for services that are
1121 normally passed to it (for example in order to mask out locale
1122 settings for specific services that can't deal with it).
1124 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1125 traffic accounting (packet count as well as byte count) is done for
1126 the service, and shown as part of "systemctl status" or "systemd-run
1129 * Service units acquired two new options IPAddressAllow= and
1130 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1131 for configuring a simple IP access control list for all sockets of
1132 the unit. These options are available also on .slice and .socket
1133 units, permitting flexible access list configuration for individual
1134 services as well as groups of services (as defined by a slice unit),
1135 including system-wide. Note that IP ACLs configured this way are
1136 enforced on every single IPv4 and IPv6 socket created by any process
1137 of the service unit, and apply to ingress as well as egress traffic.
1139 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
1140 structured log message is generated each time the unit is stopped,
1141 containing information about the consumed resources of this
1144 * A new setting KeyringMode= has been added to unit files, which may be
1145 used to control how the kernel keyring is set up for executed
1148 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1149 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1150 behaviour (that is: these commands return immediately after the
1151 operation was enqueued instead of waiting for the operation to
1152 complete). Previously, "systemctl poweroff" and "systemctl reboot"
1153 were asynchronous on systems using systemd-logind (i.e. almost
1154 always, and like they were on sysvinit), and the other three commands
1155 were unconditionally synchronous. With this release this is cleaned
1156 up, and callers will see the same asynchronous behaviour on all
1157 systems for all five operations.
1159 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1162 * .timer units now accept calendar specifications in other timezones
1163 than UTC or the local timezone.
1165 * The tmpfiles snippet var.conf has been changed to create
1166 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1167 the "utmp" group already, and it appears to be generally understood
1168 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1169 databases. Previously this was implemented correctly for all these
1170 databases excepts btmp, which has been opened up like this now
1171 too. Note that while the other databases are world-readable
1172 (i.e. 0644), btmp is not and remains more restrictive.
1174 * The systemd-resolve tool gained a new --reset-server-features
1175 switch. When invoked like this systemd-resolved will forget
1176 everything it learnt about the features supported by the configured
1177 upstream DNS servers, and restarts the feature probing logic on the
1178 next resolver look-up for them at the highest feature level
1181 * The status dump systemd-resolved sends to the logs upon receiving
1182 SIGUSR1 now also includes information about all DNS servers it is
1183 configured to use, and the features levels it probed for them.
1185 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1186 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
1187 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1188 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1189 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1190 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1191 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1192 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1193 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1194 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1195 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1196 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1197 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1198 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1199 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1200 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1201 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1202 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1203 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1204 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1206 — Berlin, 2017-10-06
1210 * Meson is now supported as build system in addition to Automake. It is
1211 our plan to remove Automake in one of our next releases, so that
1212 Meson becomes our exclusive build system. Hence, please start using
1213 the Meson build system in your downstream packaging. There's plenty
1214 of documentation around how to use Meson, the extremely brief
1217 ./autogen.sh && ./configure && make && sudo make install
1221 meson build && ninja -C build && sudo ninja -C build install
1223 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1224 which permits configuring a timeout on the time a job is
1225 running. This is particularly useful for setting timeouts on jobs for
1228 * Unit files gained two new options ConditionUser= and ConditionGroup=
1229 for conditionalizing units based on the identity of the user/group
1230 running a systemd user instance.
1232 * systemd-networkd now understands a new FlowLabel= setting in the
1233 [VXLAN] section of .network files, as well as a Priority= in
1234 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1235 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1236 gained support for configuration of GENEVE links, and IPv6 address
1237 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1239 * .link files now understand a new Port= setting.
1241 * systemd-networkd's DHCP support gained support for DHCP option 119
1242 (domain search list).
1244 * systemd-networkd gained support for serving IPv6 address ranges using
1245 the Router Advertisement protocol. The new .network configuration
1246 section [IPv6Prefix] may be used to configure the ranges to
1247 serve. This is implemented based on a new, minimal, native server
1248 implementation of RA.
1250 * journalctl's --output= switch gained support for a new parameter
1251 "short-iso-precise" for a mode where timestamps are shown as precise
1254 * systemd-udevd's "net_id" builtin may now generate stable network
1255 interface names from IBM PowerVM VIO devices as well as ACPI platform
1258 * MulticastDNS support in systemd-resolved may now be explicitly
1259 enabled/disabled using the new MulticastDNS= configuration file
1262 * systemd-resolved may now optionally use libidn2 instead of the libidn
1263 for processing internationalized domain names. Support for libidn2
1264 should be considered experimental and should not be enabled by
1267 * "machinectl pull-tar" and related call may now do verification of
1268 downloaded images using SUSE-style .sha256 checksum files in addition
1269 to the already existing support for validating using Ubuntu-style
1272 * sd-bus gained support for a new sd_bus_message_appendv() call which
1273 is va_list equivalent of sd_bus_message_append().
1275 * sd-boot gained support for validating images using SHIM/MOK.
1277 * The SMACK code learnt support for "onlycap".
1279 * systemd-mount --umount is now much smarter in figuring out how to
1280 properly unmount a device given its mount or device path.
1282 * The code to call libnss_dns as a fallback from libnss_resolve when
1283 the communication with systemd-resolved fails was removed. This
1284 fallback was redundant and interfered with the [!UNAVAIL=return]
1285 suffix. See nss-resolve(8) for the recommended configuration.
1287 * systemd-logind may now be restarted without losing state. It stores
1288 the file descriptors for devices it manages in the system manager
1289 using the FDSTORE= mechanism. Please note that further changes in
1290 other components may be required to make use of this (for example
1291 Xorg has code to listen for stops of systemd-logind and terminate
1292 itself when logind is stopped or restarted, in order to avoid using
1293 stale file descriptors for graphical devices, which is now
1294 counterproductive and must be reverted in order for restarts of
1295 systemd-logind to be safe. See
1296 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1298 * All kernel install plugins are called with the environment variable
1299 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1300 /etc/machine-id. If the file is missing or empty, the variable is
1301 empty and BOOT_DIR_ABS is the path of a temporary directory which is
1302 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
1303 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1305 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
1306 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1307 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1308 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1309 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
1310 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1311 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1312 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1313 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
1314 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1315 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1316 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1317 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1318 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1319 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1320 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1321 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1322 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1323 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1324 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1325 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1326 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1327 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1328 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1329 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
1330 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1331 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
1332 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1333 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1336 — Berlin, 2017-07-12
1340 * The "hybrid" control group mode has been modified to improve
1341 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1342 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1343 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1344 cgroups-v1 hierarchy), the only externally visible change being that
1345 the cgroups-v2 hierarchy is also mounted, to
1346 /sys/fs/cgroup/unified. This should provide a large degree of
1347 compatibility with "legacy" cgroups-v1, while taking benefit of the
1348 better management capabilities of cgroups-v2.
1350 * The default control group setup mode may be selected both a boot-time
1351 via a set of kernel command line parameters (specifically:
1352 systemd.unified_cgroup_hierarchy= and
1353 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1354 default selected on the configure command line
1355 (--with-default-hierarchy=). The upstream default is "hybrid"
1356 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1357 this will change in a future systemd version to be "unified" (pure
1358 cgroups-v2 mode). The third option for the compile time option is
1359 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1360 distributions to default to "hybrid" mode for release distributions,
1361 starting with v233. We recommend "unified" for development
1362 distributions (specifically: distributions such as Fedora's rawhide)
1363 as that's where things are headed in the long run. Use "legacy" for
1364 greatest stability and compatibility only.
1366 * Note one current limitation of "unified" and "hybrid" control group
1367 setup modes: the kernel currently does not permit the systemd --user
1368 instance (i.e. unprivileged code) to migrate processes between two
1369 disconnected cgroup subtrees, even if both are managed and owned by
1370 the user. This effectively means "systemd-run --user --scope" doesn't
1371 work when invoked from outside of any "systemd --user" service or
1372 scope. Specifically, it is not supported from session scopes. We are
1373 working on fixing this in a future systemd version. (See #3388 for
1374 further details about this.)
1376 * DBus policy files are now installed into /usr rather than /etc. Make
1377 sure your system has dbus >= 1.9.18 running before upgrading to this
1378 version, or override the install path with --with-dbuspolicydir= .
1380 * All python scripts shipped with systemd (specifically: the various
1381 tests written in Python) now require Python 3.
1383 * systemd unit tests can now run standalone (without the source or
1384 build directories), and can be installed into /usr/lib/systemd/tests/
1385 with 'make install-tests'.
1387 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1388 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1391 * Support for the %c, %r, %R specifiers in unit files has been
1392 removed. Specifiers are not supposed to be dependent on configuration
1393 in the unit file itself (so that they resolve the same regardless
1394 where used in the unit files), but these specifiers were influenced
1395 by the Slice= option.
1397 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1398 all cases. If distributions want to use a different shell for this
1399 purpose (for example Fedora's /sbin/sushell) they need to specify
1400 this explicitly at configure time using --with-debug-shell=.
1402 * The confirmation spawn prompt has been reworked to offer the
1405 (c)ontinue, proceed without asking anymore
1406 (D)ump, show the state of the unit
1407 (f)ail, don't execute the command and pretend it failed
1409 (i)nfo, show a short summary of the unit
1410 (j)obs, show jobs that are in progress
1411 (s)kip, don't execute the command and pretend it succeeded
1412 (y)es, execute the command
1414 The 'n' choice for the confirmation spawn prompt has been removed,
1415 because its meaning was confusing.
1417 The prompt may now also be redirected to an alternative console by
1418 specifying the console as parameter to systemd.confirm_spawn=.
1420 * Services of Type=notify require a READY=1 notification to be sent
1421 during startup. If no such message is sent, the service now fails,
1422 even if the main process exited with a successful exit code.
1424 * Services that fail to start up correctly now always have their
1425 ExecStopPost= commands executed. Previously, they'd enter "failed"
1426 state directly, without executing these commands.
1428 * The option MulticastDNS= of network configuration files has acquired
1429 an actual implementation. With MulticastDNS=yes a host can resolve
1430 names of remote hosts and reply to mDNS A and AAAA requests.
1432 * When units are about to be started an additional check is now done to
1433 ensure that all dependencies of type BindsTo= (when used in
1434 combination with After=) have been started.
1436 * systemd-analyze gained a new verb "syscall-filter" which shows which
1437 system call groups are defined for the SystemCallFilter= unit file
1438 setting, and which system calls they contain.
1440 * A new system call filter group "@filesystem" has been added,
1441 consisting of various file system related system calls. Group
1442 "@reboot" has been added, covering reboot, kexec and shutdown related
1443 calls. Finally, group "@swap" has been added covering swap
1444 configuration related calls.
1446 * A new unit file option RestrictNamespaces= has been added that may be
1447 used to restrict access to the various process namespace types the
1448 Linux kernel provides. Specifically, it may be used to take away the
1449 right for a service unit to create additional file system, network,
1450 user, and other namespaces. This sandboxing option is particularly
1451 relevant due to the high amount of recently discovered namespacing
1452 related vulnerabilities in the kernel.
1454 * systemd-udev's .link files gained support for a new AutoNegotiation=
1455 setting for configuring Ethernet auto-negotiation.
1457 * systemd-networkd's .network files gained support for a new
1458 ListenPort= setting in the [DHCP] section to explicitly configure the
1459 UDP client port the DHCP client shall listen on.
1461 * .network files gained a new Unmanaged= boolean setting for explicitly
1462 excluding one or more interfaces from management by systemd-networkd.
1464 * The systemd-networkd ProxyARP= option has been renamed to
1465 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1466 renamed to ReduceARPProxy=. The old names continue to be available
1469 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1470 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1472 * systemd-networkd's bonding device support gained support for two new
1473 configuration options ActiveSlave= and PrimarySlave=.
1475 * The various options in the [Match] section of .network files gained
1476 support for negative matching.
1478 * New systemd-specific mount options are now understood in /etc/fstab:
1480 x-systemd.mount-timeout= may be used to configure the maximum
1481 permitted runtime of the mount command.
1483 x-systemd.device-bound may be set to bind a mount point to its
1484 backing device unit, in order to automatically remove a mount point
1485 if its backing device is unplugged. This option may also be
1486 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1487 on the block device, which is now automatically set for all CDROM
1488 drives, so that mounted CDs are automatically unmounted when they are
1489 removed from the drive.
1491 x-systemd.after= and x-systemd.before= may be used to explicitly
1492 order a mount after or before another unit or mount point.
1494 * Enqueued start jobs for device units are now automatically garbage
1495 collected if there are no jobs waiting for them anymore.
1497 * systemctl list-jobs gained two new switches: with --after, for every
1498 queued job the jobs it's waiting for are shown; with --before the
1499 jobs which it's blocking are shown.
1501 * systemd-nspawn gained support for ephemeral boots from disk images
1502 (or in other words: --ephemeral and --image= may now be
1503 combined). Moreover, ephemeral boots are now supported for normal
1504 directories, even if the backing file system is not btrfs. Of course,
1505 if the file system does not support file system snapshots or
1506 reflinks, the initial copy operation will be relatively expensive, but
1507 this should still be suitable for many use cases.
1509 * Calendar time specifications in .timer units now support
1510 specifications relative to the end of a month by using "~" instead of
1511 "-" as separator between month and day. For example, "*-02~03" means
1512 "the third last day in February". In addition a new syntax for
1513 repeated events has been added using the "/" character. For example,
1514 "9..17/2:00" means "every two hours from 9am to 5pm".
1516 * systemd-socket-proxyd gained a new parameter --connections-max= for
1517 configuring the maximum number of concurrent connections.
1519 * sd-id128 gained a new API for generating unique IDs for the host in a
1520 way that does not leak the machine ID. Specifically,
1521 sd_id128_get_machine_app_specific() derives an ID based on the
1522 machine ID a in well-defined, non-reversible, stable way. This is
1523 useful whenever an identifier for the host is needed but where the
1524 identifier shall not be useful to identify the system beyond the
1525 scope of the application itself. (Internally this uses HMAC-SHA256 as
1526 keyed hash function using the machine ID as input.)
1528 * NotifyAccess= gained a new supported value "exec". When set
1529 notifications are accepted from all processes systemd itself invoked,
1530 including all control processes.
1532 * .nspawn files gained support for defining overlay mounts using the
1533 Overlay= and OverlayReadOnly= options. Previously this functionality
1534 was only available on the systemd-nspawn command line.
1536 * systemd-nspawn's --bind= and --overlay= options gained support for
1537 bind/overlay mounts whose source lies within the container tree by
1538 prefixing the source path with "+".
1540 * systemd-nspawn's --bind= and --overlay= options gained support for
1541 automatically allocating a temporary source directory in /var/tmp
1542 that is removed when the container dies. Specifically, if the source
1543 directory is specified as empty string this mechanism is selected. An
1544 example usage is --overlay=+/var::/var, which creates an overlay
1545 mount based on the original /var contained in the image, overlayed
1546 with a temporary directory in the host's /var/tmp. This way changes
1547 to /var are automatically flushed when the container shuts down.
1549 * systemd-nspawn --image= option does now permit raw file system block
1550 devices (in addition to images containing partition tables, as
1553 * The disk image dissection logic in systemd-nspawn gained support for
1554 automatically setting up LUKS encrypted as well as Verity protected
1555 partitions. When a container is booted from an encrypted image the
1556 passphrase is queried at start-up time. When a container with Verity
1557 data is started, the root hash is search in a ".roothash" file
1558 accompanying the disk image (alternatively, pass the root hash via
1559 the new --root-hash= command line option).
1561 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1562 be used to dissect disk images the same way as systemd-nspawn does
1563 it, following the Bootable Partition Specification. It may even be
1564 used to mount disk images with complex partition setups (including
1565 LUKS and Verity partitions) to a local host directory, in order to
1566 inspect them. This tool is not considered public API (yet), and is
1567 thus not installed into /usr/bin. Please do not rely on its
1568 existence, since it might go away or be changed in later systemd
1571 * A new generator "systemd-verity-generator" has been added, similar in
1572 style to "systemd-cryptsetup-generator", permitting automatic setup of
1573 Verity root partitions when systemd boots up. In order to make use of
1574 this your partition setup should follow the Discoverable Partitions
1575 Specification, and the GPT partition ID of the root file system
1576 partition should be identical to the upper 128bit of the Verity root
1577 hash. The GPT partition ID of the Verity partition protecting it
1578 should be the lower 128bit of the Verity root hash. If the partition
1579 image follows this model it is sufficient to specify a single
1580 "roothash=" kernel command line argument to both configure which root
1581 image and verity partition to use as well as the root hash for
1582 it. Note that systemd-nspawn's Verity support follows the same
1583 semantics, meaning that disk images with proper Verity data in place
1584 may be booted in containers with systemd-nspawn as well as on
1585 physical systems via the verity generator. Also note that the "mkosi"
1586 tool available at https://github.com/systemd/mkosi has been updated
1587 to generate Verity protected disk images following this scheme. In
1588 fact, it has been updated to generate disk images that optionally
1589 implement a complete UEFI SecureBoot trust chain, involving a signed
1590 kernel and initrd image that incorporates such a root hash as well as
1591 a Verity-enabled root partition.
1593 * The hardware database (hwdb) udev supports has been updated to carry
1594 accelerometer quirks.
1596 * All system services are now run with a fresh kernel keyring set up
1597 for them. The invocation ID is stored by default in it, thus
1598 providing a safe, non-overridable way to determine the invocation
1601 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1602 options for bind mounting arbitrary paths in a service-specific
1603 way. When these options are used, arbitrary host or service files and
1604 directories may be mounted to arbitrary locations in the service's
1607 * Documentation has been added that lists all of systemd's low-level
1608 environment variables:
1610 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
1612 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1613 whether a specific socket file descriptor matches a specified socket
1616 * systemd-firstboot has been updated to check for the
1617 systemd.firstboot= kernel command line option. It accepts a boolean
1618 and when set to false the first boot questions are skipped.
1620 * systemd-fstab-generator has been updated to check for the
1621 systemd.volatile= kernel command line option, which either takes an
1622 optional boolean parameter or the special value "state". If used the
1623 system may be booted in a "volatile" boot mode. Specifically,
1624 "systemd.volatile" is used, the root directory will be mounted as
1625 tmpfs, and only /usr is mounted from the actual root file system. If
1626 "systemd.volatile=state" is used, the root directory will be mounted
1627 as usual, but /var is mounted as tmpfs. This concept provides similar
1628 functionality as systemd-nspawn's --volatile= option, but provides it
1629 on physical boots. Use this option for implementing stateless
1630 systems, or testing systems with all state and/or configuration reset
1631 to the defaults. (Note though that many distributions are not
1632 prepared to boot up without a populated /etc or /var, though.)
1634 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1635 partitions. Previously it only supported LUKS encrypted partitions
1636 for all other uses, except for the root partition itself.
1638 * Socket units gained support for listening on AF_VSOCK sockets for
1639 communication in virtualized QEMU environments.
1641 * The "configure" script gained a new option --with-fallback-hostname=
1642 for specifying the fallback hostname to use if none is configured in
1643 /etc/hostname. For example, by specifying
1644 --with-fallback-hostname=fedora it is possible to default to a
1645 hostname of "fedora" on pristine installations.
1647 * systemd-cgls gained support for a new --unit= switch for listing only
1648 the control groups of a specific unit. Similar --user-unit= has been
1649 added for listing only the control groups of a specific user unit.
1651 * systemd-mount gained a new --umount switch for unmounting a mount or
1652 automount point (and all mount/automount points below it).
1654 * systemd will now refuse full configuration reloads (via systemctl
1655 daemon-reload and related calls) unless at least 16MiB of free space
1656 are available in /run. This is a safety precaution in order to ensure
1657 that generators can safely operate after the reload completed.
1659 * A new unit file option RootImage= has been added, which has a similar
1660 effect as RootDirectory= but mounts the service's root directory from
1661 a disk image instead of plain directory. This logic reuses the same
1662 image dissection and mount logic that systemd-nspawn already uses,
1663 and hence supports any disk images systemd-nspawn supports, including
1664 those following the Discoverable Partition Specification, as well as
1665 Verity enabled images. This option enables systemd to run system
1666 services directly off disk images acting as resource bundles,
1667 possibly even including full integrity data.
1669 * A new MountAPIVFS= unit file option has been added, taking a boolean
1670 argument. If enabled /proc, /sys and /dev (collectively called the
1671 "API VFS") will be mounted for the service. This is only relevant if
1672 RootDirectory= or RootImage= is used for the service, as these mounts
1673 are of course in place in the host mount namespace anyway.
1675 * systemd-nspawn gained support for a new --pivot-root= switch. If
1676 specified the root directory within the container image is pivoted to
1677 the specified mount point, while the original root disk is moved to a
1678 different place. This option enables booting of ostree images
1679 directly with systemd-nspawn.
1681 * The systemd build scripts will no longer complain if the NTP server
1682 addresses are not changed from the defaults. Google now supports
1683 these NTP servers officially. We still recommend downstreams to
1684 properly register an NTP pool with the NTP pool project though.
1686 * coredumpctl gained a new "--reverse" option for printing the list
1687 of coredumps in reverse order.
1689 * coredumpctl will now show additional information about truncated and
1690 inaccessible coredumps, as well as coredumps that are still being
1691 processed. It also gained a new --quiet switch for suppressing
1692 additional informational message in its output.
1694 * coredumpctl gained support for only showing coredumps newer and/or
1695 older than specific timestamps, using the new --since= and --until=
1696 options, reminiscent of journalctl's options by the same name.
1698 * The systemd-coredump logic has been improved so that it may be reused
1699 to collect backtraces in non-compiled languages, for example in
1700 scripting languages such as Python.
1702 * machinectl will now show the UID shift of local containers, if user
1703 namespacing is enabled for them.
1705 * systemd will now optionally run "environment generator" binaries at
1706 configuration load time. They may be used to add environment
1707 variables to the environment block passed to services invoked. One
1708 user environment generator is shipped by default that sets up
1709 environment variables based on files dropped into /etc/environment.d
1710 and ~/.config/environment.d/.
1712 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1715 * hostnamed has been updated to report a new chassis type of
1716 "convertible" to cover "foldable" laptops that can both act as a
1717 tablet and as a laptop, such as various Lenovo Yoga devices.
1719 Contributions from: Adrián López, Alexander Galanin, Alexander
1720 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1721 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1722 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1723 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1724 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1725 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1726 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1727 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
1728 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1729 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1730 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1731 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1732 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1733 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1734 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1735 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1736 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1737 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1738 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1739 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1740 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1741 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1742 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1743 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1744 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1745 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1746 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1749 — Berlin, 2017-03-01
1753 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1754 RestrictAddressFamilies= enabled. These sandboxing options should
1755 generally be compatible with the various external udev call-out
1756 binaries we are aware of, however there may be exceptions, in
1757 particular when exotic languages for these call-outs are used. In
1758 this case, consider turning off these settings locally.
1760 * The new RemoveIPC= option can be used to remove IPC objects owned by
1761 the user or group of a service when that service exits.
1763 * The new ProtectKernelModules= option can be used to disable explicit
1764 load and unload operations of kernel modules by a service. In
1765 addition access to /usr/lib/modules is removed if this option is set.
1767 * ProtectSystem= option gained a new value "strict", which causes the
1768 whole file system tree with the exception of /dev, /proc, and /sys,
1769 to be remounted read-only for a service.
1771 * The new ProtectKernelTunables= option can be used to disable
1772 modification of configuration files in /sys and /proc by a service.
1773 Various directories and files are remounted read-only, so access is
1774 restricted even if the file permissions would allow it.
1776 * The new ProtectControlGroups= option can be used to disable write
1777 access by a service to /sys/fs/cgroup.
1779 * Various systemd services have been hardened with
1780 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1781 RestrictAddressFamilies=.
1783 * Support for dynamically creating users for the lifetime of a service
1784 has been added. If DynamicUser=yes is specified, user and group IDs
1785 will be allocated from the range 61184..65519 for the lifetime of the
1786 service. They can be resolved using the new nss-systemd.so NSS
1787 module. The module must be enabled in /etc/nsswitch.conf. Services
1788 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1789 any resources allocated by the service will be cleaned up when the
1790 service exits. They also have ProtectHome=read-only and
1791 ProtectSystem=strict enabled, so they are not able to make any
1792 permanent modifications to the system.
1794 * The nss-systemd module also always resolves root and nobody, making
1795 it possible to have no /etc/passwd or /etc/group files in minimal
1796 container or chroot environments.
1798 * Services may be started with their own user namespace using the new
1799 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1800 under which the service is running are mapped. All other users are
1803 * Support for the cgroup namespace has been added to systemd-nspawn. If
1804 supported by kernel, the container system started by systemd-nspawn
1805 will have its own view of the cgroup hierarchy. This new behaviour
1806 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1808 * The new MemorySwapMax= option can be used to limit the maximum swap
1809 usage under the unified cgroup hierarchy.
1811 * Support for the CPU controller in the unified cgroup hierarchy has
1812 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1813 options. This controller requires out-of-tree patches for the kernel
1814 and the support is provisional.
1816 * Mount and automount units may now be created transiently
1817 (i.e. dynamically at runtime via the bus API, instead of requiring
1818 unit files in the file system).
1820 * systemd-mount is a new tool which may mount file systems – much like
1821 mount(8), optionally pulling in additional dependencies through
1822 transient .mount and .automount units. For example, this tool
1823 automatically runs fsck on a backing block device before mounting,
1824 and allows the automount logic to be used dynamically from the
1825 command line for establishing mount points. This tool is particularly
1826 useful when dealing with removable media, as it will ensure fsck is
1827 run – if necessary – before the first access and that the file system
1828 is quickly unmounted after each access by utilizing the automount
1829 logic. This maximizes the chance that the file system on the
1830 removable media stays in a clean state, and if it isn't in a clean
1831 state is fixed automatically.
1833 * LazyUnmount=yes option for mount units has been added to expose the
1834 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1837 * /efi will be used as the mount point of the EFI boot partition, if
1838 the directory is present, and the mount point was not configured
1839 through other means (e.g. fstab). If /efi directory does not exist,
1840 /boot will be used as before. This makes it easier to automatically
1841 mount the EFI partition on systems where /boot is used for something
1844 * When operating on GPT disk images for containers, systemd-nspawn will
1845 now mount the ESP to /boot or /efi according to the same rules as PID
1846 1 running on a host. This allows tools like "bootctl" to operate
1847 correctly within such containers, in order to make container images
1848 bootable on physical systems.
1850 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
1852 * Two new user session targets have been added to support running
1853 graphical sessions under the systemd --user instance:
1854 graphical-session.target and graphical-session-pre.target. See
1855 systemd.special(7) for a description of how those targets should be
1858 * The vconsole initialization code has been significantly reworked to
1859 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
1860 support unicode keymaps. Font and keymap configuration will now be
1861 copied to all allocated virtual consoles.
1863 * FreeBSD's bhyve virtualization is now detected.
1865 * Information recorded in the journal for core dumps now includes the
1866 contents of /proc/mountinfo and the command line of the process at
1867 the top of the process hierarchy (which is usually the init process
1870 * systemd-journal-gatewayd learned the --directory= option to serve
1871 files from the specified location.
1873 * journalctl --root=… can be used to peruse the journal in the
1874 /var/log/ directories inside of a container tree. This is similar to
1875 the existing --machine= option, but does not require the container to
1878 * The hardware database has been extended to support
1879 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1882 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1883 specify the click rate for mice which include a horizontal wheel with
1884 a click rate that is different than the one for the vertical wheel.
1886 * systemd-run gained a new --wait option that makes service execution
1887 synchronous. (Specifically, the command will not return until the
1888 specified service binary exited.)
1890 * systemctl gained a new --wait option that causes the start command to
1891 wait until the units being started have terminated again.
1893 * A new journal output mode "short-full" has been added which displays
1894 timestamps with abbreviated English day names and adds a timezone
1895 suffix. Those timestamps include more information than the default
1896 "short" output mode, and can be passed directly to journalctl's
1897 --since= and --until= options.
1899 * /etc/resolv.conf will be bind-mounted into containers started by
1900 systemd-nspawn, if possible, so any changes to resolv.conf contents
1901 are automatically propagated to the container.
1903 * The number of instances for socket-activated services originating
1904 from a single IP address can be limited with
1905 MaxConnectionsPerSource=, extending the existing setting of
1908 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1911 * .netdev and .network configuration can now be extended through
1914 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1915 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1916 can be enabled and disabled using the new UDPSegmentationOffload=,
1917 TCPSegmentationOffload=, GenericSegmentationOffload=,
1918 GenericReceiveOffload=, LargeReceiveOffload= options in the
1919 [Link] section of .link files.
1921 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1922 Port VLAN ID can be configured for bridge devices using the new STP=,
1923 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1924 section of .netdev files.
1926 * The route table to which routes received over DHCP or RA should be
1927 added can be configured with the new RouteTable= option in the [DHCP]
1928 and [IPv6AcceptRA] sections of .network files.
1930 * The Address Resolution Protocol can be disabled on links managed by
1931 systemd-networkd using the ARP=no setting in the [Link] section of
1934 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1935 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1936 encode information about the result and exit codes of the current
1937 service runtime cycle.
1939 * systemd-sysctl will now configure kernel parameters in the order
1940 they occur in the configuration files. This matches what sysctl
1941 has been traditionally doing.
1943 * kernel-install "plugins" that are executed to perform various
1944 tasks after a new kernel is added and before an old one is removed
1945 can now return a special value to terminate the procedure and
1946 prevent any later plugins from running.
1948 * Journald's SplitMode=login setting has been deprecated. It has been
1949 removed from documentation, and its use is discouraged. In a future
1950 release it will be completely removed, and made equivalent to current
1951 default of SplitMode=uid.
1953 * Storage=both option setting in /etc/systemd/coredump.conf has been
1954 removed. With fast LZ4 compression storing the core dump twice is not
1957 * The --share-system systemd-nspawn option has been replaced with an
1958 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1959 this functionality is discouraged. In addition the variables
1960 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1961 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1962 individual namespaces.
1964 * "machinectl list" now shows the IP address of running containers in
1965 the output, as well as OS release information.
1967 * "loginctl list" now shows the TTY of each session in the output.
1969 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1970 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1971 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1972 tracking objects in a "recursive" mode, where a single client can be
1973 counted multiple times, if it takes multiple references.
1975 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
1976 sd_bus_get_exit_on_disconnect(). They may be used to make a
1977 process using sd-bus automatically exit if the bus connection is
1980 * Bus clients of the service manager may now "pin" loaded units into
1981 memory, by taking an explicit reference on them. This is useful to
1982 ensure the client can retrieve runtime data about the service even
1983 after the service completed execution. Taking such a reference is
1984 available only for privileged clients and should be helpful to watch
1985 running services in a race-free manner, and in particular collect
1986 information about exit statuses and results.
1988 * The nss-resolve module has been changed to strictly return UNAVAIL
1989 when communication via D-Bus with resolved failed, and NOTFOUND when
1990 a lookup completed but was negative. This means it is now possible to
1991 neatly configure fallbacks using nsswitch.conf result checking
1992 expressions. Taking benefit of this, the new recommended
1993 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
1995 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
1997 * A new setting CtrlAltDelBurstAction= has been added to
1998 /etc/systemd/system.conf which may be used to configure the precise
1999 behaviour if the user on the console presses Ctrl-Alt-Del more often
2000 than 7 times in 2s. Previously this would unconditionally result in
2001 an expedited, immediate reboot. With this new setting the precise
2002 operation may be configured in more detail, and also turned off
2005 * In .netdev files two new settings RemoteChecksumTx= and
2006 RemoteChecksumRx= are now understood that permit configuring the
2007 remote checksumming logic for VXLAN networks.
2009 * The service manager learnt a new "invocation ID" concept for invoked
2010 services. Each runtime cycle of a service will get a new invocation
2011 ID (a 128bit random UUID) assigned that identifies the current
2012 run of the service uniquely and globally. A new invocation ID
2013 is generated each time a service starts up. The journal will store
2014 the invocation ID of a service along with any logged messages, thus
2015 making the invocation ID useful for matching the online runtime of a
2016 service with the offline log data it generated in a safe way without
2017 relying on synchronized timestamps. In many ways this new service
2018 invocation ID concept is similar to the kernel's boot ID concept that
2019 uniquely and globally identifies the runtime of each boot. The
2020 invocation ID of a service is passed to the service itself via an
2021 environment variable ($INVOCATION_ID). A new bus call
2022 GetUnitByInvocationID() has been added that is similar to GetUnit()
2023 but instead of retrieving the bus path for a unit by its name
2024 retrieves it by its invocation ID. The returned path is valid only as
2025 long as the passed invocation ID is current.
2027 * systemd-resolved gained a new "DNSStubListener" setting in
2028 resolved.conf. It either takes a boolean value or the special values
2029 "udp" and "tcp", and configures whether to enable the stub DNS
2030 listener on 127.0.0.53:53.
2032 * IP addresses configured via networkd may now carry additional
2033 configuration settings supported by the kernel. New options include:
2034 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2035 PrefixRoute=, AutoJoin=.
2037 * The PAM configuration fragment file for "user@.service" shipped with
2038 systemd (i.e. the --user instance of systemd) has been stripped to
2039 the minimum necessary to make the system boot. Previously, it
2040 contained Fedora-specific stanzas that did not apply to other
2041 distributions. It is expected that downstream distributions add
2042 additional configuration lines, matching their needs to this file,
2043 using it only as rough template of what systemd itself needs. Note
2044 that this reduced fragment does not even include an invocation of
2045 pam_limits which most distributions probably want to add, even though
2046 systemd itself does not need it. (There's also the new build time
2047 option --with-pamconfdir=no to disable installation of the PAM
2050 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2051 capability is now also dropped from its set (in addition to
2052 CAP_SYS_MKNOD as before).
2054 * In service unit files it is now possible to connect a specific named
2055 file descriptor with stdin/stdout/stdout of an executed service. The
2056 name may be specified in matching .socket units using the
2057 FileDescriptorName= setting.
2059 * A number of journal settings may now be configured on the kernel
2060 command line. Specifically, the following options are now understood:
2061 systemd.journald.max_level_console=,
2062 systemd.journald.max_level_store=,
2063 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2064 systemd.journald.max_level_wall=.
2066 * "systemctl is-enabled --full" will now show by which symlinks a unit
2067 file is enabled in the unit dependency tree.
2069 * Support for VeraCrypt encrypted partitions has been added to the
2070 "cryptsetup" logic and /etc/crypttab.
2072 * systemd-detect-virt gained support for a new --private-users switch
2073 that checks whether the invoking processes are running inside a user
2074 namespace. Similar, a new special value "private-users" for the
2075 existing ConditionVirtualization= setting has been added, permitting
2076 skipping of specific units in user namespace environments.
2078 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2079 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2080 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2081 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2082 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2083 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2084 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2085 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2086 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2087 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2088 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2089 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2090 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2091 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2092 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2093 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2094 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2095 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2096 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2097 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2098 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2099 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2100 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2101 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2102 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2103 Jędrzejewski-Szmek, Zeal Jagannatha
2105 — Santa Fe, 2016-11-03
2109 * In service units the various ExecXYZ= settings have been extended
2110 with an additional special character as first argument of the
2111 assigned value: if the character '+' is used the specified command
2112 line it will be run with full privileges, regardless of User=,
2113 Group=, CapabilityBoundingSet= and similar options. The effect is
2114 similar to the existing PermissionsStartOnly= option, but allows
2115 configuration of this concept for each executed command line
2118 * Services may now alter the service watchdog timeout at runtime by
2119 sending a WATCHDOG_USEC= message via sd_notify().
2121 * MemoryLimit= and related unit settings now optionally take percentage
2122 specifications. The percentage is taken relative to the amount of
2123 physical memory in the system (or in case of containers, the assigned
2124 amount of memory). This allows scaling service resources neatly with
2125 the amount of RAM available on the system. Similarly, systemd-logind's
2126 RuntimeDirectorySize= option now also optionally takes percentage
2129 * In similar fashion TasksMax= takes percentage values now, too. The
2130 value is taken relative to the configured maximum number of processes
2131 on the system. The per-service task maximum has been changed to 15%
2132 using this functionality. (Effectively this is an increase of 512 →
2133 4915 for service units, given the kernel's default pid_max setting.)
2135 * Calendar time specifications in .timer units now understand a ".."
2136 syntax for time ranges. Example: "4..7:10" may now be used for
2137 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2140 * The InaccessableDirectories=, ReadOnlyDirectories= and
2141 ReadWriteDirectories= unit file settings have been renamed to
2142 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2143 applied to all kinds of file nodes, and not just directories, with
2144 the exception of symlinks. Specifically these settings may now be
2145 used on block and character device nodes, UNIX sockets and FIFOS as
2146 well as regular files. The old names of these settings remain
2147 available for compatibility.
2149 * systemd will now log about all service processes it kills forcibly
2150 (using SIGKILL) because they remained after the clean shutdown phase
2151 of the service completed. This should help identifying services that
2152 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2153 systemd-logind's configuration a similar log message is generated for
2154 processes killed at the end of each session due to this setting.
2156 * systemd will now set the $JOURNAL_STREAM environment variable for all
2157 services whose stdout/stderr are connected to the Journal (which
2158 effectively means by default: all services). The variable contains
2159 the device and inode number of the file descriptor used for
2160 stdout/stderr. This may be used by invoked programs to detect whether
2161 their stdout/stderr is connected to the Journal, in which case they
2162 can switch over to direct Journal communication, thus being able to
2163 pass extended, structured metadata along with their log messages. As
2164 one example, this is now used by glib's logging primitives.
2166 * When using systemd's default tmp.mount unit for /tmp, the mount point
2167 will now be established with the "nosuid" and "nodev" options. This
2168 avoids privilege escalation attacks that put traps and exploits into
2169 /tmp. However, this might cause problems if you e. g. put container
2170 images or overlays into /tmp; if you need this, override tmp.mount's
2171 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2174 * systemd now supports the "memory" cgroup controller also on
2177 * The systemd-cgtop tool now optionally takes a control group path as
2178 command line argument. If specified, the control group list shown is
2179 limited to subgroups of that group.
2181 * The SystemCallFilter= unit file setting gained support for
2182 pre-defined, named system call filter sets. For example
2183 SystemCallFilter=@clock is now an effective way to make all clock
2184 changing-related system calls unavailable to a service. A number of
2185 similar pre-defined groups are defined. Writing system call filters
2186 for system services is simplified substantially with this new
2187 concept. Accordingly, all of systemd's own, long-running services now
2188 enable system call filtering based on this, by default.
2190 * A new service setting MemoryDenyWriteExecute= has been added, taking
2191 a boolean value. If turned on, a service may no longer create memory
2192 mappings that are writable and executable at the same time. This
2193 enhances security for services where this is enabled as it becomes
2194 harder to dynamically write and then execute memory in exploited
2195 service processes. This option has been enabled for all of systemd's
2196 own long-running services.
2198 * A new RestrictRealtime= service setting has been added, taking a
2199 boolean argument. If set the service's processes may no longer
2200 acquire realtime scheduling. This improves security as realtime
2201 scheduling may otherwise be used to easily freeze the system.
2203 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2204 value. This may be used for requesting that the system manager inside
2205 of the container reports start-up completion to nspawn which then
2206 propagates this notification further to the service manager
2207 supervising nspawn itself. A related option NotifyReady= in .nspawn
2208 files has been added too. This functionality allows ordering of the
2209 start-up of multiple containers using the usual systemd ordering
2212 * machinectl gained a new command "stop" that is an alias for
2215 * systemd-resolved gained support for contacting DNS servers on
2216 link-local IPv6 addresses.
2218 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2219 its caches. A method call for requesting the same operation has been
2220 added to the bus API too, and is made available via "systemd-resolve
2223 * systemd-resolve gained a new --status switch. If passed a brief
2224 summary of the used DNS configuration with per-interface information
2227 * resolved.conf gained a new Cache= boolean option, defaulting to
2228 on. If turned off local DNS caching is disabled. This comes with a
2229 performance penalty in particular when DNSSEC is enabled. Note that
2230 resolved disables its internal caching implicitly anyway, when the
2231 configured DNS server is on a host-local IP address such as ::1 or
2232 127.0.0.1, thus automatically avoiding double local caching.
2234 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2235 for DNS requests. This improves compatibility with local programs
2236 that do not use the libc NSS or systemd-resolved's bus APIs for name
2237 resolution. This minimal DNS service is only available to local
2238 programs and does not implement the full DNS protocol, but enough to
2239 cover local DNS clients. A new, static resolv.conf file, listing just
2240 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2241 now recommended to make /etc/resolv.conf a symlink to this file in
2242 order to route all DNS lookups to systemd-resolved, regardless if
2243 done via NSS, the bus API or raw DNS packets. Note that this local
2244 DNS service is not as fully featured as the libc NSS or
2245 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2246 used to deliver link-local address information (as this implies
2247 sending a local interface index along), LLMNR/mDNS support via this
2248 interface is severely restricted. It is thus strongly recommended for
2249 all applications to use the libc NSS API or native systemd-resolved
2252 * systemd-networkd's bridge support learned a new setting
2253 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2254 in .network files has been added for configuring VLAN bridging in
2255 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2257 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2258 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2259 now be acquired without relying on DHCPv6. Two new options
2260 UseDomains= and UseDNS= have been added to configure this behaviour.
2262 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2263 renamed IPv6AcceptRA=, without altering its behaviour. The old
2264 setting name remains available for compatibility reasons.
2266 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2267 Key=, InputKey= and OutputKey=.
2269 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2270 interface configuration.
2272 * "systemctl edit" may now be used to create new unit files by
2273 specifying the --force switch.
2275 * sd-event gained a new function sd_event_get_iteration() for
2276 requesting the current iteration counter of the event loop. It starts
2277 at zero and is increased by one with each event loop iteration.
2279 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2280 file. It can be used in lieu of %systemd_requires in packages which
2281 don't use any systemd functionality and are intended to be installed
2282 in minimal containers without systemd present. This macro provides
2283 ordering dependencies to ensure that if the package is installed in
2284 the same rpm transaction as systemd, systemd will be installed before
2285 the scriptlets for the package are executed, allowing unit presets
2288 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2289 been added to simplify packaging of generators.
2291 * The os-release file gained VERSION_CODENAME field for the
2292 distribution nickname (e.g. VERSION_CODENAME=woody).
2294 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2295 can be set to disable parsing of metadata and the creation
2296 of persistent symlinks for that device.
2298 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2299 to make them available to logged-in users has been reverted.
2301 * Much of the common code of the various systemd components is now
2302 built into an internal shared library libsystemd-shared-231.so
2303 (incorporating the systemd version number in the name, to be updated
2304 with future releases) that the components link to. This should
2305 decrease systemd footprint both in memory during runtime and on
2306 disk. Note that the shared library is not for public use, and is
2307 neither API nor ABI stable, but is likely to change with every new
2308 released update. Packagers need to make sure that binaries
2309 linking to libsystemd-shared.so are updated in step with the
2312 * Configuration for "mkosi" is now part of the systemd
2313 repository. mkosi is a tool to easily build legacy-free OS images,
2314 and is available on github: https://github.com/systemd/mkosi. If
2315 "mkosi" is invoked in the build tree a new raw OS image is generated
2316 incorporating the systemd sources currently being worked on and a
2317 clean, fresh distribution installation. The generated OS image may be
2318 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
2319 UEFI PC. This functionality is particularly useful to easily test
2320 local changes made to systemd in a pristine, defined environment. See
2321 doc/HACKING for details.
2323 * configure learned the --with-support-url= option to specify the
2324 distribution's bugtracker.
2326 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2327 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2328 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2329 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2330 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2331 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2332 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2333 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2334 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2335 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2336 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2337 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2338 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2339 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2340 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2341 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
2342 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2343 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
2344 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
2346 — Berlin, 2016-07-25
2350 * DNSSEC is now turned on by default in systemd-resolved (in
2351 "allow-downgrade" mode), but may be turned off during compile time by
2352 passing "--with-default-dnssec=no" to "configure" (and of course,
2353 during runtime with DNSSEC= in resolved.conf). We recommend
2354 downstreams to leave this on at least during development cycles and
2355 report any issues with the DNSSEC logic upstream. We are very
2356 interested in collecting feedback about the DNSSEC validator and its
2357 limitations in the wild. Note however, that DNSSEC support is
2358 probably nothing downstreams should turn on in stable distros just
2359 yet, as it might create incompatibilities with a few DNS servers and
2360 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2361 automatically whenever we detect such incompatible setups, but there
2362 might be systems we do not cover yet. Hence: please help us testing
2363 the DNSSEC code, leave this on where you can, report back, but then
2364 again don't consider turning this on in your stable, LTS or
2365 production release just yet. (Note that you have to enable
2366 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2367 and its DNSSEC mode for host name resolution from local
2370 * systemd-resolve conveniently resolves DANE records with the --tlsa
2371 option and OPENPGPKEY records with the --openpgp option. It also
2372 supports dumping raw DNS record data via the new --raw= switch.
2374 * systemd-logind will now by default terminate user processes that are
2375 part of the user session scope unit (session-XX.scope) when the user
2376 logs out. This behavior is controlled by the KillUserProcesses=
2377 setting in logind.conf, and the previous default of "no" is now
2378 changed to "yes". This means that user sessions will be properly
2379 cleaned up after, but additional steps are necessary to allow
2380 intentionally long-running processes to survive logout.
2382 While the user is logged in at least once, user@.service is running,
2383 and any service that should survive the end of any individual login
2384 session can be started at a user service or scope using systemd-run.
2385 systemd-run(1) man page has been extended with an example which shows
2386 how to run screen in a scope unit underneath user@.service. The same
2387 command works for tmux.
2389 After the user logs out of all sessions, user@.service will be
2390 terminated too, by default, unless the user has "lingering" enabled.
2391 To effectively allow users to run long-term tasks even if they are
2392 logged out, lingering must be enabled for them. See loginctl(1) for
2393 details. The default polkit policy was modified to allow users to
2394 set lingering for themselves without authentication.
2396 Previous defaults can be restored at compile time by the
2397 --without-kill-user-processes option to "configure".
2399 * systemd-logind gained new configuration settings SessionsMax= and
2400 InhibitorsMax=, both with a default of 8192. It will not register new
2401 user sessions or inhibitors above this limit.
2403 * systemd-logind will now reload configuration on SIGHUP.
2405 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
2406 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
2407 enable. Also, support for the "io" cgroup controller in the unified
2408 hierarchy has been added, so that the "memory", "pids" and "io" are
2409 now the controllers that are supported on the unified hierarchy.
2411 WARNING: it is not possible to use previous systemd versions with
2412 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2413 is necessary to also update systemd in the initramfs if using the
2414 unified hierarchy. An updated SELinux policy is also required.
2416 * LLDP support has been extended, and both passive (receive-only) and
2417 active (sender) modes are supported. Passive mode ("routers-only") is
2418 enabled by default in systemd-networkd. Active LLDP mode is enabled
2419 by default for containers on the internal network. The "networkctl
2420 lldp" command may be used to list information gathered. "networkctl
2421 status" will also show basic LLDP information on connected peers now.
2423 * The IAID and DUID unique identifier sent in DHCP requests may now be
2424 configured for the system and each .network file managed by
2425 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2427 * systemd-networkd gained support for configuring proxy ARP support for
2428 each interface, via the ProxyArp= setting in .network files. It also
2429 gained support for configuring the multicast querier feature of
2430 bridge devices, via the new MulticastQuerier= setting in .netdev
2431 files. Similarly, snooping on the IGMP traffic can be controlled
2432 via the new setting MulticastSnooping=.
2434 A new setting PreferredLifetime= has been added for addresses
2435 configured in .network file to configure the lifetime intended for an
2438 The systemd-networkd DHCP server gained the option EmitRouter=, which
2439 defaults to yes, to configure whether the DHCP Option 3 (Router)
2442 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
2443 systemd-socket-activate and installed into /usr/bin. It is now fully
2446 * systemd-journald now uses separate threads to flush changes to disk
2447 when closing journal files, thus reducing impact of slow disk I/O on
2448 logging performance.
2450 * The sd-journal API gained two new calls
2451 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2452 can be used to open journal files using file descriptors instead of
2453 file or directory paths. sd_journal_open_container() has been
2454 deprecated, sd_journal_open_directory_fd() should be used instead
2455 with the flag SD_JOURNAL_OS_ROOT.
2457 * journalctl learned a new output mode "-o short-unix" that outputs log
2458 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2459 UTC). It also gained support for a new --no-hostname setting to
2460 suppress the hostname column in the family of "short" output modes.
2462 * systemd-ask-password now optionally skips printing of the password to
2463 stdout with --no-output which can be useful in scripts.
2465 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2466 (devices tagged with ID_MAKER_TOOL) are now tagged with
2467 "uaccess" and are available to logged in users.
2469 * The DeviceAllow= unit setting now supports specifiers (with "%").
2471 * "systemctl show" gained a new --value switch, which allows print a
2472 only the contents of a specific unit property, without also printing
2473 the property's name. Similar support was added to "show*" verbs
2474 of loginctl and machinectl that output "key=value" lists.
2476 * A new unit type "generated" was added for files dynamically generated
2477 by generator tools. Similarly, a new unit type "transient" is used
2478 for unit files created using the runtime API. "systemctl enable" will
2479 refuse to operate on such files.
2481 * A new command "systemctl revert" has been added that may be used to
2482 revert to the vendor version of a unit file, in case local changes
2483 have been made by adding drop-ins or overriding the unit file.
2485 * "machinectl clean" gained a new verb to automatically remove all or
2486 just hidden container images.
2488 * systemd-tmpfiles gained support for a new line type "e" for emptying
2489 directories, if they exist, without creating them if they don't.
2491 * systemd-nspawn gained support for automatically patching the UID/GIDs
2492 of the owners and the ACLs of all files and directories in a
2493 container tree to match the UID/GID user namespacing range selected
2494 for the container invocation. This mode is enabled via the new
2495 --private-users-chown switch. It also gained support for
2496 automatically choosing a free, previously unused UID/GID range when
2497 starting a container, via the new --private-users=pick setting (which
2498 implies --private-users-chown). Together, these options for the first
2499 time make user namespacing for nspawn containers fully automatic and
2500 thus deployable. The systemd-nspawn@.service template unit file has
2501 been changed to use this functionality by default.
2503 * systemd-nspawn gained a new --network-zone= switch, that allows
2504 creating ad-hoc virtual Ethernet links between multiple containers,
2505 that only exist as long as at least one container referencing them is
2506 running. This allows easy connecting of multiple containers with a
2507 common link that implements an Ethernet broadcast domain. Each of
2508 these network "zones" may be named relatively freely by the user, and
2509 may be referenced by any number of containers, but each container may
2510 only reference one of these "zones". On the lower level, this is
2511 implemented by an automatically managed bridge network interface for
2512 each zone, that is created when the first container referencing its
2513 zone is created and removed when the last one referencing its zone
2516 * The default start timeout may now be configured on the kernel command
2517 line via systemd.default_timeout_start_sec=. It was already
2518 configurable via the DefaultTimeoutStartSec= option in
2519 /etc/systemd/system.conf.
2521 * Socket units gained a new TriggerLimitIntervalSec= and
2522 TriggerLimitBurst= setting to configure a limit on the activation
2523 rate of the socket unit.
2525 * The LimitNICE= setting now optionally takes normal UNIX nice values
2526 in addition to the raw integer limit value. If the specified
2527 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2528 value is understood as UNIX nice value. If not prefixed like this it
2529 is understood as raw RLIMIT_NICE limit.
2531 * Note that the effect of the PrivateDevices= unit file setting changed
2532 slightly with this release: the per-device /dev file system will be
2533 mounted read-only from this version on, and will have "noexec"
2534 set. This (minor) change of behavior might cause some (exceptional)
2535 legacy software to break, when PrivateDevices=yes is set for its
2536 service. Please leave PrivateDevices= off if you run into problems
2539 * systemd-bootchart has been split out to a separate repository:
2540 https://github.com/systemd/systemd-bootchart
2542 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2543 merged into the kernel in its current form.
2545 * The compatibility libraries libsystemd-daemon.so,
2546 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2547 which have been deprecated since systemd-209 have been removed along
2548 with the corresponding pkg-config files. All symbols provided by
2549 those libraries are provided by libsystemd.so.
2551 * The Capabilities= unit file setting has been removed (it is ignored
2552 for backwards compatibility). AmbientCapabilities= and
2553 CapabilityBoundingSet= should be used instead.
2555 * A new special target has been added, initrd-root-device.target,
2556 which creates a synchronization point for dependencies of the root
2557 device in early userspace. Initramfs builders must ensure that this
2558 target is now included in early userspace.
2560 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2561 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2562 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2563 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2564 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2565 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2566 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2567 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
2568 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2569 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
2570 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2571 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2572 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
2573 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2574 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2575 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
2576 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2577 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2578 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2579 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2580 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2581 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
2582 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2583 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2584 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2587 — Fairfax, 2016-05-21
2591 * The systemd-resolved DNS resolver service has gained a substantial
2592 set of new features, most prominently it may now act as a DNSSEC
2593 validating stub resolver. DNSSEC mode is currently turned off by
2594 default, but is expected to be turned on by default in one of the
2595 next releases. For now, we invite everybody to test the DNSSEC logic
2596 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2597 service also gained a full set of D-Bus interfaces, including calls
2598 to configure DNS and DNSSEC settings per link (for use by external
2599 network management software). systemd-resolved and systemd-networkd
2600 now distinguish between "search" and "routing" domains. The former
2601 are used to qualify single-label names, the latter are used purely
2602 for routing lookups within certain domains to specific links.
2603 resolved now also synthesizes RRs for all entries from /etc/hosts.
2605 * The systemd-resolve tool (which is a client utility for
2606 systemd-resolved) has been improved considerably and is now fully
2607 supported and documented. Hence it has moved from /usr/lib/systemd to
2610 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2613 * The coredump collection logic has been reworked: when a coredump is
2614 collected it is now written to disk, compressed and processed
2615 (including stacktrace extraction) from a new instantiated service
2616 systemd-coredump@.service, instead of directly from the
2617 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2618 processing large coredumps can take up a substantial amount of
2619 resources and time, and this previously happened entirely outside of
2620 systemd's service supervision. With the new logic the core_pattern
2621 hook only does minimal metadata collection before passing off control
2622 to the new instantiated service, which is configured with a time
2623 limit, a nice level and other settings to minimize negative impact on
2624 the rest of the system. Also note that the new logic will honour the
2625 RLIMIT_CORE setting of the crashed process, which now allows users
2626 and processes to turn off coredumping for their processes by setting
2629 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2630 and all forked processes by default. Previously, PID 1 would leave
2631 the setting at "0" for all processes, as set by the kernel. Note that
2632 the resource limit traditionally has no effect on the generated
2633 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2634 logic is used. Since the limit is now honoured (see above) its
2635 default has been changed so that the coredumping logic is enabled by
2636 default for all processes, while allowing specific opt-out.
2638 * When the stacktrace is extracted from processes of system users, this
2639 is now done as "systemd-coredump" user, in order to sandbox this
2640 potentially security sensitive parsing operation. (Note that when
2641 processing coredumps of normal users this is done under the user ID
2642 of process that crashed, as before.) Packagers should take notice
2643 that it is now necessary to create the "systemd-coredump" system user
2644 and group at package installation time.
2646 * The systemd-activate socket activation testing tool gained support
2647 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2648 and --seqpacket switches. It also has been extended to support both
2649 new-style and inetd-style file descriptor passing. Use the new
2650 --inetd switch to request inetd-style file descriptor passing.
2652 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2653 variable, which takes a boolean value. If set to false, ANSI color
2654 output is disabled in the tools even when run on a terminal that
2657 * The VXLAN support in networkd now supports two new settings
2658 DestinationPort= and PortRange=.
2660 * A new systemd.machine_id= kernel command line switch has been added,
2661 that may be used to set the machine ID in /etc/machine-id if it is
2662 not initialized yet. This command line option has no effect if the
2663 file is already initialized.
2665 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2666 specified command line as PID 2 rather than PID 1 in the
2667 container. In this mode PID 1 is a minimal stub init process that
2668 implements the special POSIX and Linux semantics of PID 1 regarding
2669 signal and child process management. Note that this stub init process
2670 is implemented in nspawn itself and requires no support from the
2671 container image. This new logic is useful to support running
2672 arbitrary commands in the container, as normal processes are
2673 generally not prepared to run as PID 1.
2675 * systemd-nspawn gained a new --chdir= switch for setting the current
2676 working directory for the process started in the container.
2678 * "journalctl /dev/sda" will now output all kernel log messages for
2679 specified device from the current boot, in addition to all devices
2680 that are parents of it. This should make log output about devices
2681 pretty useful, as long as kernel drivers attach enough metadata to
2682 the log messages. (The usual SATA drivers do.)
2684 * The sd-journal API gained two new calls
2685 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2686 that report whether log data from /run or /var has been found.
2688 * journalctl gained a new switch "--fields" that prints all journal
2689 record field names currently in use in the journal. This is backed
2690 by two new sd-journal API calls sd_journal_enumerate_fields() and
2691 sd_journal_restart_fields().
2693 * Most configurable timeouts in systemd now expect an argument of
2694 "infinity" to turn them off, instead of "0" as before. The semantics
2695 from now on is that a timeout of "0" means "now", and "infinity"
2696 means "never". To maintain backwards compatibility, "0" continues to
2697 turn off previously existing timeout settings.
2699 * "systemctl reload-or-try-restart" has been renamed to "systemctl
2700 try-reload-or-restart" to clarify what it actually does: the "try"
2701 logic applies to both reloading and restarting, not just restarting.
2702 The old name continues to be accepted for compatibility.
2704 * On boot-up, when PID 1 detects that the system clock is behind the
2705 release date of the systemd version in use, the clock is now set
2706 to the latter. Previously, this was already done in timesyncd, in order
2707 to avoid running with clocks set to the various clock epochs such as
2708 1902, 1938 or 1970. With this change the logic is now done in PID 1
2709 in addition to timesyncd during early boot-up, so that it is enforced
2710 before the first process is spawned by systemd. Note that the logic
2711 in timesyncd remains, as it is more comprehensive and ensures
2712 clock monotonicity by maintaining a persistent timestamp file in
2713 /var. Since /var is generally not available in earliest boot or the
2714 initrd, this part of the logic remains in timesyncd, and is not done
2717 * Support for tweaking details in net_cls.class_id through the
2718 NetClass= configuration directive has been removed, as the kernel
2719 people have decided to deprecate that controller in cgroup v2.
2720 Userspace tools such as nftables are moving over to setting rules
2721 that are specific to the full cgroup path of a task, which obsoletes
2722 these controllers anyway. The NetClass= directive is kept around for
2723 legacy compatibility reasons. For a more in-depth description of the
2724 kernel change, please refer to the respective upstream commit:
2726 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2728 * A new service setting RuntimeMaxSec= has been added that may be used
2729 to specify a maximum runtime for a service. If the timeout is hit, the
2730 service is terminated and put into a failure state.
2732 * A new service setting AmbientCapabilities= has been added. It allows
2733 configuration of additional Linux process capabilities that are
2734 passed to the activated processes. This is only available on very
2737 * The process resource limit settings in service units may now be used
2738 to configure hard and soft limits individually.
2740 * The various libsystemd APIs such as sd-bus or sd-event now publicly
2741 expose support for gcc's __attribute__((cleanup())) C extension.
2742 Specifically, for many object destructor functions alternative
2743 versions have been added that have names suffixed with "p" and take a
2744 pointer to a pointer to the object to destroy, instead of just a
2745 pointer to the object itself. This is useful because these destructor
2746 functions may be used directly as parameters to the cleanup
2747 construct. Internally, systemd has been a heavy user of this GCC
2748 extension for a long time, and with this change similar support is
2749 now available to consumers of the library outside of systemd. Note
2750 that by using this extension in your sources compatibility with old
2751 and strictly ANSI compatible C compilers is lost. However, all gcc or
2752 LLVM versions of recent years support this extension.
2754 * Timer units gained support for a new setting RandomizedDelaySec= that
2755 allows configuring some additional randomized delay to the configured
2756 time. This is useful to spread out timer events to avoid load peaks in
2757 clusters or larger setups.
2759 * Calendar time specifications now support sub-second accuracy.
2761 * Socket units now support listening on SCTP and UDP-lite protocol
2764 * The sd-event API now comes with a full set of man pages.
2766 * Older versions of systemd contained experimental support for
2767 compressing journal files and coredumps with the LZ4 compressor that
2768 was not compatible with the lz4 binary (due to API limitations of the
2769 lz4 library). This support has been removed; only support for files
2770 compatible with the lz4 binary remains. This LZ4 logic is now
2771 officially supported and no longer considered experimental.
2773 * The dkr image import logic has been removed again from importd. dkr's
2774 micro-services focus doesn't fit into the machine image focus of
2775 importd, and quickly got out of date with the upstream dkr API.
2777 * Creation of the /run/lock/lockdev/ directory was dropped from
2778 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
2779 been available for many years. If you still need this, you need to
2780 create your own tmpfiles.d config file with:
2782 d /run/lock/lockdev 0775 root lock -
2784 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2785 and RebootArgument= have been moved from the [Service] section of
2786 unit files to [Unit], and they are now supported on all unit types,
2787 not just service units. Of course, elogind will continue to
2788 understand these settings also at the old location, in order to
2789 maintain compatibility.
2791 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2792 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2793 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2794 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2795 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2796 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2797 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2798 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2799 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2800 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2801 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2802 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2803 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2804 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2805 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2806 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2807 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2808 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2809 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2811 — Berlin, 2016-02-11
2815 * A number of properties previously only settable in unit
2816 files are now also available as properties to set when
2817 creating transient units programmatically via the bus, as it
2818 is exposed with systemd-run's --property=
2819 setting. Specifically, these are: SyslogIdentifier=,
2820 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2821 EnvironmentFile=, ReadWriteDirectories=,
2822 ReadOnlyDirectories=, InaccessibleDirectories=,
2823 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2825 * When creating transient services via the bus API it is now
2826 possible to pass in a set of file descriptors to use as
2827 STDIN/STDOUT/STDERR for the invoked process.
2829 * Slice units may now be created transiently via the bus APIs,
2830 similar to the way service and scope units may already be
2831 created transiently.
2833 * Wherever systemd expects a calendar timestamp specification
2834 (like in journalctl's --since= and --until= switches) UTC
2835 timestamps are now supported. Timestamps suffixed with "UTC"
2836 are now considered to be in Universal Time Coordinated
2837 instead of the local timezone. Also, timestamps may now
2838 optionally be specified with sub-second accuracy. Both of
2839 these additions also apply to recurring calendar event
2840 specification, such as OnCalendar= in timer units.
2842 * journalctl gained a new "--sync" switch that asks the
2843 journal daemon to write all so far unwritten log messages to
2844 disk and sync the files, before returning.
2846 * systemd-tmpfiles learned two new line types "q" and "Q" that
2847 operate like "v", but also set up a basic btrfs quota
2848 hierarchy when used on a btrfs file system with quota
2851 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2852 instead of a subvolume (even on a btrfs file system) if the
2853 root directory is a plain directory, and not a
2854 subvolume. This should simplify things with certain chroot()
2855 environments which are not aware of the concept of btrfs
2858 * systemd-detect-virt gained a new --chroot switch to detect
2859 whether execution takes place in a chroot() environment.
2861 * CPUAffinity= now takes CPU index ranges in addition to
2864 * The various memory-related resource limit settings (such as
2865 LimitAS=) now understand the usual K, M, G, ... suffixes to
2866 the base of 1024 (IEC). Similar, the time-related resource
2867 limit settings understand the usual min, h, day, ...
2870 * There's a new system.conf setting DefaultTasksMax= to
2871 control the default TasksMax= setting for services and
2872 scopes running on the system. (TasksMax= is the primary
2873 setting that exposes the "pids" cgroup controller on systemd
2874 and was introduced in the previous systemd release.) The
2875 setting now defaults to 512, which means services that are
2876 not explicitly configured otherwise will only be able to
2877 create 512 processes or threads at maximum, from this
2878 version on. Note that this means that thread- or
2879 process-heavy services might need to be reconfigured to set
2880 TasksMax= to a higher value. It is sufficient to set
2881 TasksMax= in these specific unit files to a higher value, or
2882 even "infinity". Similar, there's now a logind.conf setting
2883 UserTasksMax= that defaults to 4096 and limits the total
2884 number of processes or tasks each user may own
2885 concurrently. nspawn containers also have the TasksMax=
2886 value set by default now, to 8192. Note that all of this
2887 only has an effect if the "pids" cgroup controller is
2888 enabled in the kernel. The general benefit of these changes
2889 should be a more robust and safer system, that provides a
2890 certain amount of per-service fork() bomb protection.
2892 * systemd-nspawn gained the new --network-veth-extra= switch
2893 to define additional and arbitrarily-named virtual Ethernet
2894 links between the host and the container.
2896 * A new service execution setting PassEnvironment= has been
2897 added that allows importing select environment variables
2898 from PID1's environment block into the environment block of
2901 * Timer units gained support for a new RemainAfterElapse=
2902 setting which takes a boolean argument. It defaults to on,
2903 exposing behaviour unchanged to previous releases. If set to
2904 off, timer units are unloaded after they elapsed if they
2905 cannot elapse again. This is particularly useful for
2906 transient timer units, which shall not stay around longer
2907 than until they first elapse.
2909 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
2910 default now (the kernel default is 16). This is beneficial
2911 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
2912 allows substantially larger numbers of queued
2913 datagrams. This should increase the capability of systemd to
2914 parallelize boot-up, as logging and sd_notify() are unlikely
2915 to stall execution anymore. If you need to change the value
2916 from the new defaults, use the usual sysctl.d/ snippets.
2918 * The compression framing format used by the journal or
2919 coredump processing has changed to be in line with what the
2920 official LZ4 tools generate. LZ4 compression support in
2921 systemd was considered unsupported previously, as the format
2922 was not compatible with the normal tools. With this release
2923 this has changed now, and it is hence safe for downstream
2924 distributions to turn it on. While not compressing as well
2925 as the XZ, LZ4 is substantially faster, which makes
2926 it a good default choice for the compression logic in the
2927 journal and in coredump handling.
2929 * Any reference to /etc/mtab has been dropped from
2930 systemd. The file has been obsolete since a while, but
2931 systemd refused to work on systems where it was incorrectly
2932 set up (it should be a symlink or non-existent). Please make
2933 sure to update to util-linux 2.27.1 or newer in conjunction
2934 with this systemd release, which also drops any reference to
2935 /etc/mtab. If you maintain a distribution make sure that no
2936 software you package still references it, as this is a
2937 likely source of bugs. There's also a glibc bug pending,
2938 asking for removal of any reference to this obsolete file:
2940 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
2942 Note that only util-linux versions built with
2943 --enable-libmount-force-mountinfo are supported.
2945 * Support for the ".snapshot" unit type has been removed. This
2946 feature turned out to be little useful and little used, and
2947 has now been removed from the core and from systemctl.
2949 * The dependency types RequiresOverridable= and
2950 RequisiteOverridable= have been removed from systemd. They
2951 have been used only very sparingly to our knowledge and
2952 other options that provide a similar effect (such as
2953 systemctl --mode=ignore-dependencies) are much more useful
2954 and commonly used. Moreover, they were only half-way
2955 implemented as the option to control behaviour regarding
2956 these dependencies was never added to systemctl. By removing
2957 these dependency types the execution engine becomes a bit
2958 simpler. Unit files that use these dependencies should be
2959 changed to use the non-Overridable dependency types
2960 instead. In fact, when parsing unit files with these
2961 options, that's what systemd will automatically convert them
2962 too, but it will also warn, asking users to fix the unit
2963 files accordingly. Removal of these dependency types should
2964 only affect a negligible number of unit files in the wild.
2966 * Behaviour of networkd's IPForward= option changed
2967 (again). It will no longer maintain a per-interface setting,
2968 but propagate one way from interfaces where this is enabled
2969 to the global kernel setting. The global setting will be
2970 enabled when requested by a network that is set up, but
2971 never be disabled again. This change was made to make sure
2972 IPv4 and IPv6 behaviour regarding packet forwarding is
2973 similar (as the Linux IPv6 stack does not support
2974 per-interface control of this setting) and to minimize
2977 * In unit files the behaviour of %u, %U, %h, %s has
2978 changed. These specifiers will now unconditionally resolve
2979 to the various user database fields of the user that the
2980 systemd instance is running as, instead of the user
2981 configured in the specific unit via User=. Note that this
2982 effectively doesn't change much, as resolving of these
2983 specifiers was already turned off in the --system instance
2984 of systemd, as we cannot do NSS lookups from PID 1. In the
2985 --user instance of systemd these specifiers where correctly
2986 resolved, but hardly made any sense, since the user instance
2987 lacks privileges to do user switches anyway, and User= is
2988 hence useless. Moreover, even in the --user instance of
2989 systemd behaviour was awkward as it would only take settings
2990 from User= assignment placed before the specifier into
2991 account. In order to unify and simplify the logic around
2992 this the specifiers will now always resolve to the
2993 credentials of the user invoking the manager (which in case
2994 of PID 1 is the root user).
2996 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
2997 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
2998 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
2999 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3000 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3001 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3002 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3003 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3004 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3005 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3006 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3007 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3008 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3009 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3012 — Berlin, 2015-11-18
3016 * systemd now depends on util-linux v2.27. More specifically,
3017 the newly added mount monitor feature in libmount now
3018 replaces systemd's former own implementation.
3020 * libmount mandates /etc/mtab not to be regular file, and
3021 systemd now enforces this condition at early boot.
3022 /etc/mtab has been deprecated and warned about for a very
3023 long time, so systems running systemd should already have
3024 stopped having this file around as anything else than a
3025 symlink to /proc/self/mounts.
3027 * Support for the "pids" cgroup controller has been added. It
3028 allows accounting the number of tasks in a cgroup and
3029 enforcing limits on it. This adds two new setting
3030 TasksAccounting= and TasksMax= to each unit, as well as a
3031 global option DefaultTasksAccounting=.
3033 * Support for the "net_cls" cgroup controller has been added.
3034 It allows assigning a net class ID to each task in the
3035 cgroup, which can then be used in firewall rules and traffic
3036 shaping configurations. Note that the kernel netfilter net
3037 class code does not currently work reliably for ingress
3038 packets on unestablished sockets.
3040 This adds a new config directive called NetClass= to CGroup
3041 enabled units. Allowed values are positive numbers for fixed
3042 assignments and "auto" for picking a free value
3045 * 'systemctl is-system-running' now returns 'offline' if the
3046 system is not booted with systemd. This command can now be
3047 used as a substitute for 'systemd-notify --booted'.
3049 * Watchdog timeouts have been increased to 3 minutes for all
3050 in-tree service files. Apparently, disk IO issues are more
3051 frequent than we hoped, and user reported >1 minute waiting
3054 * 'machine-id-commit' functionality has been merged into
3055 'machine-id-setup --commit'. The separate binary has been
3058 * The WorkingDirectory= directive in unit files may now be set
3059 to the special value '~'. In this case, the working
3060 directory is set to the home directory of the user
3061 configured in User=.
3063 * "machinectl shell" will now open the shell in the home
3064 directory of the selected user by default.
3066 * The CrashChVT= configuration file setting is renamed to
3067 CrashChangeVT=, following our usual logic of not
3068 abbreviating unnecessarily. The old directive is still
3069 supported for compat reasons. Also, this directive now takes
3070 an integer value between 1 and 63, or a boolean value. The
3071 formerly supported '-1' value for disabling stays around for
3074 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3075 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3076 RootDirectory= properties can now be set for transient
3079 * The systemd-analyze tool gained a new "set-log-target" verb
3080 to change the logging target the system manager logs to
3081 dynamically during runtime. This is similar to how
3082 "systemd-analyze set-log-level" already changes the log
3085 * In nspawn /sys is now mounted as tmpfs, with only a selected
3086 set of subdirectories mounted in from the real sysfs. This
3087 enhances security slightly, and is useful for ensuring user
3088 namespaces work correctly.
3090 * Support for USB FunctionFS activation has been added. This
3091 allows implementation of USB gadget services that are
3092 activated as soon as they are requested, so that they don't
3093 have to run continuously, similar to classic socket
3096 * The "systemctl exit" command now optionally takes an
3097 additional parameter that sets the exit code to return from
3098 the systemd manager when exiting. This is only relevant when
3099 running the systemd user instance, or when running the
3100 system instance in a container.
3102 * sd-bus gained the new API calls sd_bus_path_encode_many()
3103 and sd_bus_path_decode_many() that allow easy encoding and
3104 decoding of multiple identifier strings inside a D-Bus
3105 object path. Another new call sd_bus_default_flush_close()
3106 has been added to flush and close per-thread default
3109 * systemd-cgtop gained support for a -M/--machine= switch to
3110 show the control groups within a certain container only.
3112 * "systemctl kill" gained support for an optional --fail
3113 switch. If specified the requested operation will fail of no
3114 processes have been killed, because the unit had no
3115 processes attached, or similar.
3117 * A new systemd.crash_reboot=1 kernel command line option has
3118 been added that triggers a reboot after crashing. This can
3119 also be set through CrashReboot= in systemd.conf.
3121 * The RuntimeDirectory= setting now understands unit
3122 specifiers like %i or %f.
3124 * A new (still internal) library API sd-ipv4acd has been added,
3125 that implements address conflict detection for IPv4. It's
3126 based on code from sd-ipv4ll, and will be useful for
3127 detecting DHCP address conflicts.
3129 * File descriptors passed during socket activation may now be
3130 named. A new API sd_listen_fds_with_names() is added to
3131 access the names. The default names may be overridden,
3132 either in the .socket file using the FileDescriptorName=
3133 parameter, or by passing FDNAME= when storing the file
3134 descriptors using sd_notify().
3136 * systemd-networkd gained support for:
3138 - Setting the IPv6 Router Advertisement settings via
3139 IPv6AcceptRouterAdvertisements= in .network files.
3141 - Configuring the HelloTimeSec=, MaxAgeSec= and
3142 ForwardDelaySec= bridge parameters in .netdev files.
3144 - Configuring PreferredSource= for static routes in
3147 * The "ask-password" framework used to query for LUKS harddisk
3148 passwords or SSL passwords during boot gained support for
3149 caching passwords in the kernel keyring, if it is
3150 available. This makes sure that the user only has to type in
3151 a passphrase once if there are multiple objects to unlock
3152 with the same one. Previously, such password caching was
3153 available only when Plymouth was used; this moves the
3154 caching logic into the systemd codebase itself. The
3155 "systemd-ask-password" utility gained a new --keyname=
3156 switch to control which kernel keyring key to use for
3157 caching a password in. This functionality is also useful for
3158 enabling display managers such as gdm to automatically
3159 unlock the user's GNOME keyring if its passphrase, the
3160 user's password and the harddisk password are the same, if
3161 gdm-autologin is used.
3163 * When downloading tar or raw images using "machinectl
3164 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3165 file is now also downloaded, if it is available and stored
3166 next to the image file.
3168 * Units of type ".socket" gained a new boolean setting
3169 Writable= which is only useful in conjunction with
3170 ListenSpecial=. If true, enables opening the specified
3171 special file in O_RDWR mode rather than O_RDONLY mode.
3173 * systemd-rfkill has been reworked to become a singleton
3174 service that is activated through /dev/rfkill on each rfkill
3175 state change and saves the settings to disk. This way,
3176 systemd-rfkill is now compatible with devices that exist
3177 only intermittendly, and even restores state if the previous
3178 system shutdown was abrupt rather than clean.
3180 * The journal daemon gained support for vacuuming old journal
3181 files controlled by the number of files that shall remain,
3182 in addition to the already existing control by size and by
3183 date. This is useful as journal interleaving performance
3184 degrades with too many separate journal files, and allows
3185 putting an effective limit on them. The new setting defaults
3186 to 100, but this may be changed by setting SystemMaxFiles=
3187 and RuntimeMaxFiles= in journald.conf. Also, the
3188 "journalctl" tool gained the new --vacuum-files= switch to
3189 manually vacuum journal files to leave only the specified
3190 number of files in place.
3192 * udev will now create /dev/disk/by-path links for ATA devices
3193 on kernels where that is supported.
3195 * Galician, Serbian, Turkish and Korean translations were added.
3197 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3198 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3199 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3200 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3201 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3202 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3203 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3204 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3205 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3206 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3207 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3208 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3209 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3210 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3211 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3212 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3213 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3214 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3216 — Berlin, 2015-10-07
3220 * The DHCP implementation of systemd-networkd gained a set of
3223 - The DHCP server now supports emitting DNS and NTP
3224 information. It may be enabled and configured via
3225 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3226 and NTP information is enabled, but no servers are
3227 configured, the corresponding uplink information (if there
3228 is any) is propagated.
3230 - Server and client now support transmission and reception
3231 of timezone information. It can be configured via the
3232 newly introduced network options UseTimezone=,
3233 EmitTimezone=, and Timezone=. Transmission of timezone
3234 information is enabled between host and containers by
3235 default now: the container will change its local timezone
3236 to what the host has set.
3238 - Lease timeouts can now be configured via
3239 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3241 - The DHCP server improved on the stability of
3242 leases. Clients are more likely to get the same lease
3243 information back, even if the server loses state.
3245 - The DHCP server supports two new configuration options to
3246 control the lease address pool metrics, PoolOffset= and
3249 * The encapsulation limit of tunnels in systemd-networkd may
3250 now be configured via 'EncapsulationLimit='. It allows
3251 modifying the maximum additional levels of encapsulation
3252 that are permitted to be prepended to a packet.
3254 * systemd now supports the concept of user buses replacing
3255 session buses, if used with dbus-1.10 (and enabled via dbus
3256 --enable-user-session). It previously only supported this on
3257 kdbus-enabled systems, and this release expands this to
3258 'dbus-daemon' systems.
3260 * systemd-networkd now supports predictable interface names
3263 * systemd now optionally supports the new Linux kernel
3264 "unified" control group hierarchy. If enabled via the kernel
3265 command-line option 'systemd.unified_cgroup_hierarchy=1',
3266 systemd will try to mount the unified cgroup hierarchy
3267 directly on /sys/fs/cgroup. If not enabled, or not
3268 available, systemd will fall back to the legacy cgroup
3269 hierarchy setup, as before. Host system and containers can
3270 mix and match legacy and unified hierarchies as they
3271 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
3272 environment variable to individually select the hierarchy to
3273 use for executed containers. By default, nspawn will use the
3274 unified hierarchy for the containers if the host uses the
3275 unified hierarchy, and the legacy hierarchy otherwise.
3276 Please note that at this point the unified hierarchy is an
3277 experimental kernel feature and is likely to change in one
3278 of the next kernel releases. Therefore, it should not be
3279 enabled by default in downstream distributions yet. The
3280 minimum required kernel version for the unified hierarchy to
3281 work is 4.2. Note that when the unified hierarchy is used
3282 for the first time delegated access to controllers is
3283 safe. Because of this systemd-nspawn containers will get
3284 access to controllers now, as will systemd user
3285 sessions. This means containers and user sessions may now
3286 manage their own resources, partitioning up what the system
3289 * A new special scope unit "init.scope" has been introduced
3290 that encapsulates PID 1 of the system. It may be used to
3291 determine resource usage and enforce resource limits on PID
3292 1 itself. PID 1 hence moved out of the root of the control
3295 * The cgtop tool gained support for filtering out kernel
3296 threads when counting tasks in a control group. Also, the
3297 count of processes is now recursively summed up by
3298 default. Two options -k and --recursive= have been added to
3299 revert to old behaviour. The tool has also been updated to
3300 work correctly in containers now.
3302 * systemd-nspawn's --bind= and --bind-ro= options have been
3303 extended to allow creation of non-recursive bind mounts.
3305 * libsystemd gained two new calls sd_pid_get_cgroup() and
3306 sd_peer_get_cgroup() which return the control group path of
3307 a process or peer of a connected AF_UNIX socket. This
3308 function call is particularly useful when implementing
3309 delegated subtrees support in the control group hierarchy.
3311 * The "sd-event" event loop API of libsystemd now supports
3312 correct dequeuing of real-time signals, without losing
3315 * When systemd requests a PolicyKit decision when managing
3316 units it will now add additional fields to the request,
3317 including unit name and desired operation. This enables more
3318 powerful PolicyKit policies, that make decisions depending
3319 on these parameters.
3321 * nspawn learnt support for .nspawn settings files, that may
3322 accompany the image files or directories of containers, and
3323 may contain additional settings for the container. This is
3324 an alternative to configuring container parameters via the
3325 nspawn command line.
3327 Contributions from: Cristian Rodríguez, Daniel Mack, David
3328 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3329 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3330 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3331 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3332 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3333 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
3334 Andersen, Tom Gundersen, Torstein Husebø
3336 — Berlin, 2015-09-08
3340 * machinectl gained a new verb 'shell' which opens a fresh
3341 shell on the target container or the host. It is similar to
3342 the existing 'login' command of machinectl, but spawns the
3343 shell directly without prompting for username or
3344 password. The pseudo machine '.host' now refers to the local
3345 host and is used by default. Hence, 'machinectl shell' can
3346 be used as replacement for 'su -' which spawns a session as
3347 a fresh systemd unit in a way that is fully isolated from
3348 the originating session.
3350 * systemd-networkd learned to cope with private-zone DHCP
3351 options and allows other programs to query the values.
3353 * SELinux access control when enabling/disabling units is no
3354 longer enforced with this release. The previous
3355 implementation was incorrect, and a new corrected
3356 implementation is not yet available. As unit file operations
3357 are still protected via PolicyKit and D-Bus policy this is
3358 not a security problem. Yet, distributions which care about
3359 optimal SELinux support should probably not stabilize on
3362 * sd-bus gained support for matches of type "arg0has=", that
3363 test for membership of strings in string arrays sent in bus
3366 * systemd-resolved now dumps the contents of its DNS and LLMNR
3367 caches to the logs on reception of the SIGUSR1 signal. This
3368 is useful to debug DNS behaviour.
3370 * The coredumpctl tool gained a new --directory= option to
3371 operate on journal files in a specific directory.
3373 * "systemctl reboot" and related commands gained a new
3374 "--message=" option which may be used to set a free-text
3375 wall message when shutting down or rebooting the
3376 system. This message is also logged, which is useful for
3377 figuring out the reason for a reboot or shutdown a
3380 * The "systemd-resolve-host" tool's -i switch now takes
3381 network interface numbers as alternative to interface names.
3383 * A new unit file setting for services has been introduced:
3384 UtmpMode= allows configuration of how precisely systemd
3385 handles utmp and wtmp entries for the service if this is
3386 enabled. This allows writing services that appear similar to
3387 user sessions in the output of the "w", "who", "last" and
3390 * systemd-resolved will now locally synthesize DNS resource
3391 records for the "localhost" and "gateway" domains as well as
3392 the local hostname. This should ensure that clients querying
3393 RRs via resolved will get similar results as those going via
3394 NSS, if nss-myhostname is enabled.
3396 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3397 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3398 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3399 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3400 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3401 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3402 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3403 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3404 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3405 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3406 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3407 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3409 — Berlin, 2015-08-27
3413 * The systemd-efi-boot-generator functionality was merged into
3414 systemd-gpt-auto-generator.
3416 * systemd-networkd now supports Group Policy for vxlan
3417 devices. It can be enabled via the new boolean configuration
3418 option called 'GroupPolicyExtension='.
3420 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3421 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3422 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3424 — Berlin, 2015-07-31
3428 * The python-systemd code has been removed from the systemd repository.
3429 A new repository has been created which accommodates the code from
3430 now on, and we kindly ask distributions to create a separate package
3431 for this: https://github.com/systemd/python-systemd
3433 * The systemd daemon will now reload its main configuration
3434 (/etc/systemd/system.conf) on daemon-reload.
3436 * sd-dhcp now exposes vendor specific extensions via
3437 sd_dhcp_lease_get_vendor_specific().
3439 * systemd-networkd gained a number of new configuration options.
3441 - A new boolean configuration option for TAP devices called
3442 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
3443 device, thus allowing to send and receive GSO packets.
3445 - A new tunnel configuration option called 'CopyDSCP='.
3446 If enabled, the DSCP field of ip6 tunnels is copied into the
3447 decapsulated packet.
3449 - A set of boolean bridge configuration options were added.
3450 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3451 and 'UnicastFlood=' are now parsed by networkd and applied to the
3452 respective bridge link device via the respective IFLA_BRPORT_*
3455 - A new string configuration option to override the hostname sent
3456 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3457 is true, networkd will use the configured hostname instead of the
3458 system hostname when sending DHCP requests.
3460 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3461 networkd will configure the IPv6 flow-label of the tunnel device
3462 according to RFC2460.
3464 - The 'macvtap' virtual network devices are now supported, similar to
3465 the already supported 'macvlan' devices.
3467 * systemd-resolved now implements RFC5452 to improve resilience against
3468 cache poisoning. Additionally, source port randomization is enabled
3469 by default to further protect against DNS spoofing attacks.
3471 * nss-mymachines now supports translating UIDs and GIDs of running
3472 containers with user-namespaces enabled. If a container 'foo'
3473 translates a host uid 'UID' to the container uid 'TUID', then
3474 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3475 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3476 mapped as 'vg-foo-TGID'.
3478 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
3479 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3480 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3481 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3482 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3483 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3484 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3485 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3486 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3487 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3489 — Berlin, 2015-07-29
3493 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3494 There are no known issues with current sysfs, and udev does not need
3495 or should be used to work around such bugs.
3497 * udev does no longer enable USB HID power management. Several reports
3498 indicate, that some devices cannot handle that setting.
3500 * The udev accelerometer helper was removed. The functionality
3501 is now fully included in iio-sensor-proxy. But this means,
3502 older iio-sensor-proxy versions will no longer provide
3503 accelerometer/orientation data with this systemd version.
3504 Please upgrade iio-sensor-proxy to version 1.0.
3506 * networkd gained a new configuration option IPv6PrivacyExtensions=
3507 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3508 for Stateless Address") on selected networks.
3510 * For the sake of fewer build-time dependencies and less code in the
3511 main repository, the python bindings are about to be removed in the
3512 next release. A new repository has been created which accommodates
3513 the code from now on, and we kindly ask distributions to create a
3514 separate package for this. The removal will take place in v223.
3516 https://github.com/systemd/python-systemd
3518 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3519 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3520 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3521 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
3522 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3523 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
3524 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3525 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
3526 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3527 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3529 — Berlin, 2015-07-07
3533 * The sd-bus.h and sd-event.h APIs have now been declared
3534 stable and have been added to the official interface of
3535 libsystemd.so. sd-bus implements an alternative D-Bus client
3536 library, that is relatively easy to use, very efficient and
3537 supports both classic D-Bus as well as kdbus as transport
3538 backend. sd-event is a generic event loop abstraction that
3539 is built around Linux epoll, but adds features such as event
3540 prioritization or efficient timer handling. Both APIs are good
3541 choices for C programs looking for a bus and/or event loop
3542 implementation that is minimal and does not have to be
3543 portable to other kernels.
3545 * kdbus support is no longer compile-time optional. It is now
3546 always built-in. However, it can still be disabled at
3547 runtime using the kdbus=0 kernel command line setting, and
3548 that setting may be changed to default to off, by specifying
3549 --disable-kdbus at build-time. Note though that the kernel
3550 command line setting has no effect if the kdbus.ko kernel
3551 module is not installed, in which case kdbus is (obviously)
3552 also disabled. We encourage all downstream distributions to
3553 begin testing kdbus by adding it to the kernel images in the
3554 development distributions, and leaving kdbus support in
3557 * The minimal required util-linux version has been bumped to
3560 * Support for chkconfig (--enable-chkconfig) was removed in
3561 favor of calling an abstraction tool
3562 /lib/systemd/systemd-sysv-install. This needs to be
3563 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3564 in README for details.
3566 * If there's a systemd unit and a SysV init script for the
3567 same service name, and the user executes "systemctl enable"
3568 for it (or a related call), then this will now enable both
3569 (or execute the related operation on both), not just the
3572 * The libudev API documentation has been converted from gtkdoc
3575 * gudev has been removed from the systemd tree, it is now an
3578 * The systemd-cgtop tool learnt a new --raw switch to generate
3579 "raw" (machine parsable) output.
3581 * networkd's IPForwarding= .network file setting learnt the
3582 new setting "kernel", which ensures that networkd does not
3583 change the IP forwarding sysctl from the default kernel
3586 * The systemd-logind bus API now exposes a new boolean
3587 property "Docked" that reports whether logind considers the
3588 system "docked", i.e. connected to a docking station or not.
3590 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3591 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3592 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3593 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3594 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3595 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3596 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3597 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3598 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3599 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3600 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
3601 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3602 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3603 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3604 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3605 Fink, Zbigniew Jędrzejewski-Szmek
3607 — Berlin, 2015-06-19
3611 * The gudev library has been extracted into a separate repository
3612 available at: https://git.gnome.org/browse/libgudev/
3613 It is now managed as part of the Gnome project. Distributions
3614 are recommended to pass --disable-gudev to systemd and use
3615 gudev from the Gnome project instead. gudev is still included
3616 in systemd, for now. It will be removed soon, though. Please
3617 also see the announcement-thread on systemd-devel:
3618 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
3620 * systemd now exposes a CPUUsageNSec= property for each
3621 service unit on the bus, that contains the overall consumed
3622 CPU time of a service (the sum of what each process of the
3623 service consumed). This value is only available if
3624 CPUAccounting= is turned on for a service, and is then shown
3625 in the "systemctl status" output.
3627 * Support for configuring alternative mappings of the old SysV
3628 runlevels to systemd targets has been removed. They are now
3629 hardcoded in a way that runlevels 2, 3, 4 all map to
3630 multi-user.target and 5 to graphical.target (which
3631 previously was already the default behaviour).
3633 * The auto-mounter logic gained support for mount point
3634 expiry, using a new TimeoutIdleSec= setting in .automount
3635 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3637 * The EFI System Partition (ESP) as mounted to /boot by
3638 systemd-efi-boot-generator will now be unmounted
3639 automatically after 2 minutes of not being used. This should
3640 minimize the risk of ESP corruptions.
3642 * New /etc/fstab options x-systemd.requires= and
3643 x-systemd.requires-mounts-for= are now supported to express
3644 additional dependencies for mounts. This is useful for
3645 journalling file systems that support external journal
3646 devices or overlay file systems that require underlying file
3647 systems to be mounted.
3649 * systemd does not support direct live-upgrades (via systemctl
3650 daemon-reexec) from versions older than v44 anymore. As no
3651 distribution we are aware of shipped such old versions in a
3652 stable release this should not be problematic.
3654 * When systemd forks off a new per-connection service instance
3655 it will now set the $REMOTE_ADDR environment variable to the
3656 remote IP address, and $REMOTE_PORT environment variable to
3657 the remote IP port. This behaviour is similar to the
3658 corresponding environment variables defined by CGI.
3660 * systemd-networkd gained support for uplink failure
3661 detection. The BindCarrier= option allows binding interface
3662 configuration dynamically to the link sense of other
3663 interfaces. This is useful to achieve behaviour like in
3666 * systemd-networkd gained support for configuring the DHCP
3667 client identifier to use when requesting leases.
3669 * systemd-networkd now has a per-network UseNTP= option to
3670 configure whether NTP server information acquired via DHCP
3671 is passed on to services like systemd-timesyncd.
3673 * systemd-networkd gained support for vti6 tunnels.
3675 * Note that systemd-networkd manages the sysctl variable
3676 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3677 it is configured for since v219. The variable controls IP
3678 forwarding, and is a per-interface alternative to the global
3679 /proc/sys/net/ipv[46]/ip_forward. This setting is
3680 configurable in the IPForward= option, which defaults to
3681 "no". This means if networkd is used for an interface it is
3682 no longer sufficient to set the global sysctl option to turn
3683 on IP forwarding! Instead, the .network file option
3684 IPForward= needs to be turned on! Note that the
3685 implementation of this behaviour was broken in v219 and has
3688 * Many bonding and vxlan options are now configurable in
3691 * systemd-nspawn gained a new --property= setting to set unit
3692 properties for the container scope. This is useful for
3693 setting resource parameters (e.g. "CPUShares=500") on
3694 containers started from the command line.
3696 * systemd-nspawn gained a new --private-users= switch to make
3697 use of user namespacing available on recent Linux kernels.
3699 * systemd-nspawn may now be called as part of a shell pipeline
3700 in which case the pipes used for stdin and stdout are passed
3701 directly to the process invoked in the container, without
3702 indirection via a pseudo tty.
3704 * systemd-nspawn gained a new switch to control the UNIX
3705 signal to use when killing the init process of the container
3708 * systemd-nspawn gained a new --overlay= switch for mounting
3709 overlay file systems into the container using the new kernel
3712 * When a container image is imported via systemd-importd and
3713 the host file system is not btrfs, a loopback block device
3714 file is created in /var/lib/machines.raw with a btrfs file
3715 system inside. It is then mounted to /var/lib/machines to
3716 enable btrfs features for container management. The loopback
3717 file and btrfs file system is grown as needed when container
3718 images are imported via systemd-importd.
3720 * systemd-machined/systemd-importd gained support for btrfs
3721 quota, to enforce container disk space limits on disk. This
3722 is exposed in "machinectl set-limit".
3724 * systemd-importd now can import containers from local .tar,
3725 .raw and .qcow2 images, and export them to .tar and .raw. It
3726 can also import dkr v2 images now from the network (on top
3729 * systemd-importd gained support for verifying downloaded
3730 images with gpg2 (previously only gpg1 was supported).
3732 * systemd-machined, systemd-logind, systemd: most bus calls
3733 are now accessible to unprivileged processes via
3734 PolicyKit. Also, systemd-logind will now allow users to kill
3735 their own sessions without further privileges or
3738 * systemd-shutdownd has been removed. This service was
3739 previously responsible for implementing scheduled shutdowns
3740 as exposed in /usr/bin/shutdown's time parameter. This
3741 functionality has now been moved into systemd-logind and is
3742 accessible via a bus interface.
3744 * "systemctl reboot" gained a new switch --firmware-setup that
3745 can be used to reboot into the EFI firmware setup, if that
3746 is available. systemd-logind now exposes an API on the bus
3747 to trigger such reboots, in case graphical desktop UIs want
3748 to cover this functionality.
3750 * "systemctl enable", "systemctl disable" and "systemctl mask"
3751 now support a new "--now" switch. If specified the units
3752 that are enabled will also be started, and the ones
3753 disabled/masked also stopped.
3755 * The Gummiboot EFI boot loader tool has been merged into
3756 systemd, and renamed to "systemd-boot". The bootctl tool has been
3757 updated to support systemd-boot.
3759 * An EFI kernel stub has been added that may be used to create
3760 kernel EFI binaries that contain not only the actual kernel,
3761 but also an initrd, boot splash, command line and OS release
3762 information. This combined binary can then be signed as a
3763 single image, so that the firmware can verify it all in one
3764 step. systemd-boot has special support for EFI binaries created
3765 like this and can extract OS release information from them
3766 and show them in the boot menu. This functionality is useful
3767 to implement cryptographically verified boot schemes.
3769 * Optional support has been added to systemd-fsck to pass
3770 fsck's progress report to an AF_UNIX socket in the file
3773 * udev will no longer create device symlinks for all block
3774 devices by default. A blacklist for excluding special block
3775 devices from this logic has been turned into a whitelist
3776 that requires picking block devices explicitly that require
3779 * A new (currently still internal) API sd-device.h has been
3780 added to libsystemd. This modernized API is supposed to
3781 replace libudev eventually. In fact, already much of libudev
3782 is now just a wrapper around sd-device.h.
3784 * A new hwdb database for storing metadata about pointing
3785 stick devices has been added.
3787 * systemd-tmpfiles gained support for setting file attributes
3788 similar to the "chattr" tool with new 'h' and 'H' lines.
3790 * systemd-journald will no longer unconditionally set the
3791 btrfs NOCOW flag on new journal files. This is instead done
3792 with tmpfiles snippet using the new 'h' line type. This
3793 allows easy disabling of this logic, by masking the
3794 journal-nocow.conf tmpfiles file.
3796 * systemd-journald will now translate audit message types to
3797 human readable identifiers when writing them to the
3798 journal. This should improve readability of audit messages.
3800 * The LUKS logic gained support for the offset= and skip=
3801 options in /etc/crypttab, as previously implemented by
3804 * /usr/lib/os-release gained a new optional field VARIANT= for
3805 distributions that support multiple variants (such as a
3806 desktop edition, a server edition, ...)
3808 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3809 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3810 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3811 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3812 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3813 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3814 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3815 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3816 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3817 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3818 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3819 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3820 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3821 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3822 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3823 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3824 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3825 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3826 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3827 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3828 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3829 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3830 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3831 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3832 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3833 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3834 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3836 — Berlin, 2015-05-22
3840 * Introduce a new API "sd-hwdb.h" for querying the hardware
3841 metadata database. With this minimal interface one can query
3842 and enumerate the udev hwdb, decoupled from the old libudev
3843 library. libudev's interface for this is now only a wrapper
3844 around sd-hwdb. A new tool systemd-hwdb has been added to
3845 interface with and update the database.
3847 * When any of systemd's tools copies files (for example due to
3848 tmpfiles' C lines) a btrfs reflink will attempted first,
3849 before bytewise copying is done.
3851 * systemd-nspawn gained a new --ephemeral switch. When
3852 specified a btrfs snapshot is taken of the container's root
3853 directory, and immediately removed when the container
3854 terminates again. Thus, a container can be started whose
3855 changes never alter the container's root directory, and are
3856 lost on container termination. This switch can also be used
3857 for starting a container off the root file system of the
3858 host without affecting the host OS. This switch is only
3859 available on btrfs file systems.
3861 * systemd-nspawn gained a new --template= switch. It takes the
3862 path to a container tree to use as template for the tree
3863 specified via --directory=, should that directory be
3864 missing. This allows instantiating containers dynamically,
3865 on first run. This switch is only available on btrfs file
3868 * When a .mount unit refers to a mount point on which multiple
3869 mounts are stacked, and the .mount unit is stopped all of
3870 the stacked mount points will now be unmounted until no
3871 mount point remains.
3873 * systemd now has an explicit notion of supported and
3874 unsupported unit types. Jobs enqueued for unsupported unit
3875 types will now fail with an "unsupported" error code. More
3876 specifically .swap, .automount and .device units are not
3877 supported in containers, .busname units are not supported on
3878 non-kdbus systems. .swap and .automount are also not
3879 supported if their respective kernel compile time options
3882 * machinectl gained support for two new "copy-from" and
3883 "copy-to" commands for copying files from a running
3884 container to the host or vice versa.
3886 * machinectl gained support for a new "bind" command to bind
3887 mount host directories into local containers. This is
3888 currently only supported for nspawn containers.
3890 * networkd gained support for configuring bridge forwarding
3891 database entries (fdb) from .network files.
3893 * A new tiny daemon "systemd-importd" has been added that can
3894 download container images in tar, raw, qcow2 or dkr formats,
3895 and make them available locally in /var/lib/machines, so
3896 that they can run as nspawn containers. The daemon can GPG
3897 verify the downloads (not supported for dkr, since it has no
3898 provisions for verifying downloads). It will transparently
3899 decompress bz2, xz, gzip compressed downloads if necessary,
3900 and restore sparse files on disk. The daemon uses privilege
3901 separation to ensure the actual download logic runs with
3902 fewer privileges than the daemon itself. machinectl has
3903 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3904 make the functionality of importd available to the
3905 user. With this in place the Fedora and Ubuntu "Cloud"
3906 images can be downloaded and booted as containers unmodified
3907 (the Fedora images lack the appropriate GPG signature files
3908 currently, so they cannot be verified, but this will change
3909 soon, hopefully). Note that downloading images is currently
3910 only fully supported on btrfs.
3912 * machinectl is now able to list container images found in
3913 /var/lib/machines, along with some metadata about sizes of
3914 disk and similar. If the directory is located on btrfs and
3915 quota is enabled, this includes quota display. A new command
3916 "image-status" has been added that shows additional
3917 information about images.
3919 * machinectl is now able to clone container images
3920 efficiently, if the underlying file system (btrfs) supports
3921 it, with the new "machinectl clone" command. It also
3922 gained commands for renaming and removing images, as well as
3923 marking them read-only or read-write (supported also on
3924 legacy file systems).
3926 * networkd gained support for collecting LLDP network
3927 announcements, from hardware that supports this. This is
3928 shown in networkctl output.
3930 * systemd-run gained support for a new -t (--pty) switch for
3931 invoking a binary on a pty whose input and output is
3932 connected to the invoking terminal. This allows executing
3933 processes as system services while interactively
3934 communicating with them via the terminal. Most interestingly
3935 this is supported across container boundaries. Invoking
3936 "systemd-run -t /bin/bash" is an alternative to running a
3937 full login session, the difference being that the former
3938 will not register a session, nor go through the PAM session
3941 * tmpfiles gained support for a new "v" line type for creating
3942 btrfs subvolumes. If the underlying file system is a legacy
3943 file system, this automatically degrades to creating a
3944 normal directory. Among others /var/lib/machines is now
3945 created like this at boot, should it be missing.
3947 * The directory /var/lib/containers/ has been deprecated and
3948 been replaced by /var/lib/machines. The term "machines" has
3949 been used in the systemd context as generic term for both
3950 VMs and containers, and hence appears more appropriate for
3951 this, as the directory can also contain raw images bootable
3954 * systemd-nspawn when invoked with -M but without --directory=
3955 or --image= is now capable of searching for the container
3956 root directory, subvolume or disk image automatically, in
3957 /var/lib/machines. systemd-nspawn@.service has been updated
3958 to make use of this, thus allowing it to be used for raw
3961 * A new machines.target unit has been introduced that is
3962 supposed to group all containers/VMs invoked as services on
3963 the system. systemd-nspawn@.service has been updated to
3964 integrate with that.
3966 * machinectl gained a new "start" command, for invoking a
3967 container as a service. "machinectl start foo" is mostly
3968 equivalent to "systemctl start systemd-nspawn@foo.service",
3969 but handles escaping in a nicer way.
3971 * systemd-nspawn will now mount most of the cgroupfs tree
3972 read-only into each container, with the exception of the
3973 container's own subtree in the name=systemd hierarchy.
3975 * journald now sets the special FS_NOCOW file flag for its
3976 journal files. This should improve performance on btrfs, by
3977 avoiding heavy fragmentation when journald's write-pattern
3978 is used on COW file systems. It degrades btrfs' data
3979 integrity guarantees for the files to the same levels as for
3980 ext3/ext4 however. This should be OK though as journald does
3981 its own data integrity checks and all its objects are
3982 checksummed on disk. Also, journald should handle btrfs disk
3983 full events a lot more gracefully now, by processing SIGBUS
3984 errors, and not relying on fallocate() anymore.
3986 * When journald detects that journal files it is writing to
3987 have been deleted it will immediately start new journal
3990 * systemd now provides a way to store file descriptors
3991 per-service in PID 1. This is useful for daemons to ensure
3992 that fds they require are not lost during a daemon
3993 restart. The fds are passed to the daemon on the next
3994 invocation in the same way socket activation fds are
3995 passed. This is now used by journald to ensure that the
3996 various sockets connected to all the system's stdout/stderr
3997 are not lost when journald is restarted. File descriptors
3998 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
3999 an extension to sd_notify(). Note that a limit is enforced
4000 on the number of fds a service can store in PID 1, and it
4001 defaults to 0, so that no fds may be stored, unless this is
4002 explicitly turned on.
4004 * The default TERM variable to use for units connected to a
4005 terminal, when no other value is explicitly is set is now
4006 vt220 rather than vt102. This should be fairly safe still,
4007 but allows PgUp/PgDn work.
4009 * The /etc/crypttab option header= as known from Debian is now
4012 * "loginctl user-status" and "loginctl session-status" will
4013 now show the last 10 lines of log messages of the
4014 user/session following the status output. Similar,
4015 "machinectl status" will show the last 10 log lines
4016 associated with a virtual machine or container
4017 service. (Note that this is usually not the log messages
4018 done in the VM/container itself, but simply what the
4019 container manager logs. For nspawn this includes all console
4022 * "loginctl session-status" without further argument will now
4023 show the status of the session of the caller. Similar,
4024 "lock-session", "unlock-session", "activate",
4025 "enable-linger", "disable-linger" may now be called without
4026 session/user parameter in which case they apply to the
4027 caller's session/user.
4029 * An X11 session scriptlet is now shipped that uploads
4030 $DISPLAY and $XAUTHORITY into the environment of the systemd
4031 --user daemon if a session begins. This should improve
4032 compatibility with X11 enabled applications run as systemd
4035 * Generators are now subject to masking via /etc and /run, the
4036 same way as unit files.
4038 * networkd .network files gained support for configuring
4039 per-link IPv4/IPv6 packet forwarding as well as IPv4
4040 masquerading. This is by default turned on for veth links to
4041 containers, as registered by systemd-nspawn. This means that
4042 nspawn containers run with --network-veth will now get
4043 automatic routed access to the host's networks without any
4044 further configuration or setup, as long as networkd runs on
4047 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4048 or UDP posts of a container on the host. With this in place
4049 it is possible to run containers with private veth links
4050 (--network-veth), and have their functionality exposed on
4051 the host as if their services were running directly on the
4054 * systemd-nspawn's --network-veth switch now gained a short
4055 version "-n", since with the changes above it is now truly
4056 useful out-of-the-box. The systemd-nspawn@.service has been
4057 updated to make use of it too by default.
4059 * systemd-nspawn will now maintain a per-image R/W lock, to
4060 ensure that the same image is not started more than once
4061 writable. (It's OK to run an image multiple times
4062 simultaneously in read-only mode.)
4064 * systemd-nspawn's --image= option is now capable of
4065 dissecting and booting MBR and GPT disk images that contain
4066 only a single active Linux partition. Previously it
4067 supported only GPT disk images with proper GPT type
4068 IDs. This allows running cloud images from major
4069 distributions directly with systemd-nspawn, without
4072 * In addition to collecting mouse dpi data in the udev
4073 hardware database, there's now support for collecting angle
4074 information for mouse scroll wheels. The database is
4075 supposed to guarantee similar scrolling behavior on mice
4076 that it knows about. There's also support for collecting
4077 information about Touchpad types.
4079 * udev's input_id built-in will now also collect touch screen
4080 dimension data and attach it to probed devices.
4082 * /etc/os-release gained support for a Distribution Privacy
4085 * networkd gained support for creating "ipvlan", "gretap",
4086 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4088 * systemd-tmpfiles gained support for "a" lines for setting
4091 * systemd-nspawn will now mount /tmp in the container to
4092 tmpfs, automatically.
4094 * systemd now exposes the memory.usage_in_bytes cgroup
4095 attribute and shows it for each service in the "systemctl
4096 status" output, if available.
4098 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4099 immediate reboot is triggered. This useful if shutdown is
4100 hung and is unable to complete, to expedite the
4101 operation. Note that this kind of reboot will still unmount
4102 all file systems, and hence should not result in fsck being
4105 * A .device unit for an optical block device will now be
4106 considered active only when a medium is in the drive. Also,
4107 mount units are now bound to their backing devices thus
4108 triggering automatic unmounting when devices become
4109 unavailable. With this in place systemd will now
4110 automatically unmount left-over mounts when a CD-ROM is
4111 ejected or an USB stick is yanked from the system.
4113 * networkd-wait-online now has support for waiting for
4114 specific interfaces only (with globbing), and for giving up
4115 after a configurable timeout.
4117 * networkd now exits when idle. It will be automatically
4118 restarted as soon as interfaces show up, are removed or
4119 change state. networkd will stay around as long as there is
4120 at least one DHCP state machine or similar around, that keep
4123 * networkd may now configure IPv6 link-local addressing in
4124 addition to IPv4 link-local addressing.
4126 * The IPv6 "token" for use in SLAAC may now be configured for
4127 each .network interface in networkd.
4129 * Routes configured with networkd may now be assigned a scope
4132 * networkd's [Match] sections now support globbing and lists
4133 of multiple space-separated matches per item.
4135 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4136 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4137 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4138 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4139 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4140 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4141 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4142 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4143 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4144 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4145 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4146 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4147 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4148 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4149 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
4150 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4151 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4152 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4153 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4154 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4155 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4156 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
4157 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4158 Hoffmann, Zbigniew Jędrzejewski-Szmek
4160 — Berlin, 2015-02-16
4164 * When querying unit file enablement status (for example via
4165 "systemctl is-enabled"), a new state "indirect" is now known
4166 which indicates that a unit might not be enabled itself, but
4167 another unit listed in its Also= setting might be.
4169 * Similar to the various existing ConditionXYZ= settings for
4170 units, there are now matching AssertXYZ= settings. While
4171 failing conditions cause a unit to be skipped, but its job
4172 to succeed, failing assertions declared like this will cause
4173 a unit start operation and its job to fail.
4175 * hostnamed now knows a new chassis type "embedded".
4177 * systemctl gained a new "edit" command. When used on a unit
4178 file, this allows extending unit files with .d/ drop-in
4179 configuration snippets or editing the full file (after
4180 copying it from /usr/lib to /etc). This will invoke the
4181 user's editor (as configured with $EDITOR), and reload the
4182 modified configuration after editing.
4184 * "systemctl status" now shows the suggested enablement state
4185 for a unit, as declared in the (usually vendor-supplied)
4186 system preset files.
4188 * nss-myhostname will now resolve the single-label host name
4189 "gateway" to the locally configured default IP routing
4190 gateways, ordered by their metrics. This assigns a stable
4191 name to the used gateways, regardless which ones are
4192 currently configured. Note that the name will only be
4193 resolved after all other name sources (if nss-myhostname is
4194 configured properly) and should hence not negatively impact
4195 systems that use the single-label host name "gateway" in
4198 * systemd-inhibit now allows filtering by mode when listing
4201 * Scope and service units gained a new "Delegate" boolean
4202 property, which, when set, allows processes running inside the
4203 unit to further partition resources. This is primarily
4204 useful for systemd user instances as well as container
4207 * journald will now pick up audit messages directly from
4208 the kernel, and log them like any other log message. The
4209 audit fields are split up and fully indexed. This means that
4210 journalctl in many ways is now a (nicer!) alternative to
4211 ausearch, the traditional audit client. Note that this
4212 implements only a minimal audit client. If you want the
4213 special audit modes like reboot-on-log-overflow, please use
4214 the traditional auditd instead, which can be used in
4215 parallel to journald.
4217 * The ConditionSecurity= unit file option now understands the
4218 special string "audit" to check whether auditing is
4221 * journalctl gained two new commands --vacuum-size= and
4222 --vacuum-time= to delete old journal files until the
4223 remaining ones take up no more than the specified size on disk,
4224 or are not older than the specified time.
4226 * A new, native PPPoE library has been added to sd-network,
4227 systemd's library of light-weight networking protocols. This
4228 library will be used in a future version of networkd to
4229 enable PPPoE communication without an external pppd daemon.
4231 * The busctl tool now understands a new "capture" verb that
4232 works similar to "monitor", but writes a packet capture
4233 trace to STDOUT that can be redirected to a file which is
4234 compatible with libcap's capture file format. This can then
4235 be loaded in Wireshark and similar tools to inspect bus
4238 * The busctl tool now understands a new "tree" verb that shows
4239 the object trees of a specific service on the bus, or of all
4242 * The busctl tool now understands a new "introspect" verb that
4243 shows all interfaces and members of objects on the bus,
4244 including their signature and values. This is particularly
4245 useful to get more information about bus objects shown by
4246 the new "busctl tree" command.
4248 * The busctl tool now understands new verbs "call",
4249 "set-property" and "get-property" for invoking bus method
4250 calls, setting and getting bus object properties in a
4253 * busctl gained a new --augment-creds= argument that controls
4254 whether the tool shall augment credential information it
4255 gets from the bus with data from /proc, in a possibly
4258 * nspawn's --link-journal= switch gained two new values
4259 "try-guest" and "try-host" that work like "guest" and
4260 "host", but do not fail if the host has no persistent
4261 journalling enabled. -j is now equivalent to
4262 --link-journal=try-guest.
4264 * macvlan network devices created by nspawn will now have
4265 stable MAC addresses.
4267 * A new SmackProcessLabel= unit setting has been added, which
4268 controls the SMACK security label processes forked off by
4269 the respective unit shall use.
4271 * If compiled with --enable-xkbcommon, systemd-localed will
4272 verify x11 keymap settings by compiling the given keymap. It
4273 will spew out warnings if the compilation fails. This
4274 requires libxkbcommon to be installed.
4276 * When a coredump is collected, a larger number of metadata
4277 fields is now collected and included in the journal records
4278 created for it. More specifically, control group membership,
4279 environment variables, memory maps, working directory,
4280 chroot directory, /proc/$PID/status, and a list of open file
4281 descriptors is now stored in the log entry.
4283 * The udev hwdb now contains DPI information for mice. For
4286 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4288 * All systemd programs that read standalone configuration
4289 files in /etc now also support a corresponding series of
4290 .conf.d configuration directories in /etc/, /run/,
4291 /usr/local/lib/, /usr/lib/, and (if configured with
4292 --enable-split-usr) /lib/. In particular, the following
4293 configuration files now have corresponding configuration
4294 directories: system.conf user.conf, logind.conf,
4295 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4296 resolved.conf, timesyncd.conf, journal-remote.conf, and
4297 journal-upload.conf. Note that distributions should use the
4298 configuration directories in /usr/lib/; the directories in
4299 /etc/ are reserved for the system administrator.
4301 * systemd-rfkill will no longer take the rfkill device name
4302 into account when storing rfkill state on disk, as the name
4303 might be dynamically assigned and not stable. Instead, the
4304 ID_PATH udev variable combined with the rfkill type (wlan,
4305 bluetooth, ...) is used.
4307 * A new service systemd-machine-id-commit.service has been
4308 added. When used on systems where /etc is read-only during
4309 boot, and /etc/machine-id is not initialized (but an empty
4310 file), this service will copy the temporary machine ID
4311 created as replacement into /etc after the system is fully
4312 booted up. This is useful for systems that are freshly
4313 installed with a non-initialized machine ID, but should get
4314 a fixed machine ID for subsequent boots.
4316 * networkd's .netdev files now provide a large set of
4317 configuration parameters for VXLAN devices. Similarly, the
4318 bridge port cost parameter is now configurable in .network
4319 files. There's also new support for configuring IP source
4320 routing. networkd .link files gained support for a new
4321 OriginalName= match that is useful to match against the
4322 original interface name the kernel assigned. .network files
4323 may include MTU= and MACAddress= fields for altering the MTU
4324 and MAC address while being connected to a specific network
4327 * The LUKS logic gained supported for configuring
4328 UUID-specific key files. There's also new support for naming
4329 LUKS device from the kernel command line, using the new
4330 luks.name= argument.
4332 * Timer units may now be transiently created via the bus API
4333 (this was previously already available for scope and service
4334 units). In addition it is now possible to create multiple
4335 transient units at the same time with a single bus call. The
4336 "systemd-run" tool has been updated to make use of this for
4337 running commands on a specified time, in at(1)-style.
4339 * tmpfiles gained support for "t" lines, for assigning
4340 extended attributes to files. Among other uses this may be
4341 used to assign SMACK labels to files.
4343 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4344 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4345 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4346 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4347 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4348 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4349 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4350 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4351 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4352 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4353 Michael Biebl, Michael Chapman, Michael Marineau, Michal
4354 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4355 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4356 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4357 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4358 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4359 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4360 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
4362 — Berlin, 2014-12-10
4366 * journalctl gained the new options -t/--identifier= to match
4367 on the syslog identifier (aka "tag"), as well as --utc to
4368 show log timestamps in the UTC timezone. journalctl now also
4369 accepts -n/--lines=all to disable line capping in a pager.
4371 * journalctl gained a new switch, --flush, that synchronously
4372 flushes logs from /run/log/journal to /var/log/journal if
4373 persistent storage is enabled. systemd-journal-flush.service
4374 now waits until the operation is complete.
4376 * Services can notify the manager before they start a reload
4377 (by sending RELOADING=1) or shutdown (by sending
4378 STOPPING=1). This allows the manager to track and show the
4379 internal state of daemons and closes a race condition when
4380 the process is still running but has closed its D-Bus
4383 * Services with Type=oneshot do not have to have any ExecStart
4386 * User units are now loaded also from
4387 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4388 /run/systemd/user directory that was already previously
4389 supported, but is under the control of the user.
4391 * Job timeouts (i.e. time-outs on the time a job that is
4392 queued stays in the run queue) can now optionally result in
4393 immediate reboot or power-off actions (JobTimeoutAction= and
4394 JobTimeoutRebootArgument=). This is useful on ".target"
4395 units, to limit the maximum time a target remains
4396 undispatched in the run queue, and to trigger an emergency
4397 operation in such a case. This is now used by default to
4398 turn off the system if boot-up (as defined by everything in
4399 basic.target) hangs and does not complete for at least
4400 15min. Also, if power-off or reboot hang for at least 30min
4401 an immediate power-off/reboot operation is triggered. This
4402 functionality is particularly useful to increase reliability
4403 on embedded devices, but also on laptops which might
4404 accidentally get powered on when carried in a backpack and
4405 whose boot stays stuck in a hard disk encryption passphrase
4408 * systemd-logind can be configured to also handle lid switch
4409 events even when the machine is docked or multiple displays
4410 are attached (HandleLidSwitchDocked= option).
4412 * A helper binary and a service have been added which can be
4413 used to resume from hibernation in the initramfs. A
4414 generator will parse the resume= option on the kernel
4415 command line to trigger resume.
4417 * A user console daemon systemd-consoled has been
4418 added. Currently, it is a preview, and will so far open a
4419 single terminal on each session of the user marked as
4420 Desktop=systemd-console.
4422 * Route metrics can be specified for DHCP routes added by
4425 * The SELinux context of socket-activated services can be set
4426 from the information provided by the networking stack
4427 (SELinuxContextFromNet= option).
4429 * Userspace firmware loading support has been removed and
4430 the minimum supported kernel version is thus bumped to 3.7.
4432 * Timeout for udev workers has been increased from 1 to 3
4433 minutes, but a warning will be printed after 1 minute to
4434 help diagnose kernel modules that take a long time to load.
4436 * Udev rules can now remove tags on devices with TAG-="foobar".
4438 * systemd's readahead implementation has been removed. In many
4439 circumstances it didn't give expected benefits even for
4440 rotational disk drives and was becoming less relevant in the
4441 age of SSDs. As none of the developers has been using
4442 rotating media anymore, and nobody stepped up to actively
4443 maintain this component of systemd it has now been removed.
4445 * Swap units can use Options= to specify discard options.
4446 Discard options specified for swaps in /etc/fstab are now
4449 * Docker containers are now detected as a separate type of
4452 * The Password Agent protocol gained support for queries where
4453 the user input is shown, useful e.g. for user names.
4454 systemd-ask-password gained a new --echo option to turn that
4457 * The default sysctl.d/ snippets will now set:
4459 net.core.default_qdisc = fq_codel
4461 This selects Fair Queuing Controlled Delay as the default
4462 queuing discipline for network interfaces. fq_codel helps
4463 fight the network bufferbloat problem. It is believed to be
4464 a good default with no tuning required for most workloads.
4465 Downstream distributions may override this choice. On 10Gbit
4466 servers that do not do forwarding, "fq" may perform better.
4467 Systems without a good clocksource should use "pfifo_fast".
4469 * If kdbus is enabled during build a new option BusPolicy= is
4470 available for service units, that allows locking all service
4471 processes into a stricter bus policy, in order to limit
4472 access to various bus services, or even hide most of them
4473 from the service's view entirely.
4475 * networkctl will now show the .network and .link file
4476 networkd has applied to a specific interface.
4478 * sd-login gained a new API call sd_session_get_desktop() to
4479 query which desktop environment has been selected for a
4482 * UNIX utmp support is now compile-time optional to support
4483 legacy-free systems.
4485 * systemctl gained two new commands "add-wants" and
4486 "add-requires" for pulling in units from specific targets
4489 * If the word "rescue" is specified on the kernel command line
4490 the system will now boot into rescue mode (aka
4491 rescue.target), which was previously available only by
4492 specifying "1" or "systemd.unit=rescue.target" on the kernel
4493 command line. This new kernel command line option nicely
4494 mirrors the already existing "emergency" kernel command line
4497 * New kernel command line options mount.usr=, mount.usrflags=,
4498 mount.usrfstype= have been added that match root=, rootflags=,
4499 rootfstype= but allow mounting a specific file system to
4502 * The $NOTIFY_SOCKET is now also passed to control processes of
4503 services, not only the main process.
4505 * This version reenables support for fsck's -l switch. This
4506 means at least version v2.25 of util-linux is required for
4507 operation, otherwise dead-locks on device nodes may
4508 occur. Again: you need to update util-linux to at least
4509 v2.25 when updating systemd to v217.
4511 * The "multi-seat-x" tool has been removed from systemd, as
4512 its functionality has been integrated into X servers 1.16,
4513 and the tool is hence redundant. It is recommended to update
4514 display managers invoking this tool to simply invoke X
4515 directly from now on, again.
4517 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4518 message flag has been added for all of systemd's PolicyKit
4519 authenticated method calls has been added. In particular
4520 this now allows optional interactive authorization via
4521 PolicyKit for many of PID1's privileged operations such as
4522 unit file enabling and disabling.
4524 * "udevadm hwdb --update" learnt a new switch "--usr" for
4525 placing the rebuilt hardware database in /usr instead of
4526 /etc. When used only hardware database entries stored in
4527 /usr will be used, and any user database entries in /etc are
4528 ignored. This functionality is useful for vendors to ship a
4529 pre-built database on systems where local configuration is
4530 unnecessary or unlikely.
4532 * Calendar time specifications in .timer units now also
4533 understand the strings "semi-annually", "quarterly" and
4534 "minutely" as shortcuts (in addition to the preexisting
4535 "anually", "hourly", ...).
4537 * systemd-tmpfiles will now correctly create files in /dev
4538 at boot which are marked for creation only at boot. It is
4539 recommended to always create static device nodes with 'c!'
4540 and 'b!', so that they are created only at boot and not
4541 overwritten at runtime.
4543 * When the watchdog logic is used for a service (WatchdogSec=)
4544 and the watchdog timeout is hit the service will now be
4545 terminated with SIGABRT (instead of just SIGTERM), in order
4546 to make sure a proper coredump and backtrace is
4547 generated. This ensures that hanging services will result in
4548 similar coredump/backtrace behaviour as services that hit a
4551 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4552 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4553 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4554 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4555 Herrmann, David Sommerseth, David Strauss, Emil Renner
4556 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4557 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4558 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4559 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4560 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4561 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4562 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4563 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4564 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4565 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4566 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4567 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4568 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4569 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4570 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4571 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
4574 — Berlin, 2014-10-28
4578 * timedated no longer reads NTP implementation unit names from
4579 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
4580 implementations should add a
4582 Conflicts=systemd-timesyncd.service
4584 to their unit files to take over and replace systemd's NTP
4585 default functionality.
4587 * systemd-sysusers gained a new line type "r" for configuring
4588 which UID/GID ranges to allocate system users/groups
4589 from. Lines of type "u" may now add an additional column
4590 that specifies the home directory for the system user to be
4591 created. Also, systemd-sysusers may now optionally read user
4592 information from STDIN instead of a file. This is useful for
4593 invoking it from RPM preinst scriptlets that need to create
4594 users before the first RPM file is installed since these
4595 files might need to be owned by them. A new
4596 %sysusers_create_inline RPM macro has been introduced to do
4597 just that. systemd-sysusers now updates the shadow files as
4598 well as the user/group databases, which should enhance
4599 compatibility with certain tools like grpck.
4601 * A number of bus APIs of PID 1 now optionally consult
4602 PolicyKit to permit access for otherwise unprivileged
4603 clients under certain conditions. Note that this currently
4604 doesn't support interactive authentication yet, but this is
4605 expected to be added eventually, too.
4607 * /etc/machine-info now has new fields for configuring the
4608 deployment environment of the machine, as well as the
4609 location of the machine. hostnamectl has been updated with
4610 new command to update these fields.
4612 * systemd-timesyncd has been updated to automatically acquire
4613 NTP server information from systemd-networkd, which might
4614 have been discovered via DHCP.
4616 * systemd-resolved now includes a caching DNS stub resolver
4617 and a complete LLMNR name resolution implementation. A new
4618 NSS module "nss-resolve" has been added which can be used
4619 instead of glibc's own "nss-dns" to resolve hostnames via
4620 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4621 be resolved via systemd-resolved D-Bus APIs. In contrast to
4622 the glibc internal resolver systemd-resolved is aware of
4623 multi-homed system, and keeps DNS server and caches separate
4624 and per-interface. Queries are sent simultaneously on all
4625 interfaces that have DNS servers configured, in order to
4626 properly handle VPNs and local LANs which might resolve
4627 separate sets of domain names. systemd-resolved may acquire
4628 DNS server information from systemd-networkd automatically,
4629 which in turn might have discovered them via DHCP. A tool
4630 "systemd-resolve-host" has been added that may be used to
4631 query the DNS logic in resolved. systemd-resolved implements
4632 IDNA and automatically uses IDNA or UTF-8 encoding depending
4633 on whether classic DNS or LLMNR is used as transport. In the
4634 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4635 implementation to systemd-resolved.
4637 * A new NSS module nss-mymachines has been added, that
4638 automatically resolves the names of all local registered
4639 containers to their respective IP addresses.
4641 * A new client tool "networkctl" for systemd-networkd has been
4642 added. It currently is entirely passive and will query
4643 networking configuration from udev, rtnetlink and networkd,
4644 and present it to the user in a very friendly
4645 way. Eventually, we hope to extend it to become a full
4646 control utility for networkd.
4648 * .socket units gained a new DeferAcceptSec= setting that
4649 controls the kernels' TCP_DEFER_ACCEPT sockopt for
4650 TCP. Similarly, support for controlling TCP keep-alive
4651 settings has been added (KeepAliveTimeSec=,
4652 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4653 turning off Nagle's algorithm on TCP has been added
4656 * logind learned a new session type "web", for use in projects
4657 like Cockpit which register web clients as PAM sessions.
4659 * timer units with at least one OnCalendar= setting will now
4660 be started only after time-sync.target has been
4661 reached. This way they will not elapse before the system
4662 clock has been corrected by a local NTP client or
4663 similar. This is particular useful on RTC-less embedded
4664 machines, that come up with an invalid system clock.
4666 * systemd-nspawn's --network-veth= switch should now result in
4667 stable MAC addresses for both the outer and the inner side
4670 * systemd-nspawn gained a new --volatile= switch for running
4671 container instances with /etc or /var unpopulated.
4673 * The kdbus client code has been updated to use the new Linux
4674 3.17 memfd subsystem instead of the old kdbus-specific one.
4676 * systemd-networkd's DHCP client and server now support
4677 FORCERENEW. There are also new configuration options to
4678 configure the vendor client identifier and broadcast mode
4681 * systemd will no longer inform the kernel about the current
4682 timezone, as this is necessarily incorrect and racy as the
4683 kernel has no understanding of DST and similar
4684 concepts. This hence means FAT timestamps will be always
4685 considered UTC, similar to what Android is already
4686 doing. Also, when the RTC is configured to the local time
4687 (rather than UTC) systemd will never synchronize back to it,
4688 as this might confuse Windows at a later boot.
4690 * systemd-analyze gained a new command "verify" for offline
4691 validation of unit files.
4693 * systemd-networkd gained support for a couple of additional
4694 settings for bonding networking setups. Also, the metric for
4695 statically configured routes may now be configured. For
4696 network interfaces where this is appropriate the peer IP
4697 address may now be configured.
4699 * systemd-networkd's DHCP client will no longer request
4700 broadcasting by default, as this tripped up some networks.
4701 For hardware where broadcast is required the feature should
4702 be switched back on using RequestBroadcast=yes.
4704 * systemd-networkd will now set up IPv4LL addresses (when
4705 enabled) even if DHCP is configured successfully.
4707 * udev will now default to respect network device names given
4708 by the kernel when the kernel indicates that these are
4709 predictable. This behavior can be tweaked by changing
4710 NamePolicy= in the relevant .link file.
4712 * A new library systemd-terminal has been added that
4713 implements full TTY stream parsing and rendering. This
4714 library is supposed to be used later on for implementing a
4715 full userspace VT subsystem, replacing the current kernel
4718 * A new tool systemd-journal-upload has been added to push
4719 journal data to a remote system running
4720 systemd-journal-remote.
4722 * journald will no longer forward all local data to another
4723 running syslog daemon. This change has been made because
4724 rsyslog (which appears to be the most commonly used syslog
4725 implementation these days) no longer makes use of this, and
4726 instead pulls the data out of the journal on its own. Since
4727 forwarding the messages to a non-existent syslog server is
4728 more expensive than we assumed we have now turned this
4729 off. If you run a syslog server that is not a recent rsyslog
4730 version, you have to turn this option on again
4731 (ForwardToSyslog= in journald.conf).
4733 * journald now optionally supports the LZ4 compressor for
4734 larger journal fields. This compressor should perform much
4735 better than XZ which was the previous default.
4737 * machinectl now shows the IP addresses of local containers,
4738 if it knows them, plus the interface name of the container.
4740 * A new tool "systemd-escape" has been added that makes it
4741 easy to escape strings to build unit names and similar.
4743 * sd_notify() messages may now include a new ERRNO= field
4744 which is parsed and collected by systemd and shown among the
4745 "systemctl status" output for a service.
4747 * A new component "systemd-firstboot" has been added that
4748 queries the most basic systemd information (timezone,
4749 hostname, root password) interactively on first
4750 boot. Alternatively it may also be used to provision these
4751 things offline on OS images installed into directories.
4753 * The default sysctl.d/ snippets will now set
4755 net.ipv4.conf.default.promote_secondaries=1
4757 This has the benefit of no flushing secondary IP addresses
4758 when primary addresses are removed.
4760 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4761 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4762 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4763 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4764 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4765 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4766 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4767 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4768 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4769 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4770 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4771 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4772 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4773 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4774 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4776 — Berlin, 2014-08-19
4780 * A new tool systemd-sysusers has been added. This tool
4781 creates system users and groups in /etc/passwd and
4782 /etc/group, based on static declarative system user/group
4783 definitions in /usr/lib/sysusers.d/. This is useful to
4784 enable factory resets and volatile systems that boot up with
4785 an empty /etc directory, and thus need system users and
4786 groups created during early boot. systemd now also ships
4787 with two default sysusers.d/ files for the most basic
4788 users and groups systemd and the core operating system
4791 * A new tmpfiles snippet has been added that rebuilds the
4792 essential files in /etc on boot, should they be missing.
4794 * A directive for ensuring automatic clean-up of
4795 /var/cache/man/ has been removed from the default
4796 configuration. This line should now be shipped by the man
4797 implementation. The necessary change has been made to the
4798 man-db implementation. Note that you need to update your man
4799 implementation to one that ships this line, otherwise no
4800 automatic clean-up of /var/cache/man will take place.
4802 * A new condition ConditionNeedsUpdate= has been added that
4803 may conditionalize services to only run when /etc or /var
4804 are "older" than the vendor operating system resources in
4805 /usr. This is useful for reconstructing or updating /etc
4806 after an offline update of /usr or a factory reset, on the
4807 next reboot. Services that want to run once after such an
4808 update or reset should use this condition and order
4809 themselves before the new systemd-update-done.service, which
4810 will mark the two directories as fully updated. A number of
4811 service files have been added making use of this, to rebuild
4812 the udev hardware database, the journald message catalog and
4813 dynamic loader cache (ldconfig). The systemd-sysusers tool
4814 described above also makes use of this now. With this in
4815 place it is now possible to start up a minimal operating
4816 system with /etc empty cleanly. For more information on the
4817 concepts involved see this recent blog story:
4819 http://0pointer.de/blog/projects/stateless.html
4821 * A new system group "input" has been introduced, and all
4822 input device nodes get this group assigned. This is useful
4823 for system-level software to get access to input devices. It
4824 complements what is already done for "audio" and "video".
4826 * systemd-networkd learnt minimal DHCPv4 server support in
4827 addition to the existing DHCPv4 client support. It also
4828 learnt DHCPv6 client and IPv6 Router Solicitation client
4829 support. The DHCPv4 client gained support for static routes
4830 passed in from the server. Note that the [DHCPv4] section
4831 known in older systemd-networkd versions has been renamed to
4832 [DHCP] and is now also used by the DHCPv6 client. Existing
4833 .network files using settings of this section should be
4834 updated, though compatibility is maintained. Optionally, the
4835 client hostname may now be sent to the DHCP server.
4837 * networkd gained support for vxlan virtual networks as well
4838 as tun/tap and dummy devices.
4840 * networkd gained support for automatic allocation of address
4841 ranges for interfaces from a system-wide pool of
4842 addresses. This is useful for dynamically managing a large
4843 number of interfaces with a single network configuration
4844 file. In particular this is useful to easily assign
4845 appropriate IP addresses to the veth links of a large number
4846 of nspawn instances.
4848 * RPM macros for processing sysusers, sysctl and binfmt
4849 drop-in snippets at package installation time have been
4852 * The /etc/os-release file should now be placed in
4853 /usr/lib/os-release. The old location is automatically
4854 created as symlink. /usr/lib is the more appropriate
4855 location of this file, since it shall actually describe the
4856 vendor operating system shipped in /usr, and not the
4857 configuration stored in /etc.
4859 * .mount units gained a new boolean SloppyOptions= setting
4860 that maps to mount(8)'s -s option which enables permissive
4861 parsing of unknown mount options.
4863 * tmpfiles learnt a new "L+" directive which creates a symlink
4864 but (unlike "L") deletes a pre-existing file first, should
4865 it already exist and not already be the correct
4866 symlink. Similarly, "b+", "c+" and "p+" directives have been
4867 added as well, which create block and character devices, as
4868 well as fifos in the filesystem, possibly removing any
4869 pre-existing files of different types.
4871 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4872 'argument' field (which so far specified the source to
4873 symlink/copy the files from) is now optional. If omitted the
4874 same file os copied from /usr/share/factory/ suffixed by the
4875 full destination path. This is useful for populating /etc
4876 with essential files, by copying them from vendor defaults
4877 shipped in /usr/share/factory/etc.
4879 * A new command "systemctl preset-all" has been added that
4880 applies the service preset settings to all installed unit
4881 files. A new switch --preset-mode= has been added that
4882 controls whether only enable or only disable operations
4885 * A new command "systemctl is-system-running" has been added
4886 that allows checking the overall state of the system, for
4887 example whether it is fully up and running.
4889 * When the system boots up with an empty /etc, the equivalent
4890 to "systemctl preset-all" is executed during early boot, to
4891 make sure all default services are enabled after a factory
4894 * systemd now contains a minimal preset file that enables the
4895 most basic services systemd ships by default.
4897 * Unit files' [Install] section gained a new DefaultInstance=
4898 field for defining the default instance to create if a
4899 template unit is enabled with no instance specified.
4901 * A new passive target cryptsetup-pre.target has been added
4902 that may be used by services that need to make they run and
4903 finish before the first LUKS cryptographic device is set up.
4905 * The /dev/loop-control and /dev/btrfs-control device nodes
4906 are now owned by the "disk" group by default, opening up
4907 access to this group.
4909 * systemd-coredump will now automatically generate a
4910 stack trace of all core dumps taking place on the system,
4911 based on elfutils' libdw library. This stack trace is logged
4914 * systemd-coredump may now optionally store coredumps directly
4915 on disk (in /var/lib/systemd/coredump, possibly compressed),
4916 instead of storing them unconditionally in the journal. This
4917 mode is the new default. A new configuration file
4918 /etc/systemd/coredump.conf has been added to configure this
4919 and other parameters of systemd-coredump.
4921 * coredumpctl gained a new "info" verb to show details about a
4922 specific coredump. A new switch "-1" has also been added
4923 that makes sure to only show information about the most
4924 recent entry instead of all entries. Also, as the tool is
4925 generally useful now the "systemd-" prefix of the binary
4926 name has been removed. Distributions that want to maintain
4927 compatibility with the old name should add a symlink from
4928 the old name to the new name.
4930 * journald's SplitMode= now defaults to "uid". This makes sure
4931 that unprivileged users can access their own coredumps with
4932 coredumpctl without restrictions.
4934 * New kernel command line options "systemd.wants=" (for
4935 pulling an additional unit during boot), "systemd.mask="
4936 (for masking a specific unit for the boot), and
4937 "systemd.debug-shell" (for enabling the debug shell on tty9)
4938 have been added. This is implemented in the new generator
4939 "systemd-debug-generator".
4941 * systemd-nspawn will now by default filter a couple of
4942 syscalls for containers, among them those required for
4943 kernel module loading, direct x86 IO port access, swap
4944 management, and kexec. Most importantly though
4945 open_by_handle_at() is now prohibited for containers,
4946 closing a hole similar to a recently discussed vulnerability
4947 in docker regarding access to files on file hierarchies the
4948 container should normally not have access to. Note that, for
4949 nspawn, we generally make no security claims anyway (and
4950 this is explicitly documented in the man page), so this is
4951 just a fix for one of the most obvious problems.
4953 * A new man page file-hierarchy(7) has been added that
4954 contains a minimized, modernized version of the file system
4955 layout systemd expects, similar in style to the FHS
4956 specification or hier(5). A new tool systemd-path(1) has
4957 been added to query many of these paths for the local
4960 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4961 longer done. Since the directory now has a per-user size
4962 limit, and is cleaned on logout this appears unnecessary,
4963 in particular since this now brings the lifecycle of this
4964 directory closer in line with how IPC objects are handled.
4966 * systemd.pc now exports a number of additional directories,
4967 including $libdir (which is useful to identify the library
4968 path for the primary architecture of the system), and a
4969 couple of drop-in directories.
4971 * udev's predictable network interface names now use the dev_port
4972 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4973 distinguish between ports of the same PCI function. dev_id should
4974 only be used for ports using the same HW address, hence the need
4977 * machined has been updated to export the OS version of a
4978 container (read from /etc/os-release and
4979 /usr/lib/os-release) on the bus. This is now shown in
4980 "machinectl status" for a machine.
4982 * A new service setting RestartForceExitStatus= has been
4983 added. If configured to a set of exit signals or process
4984 return values, the service will be restarted when the main
4985 daemon process exits with any of them, regardless of the
4988 * systemctl's -H switch for connecting to remote systemd
4989 machines has been extended so that it may be used to
4990 directly connect to a specific container on the
4991 host. "systemctl -H root@foobar:waldi" will now connect as
4992 user "root" to host "foobar", and then proceed directly to
4993 the container named "waldi". Note that currently you have to
4994 authenticate as user "root" for this to work, as entering
4995 containers is a privileged operation.
4997 Contributions from: Andreas Henriksson, Benjamin Steinwender,
4998 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
4999 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5000 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5001 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5002 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5003 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5004 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5005 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5006 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5007 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5008 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5010 — Berlin, 2014-07-03
5014 * As an experimental feature, udev now tries to lock the
5015 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5016 executes events for the disk or any of its partitions.
5017 Applications like partitioning programs can lock the
5018 disk device node (flock(LOCK_EX)) and claim temporary
5019 device ownership that way; udev will entirely skip all event
5020 handling for this disk and its partitions. If the disk
5021 was opened for writing, the close will trigger a partition
5022 table rescan in udev's "watch" facility, and if needed
5023 synthesize "change" events for the disk and all its partitions.
5024 This is now unconditionally enabled, and if it turns out to
5025 cause major problems, we might turn it on only for specific
5026 devices, or might need to disable it entirely. Device Mapper
5027 devices are excluded from this logic.
5029 * We temporarily dropped the "-l" switch for fsck invocations,
5030 since they collide with the flock() logic above. util-linux
5031 upstream has been changed already to avoid this conflict,
5032 and we will readd "-l" as soon as util-linux with this
5033 change has been released.
5035 * The dependency on libattr has been removed. Since a long
5036 time, the extended attribute calls have moved to glibc, and
5037 libattr is thus unnecessary.
5039 * Virtualization detection works without privileges now. This
5040 means the systemd-detect-virt binary no longer requires
5041 CAP_SYS_PTRACE file capabilities, and our daemons can run
5042 with fewer privileges.
5044 * systemd-networkd now runs under its own "systemd-network"
5045 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5046 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5047 loses the ability to write to files owned by root this way.
5049 * Similarly, systemd-resolved now runs under its own
5050 "systemd-resolve" user with no capabilities remaining.
5052 * Similarly, systemd-bus-proxyd now runs under its own
5053 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5055 * systemd-networkd gained support for setting up "veth"
5056 virtual Ethernet devices for container connectivity, as well
5057 as GRE and VTI tunnels.
5059 * systemd-networkd will no longer automatically attempt to
5060 manually load kernel modules necessary for certain tunnel
5061 transports. Instead, it is assumed the kernel loads them
5062 automatically when required. This only works correctly on
5063 very new kernels. On older kernels, please consider adding
5064 the kernel modules to /etc/modules-load.d/ as a work-around.
5066 * The resolv.conf file systemd-resolved generates has been
5067 moved to /run/systemd/resolve/. If you have a symlink from
5068 /etc/resolv.conf, it might be necessary to correct it.
5070 * Two new service settings, ProtectHome= and ProtectSystem=,
5071 have been added. When enabled, they will make the user data
5072 (such as /home) inaccessible or read-only and the system
5073 (such as /usr) read-only, for specific services. This allows
5074 very light-weight per-service sandboxing to avoid
5075 modifications of user data or system files from
5076 services. These two new switches have been enabled for all
5077 of systemd's long-running services, where appropriate.
5079 * Socket units gained new SocketUser= and SocketGroup=
5080 settings to set the owner user and group of AF_UNIX sockets
5081 and FIFOs in the file system.
5083 * Socket units gained a new RemoveOnStop= setting. If enabled,
5084 all FIFOS and sockets in the file system will be removed
5085 when the specific socket unit is stopped.
5087 * Socket units gained a new Symlinks= setting. It takes a list
5088 of symlinks to create to file system sockets or FIFOs
5089 created by the specific Unix sockets. This is useful to
5090 manage symlinks to socket nodes with the same life-cycle as
5093 * The /dev/log socket and /dev/initctl FIFO have been moved to
5094 /run, and have been replaced by symlinks. This allows
5095 connecting to these facilities even if PrivateDevices=yes is
5096 used for a service (which makes /dev/log itself unavailable,
5097 but /run is left). This also has the benefit of ensuring
5098 that /dev only contains device nodes, directories and
5099 symlinks, and nothing else.
5101 * sd-daemon gained two new calls sd_pid_notify() and
5102 sd_pid_notifyf(). They are similar to sd_notify() and
5103 sd_notifyf(), but allow overriding of the source PID of
5104 notification messages if permissions permit this. This is
5105 useful to send notify messages on behalf of a different
5106 process (for example, the parent process). The
5107 systemd-notify tool has been updated to make use of this
5108 when sending messages (so that notification messages now
5109 originate from the shell script invoking systemd-notify and
5110 not the systemd-notify process itself. This should minimize
5111 a race where systemd fails to associate notification
5112 messages to services when the originating process already
5115 * A new "on-abnormal" setting for Restart= has been added. If
5116 set, it will result in automatic restarts on all "abnormal"
5117 reasons for a process to exit, which includes unclean
5118 signals, core dumps, timeouts and watchdog timeouts, but
5119 does not include clean and unclean exit codes or clean
5120 signals. Restart=on-abnormal is an alternative for
5121 Restart=on-failure for services that shall be able to
5122 terminate and avoid restarts on certain errors, by
5123 indicating so with an unclean exit code. Restart=on-failure
5124 or Restart=on-abnormal is now the recommended setting for
5125 all long-running services.
5127 * If the InaccessibleDirectories= service setting points to a
5128 mount point (or if there are any submounts contained within
5129 it), it is now attempted to completely unmount it, to make
5130 the file systems truly unavailable for the respective
5133 * The ReadOnlyDirectories= service setting and
5134 systemd-nspawn's --read-only parameter are now recursively
5135 applied to all submounts, too.
5137 * Mount units may now be created transiently via the bus APIs.
5139 * The support for SysV and LSB init scripts has been removed
5140 from the systemd daemon itself. Instead, it is now
5141 implemented as a generator that creates native systemd units
5142 from these scripts when needed. This enables us to remove a
5143 substantial amount of legacy code from PID 1, following the
5144 fact that many distributions only ship a very small number
5145 of LSB/SysV init scripts nowadays.
5147 * Privileged Xen (dom0) domains are not considered
5148 virtualization anymore by the virtualization detection
5149 logic. After all, they generally have unrestricted access to
5150 the hardware and usually are used to manage the unprivileged
5153 * systemd-tmpfiles gained a new "C" line type, for copying
5154 files or entire directories.
5156 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
5157 lines. So far, they have been non-globbing versions of the
5158 latter, and have thus been redundant. In future, it is
5159 recommended to only use "z". "m" has hence been removed
5160 from the documentation, even though it stays supported.
5162 * A tmpfiles snippet to recreate the most basic structure in
5163 /var has been added. This is enough to create the /var/run →
5164 /run symlink and create a couple of structural
5165 directories. This allows systems to boot up with an empty or
5166 volatile /var. Of course, while with this change, the core OS
5167 now is capable with dealing with a volatile /var, not all
5168 user services are ready for it. However, we hope that sooner
5169 or later, many service daemons will be changed upstream so
5170 that they are able to automatically create their necessary
5171 directories in /var at boot, should they be missing. This is
5172 the first step to allow state-less systems that only require
5173 the vendor image for /usr to boot.
5175 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5176 empty tmpfs instance to a specific directory. This is
5177 particularly useful for making use of the automatic
5178 reconstruction of /var (see above), by passing --tmpfs=/var.
5180 * Access modes specified in tmpfiles snippets may now be
5181 prefixed with "~", which indicates that they shall be masked
5182 by whether the existing file or directory is currently
5183 writable, readable or executable at all. Also, if specified,
5184 the sgid/suid/sticky bits will be masked for all
5187 * A new passive target unit "network-pre.target" has been
5188 added which is useful for services that shall run before any
5189 network is configured, for example firewall scripts.
5191 * The "floppy" group that previously owned the /dev/fd*
5192 devices is no longer used. The "disk" group is now used
5193 instead. Distributions should probably deprecate usage of
5196 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5197 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5198 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5199 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5200 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5201 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5204 — Berlin, 2014-06-11
5208 * A new "systemd-timesyncd" daemon has been added for
5209 synchronizing the system clock across the network. It
5210 implements an SNTP client. In contrast to NTP
5211 implementations such as chrony or the NTP reference server,
5212 this only implements a client side, and does not bother with
5213 the full NTP complexity, focusing only on querying time from
5214 one remote server and synchronizing the local clock to
5215 it. Unless you intend to serve NTP to networked clients or
5216 want to connect to local hardware clocks, this simple NTP
5217 client should be more than appropriate for most
5218 installations. The daemon runs with minimal privileges, and
5219 has been hooked up with networkd to only operate when
5220 network connectivity is available. The daemon saves the
5221 current clock to disk every time a new NTP sync has been
5222 acquired, and uses this to possibly correct the system clock
5223 early at bootup, in order to accommodate for systems that
5224 lack an RTC such as the Raspberry Pi and embedded devices,
5225 and to make sure that time monotonically progresses on these
5226 systems, even if it is not always correct. To make use of
5227 this daemon, a new system user and group "systemd-timesync"
5228 needs to be created on installation of systemd.
5230 * The queue "seqnum" interface of libudev has been disabled, as
5231 it was generally incompatible with device namespacing as
5232 sequence numbers of devices go "missing" if the devices are
5233 part of a different namespace.
5235 * "systemctl list-timers" and "systemctl list-sockets" gained
5236 a --recursive switch for showing units of these types also
5237 for all local containers, similar in style to the already
5238 supported --recursive switch for "systemctl list-units".
5240 * A new RebootArgument= setting has been added for service
5241 units, which may be used to specify a kernel reboot argument
5242 to use when triggering reboots with StartLimitAction=.
5244 * A new FailureAction= setting has been added for service
5245 units which may be used to specify an operation to trigger
5246 when a service fails. This works similarly to
5247 StartLimitAction=, but unlike it, controls what is done
5248 immediately rather than only after several attempts to
5249 restart the service in question.
5251 * hostnamed got updated to also expose the kernel name,
5252 release, and version on the bus. This is useful for
5253 executing commands like hostnamectl with the -H switch.
5254 systemd-analyze makes use of this to properly display
5255 details when running non-locally.
5257 * The bootchart tool can now show cgroup information in the
5258 graphs it generates.
5260 * The CFS CPU quota cgroup attribute is now exposed for
5261 services. The new CPUQuota= switch has been added for this
5262 which takes a percentage value. Setting this will have the
5263 result that a service may never get more CPU time than the
5264 specified percentage, even if the machine is otherwise idle.
5266 * systemd-networkd learned IPIP and SIT tunnel support.
5268 * LSB init scripts exposing a dependency on $network will now
5269 get a dependency on network-online.target rather than simply
5270 network.target. This should bring LSB handling closer to
5271 what it was on SysV systems.
5273 * A new fsck.repair= kernel option has been added to control
5274 how fsck shall deal with unclean file systems at boot.
5276 * The (.ini) configuration file parser will now silently
5277 ignore sections whose name begins with "X-". This may be
5278 used to maintain application-specific extension sections in unit
5281 * machined gained a new API to query the IP addresses of
5282 registered containers. "machinectl status" has been updated
5283 to show these addresses in its output.
5285 * A new call sd_uid_get_display() has been added to the
5286 sd-login APIs for querying the "primary" session of a
5287 user. The "primary" session of the user is elected from the
5288 user's sessions and generally a graphical session is
5289 preferred over a text one.
5291 * A minimal systemd-resolved daemon has been added. It
5292 currently simply acts as a companion to systemd-networkd and
5293 manages resolv.conf based on per-interface DNS
5294 configuration, possibly supplied via DHCP. In the long run
5295 we hope to extend this into a local DNSSEC enabled DNS and
5298 * The systemd-networkd-wait-online tool is now enabled by
5299 default. It will delay network-online.target until a network
5300 connection has been configured. The tool primarily integrates
5301 with networkd, but will also make a best effort to make sense
5302 of network configuration performed in some other way.
5304 * Two new service options StartupCPUShares= and
5305 StartupBlockIOWeight= have been added that work similarly to
5306 CPUShares= and BlockIOWeight= however only apply during
5307 system startup. This is useful to prioritize certain services
5308 differently during bootup than during normal runtime.
5310 * hostnamed has been changed to prefer the statically
5311 configured hostname in /etc/hostname (unless set to
5312 'localhost' or empty) over any dynamic one supplied by
5313 dhcp. With this change, the rules for picking the hostname
5314 match more closely the rules of other configuration settings
5315 where the local administrator's configuration in /etc always
5316 overrides any other settings.
5318 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
5319 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5320 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5321 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5322 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5323 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5324 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5325 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5326 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
5327 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5328 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5329 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5330 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5331 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5332 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5333 Lindskog, WaLyong Cho, Will Woods, Zbigniew
5336 — Beijing, 2014-05-28
5340 * When restoring the screen brightness at boot, stay away from
5341 the darkest setting or from the lowest 5% of the available
5342 range, depending on which is the larger value of both. This
5343 should effectively protect the user from rebooting into a
5344 black screen, should the brightness have been set to minimum
5347 * sd-login gained a new sd_machine_get_class() call to
5348 determine the class ("vm" or "container") of a machine
5349 registered with machined.
5351 * sd-login gained new calls
5352 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5353 to query the identity of the peer of a local AF_UNIX
5354 connection. They operate similarly to their sd_pid_get_xyz()
5357 * PID 1 will now maintain a system-wide system state engine
5358 with the states "starting", "running", "degraded",
5359 "maintenance", "stopping". These states are bound to system
5360 startup, normal runtime, runtime with at least one failed
5361 service, rescue/emergency mode and system shutdown. This
5362 state is shown in the "systemctl status" output when no unit
5363 name is passed. It is useful to determine system state, in
5364 particularly when doing so for many systems or containers at
5367 * A new command "list-machines" has been added to "systemctl"
5368 that lists all local OS containers and shows their system
5369 state (see above), if systemd runs inside of them.
5371 * systemctl gained a new "-r" switch to recursively enumerate
5372 units on all local containers, when used with the
5373 "list-unit" command (which is the default one that is
5374 executed when no parameters are specified).
5376 * The GPT automatic partition discovery logic will now honour
5377 two GPT partition flags: one may be set on a partition to
5378 cause it to be mounted read-only, and the other may be set
5379 on a partition to ignore it during automatic discovery.
5381 * Two new GPT type UUIDs have been added for automatic root
5382 partition discovery, for 32-bit and 64-bit ARM. This is not
5383 particularly useful for discovering the root directory on
5384 these architectures during bare-metal boots (since UEFI is
5385 not common there), but still very useful to allow booting of
5386 ARM disk images in nspawn with the -i option.
5388 * MAC addresses of interfaces created with nspawn's
5389 --network-interface= switch will now be generated from the
5390 machine name, and thus be stable between multiple invocations
5393 * logind will now automatically remove all IPC objects owned
5394 by a user if she or he fully logs out. This makes sure that
5395 users who are logged out cannot continue to consume IPC
5396 resources. This covers SysV memory, semaphores and message
5397 queues as well as POSIX shared memory and message
5398 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5399 limits. With this functionality, that is corrected. This may
5400 be turned off by using the RemoveIPC= switch of logind.conf.
5402 * The systemd-machine-id-setup and tmpfiles tools gained a
5403 --root= switch to operate on a specific root directory,
5406 * journald can now forward logged messages to the TTYs of all
5407 logged in users ("wall"). This is the default for all
5408 emergency messages now.
5410 * A new tool systemd-journal-remote has been added to stream
5411 journal log messages across the network.
5413 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5414 controller trees are mounted into it. Note that the
5415 directories mounted beneath it are not read-only. This is a
5416 security measure and is particularly useful because glibc
5417 actually includes a search logic to pick any tmpfs it can
5418 find to implement shm_open() if /dev/shm is not available
5419 (which it might very well be in namespaced setups).
5421 * machinectl gained a new "poweroff" command to cleanly power
5422 down a local OS container.
5424 * The PrivateDevices= unit file setting will now also drop the
5425 CAP_MKNOD capability from the capability bound set, and
5426 imply DevicePolicy=closed.
5428 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5429 comprehensively on all long-running systemd services where
5430 this is appropriate.
5432 * systemd-udevd will now run in a disassociated mount
5433 namespace. To mount directories from udev rules, make sure to
5434 pull in mount units via SYSTEMD_WANTS properties.
5436 * The kdbus support gained support for uploading policy into
5437 the kernel. sd-bus gained support for creating "monitoring"
5438 connections that can eavesdrop into all bus communication
5439 for debugging purposes.
5441 * Timestamps may now be specified in seconds since the UNIX
5442 epoch Jan 1st, 1970 by specifying "@" followed by the value
5445 * Native tcpwrap support in systemd has been removed. tcpwrap
5446 is old code, not really maintained anymore and has serious
5447 shortcomings, and better options such as firewalls
5448 exist. For setups that require tcpwrap usage, please
5449 consider invoking your socket-activated service via tcpd,
5450 like on traditional inetd.
5452 * A new system.conf configuration option
5453 DefaultTimerAccuracySec= has been added that controls the
5454 default AccuracySec= setting of .timer units.
5456 * Timer units gained a new WakeSystem= switch. If enabled,
5457 timers configured this way will cause the system to resume
5458 from system suspend (if the system supports that, which most
5461 * Timer units gained a new Persistent= switch. If enabled,
5462 timers configured this way will save to disk when they have
5463 been last triggered. This information is then used on next
5464 reboot to possible execute overdue timer events, that
5465 could not take place because the system was powered off.
5466 This enables simple anacron-like behaviour for timer units.
5468 * systemctl's "list-timers" will now also list the time a
5469 timer unit was last triggered in addition to the next time
5470 it will be triggered.
5472 * systemd-networkd will now assign predictable IPv4LL
5473 addresses to its local interfaces.
5475 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5476 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5477 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5478 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5479 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5480 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5481 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5482 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5485 — Berlin, 2014-03-25
5489 * A new unit file setting RestrictAddressFamilies= has been
5490 added to restrict which socket address families unit
5491 processes gain access to. This takes address family names
5492 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5493 attack surface of services via exotic protocol stacks. This
5494 is built on seccomp system call filters.
5496 * Two new unit file settings RuntimeDirectory= and
5497 RuntimeDirectoryMode= have been added that may be used to
5498 manage a per-daemon runtime directories below /run. This is
5499 an alternative for setting up directory permissions with
5500 tmpfiles snippets, and has the advantage that the runtime
5501 directory's lifetime is bound to the daemon runtime and that
5502 the daemon starts up with an empty directory each time. This
5503 is particularly useful when writing services that drop
5504 privileges using the User= or Group= setting.
5506 * The DeviceAllow= unit setting now supports globbing for
5507 matching against device group names.
5509 * The systemd configuration file system.conf gained new
5510 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5511 DefaultMemoryAccounting= to globally turn on/off accounting
5512 for specific resources (cgroups) for all units. These
5513 settings may still be overridden individually in each unit
5516 * systemd-gpt-auto-generator is now able to discover /srv and
5517 root partitions in addition to /home and swap partitions. It
5518 also supports LUKS-encrypted partitions now. With this in
5519 place, automatic discovery of partitions to mount following
5520 the Discoverable Partitions Specification
5521 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
5522 is now a lot more complete. This allows booting without
5523 /etc/fstab and without root= on the kernel command line on
5524 systems prepared appropriately.
5526 * systemd-nspawn gained a new --image= switch which allows
5527 booting up disk images and Linux installations on any block
5528 device that follow the Discoverable Partitions Specification
5529 (see above). This means that installations made with
5530 appropriately updated installers may now be started and
5531 deployed using container managers, completely
5532 unmodified. (We hope that libvirt-lxc will add support for
5533 this feature soon, too.)
5535 * systemd-nspawn gained a new --network-macvlan= setting to
5536 set up a private macvlan interface for the
5537 container. Similarly, systemd-networkd gained a new
5538 Kind=macvlan setting in .netdev files.
5540 * systemd-networkd now supports configuring local addresses
5543 * A new tool systemd-network-wait-online has been added to
5544 synchronously wait for network connectivity using
5547 * The sd-bus.h bus API gained a new sd_bus_track object for
5548 tracking the life-cycle of bus peers. Note that sd-bus.h is
5549 still not a public API though (unless you specify
5550 --enable-kdbus on the configure command line, which however
5551 voids your warranty and you get no API stability guarantee).
5553 * The $XDG_RUNTIME_DIR runtime directories for each user are
5554 now individual tmpfs instances, which has the benefit of
5555 introducing separate pools for each user, with individual
5556 size limits, and thus making sure that unprivileged clients
5557 can no longer negatively impact the system or other users by
5558 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5559 RuntimeDirectorySize= has been introduced that allows
5560 controlling the default size limit for all users. It
5561 defaults to 10% of the available physical memory. This is no
5562 replacement for quotas on tmpfs though (which the kernel
5563 still does not support), as /dev/shm and /tmp are still
5564 shared resources used by both the system and unprivileged
5567 * logind will now automatically turn off automatic suspending
5568 on laptop lid close when more than one display is
5569 connected. This was previously expected to be implemented
5570 individually in desktop environments (such as GNOME),
5571 however has been added to logind now, in order to fix a
5572 boot-time race where a desktop environment might not have
5573 been started yet and thus not been able to take an inhibitor
5574 lock at the time where logind already suspends the system
5575 due to a closed lid.
5577 * logind will now wait at least 30s after each system
5578 suspend/resume cycle, and 3min after system boot before
5579 suspending the system due to a closed laptop lid. This
5580 should give USB docking stations and similar enough time to
5581 be probed and configured after system resume and boot in
5582 order to then act as suspend blocker.
5584 * systemd-run gained a new --property= setting which allows
5585 initialization of resource control properties (and others)
5586 for the created scope or service unit. Example: "systemd-run
5587 --property=BlockIOWeight=10 updatedb" may be used to run
5588 updatedb at a low block IO scheduling weight.
5590 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5591 now also work in --scope mode.
5593 * When systemd is compiled with kdbus support, basic support
5594 for enforced policies is now in place. (Note that enabling
5595 kdbus still voids your warranty and no API compatibility
5598 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5599 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5600 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5601 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5602 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5603 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5604 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5605 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5606 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5607 Zbigniew Jędrzejewski-Szmek
5609 — Berlin, 2014-03-12
5613 * systemd will now relabel /dev after loading the SMACK policy
5614 according to SMACK rules.
5616 * A new unit file option AppArmorProfile= has been added to
5617 set the AppArmor profile for the processes of a unit.
5619 * A new condition check ConditionArchitecture= has been added
5620 to conditionalize units based on the system architecture, as
5621 reported by uname()'s "machine" field.
5623 * systemd-networkd now supports matching on the system
5624 virtualization, architecture, kernel command line, host name
5627 * logind is now a lot more aggressive when suspending the
5628 machine due to a closed laptop lid. Instead of acting only
5629 on the lid close action, it will continuously watch the lid
5630 status and act on it. This is useful for laptops where the
5631 power button is on the outside of the chassis so that it can
5632 be reached without opening the lid (such as the Lenovo
5633 Yoga). On those machines, logind will now immediately
5634 re-suspend the machine if the power button has been
5635 accidentally pressed while the laptop was suspended and in a
5636 backpack or similar.
5638 * logind will now watch SW_DOCK switches and inhibit reaction
5639 to the lid switch if it is pressed. This means that logind
5640 will not suspend the machine anymore if the lid is closed
5641 and the system is docked, if the laptop supports SW_DOCK
5642 notifications via the input layer. Note that ACPI docking
5643 stations do not generate this currently. Also note that this
5644 logic is usually not fully sufficient and Desktop
5645 Environments should take a lid switch inhibitor lock when an
5646 external display is connected, as systemd will not watch
5649 * nspawn will now make use of the devices cgroup controller by
5650 default, and only permit creation of and access to the usual
5651 API device nodes like /dev/null or /dev/random, as well as
5652 access to (but not creation of) the pty devices.
5654 * We will now ship a default .network file for
5655 systemd-networkd that automatically configures DHCP for
5656 network interfaces created by nspawn's --network-veth or
5657 --network-bridge= switches.
5659 * systemd will now understand the usual M, K, G, T suffixes
5660 according to SI conventions (i.e. to the base 1000) when
5661 referring to throughput and hardware metrics. It will stay
5662 with IEC conventions (i.e. to the base 1024) for software
5663 metrics, according to what is customary according to
5664 Wikipedia. We explicitly document which base applies for
5665 each configuration option.
5667 * The DeviceAllow= setting in unit files now supports a syntax
5668 to whitelist an entire group of devices node majors at once,
5669 based on the /proc/devices listing. For example, with the
5670 string "char-pts", it is now possible to whitelist all
5671 current and future pseudo-TTYs at once.
5673 * sd-event learned a new "post" event source. Event sources of
5674 this type are triggered by the dispatching of any event
5675 source of a type that is not "post". This is useful for
5676 implementing clean-up and check event sources that are
5677 triggered by other work being done in the program.
5679 * systemd-networkd is no longer statically enabled, but uses
5680 the usual [Install] sections so that it can be
5681 enabled/disabled using systemctl. It still is enabled by
5684 * When creating a veth interface pair with systemd-nspawn, the
5685 host side will now be prefixed with "vb-" if
5686 --network-bridge= is used, and with "ve-" if --network-veth
5687 is used. This way, it is easy to distinguish these cases on
5688 the host, for example to apply different configuration to
5689 them with systemd-networkd.
5691 * The compatibility libraries for libsystemd-journal.so,
5692 libsystem-id128.so, libsystemd-login.so and
5693 libsystemd-daemon.so do not make use of IFUNC
5694 anymore. Instead, we now build libsystemd.so multiple times
5695 under these alternative names. This means that the footprint
5696 is drastically increased, but given that these are
5697 transitional compatibility libraries, this should not matter
5698 much. This change has been made necessary to support the ARM
5699 platform for these compatibility libraries, as the ARM
5700 toolchain is not really at the same level as the toolchain
5701 for other architectures like x86 and does not support
5702 IFUNC. Please make sure to use --enable-compat-libs only
5703 during a transitional period!
5705 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
5706 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5707 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5708 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5709 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5710 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5711 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5712 Zbigniew Jędrzejewski-Szmek
5714 — Berlin, 2014-02-24
5718 * A new component "systemd-networkd" has been added that can
5719 be used to configure local network interfaces statically or
5720 via DHCP. It is capable of bringing up bridges, VLANs, and
5721 bonding. Currently, no hook-ups for interactive network
5722 configuration are provided. Use this for your initrd,
5723 container, embedded, or server setup if you need a simple,
5724 yet powerful, network configuration solution. This
5725 configuration subsystem is quite nifty, as it allows wildcard
5726 hotplug matching in interfaces. For example, with a single
5727 configuration snippet, you can configure that all Ethernet
5728 interfaces showing up are automatically added to a bridge,
5729 or similar. It supports link-sensing and more.
5731 * A new tool "systemd-socket-proxyd" has been added which can
5732 act as a bidirectional proxy for TCP sockets. This is
5733 useful for adding socket activation support to services that
5734 do not actually support socket activation, including virtual
5735 machines and the like.
5737 * Add a new tool to save/restore rfkill state on
5740 * Save/restore state of keyboard backlights in addition to
5741 display backlights on shutdown/boot.
5743 * udev learned a new SECLABEL{} construct to label device
5744 nodes with a specific security label when they appear. For
5745 now, only SECLABEL{selinux} is supported, but the syntax is
5746 prepared for additional security frameworks.
5748 * udev gained a new scheme to configure link-level attributes
5749 from files in /etc/systemd/network/*.link. These files can
5750 match against MAC address, device path, driver name and type,
5751 and will apply attributes like the naming policy, link speed,
5752 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
5753 address assignment policy (randomized, ...).
5755 * The configuration of network interface naming rules for
5756 "permanent interface names" has changed: a new NamePolicy=
5757 setting in the [Link] section of .link files determines the
5758 priority of possible naming schemes (onboard, slot, MAC,
5759 path). The default value of this setting is determined by
5760 /usr/lib/net/links/99-default.link. Old
5761 80-net-name-slot.rules udev configuration file has been
5762 removed, so local configuration overriding this file should
5763 be adapted to override 99-default.link instead.
5765 * When the User= switch is used in a unit file, also
5766 initialize $SHELL= based on the user database entry.
5768 * systemd no longer depends on libdbus. All communication is
5769 now done with sd-bus, systemd's low-level bus library
5772 * kdbus support has been added to PID 1 itself. When kdbus is
5773 enabled, this causes PID 1 to set up the system bus and
5774 enable support for a new ".busname" unit type that
5775 encapsulates bus name activation on kdbus. It works a little
5776 bit like ".socket" units, except for bus names. A new
5777 generator has been added that converts classic dbus1 service
5778 activation files automatically into native systemd .busname
5781 * sd-bus: add a light-weight vtable implementation that allows
5782 defining objects on the bus with a simple static const
5783 vtable array of its methods, signals and properties.
5785 * systemd will not generate or install static dbus
5786 introspection data anymore to /usr/share/dbus-1/interfaces,
5787 as the precise format of these files is unclear, and
5788 nothing makes use of it.
5790 * A proxy daemon is now provided to proxy clients connecting
5791 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5792 compatibility with classic D-Bus.
5794 * A bus driver implementation has been added that supports the
5795 classic D-Bus bus driver calls on kdbus, also for
5796 compatibility purposes.
5798 * A new API "sd-event.h" has been added that implements a
5799 minimal event loop API built around epoll. It provides a
5800 couple of features that direct epoll usage is lacking:
5801 prioritization of events, scales to large numbers of timer
5802 events, per-event timer slack (accuracy), system-wide
5803 coalescing of timer events, exit handlers, watchdog
5804 supervision support using systemd's sd_notify() API, child
5807 * A new API "sd-rntl.h" has been added that provides an API
5808 around the route netlink interface of the kernel, similar in
5809 style to "sd-bus.h".
5811 * A new API "sd-dhcp-client.h" has been added that provides a
5812 small DHCPv4 client-side implementation. This is used by
5815 * There is a new kernel command line option
5816 "systemd.restore_state=0|1". When set to "0", none of the
5817 systemd tools will restore saved runtime state to hardware
5818 devices. More specifically, the rfkill and backlight states
5821 * The FsckPassNo= compatibility option in mount/service units
5822 has been removed. The fstab generator will now add the
5823 necessary dependencies automatically, and does not require
5824 PID1's support for that anymore.
5826 * journalctl gained a new switch, --list-boots, that lists
5827 recent boots with their times and boot IDs.
5829 * The various tools like systemctl, loginctl, timedatectl,
5830 busctl, systemd-run, ... have gained a new switch "-M" to
5831 connect to a specific, local OS container (as direct
5832 connection, without requiring SSH). This works on any
5833 container that is registered with machined, such as those
5834 created by libvirt-lxc or nspawn.
5836 * systemd-run and systemd-analyze also gained support for "-H"
5837 to connect to remote hosts via SSH. This is particularly
5838 useful for systemd-run because it enables queuing of jobs
5839 onto remote systems.
5841 * machinectl gained a new command "login" to open a getty
5842 login in any local container. This works with any container
5843 that is registered with machined (such as those created by
5844 libvirt-lxc or nspawn), and which runs systemd inside.
5846 * machinectl gained a new "reboot" command that may be used to
5847 trigger a reboot on a specific container that is registered
5848 with machined. This works on any container that runs an init
5849 system of some kind.
5851 * systemctl gained a new "list-timers" command to print a nice
5852 listing of installed timer units with the times they elapse
5855 * Alternative reboot() parameters may now be specified on the
5856 "systemctl reboot" command line and are passed to the
5857 reboot() system call.
5859 * systemctl gained a new --job-mode= switch to configure the
5860 mode to queue a job with. This is a more generic version of
5861 --fail, --irreversible, and --ignore-dependencies, which are
5862 still available but not advertised anymore.
5864 * /etc/systemd/system.conf gained new settings to configure
5865 various default timeouts of units, as well as the default
5866 start limit interval and burst. These may still be overridden
5869 * PID1 will now export on the bus profile data of the security
5870 policy upload process (such as the SELinux policy upload to
5873 * journald: when forwarding logs to the console, include
5874 timestamps (following the setting in
5875 /sys/module/printk/parameters/time).
5877 * OnCalendar= in timer units now understands the special
5878 strings "yearly" and "annually". (Both are equivalent)
5880 * The accuracy of timer units is now configurable with the new
5881 AccuracySec= setting. It defaults to 1min.
5883 * A new dependency type JoinsNamespaceOf= has been added that
5884 allows running two services within the same /tmp and network
5885 namespace, if PrivateNetwork= or PrivateTmp= are used.
5887 * A new command "cat" has been added to systemctl. It outputs
5888 the original unit file of a unit, and concatenates the
5889 contents of additional "drop-in" unit file snippets, so that
5890 the full configuration is shown.
5892 * systemctl now supports globbing on the various "list-xyz"
5893 commands, like "list-units" or "list-sockets", as well as on
5894 those commands which take multiple unit names.
5896 * journalctl's --unit= switch gained support for globbing.
5898 * All systemd daemons now make use of the watchdog logic so
5899 that systemd automatically notices when they hang.
5901 * If the $container_ttys environment variable is set,
5902 getty-generator will automatically spawn a getty for each
5903 listed tty. This is useful for container managers to request
5904 login gettys to be spawned on as many ttys as needed.
5906 * %h, %s, %U specifier support is not available anymore when
5907 used in unit files for PID 1. This is because NSS calls are
5908 not safe from PID 1. They stay available for --user
5909 instances of systemd, and as special case for the root user.
5911 * loginctl gained a new "--no-legend" switch to turn off output
5914 * The "sd-login.h" API gained three new calls:
5915 sd_session_is_remote(), sd_session_get_remote_user(),
5916 sd_session_get_remote_host() to query information about
5919 * The udev hardware database now also carries vendor/product
5920 information of SDIO devices.
5922 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5923 determine whether watchdog notifications are requested by
5926 * Socket-activated per-connection services now include a
5927 short description of the connection parameters in the
5930 * tmpfiles gained a new "--boot" option. When this is not used,
5931 only lines where the command character is not suffixed with
5932 "!" are executed. When this option is specified, those
5933 options are executed too. This partitions tmpfiles
5934 directives into those that can be safely executed at any
5935 time, and those which should be run only at boot (for
5936 example, a line that creates /run/nologin).
5938 * A new API "sd-resolve.h" has been added which provides a simple
5939 asynchronous wrapper around glibc NSS host name resolution
5940 calls, such as getaddrinfo(). In contrast to glibc's
5941 getaddrinfo_a(), it does not use signals. In contrast to most
5942 other asynchronous name resolution libraries, this one does
5943 not reimplement DNS, but reuses NSS, so that alternate
5944 host name resolution systems continue to work, such as mDNS,
5945 LDAP, etc. This API is based on libasyncns, but it has been
5946 cleaned up for inclusion in systemd.
5948 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5949 "sd-daemon.h" are no longer found in individual libraries
5950 libsystemd-journal.so, libsystemd-login.so,
5951 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
5952 merged them into a single library, libsystemd.so, which
5953 provides all symbols. The reason for this is cyclic
5954 dependencies, as these libraries tend to use each other's
5955 symbols. So far, we have managed to workaround that by linking
5956 a copy of a good part of our code into each of these
5957 libraries again and again, which, however, makes certain
5958 things hard to do, like sharing static variables. Also, it
5959 substantially increases footprint. With this change, there
5960 is only one library for the basic APIs systemd
5961 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5962 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5963 library as well, however are subject to the --enable-kdbus
5964 switch (see below). Note that "sd-dhcp-client.h" is not part
5965 of this library (this is because it only consumes, never
5966 provides, services of/to other APIs). To make the transition
5967 easy from the separate libraries to the unified one, we
5968 provide the --enable-compat-libs compile-time switch which
5969 will generate stub libraries that are compatible with the
5970 old ones but redirect all calls to the new one.
5972 * All of the kdbus logic and the new APIs "sd-bus.h",
5973 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
5974 and "sd-utf8.h" are compile-time optional via the
5975 "--enable-kdbus" switch, and they are not compiled in by
5976 default. To make use of kdbus, you have to explicitly enable
5977 the switch. Note however, that neither the kernel nor the
5978 userspace API for all of this is considered stable yet. We
5979 want to maintain the freedom to still change the APIs for
5980 now. By specifying this build-time switch, you acknowledge
5981 that you are aware of the instability of the current
5984 * Also, note that while kdbus is pretty much complete,
5985 it lacks one thing: proper policy support. This means you
5986 can build a fully working system with all features; however,
5987 it will be highly insecure. Policy support will be added in
5988 one of the next releases, at the same time that we will
5989 declare the APIs stable.
5991 * When the kernel command line argument "kdbus" is specified,
5992 systemd will automatically load the kdbus.ko kernel module. At
5993 this stage of development, it is only useful for testing kdbus
5994 and should not be used in production. Note: if "--enable-kdbus"
5995 is specified, and the kdbus.ko kernel module is available, and
5996 "kdbus" is added to the kernel command line, the entire system
5997 runs with kdbus instead of dbus-daemon, with the above mentioned
5998 problem of missing the system policy enforcement. Also a future
5999 version of kdbus.ko or a newer systemd will not be compatible with
6000 each other, and will unlikely be able to boot the machine if only
6001 one of them is updated.
6003 * systemctl gained a new "import-environment" command which
6004 uploads the caller's environment (or parts thereof) into the
6005 service manager so that it is inherited by services started
6006 by the manager. This is useful to upload variables like
6007 $DISPLAY into the user service manager.
6009 * A new PrivateDevices= switch has been added to service units
6010 which allows running a service with a namespaced /dev
6011 directory that does not contain any device nodes for
6012 physical devices. More specifically, it only includes devices
6013 such as /dev/null, /dev/urandom, and /dev/zero which are API
6016 * logind has been extended to support behaviour like VT
6017 switching on seats that do not support a VT. This makes
6018 multi-session available on seats that are not the first seat
6019 (seat0), and on systems where kernel support for VTs has
6020 been disabled at compile-time.
6022 * If a process holds a delay lock for system sleep or shutdown
6023 and fails to release it in time, we will now log its
6024 identity. This makes it easier to identify processes that
6025 cause slow suspends or power-offs.
6027 * When parsing /etc/crypttab, support for a new key-slot=
6028 option as supported by Debian is added. It allows indicating
6029 which LUKS slot to use on disk, speeding up key loading.
6031 * The sd_journal_sendv() API call has been checked and
6032 officially declared to be async-signal-safe so that it may
6033 be invoked from signal handlers for logging purposes.
6035 * Boot-time status output is now enabled automatically after a
6036 short timeout if boot does not progress, in order to give
6037 the user an indication what she or he is waiting for.
6039 * The boot-time output has been improved to show how much time
6040 remains until jobs expire.
6042 * The KillMode= switch in service units gained a new possible
6043 value "mixed". If set, and the unit is shut down, then the
6044 initial SIGTERM signal is sent only to the main daemon
6045 process, while the following SIGKILL signal is sent to
6046 all remaining processes of the service.
6048 * When a scope unit is registered, a new property "Controller"
6049 may be set. If set to a valid bus name, systemd will send a
6050 RequestStop() signal to this name when it would like to shut
6051 down the scope. This may be used to hook manager logic into
6052 the shutdown logic of scope units. Also, scope units may now
6053 be put in a special "abandoned" state, in which case the
6054 manager process which created them takes no further
6055 responsibilities for it.
6057 * When reading unit files, systemd will now verify
6058 the access mode of these files, and warn about certain
6059 suspicious combinations. This has been added to make it
6060 easier to track down packaging bugs where unit files are
6061 marked executable or world-writable.
6063 * systemd-nspawn gained a new "--setenv=" switch to set
6064 container-wide environment variables. The similar option in
6065 systemd-activate was renamed from "--environment=" to
6066 "--setenv=" for consistency.
6068 * systemd-nspawn has been updated to create a new kdbus domain
6069 for each container that is invoked, thus allowing each
6070 container to have its own set of system and user buses,
6071 independent of the host.
6073 * systemd-nspawn gained a new --drop-capability= switch to run
6074 the container with less capabilities than the default. Both
6075 --drop-capability= and --capability= now take the special
6076 string "all" for dropping or keeping all capabilities.
6078 * systemd-nspawn gained new switches for executing containers
6079 with specific SELinux labels set.
6081 * systemd-nspawn gained a new --quiet switch to not generate
6082 any additional output but the container's own console
6085 * systemd-nspawn gained a new --share-system switch to run a
6086 container without PID namespacing enabled.
6088 * systemd-nspawn gained a new --register= switch to control
6089 whether the container is registered with systemd-machined or
6090 not. This is useful for containers that do not run full
6091 OS images, but only specific apps.
6093 * systemd-nspawn gained a new --keep-unit which may be used
6094 when invoked as the only program from a service unit, and
6095 results in registration of the unit service itself in
6096 systemd-machined, instead of a newly opened scope unit.
6098 * systemd-nspawn gained a new --network-interface= switch for
6099 moving arbitrary interfaces to the container. The new
6100 --network-veth switch creates a virtual Ethernet connection
6101 between host and container. The new --network-bridge=
6102 switch then allows assigning the host side of this virtual
6103 Ethernet connection to a bridge device.
6105 * systemd-nspawn gained a new --personality= switch for
6106 setting the kernel personality for the container. This is
6107 useful when running a 32-bit container on a 64-bit host. A
6108 similar option Personality= is now also available for service
6111 * logind will now also track a "Desktop" identifier for each
6112 session which encodes the desktop environment of it. This is
6113 useful for desktop environments that want to identify
6114 multiple running sessions of itself easily.
6116 * A new SELinuxContext= setting for service units has been
6117 added that allows setting a specific SELinux execution
6118 context for a service.
6120 * Most systemd client tools will now honour $SYSTEMD_LESS for
6121 settings of the "less" pager. By default, these tools will
6122 override $LESS to allow certain operations to work, such as
6123 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6124 influence this logic.
6126 * systemd's "seccomp" hook-up has been changed to make use of
6127 the libseccomp library instead of using its own
6128 implementation. This has benefits for portability among
6131 * For usage together with SystemCallFilter=, a new
6132 SystemCallErrorNumber= setting has been introduced that
6133 allows configuration of a system error number to be returned
6134 on filtered system calls, instead of immediately killing the
6135 process. Also, SystemCallArchitectures= has been added to
6136 limit access to system calls of a particular architecture
6137 (in order to turn off support for unused secondary
6138 architectures). There is also a global
6139 SystemCallArchitectures= setting in system.conf now to turn
6140 off support for non-native system calls system-wide.
6142 * systemd requires a kernel with a working name_to_handle_at(),
6143 please see the kernel config requirements in the README file.
6145 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6146 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6147 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6148 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6149 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6150 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6151 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6152 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6153 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6154 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6155 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6156 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6157 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6158 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6159 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6160 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6161 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6162 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6163 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6164 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6165 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6166 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6167 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6168 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6170 — Berlin, 2014-02-20
6174 * logind has gained support for facilitating privileged input
6175 and drm device access for unprivileged clients. This work is
6176 useful to allow Wayland display servers (and similar
6177 programs, such as kmscon) to run under the user's ID and
6178 access input and drm devices which are normally
6179 protected. When this is used (and the kernel is new enough)
6180 logind will "mute" IO on the file descriptors passed to
6181 Wayland as long as it is in the background and "unmute" it
6182 if it returns into the foreground. This allows secure
6183 session switching without allowing background sessions to
6184 eavesdrop on input and display data. This also introduces
6185 session switching support if VT support is turned off in the
6186 kernel, and on seats that are not seat0.
6188 * A new kernel command line option luks.options= is understood
6189 now which allows specifying LUKS options for usage for LUKS
6190 encrypted partitions specified with luks.uuid=.
6192 * tmpfiles.d(5) snippets may now use specifier expansion in
6193 path names. More specifically %m, %b, %H, %v, are now
6194 replaced by the local machine id, boot id, hostname, and
6195 kernel version number.
6197 * A new tmpfiles.d(5) command "m" has been introduced which
6198 may be used to change the owner/group/access mode of a file
6199 or directory if it exists, but do nothing if it does not.
6201 * This release removes high-level support for the
6202 MemorySoftLimit= cgroup setting. The underlying kernel
6203 cgroup attribute memory.soft_limit= is currently badly
6204 designed and likely to be removed from the kernel API in its
6205 current form, hence we should not expose it for now.
6207 * The memory.use_hierarchy cgroup attribute is now enabled for
6208 all cgroups systemd creates in the memory cgroup
6209 hierarchy. This option is likely to be come the built-in
6210 default in the kernel anyway, and the non-hierarchical mode
6211 never made much sense in the intrinsically hierarchical
6214 * A new field _SYSTEMD_SLICE= is logged along with all journal
6215 messages containing the slice a message was generated
6216 from. This is useful to allow easy per-customer filtering of
6217 logs among other things.
6219 * systemd-journald will no longer adjust the group of journal
6220 files it creates to the "systemd-journal" group. Instead we
6221 rely on the journal directory to be owned by the
6222 "systemd-journal" group, and its setgid bit set, so that the
6223 kernel file system layer will automatically enforce that
6224 journal files inherit this group assignment. The reason for
6225 this change is that we cannot allow NSS look-ups from
6226 journald which would be necessary to resolve
6227 "systemd-journal" to a numeric GID, because this might
6228 create deadlocks if NSS involves synchronous queries to
6229 other daemons (such as nscd, or sssd) which in turn are
6230 logging clients of journald and might block on it, which
6231 would then dead lock. A tmpfiles.d(5) snippet included in
6232 systemd will make sure the setgid bit and group are
6233 properly set on the journal directory if it exists on every
6234 boot. However, we recommend adjusting it manually after
6235 upgrades too (or from RPM scriptlets), so that the change is
6236 not delayed until next reboot.
6238 * Backlight and random seed files in /var/lib/ have moved into
6239 the /var/lib/systemd/ directory, in order to centralize all
6240 systemd generated files in one directory.
6242 * Boot time performance measurements (as displayed by
6243 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6244 performance information if that's available to determine how
6245 much time BIOS and boot loader initialization required. With
6246 a sufficiently new BIOS you hence no longer need to boot
6247 with Gummiboot to get access to such information.
6249 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6250 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6251 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6252 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6253 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6254 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6255 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6257 — Berlin, 2013-10-02
6261 * The Restart= option for services now understands a new
6262 on-watchdog setting, which will restart the service
6263 automatically if the service stops sending out watchdog keep
6264 alive messages (as configured with WatchdogSec=).
6266 * The getty generator (which is responsible for bringing up a
6267 getty on configured serial consoles) will no longer only
6268 start a getty on the primary kernel console but on all
6269 others, too. This makes the order in which console= is
6270 specified on the kernel command line less important.
6272 * libsystemd-logind gained a new sd_session_get_vt() call to
6273 retrieve the VT number of a session.
6275 * If the option "tries=0" is set for an entry of /etc/crypttab
6276 its passphrase is queried indefinitely instead of any
6277 maximum number of tries.
6279 * If a service with a configure PID file terminates its PID
6280 file will now be removed automatically if it still exists
6281 afterwards. This should put an end to stale PID files.
6283 * systemd-run will now also take relative binary path names
6284 for execution and no longer insists on absolute paths.
6286 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6287 paths that are optionally prefixed with "-" to indicate that
6288 it should not be considered a failure if they do not exist.
6290 * journalctl -o (and similar commands) now understands a new
6291 output mode "short-precise", it is similar to "short" but
6292 shows timestamps with usec accuracy.
6294 * The option "discard" (as known from Debian) is now
6295 synonymous to "allow-discards" in /etc/crypttab. In fact,
6296 "discard" is preferred now (since it is easier to remember
6299 * Some licensing clean-ups were made, so that more code is now
6300 LGPL-2.1 licensed than before.
6302 * A minimal tool to save/restore the display backlight
6303 brightness across reboots has been added. It will store the
6304 backlight setting as late as possible at shutdown, and
6305 restore it as early as possible during reboot.
6307 * A logic to automatically discover and enable home and swap
6308 partitions on GPT disks has been added. With this in place
6309 /etc/fstab becomes optional for many setups as systemd can
6310 discover certain partitions located on the root disk
6311 automatically. Home partitions are recognized under their
6312 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6313 partitions are recognized under their GPT type ID
6314 0657fd6da4ab43c484e50933c84b4f4f.
6316 * systemd will no longer pass any environment from the kernel
6317 or initrd to system services. If you want to set an
6318 environment for all services, do so via the kernel command
6319 line systemd.setenv= assignment.
6321 * The systemd-sysctl tool no longer natively reads the file
6322 /etc/sysctl.conf. If desired, the file should be symlinked
6323 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6324 legacy support by a symlink rather than built-in code, it
6325 also makes the otherwise hidden order of application of the
6326 different files visible. (Note that this partly reverts to a
6327 pre-198 application order of sysctl knobs!)
6329 * The "systemctl set-log-level" and "systemctl dump" commands
6330 have been moved to systemd-analyze.
6332 * systemd-run learned the new --remain-after-exit switch,
6333 which causes the scope unit not to be cleaned up
6334 automatically after the process terminated.
6336 * tmpfiles learned a new --exclude-prefix= switch to exclude
6337 certain paths from operation.
6339 * journald will now automatically flush all messages to disk
6340 as soon as a message at the log level CRIT, ALERT or EMERG
6343 Contributions from: Andrew Cook, Brandon Philips, Christian
6344 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6345 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6346 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6347 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6348 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6349 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6350 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6351 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6352 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6353 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6354 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6355 William Giokas, Zbigniew Jędrzejewski-Szmek
6357 — Berlin, 2013-09-13
6361 * The documentation has been updated to cover the various new
6362 concepts introduced with 205.
6364 * Unit files now understand the new %v specifier which
6365 resolves to the kernel version string as returned by "uname
6368 * systemctl now supports filtering the unit list output by
6369 load state, active state and sub state, using the new
6372 * "systemctl status" will now show the results of the
6373 condition checks (like ConditionPathExists= and similar) of
6374 the last start attempts of the unit. They are also logged to
6377 * "journalctl -b" may now be used to look for boot output of a
6378 specific boot. Try "journalctl -b -1" for the previous boot,
6379 but the syntax is substantially more powerful.
6381 * "journalctl --show-cursor" has been added which prints the
6382 cursor string the last shown log line. This may then be used
6383 with the new "journalctl --after-cursor=" switch to continue
6384 browsing logs from that point on.
6386 * "journalctl --force" may now be used to force regeneration
6389 * Creation of "dead" device nodes has been moved from udev
6390 into kmod and tmpfiles. Previously, udev would read the kmod
6391 databases to pre-generate dead device nodes based on meta
6392 information contained in kernel modules, so that these would
6393 be auto-loaded on access rather then at boot. As this
6394 does not really have much to do with the exposing actual
6395 kernel devices to userspace this has always been slightly
6396 alien in the udev codebase. Following the new scheme kmod
6397 will now generate a runtime snippet for tmpfiles from the
6398 module meta information and it now is tmpfiles' job to the
6399 create the nodes. This also allows overriding access and
6400 other parameters for the nodes using the usual tmpfiles
6401 facilities. As side effect this allows us to remove the
6402 CAP_SYS_MKNOD capability bit from udevd entirely.
6404 * logind's device ACLs may now be applied to these "dead"
6405 devices nodes too, thus finally allowing managed access to
6406 devices such as /dev/snd/sequencer without loading the
6407 backing module right-away.
6409 * A new RPM macro has been added that may be used to apply
6410 tmpfiles configuration during package installation.
6412 * systemd-detect-virt and ConditionVirtualization= now can
6413 detect User-Mode-Linux machines (UML).
6415 * journald will now implicitly log the effective capabilities
6416 set of processes in the message metadata.
6418 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6420 * The initrd interface has been simplified (more specifically,
6421 support for passing performance data via environment
6422 variables and fsck results via files in /run has been
6423 removed). These features were non-essential, and are
6424 nowadays available in a much nicer way by having systemd in
6425 the initrd serialize its state and have the hosts systemd
6426 deserialize it again.
6428 * The udev "keymap" data files and tools to apply keyboard
6429 specific mappings of scan to key codes, and force-release
6430 scan code lists have been entirely replaced by a udev
6431 "keyboard" builtin and a hwdb data file.
6433 * systemd will now honour the kernel's "quiet" command line
6434 argument also during late shutdown, resulting in a
6435 completely silent shutdown when used.
6437 * There's now an option to control the SO_REUSEPORT socket
6438 option in .socket units.
6440 * Instance units will now automatically get a per-template
6441 subslice of system.slice unless something else is explicitly
6442 configured. For example, instances of sshd@.service will now
6443 implicitly be placed in system-sshd.slice rather than
6444 system.slice as before.
6446 * Test coverage support may now be enabled at build time.
6448 Contributions from: Dave Reisner, Frederic Crozat, Harald
6449 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6450 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6451 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6452 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6453 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6454 Giokas, Zbigniew Jędrzejewski-Szmek
6456 — Berlin, 2013-07-23
6460 * Two new unit types have been introduced:
6462 Scope units are very similar to service units, however, are
6463 created out of pre-existing processes — instead of PID 1
6464 forking off the processes. By using scope units it is
6465 possible for system services and applications to group their
6466 own child processes (worker processes) in a powerful way
6467 which then maybe used to organize them, or kill them
6468 together, or apply resource limits on them.
6470 Slice units may be used to partition system resources in an
6471 hierarchical fashion and then assign other units to them. By
6472 default there are now three slices: system.slice (for all
6473 system services), user.slice (for all user sessions),
6474 machine.slice (for VMs and containers).
6476 Slices and scopes have been introduced primarily in
6477 context of the work to move cgroup handling to a
6478 single-writer scheme, where only PID 1
6479 creates/removes/manages cgroups.
6481 * There's a new concept of "transient" units. In contrast to
6482 normal units these units are created via an API at runtime,
6483 not from configuration from disk. More specifically this
6484 means it is now possible to run arbitrary programs as
6485 independent services, with all execution parameters passed
6486 in via bus APIs rather than read from disk. Transient units
6487 make systemd substantially more dynamic then it ever was,
6488 and useful as a general batch manager.
6490 * logind has been updated to make use of scope and slice units
6491 for managing user sessions. As a user logs in he will get
6492 his own private slice unit, to which all sessions are added
6493 as scope units. We also added support for automatically
6494 adding an instance of user@.service for the user into the
6495 slice. Effectively logind will no longer create cgroup
6496 hierarchies on its own now, it will defer entirely to PID 1
6497 for this by means of scope, service and slice units. Since
6498 user sessions this way become entities managed by PID 1
6499 the output of "systemctl" is now a lot more comprehensive.
6501 * A new mini-daemon "systemd-machined" has been added which
6502 may be used by virtualization managers to register local
6503 VMs/containers. nspawn has been updated accordingly, and
6504 libvirt will be updated shortly. machined will collect a bit
6505 of meta information about the VMs/containers, and assign
6506 them their own scope unit (see above). The collected
6507 meta-data is then made available via the "machinectl" tool,
6508 and exposed in "ps" and similar tools. machined/machinectl
6509 is compile-time optional.
6511 * As discussed earlier, the low-level cgroup configuration
6512 options ControlGroup=, ControlGroupModify=,
6513 ControlGroupPersistent=, ControlGroupAttribute= have been
6514 removed. Please use high-level attribute settings instead as
6515 well as slice units.
6517 * A new bus call SetUnitProperties() has been added to alter
6518 various runtime parameters of a unit. This is primarily
6519 useful to alter cgroup parameters dynamically in a nice way,
6520 but will be extended later on to make more properties
6521 modifiable at runtime. systemctl gained a new set-properties
6522 command that wraps this call.
6524 * A new tool "systemd-run" has been added which can be used to
6525 run arbitrary command lines as transient services or scopes,
6526 while configuring a number of settings via the command
6527 line. This tool is currently very basic, however already
6528 very useful. We plan to extend this tool to even allow
6529 queuing of execution jobs with time triggers from the
6530 command line, similar in fashion to "at".
6532 * nspawn will now inform the user explicitly that kernels with
6533 audit enabled break containers, and suggest the user to turn
6536 * Support for detecting the IMA and AppArmor security
6537 frameworks with ConditionSecurity= has been added.
6539 * journalctl gained a new "-k" switch for showing only kernel
6540 messages, mimicking dmesg output; in addition to "--user"
6541 and "--system" switches for showing only user's own logs
6544 * systemd-delta can now show information about drop-in
6545 snippets extending unit files.
6547 * libsystemd-bus has been substantially updated but is still
6548 not available as public API.
6550 * systemd will now look for the "debug" argument on the kernel
6551 command line and enable debug logging, similar to what
6552 "systemd.log_level=debug" already did before.
6554 * "systemctl set-default", "systemctl get-default" has been
6555 added to configure the default.target symlink, which
6556 controls what to boot into by default.
6558 * "systemctl set-log-level" has been added as a convenient
6559 way to raise and lower systemd logging threshold.
6561 * "systemd-analyze plot" will now show the time the various
6562 generators needed for execution, as well as information
6563 about the unit file loading.
6565 * libsystemd-journal gained a new sd_journal_open_files() call
6566 for opening specific journal files. journactl also gained a
6567 new switch to expose this new functionality. Previously we
6568 only supported opening all files from a directory, or all
6569 files from the system, as opening individual files only is
6570 racy due to journal file rotation.
6572 * systemd gained the new DefaultEnvironment= setting in
6573 /etc/systemd/system.conf to set environment variables for
6576 * If a privileged process logs a journal message with the
6577 OBJECT_PID= field set, then journald will automatically
6578 augment this with additional OBJECT_UID=, OBJECT_GID=,
6579 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6580 system services want to log events about specific client
6581 processes. journactl/systemctl has been updated to make use
6582 of this information if all log messages regarding a specific
6585 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6586 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6587 Reisner, David Coppa, David King, David Strauss, Eelco
6588 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6589 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6590 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6591 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6592 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6593 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6594 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6595 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6596 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6597 Łukasz Stelmach, 장동준
6601 * The Python bindings gained some minimal support for the APIs
6602 exposed by libsystemd-logind.
6604 * ConditionSecurity= gained support for detecting SMACK. Since
6605 this condition already supports SELinux and AppArmor we only
6606 miss IMA for this. Patches welcome!
6608 Contributions from: Karol Lewandowski, Lennart Poettering,
6609 Zbigniew Jędrzejewski-Szmek
6613 * systemd-nspawn will now create /etc/resolv.conf if
6614 necessary, before bind-mounting the host's file onto it.
6616 * systemd-nspawn will now store meta information about a
6617 container on the container's cgroup as extended attribute
6618 fields, including the root directory.
6620 * The cgroup hierarchy has been reworked in many ways. All
6621 objects any of the components systemd creates in the cgroup
6622 tree are now suffixed. More specifically, user sessions are
6623 now placed in cgroups suffixed with ".session", users in
6624 cgroups suffixed with ".user", and nspawn containers in
6625 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6626 names are now escaped in a simple scheme to avoid collision
6627 of userspace object names with kernel filenames. This work
6628 is preparation for making these objects relocatable in the
6629 cgroup tree, in order to allow easy resource partitioning of
6630 these objects without causing naming conflicts.
6632 * systemctl list-dependencies gained the new switches
6633 --plain, --reverse, --after and --before.
6635 * systemd-inhibit now shows the process name of processes that
6636 have taken an inhibitor lock.
6638 * nss-myhostname will now also resolve "localhost"
6639 implicitly. This makes /etc/hosts an optional file and
6640 nicely handles that on IPv6 ::1 maps to both "localhost" and
6643 * libsystemd-logind.so gained a new call
6644 sd_get_machine_names() to enumerate running containers and
6645 VMs (currently only supported by very new libvirt and
6646 nspawn). sd_login_monitor can now be used to watch
6647 VMs/containers coming and going.
6649 * .include is not allowed recursively anymore, and only in
6650 unit files. Usually it is better to use drop-in snippets in
6651 .d/*.conf anyway, as introduced with systemd 198.
6653 * systemd-analyze gained a new "critical-chain" command that
6654 determines the slowest chain of units run during system
6655 boot-up. It is very useful for tracking down where
6656 optimizing boot time is the most beneficial.
6658 * systemd will no longer allow manipulating service paths in
6659 the name=systemd:/system cgroup tree using ControlGroup= in
6660 units. (But is still fine with it in all other dirs.)
6662 * There's a new systemd-nspawn@.service service file that may
6663 be used to easily run nspawn containers as system
6664 services. With the container's root directory in
6665 /var/lib/container/foobar it is now sufficient to run
6666 "systemctl start systemd-nspawn@foobar.service" to boot it.
6668 * systemd-cgls gained a new parameter "--machine" to list only
6669 the processes within a certain container.
6671 * ConditionSecurity= now can check for "apparmor". We still
6672 are lacking checks for SMACK and IMA for this condition
6673 check though. Patches welcome!
6675 * A new configuration file /etc/systemd/sleep.conf has been
6676 added that may be used to configure which kernel operation
6677 systemd is supposed to execute when "suspend", "hibernate"
6678 or "hybrid-sleep" is requested. This makes the new kernel
6679 "freeze" state accessible to the user.
6681 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6682 the passed argument if applicable.
6684 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6685 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6686 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6687 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6688 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6689 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6690 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6695 * The output of 'systemctl list-jobs' got some polishing. The
6696 '--type=' argument may now be passed more than once. A new
6697 command 'systemctl list-sockets' has been added which shows
6698 a list of kernel sockets systemd is listening on with the
6699 socket units they belong to, plus the units these socket
6702 * The experimental libsystemd-bus library got substantial
6703 updates to work in conjunction with the (also experimental)
6704 kdbus kernel project. It works well enough to exchange
6705 messages with some sophistication. Note that kdbus is not
6706 ready yet, and the library is mostly an elaborate test case
6707 for now, and not installable.
6709 * systemd gained a new unit 'systemd-static-nodes.service'
6710 that generates static device nodes earlier during boot, and
6711 can run in conjunction with udev.
6713 * libsystemd-login gained a new call sd_pid_get_user_unit()
6714 to retrieve the user systemd unit a process is running
6715 in. This is useful for systems where systemd is used as
6718 * systemd-nspawn now places all containers in the new /machine
6719 top-level cgroup directory in the name=systemd
6720 hierarchy. libvirt will soon do the same, so that we get a
6721 uniform separation of /system, /user and /machine for system
6722 services, user processes and containers/virtual
6723 machines. This new cgroup hierarchy is also useful to stick
6724 stable names to specific container instances, which can be
6725 recognized later this way (this name may be controlled
6726 via systemd-nspawn's new -M switch). libsystemd-login also
6727 gained a new call sd_pid_get_machine_name() to retrieve the
6728 name of the container/VM a specific process belongs to.
6730 * bootchart can now store its data in the journal.
6732 * libsystemd-journal gained a new call
6733 sd_journal_add_conjunction() for AND expressions to the
6734 matching logic. This can be used to express more complex
6735 logical expressions.
6737 * journactl can now take multiple --unit= and --user-unit=
6740 * The cryptsetup logic now understands the "luks.key=" kernel
6741 command line switch for specifying a file to read the
6742 decryption key from. Also, if a configured key file is not
6743 found the tool will now automatically fall back to prompting
6746 * Python systemd.journal module was updated to wrap recently
6747 added functions from libsystemd-journal. The interface was
6748 changed to bring the low level interface in s.j._Reader
6749 closer to the C API, and the high level interface in
6750 s.j.Reader was updated to wrap and convert all data about
6753 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6754 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6755 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6756 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6757 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6758 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6762 * journalctl --update-catalog now understands a new --root=
6763 option to operate on catalogs found in a different root
6766 * During shutdown after systemd has terminated all running
6767 services a final killing loop kills all remaining left-over
6768 processes. We will now print the name of these processes
6769 when we send SIGKILL to them, since this usually indicates a
6772 * If /etc/crypttab refers to password files stored on
6773 configured mount points automatic dependencies will now be
6774 generated to ensure the specific mount is established first
6775 before the key file is attempted to be read.
6777 * 'systemctl status' will now show information about the
6778 network sockets a socket unit is listening on.
6780 * 'systemctl status' will also shown information about any
6781 drop-in configuration file for units. (Drop-In configuration
6782 files in this context are files such as
6783 /etc/systemd/systemd/foobar.service.d/*.conf)
6785 * systemd-cgtop now optionally shows summed up CPU times of
6786 cgroups. Press '%' while running cgtop to switch between
6787 percentage and absolute mode. This is useful to determine
6788 which cgroups use up the most CPU time over the entire
6789 runtime of the system. systemd-cgtop has also been updated
6790 to be 'pipeable' for processing with further shell tools.
6792 * 'hostnamectl set-hostname' will now allow setting of FQDN
6795 * The formatting and parsing of time span values has been
6796 changed. The parser now understands fractional expressions
6797 such as "5.5h". The formatter will now output fractional
6798 expressions for all time spans under 1min, i.e. "5.123456s"
6799 rather than "5s 123ms 456us". For time spans under 1s
6800 millisecond values are shown, for those under 1ms
6801 microsecond values are shown. This should greatly improve
6802 all time-related output of systemd.
6804 * libsystemd-login and libsystemd-journal gained new
6805 functions for querying the poll() events mask and poll()
6806 timeout value for integration into arbitrary event
6809 * localectl gained the ability to list available X11 keymaps
6810 (models, layouts, variants, options).
6812 * 'systemd-analyze dot' gained the ability to filter for
6813 specific units via shell-style globs, to create smaller,
6814 more useful graphs. I.e. it is now possible to create simple
6815 graphs of all the dependencies between only target units, or
6816 of all units that Avahi has dependencies with.
6818 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6819 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6820 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6821 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6822 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6823 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6824 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6828 * The boot-time readahead implementation for rotating media
6829 will now read the read-ahead data in multiple passes which
6830 consist of all read requests made in equidistant time
6831 intervals. This means instead of strictly reading read-ahead
6832 data in its physical order on disk we now try to find a
6833 middle ground between physical and access time order.
6835 * /etc/os-release files gained a new BUILD_ID= field for usage
6836 on operating systems that provide continuous builds of OS
6839 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6840 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6841 William Douglas, Zbigniew Jędrzejewski-Szmek
6845 * systemd-python gained an API exposing libsystemd-daemon.
6847 * The SMACK setup logic gained support for uploading CIPSO
6850 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6851 ReadOnlyDirectories= and InaccessibleDirectories= has
6852 changed. The private /tmp and /var/tmp directories are now
6853 shared by all processes of a service (which means
6854 ExecStartPre= may now leave data in /tmp that ExecStart= of
6855 the same service can still access). When a service is
6856 stopped its temporary directories are immediately deleted
6857 (normal clean-up with tmpfiles is still done in addition to
6860 * By default, systemd will now set a couple of sysctl
6861 variables in the kernel: the safe sysrq options are turned
6862 on, IP route verification is turned on, and source routing
6863 disabled. The recently added hardlink and softlink
6864 protection of the kernel is turned on. These settings should
6865 be reasonably safe, and good defaults for all new systems.
6867 * The predictable network naming logic may now be turned off
6868 with a new kernel command line switch: net.ifnames=0.
6870 * A new libsystemd-bus module has been added that implements a
6871 pretty complete D-Bus client library. For details see:
6873 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
6875 * journald will now explicitly flush the journal files to disk
6876 at the latest 5min after each write. The file will then also
6877 be marked offline until the next write. This should increase
6878 reliability in case of a crash. The synchronization delay
6879 can be configured via SyncIntervalSec= in journald.conf.
6881 * There's a new remote-fs-setup.target unit that can be used
6882 to pull in specific services when at least one remote file
6883 system is to be mounted.
6885 * There are new targets timers.target and paths.target as
6886 canonical targets to pull user timer and path units in
6887 from. This complements sockets.target with a similar
6888 purpose for socket units.
6890 * libudev gained a new call udev_device_set_attribute_value()
6891 to set sysfs attributes of a device.
6893 * The udev daemon now sets the default number of worker
6894 processes executed in parallel based on the number of available
6895 CPUs instead of the amount of available RAM. This is supposed
6896 to provide a more reliable default and limit a too aggressive
6897 parallelism for setups with 1000s of devices connected.
6899 Contributions from: Auke Kok, Colin Walters, Cristian
6900 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6901 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6902 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6903 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6904 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6905 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6906 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6907 Zbigniew Jędrzejewski-Szmek
6911 * Configuration of unit files may now be extended via drop-in
6912 files without having to edit/override the unit files
6913 themselves. More specifically, if the administrator wants to
6914 change one value for a service file foobar.service he can
6915 now do so by dropping in a configuration snippet into
6916 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
6917 will load all these snippets and apply them on top of the
6918 main unit configuration file, possibly extending or
6919 overriding its settings. Using these drop-in snippets is
6920 generally nicer than the two earlier options for changing
6921 unit files locally: copying the files from
6922 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6923 them there; or creating a new file in /etc/systemd/system/
6924 that incorporates the original one via ".include". Drop-in
6925 snippets into these .d/ directories can be placed in any
6926 directory systemd looks for units in, and the usual
6927 overriding semantics between /usr/lib, /etc and /run apply
6930 * Most unit file settings which take lists of items can now be
6931 reset by assigning the empty string to them. For example,
6932 normally, settings such as Environment=FOO=BAR append a new
6933 environment variable assignment to the environment block,
6934 each time they are used. By assigning Environment= the empty
6935 string the environment block can be reset to empty. This is
6936 particularly useful with the .d/*.conf drop-in snippets
6937 mentioned above, since this adds the ability to reset list
6938 settings from vendor unit files via these drop-ins.
6940 * systemctl gained a new "list-dependencies" command for
6941 listing the dependencies of a unit recursively.
6943 * Inhibitors are now honored and listed by "systemctl
6944 suspend", "systemctl poweroff" (and similar) too, not only
6945 GNOME. These commands will also list active sessions by
6948 * Resource limits (as exposed by the various control group
6949 controllers) can now be controlled dynamically at runtime
6950 for all units. More specifically, you can now use a command
6951 like "systemctl set-cgroup-attr foobar.service cpu.shares
6952 2000" to alter the CPU shares a specific service gets. These
6953 settings are stored persistently on disk, and thus allow the
6954 administrator to easily adjust the resource usage of
6955 services with a few simple commands. This dynamic resource
6956 management logic is also available to other programs via the
6957 bus. Almost any kernel cgroup attribute and controller is
6960 * systemd-vconsole-setup will now copy all font settings to
6961 all allocated VTs, where it previously applied them only to
6964 * libsystemd-login gained the new sd_session_get_tty() API
6967 * This release drops support for a few legacy or
6968 distribution-specific LSB facility names when parsing init
6969 scripts: $x-display-manager, $mail-transfer-agent,
6970 $mail-transport-agent, $mail-transfer-agent, $smtp,
6971 $null. Also, the mail-transfer-agent.target unit backing
6972 this has been removed. Distributions which want to retain
6973 compatibility with this should carry the burden for
6974 supporting this themselves and patch support for these back
6975 in, if they really need to. Also, the facilities $syslog and
6976 $local_fs are now ignored, since systemd does not support
6977 early-boot LSB init scripts anymore, and these facilities
6978 are implied anyway for normal services. syslog.target has
6981 * There are new bus calls on PID1's Manager object for
6982 cancelling jobs, and removing snapshot units. Previously,
6983 both calls were only available on the Job and Snapshot
6986 * systemd-journal-gatewayd gained SSL support.
6988 * The various "environment" files, such as /etc/locale.conf
6989 now support continuation lines with a backslash ("\") as
6990 last character in the line, similarly in style (but different)
6991 to how this is supported in shells.
6993 * For normal user processes the _SYSTEMD_USER_UNIT= field is
6994 now implicitly appended to every log entry logged. systemctl
6995 has been updated to filter by this field when operating on a
6996 user systemd instance.
6998 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
6999 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7000 the container. This makes it easier to boot unmodified
7001 Fedora systems in a container, which however still requires
7002 audit=0 to be passed on the kernel command line. Auditing in
7003 kernel and userspace is unfortunately still too broken in
7004 context of containers, hence we recommend compiling it out
7005 of the kernel or using audit=0. Hopefully this will be fixed
7006 one day for good in the kernel.
7008 * nspawn gained the new --bind= and --bind-ro= parameters to
7009 bind mount specific directories from the host into the
7012 * nspawn will now mount its own devpts file system instance
7013 into the container, in order not to leak pty devices from
7014 the host into the container.
7016 * systemd will now read the firmware boot time performance
7017 information from the EFI variables, if the used boot loader
7018 supports this, and takes it into account for boot performance
7019 analysis via "systemd-analyze". This is currently supported
7020 only in conjunction with Gummiboot, but could be supported
7021 by other boot loaders too. For details see:
7023 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
7025 * A new generator has been added that automatically mounts the
7026 EFI System Partition (ESP) to /boot, if that directory
7027 exists, is empty, and no other file system has been
7028 configured to be mounted there.
7030 * logind will now send out PrepareForSleep(false) out
7031 unconditionally, after coming back from suspend. This may be
7032 used by applications as asynchronous notification for
7033 system resume events.
7035 * "systemctl unlock-sessions" has been added, that allows
7036 unlocking the screens of all user sessions at once, similar
7037 to how "systemctl lock-sessions" already locked all users
7038 sessions. This is backed by a new D-Bus call UnlockSessions().
7040 * "loginctl seat-status" will now show the master device of a
7041 seat. (i.e. the device of a seat that needs to be around for
7042 the seat to be considered available, usually the graphics
7045 * tmpfiles gained a new "X" line type, that allows
7046 configuration of files and directories (with wildcards) that
7047 shall be excluded from automatic cleanup ("aging").
7049 * udev default rules set the device node permissions now only
7050 at "add" events, and do not change them any longer with a
7051 later "change" event.
7053 * The log messages for lid events and power/sleep keypresses
7054 now carry a message ID.
7056 * We now have a substantially larger unit test suite, but this
7057 continues to be work in progress.
7059 * udevadm hwdb gained a new --root= parameter to change the
7060 root directory to operate relative to.
7062 * logind will now issue a background sync() request to the kernel
7063 early at shutdown, so that dirty buffers are flushed to disk early
7064 instead of at the last moment, in order to optimize shutdown
7067 * A new bootctl tool has been added that is an interface for
7068 certain boot loader operations. This is currently a preview
7069 and is likely to be extended into a small mechanism daemon
7070 like timedated, localed, hostnamed, and can be used by
7071 graphical UIs to enumerate available boot options, and
7072 request boot into firmware operations.
7074 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7075 the rest of the package. It also has been updated to work
7076 correctly in initrds.
7078 * Policykit previously has been runtime optional, and is now
7079 also compile time optional via a configure switch.
7081 * systemd-analyze has been reimplemented in C. Also "systemctl
7082 dot" has moved into systemd-analyze.
7084 * "systemctl status" with no further parameters will now print
7085 the status of all active or failed units.
7087 * Operations such as "systemctl start" can now be executed
7088 with a new mode "--irreversible" which may be used to queue
7089 operations that cannot accidentally be reversed by a later
7090 job queuing. This is by default used to make shutdown
7091 requests more robust.
7093 * The Python API of systemd now gained a new module for
7094 reading journal files.
7096 * A new tool kernel-install has been added that can install
7097 kernel images according to the Boot Loader Specification:
7099 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7101 * Boot time console output has been improved to provide
7102 animated boot time output for hanging jobs.
7104 * A new tool systemd-activate has been added which can be used
7105 to test socket activation with, directly from the command
7106 line. This should make it much easier to test and debug
7107 socket activation in daemons.
7109 * journalctl gained a new "--reverse" (or -r) option to show
7110 journal output in reverse order (i.e. newest line first).
7112 * journalctl gained a new "--pager-end" (or -e) option to jump
7113 to immediately jump to the end of the journal in the
7114 pager. This is only supported in conjunction with "less".
7116 * journalctl gained a new "--user-unit=" option, that works
7117 similarly to "--unit=" but filters for user units rather than
7120 * A number of unit files to ease adoption of systemd in
7121 initrds has been added. This moves some minimal logic from
7122 the various initrd implementations into systemd proper.
7124 * The journal files are now owned by a new group
7125 "systemd-journal", which exists specifically to allow access
7126 to the journal, and nothing else. Previously, we used the
7127 "adm" group for that, which however possibly covers more
7128 than just journal/log file access. This new group is now
7129 already used by systemd-journal-gatewayd to ensure this
7130 daemon gets access to the journal files and as little else
7131 as possible. Note that "make install" will also set FS ACLs
7132 up for /var/log/journal to give "adm" and "wheel" read
7133 access to it, in addition to "systemd-journal" which owns
7134 the journal files. We recommend that packaging scripts also
7135 add read access to "adm" + "wheel" to /var/log/journal, and
7136 all existing/future journal files. To normal users and
7137 administrators little changes, however packagers need to
7138 ensure to create the "systemd-journal" system group at
7139 package installation time.
7141 * The systemd-journal-gatewayd now runs as unprivileged user
7142 systemd-journal-gateway:systemd-journal-gateway. Packaging
7143 scripts need to create these system user/group at
7146 * timedated now exposes a new boolean property CanNTP that
7147 indicates whether a local NTP service is available or not.
7149 * systemd-detect-virt will now also detect xen PVs
7151 * The pstore file system is now mounted by default, if it is
7154 * In addition to the SELinux and IMA policies we will now also
7155 load SMACK policies at early boot.
7157 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7158 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7159 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7160 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7161 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7162 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7163 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7164 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7165 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7166 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7167 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7168 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7169 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7170 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7174 * Timer units now support calendar time events in addition to
7175 monotonic time events. That means you can now trigger a unit
7176 based on a calendar time specification such as "Thu,Fri
7177 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7178 or fifth day of any month of the year 2013, given that it is
7179 a thursday or friday. This brings timer event support
7180 considerably closer to cron's capabilities. For details on
7181 the supported calendar time specification language see
7184 * udev now supports a number of different naming policies for
7185 network interfaces for predictable names, and a combination
7186 of these policies is now the default. Please see this wiki
7187 document for details:
7189 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
7191 * Auke Kok's bootchart implementation has been added to the
7192 systemd tree. It is an optional component that can graph the
7193 boot in quite some detail. It is one of the best bootchart
7194 implementations around and minimal in its code and
7197 * nss-myhostname has been integrated into the systemd source
7198 tree. nss-myhostname guarantees that the local hostname
7199 always stays resolvable via NSS. It has been a weak
7200 requirement of systemd-hostnamed since a long time, and
7201 since its code is actually trivial we decided to just
7202 include it in systemd's source tree. It can be turned off
7203 with a configure switch.
7205 * The read-ahead logic is now capable of properly detecting
7206 whether a btrfs file system is on SSD or rotating media, in
7207 order to optimize the read-ahead scheme. Previously, it was
7208 only capable of detecting this on traditional file systems
7211 * In udev, additional device properties are now read from the
7212 IAB in addition to the OUI database. Also, Bluetooth company
7213 identities are attached to the devices as well.
7215 * In service files %U may be used as specifier that is
7216 replaced by the configured user name of the service.
7218 * nspawn may now be invoked without a controlling TTY. This
7219 makes it suitable for invocation as its own service. This
7220 may be used to set up a simple containerized server system
7221 using only core OS tools.
7223 * systemd and nspawn can now accept socket file descriptors
7224 when they are started for socket activation. This enables
7225 implementation of socket activated nspawn
7226 containers. i.e. think about autospawning an entire OS image
7227 when the first SSH or HTTP connection is received. We expect
7228 that similar functionality will also be added to libvirt-lxc
7231 * journalctl will now suppress ANSI color codes when
7232 presenting log data.
7234 * systemctl will no longer show control group information for
7235 a unit if the control group is empty anyway.
7237 * logind can now automatically suspend/hibernate/shutdown the
7240 * /etc/machine-info and hostnamed now also expose the chassis
7241 type of the system. This can be used to determine whether
7242 the local system is a laptop, desktop, handset or
7243 tablet. This information may either be configured by the
7244 user/vendor or is automatically determined from ACPI and DMI
7245 information if possible.
7247 * A number of PolicyKit actions are now bound together with
7248 "imply" rules. This should simplify creating UIs because
7249 many actions will now authenticate similar ones as well.
7251 * Unit files learnt a new condition ConditionACPower= which
7252 may be used to conditionalize a unit depending on whether an
7253 AC power source is connected or not, of whether the system
7254 is running on battery power.
7256 * systemctl gained a new "is-failed" verb that may be used in
7257 shell scripts and suchlike to check whether a specific unit
7258 is in the "failed" state.
7260 * The EnvironmentFile= setting in unit files now supports file
7261 globbing, and can hence be used to easily read a number of
7262 environment files at once.
7264 * systemd will no longer detect and recognize specific
7265 distributions. All distribution-specific #ifdeffery has been
7266 removed, systemd is now fully generic and
7267 distribution-agnostic. Effectively, not too much is lost as
7268 a lot of the code is still accessible via explicit configure
7269 switches. However, support for some distribution specific
7270 legacy configuration file formats has been dropped. We
7271 recommend distributions to simply adopt the configuration
7272 files everybody else uses now and convert the old
7273 configuration from packaging scripts. Most distributions
7274 already did that. If that's not possible or desirable,
7275 distributions are welcome to forward port the specific
7276 pieces of code locally from the git history.
7278 * When logging a message about a unit systemd will now always
7279 log the unit name in the message meta data.
7281 * localectl will now also discover system locale data that is
7282 not stored in locale archives, but directly unpacked.
7284 * logind will no longer unconditionally use framebuffer
7285 devices as seat masters, i.e. as devices that are required
7286 to be existing before a seat is considered preset. Instead,
7287 it will now look for all devices that are tagged as
7288 "seat-master" in udev. By default, framebuffer devices will
7289 be marked as such, but depending on local systems, other
7290 devices might be marked as well. This may be used to
7291 integrate graphics cards using closed source drivers (such
7292 as NVidia ones) more nicely into logind. Note however, that
7293 we recommend using the open source NVidia drivers instead,
7294 and no udev rules for the closed-source drivers will be
7295 shipped from us upstream.
7297 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7298 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7299 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7300 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7301 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7302 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7303 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7304 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7305 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7306 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7307 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7308 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7313 * udev gained support for loading additional device properties
7314 from an indexed database that is keyed by vendor/product IDs
7315 and similar device identifiers. For the beginning this
7316 "hwdb" is populated with data from the well-known PCI and
7317 USB database, but also includes PNP, ACPI and OID data. In
7318 the longer run this indexed database shall grow into
7319 becoming the one central database for non-essential
7320 userspace device metadata. Previously, data from the PCI/USB
7321 database was only attached to select devices, since the
7322 lookup was a relatively expensive operation due to O(n) time
7323 complexity (with n being the number of entries in the
7324 database). Since this is now O(1), we decided to add in this
7325 data for all devices where this is available, by
7326 default. Note that the indexed database needs to be rebuilt
7327 when new data files are installed. To achieve this you need
7328 to update your packaging scripts to invoke "udevadm hwdb
7329 --update" after installation of hwdb data files. For
7330 RPM-based distributions we introduced the new
7331 %udev_hwdb_update macro for this purpose.
7333 * The Journal gained support for the "Message Catalog", an
7334 indexed database to link up additional information with
7335 journal entries. For further details please check:
7337 https://www.freedesktop.org/wiki/Software/systemd/catalog
7339 The indexed message catalog database also needs to be
7340 rebuilt after installation of message catalog files. Use
7341 "journalctl --update-catalog" for this. For RPM-based
7342 distributions we introduced the %journal_catalog_update
7343 macro for this purpose.
7345 * The Python Journal bindings gained support for the standard
7346 Python logging framework.
7348 * The Journal API gained new functions for checking whether
7349 the underlying file system of a journal file is capable of
7350 properly reporting file change notifications, or whether
7351 applications that want to reflect journal changes "live"
7352 need to recheck journal files continuously in appropriate
7355 * It is now possible to set the "age" field for tmpfiles
7356 entries to 0, indicating that files matching this entry
7357 shall always be removed when the directories are cleaned up.
7359 * coredumpctl gained a new "gdb" verb which invokes gdb
7360 right-away on the selected coredump.
7362 * There's now support for "hybrid sleep" on kernels that
7363 support this, in addition to "suspend" and "hibernate". Use
7364 "systemctl hybrid-sleep" to make use of this.
7366 * logind's HandleSuspendKey= setting (and related settings)
7367 now gained support for a new "lock" setting to simply
7368 request the screen lock on all local sessions, instead of
7369 actually executing a suspend or hibernation.
7371 * systemd will now mount the EFI variables file system by
7374 * Socket units now gained support for configuration of the
7375 SMACK security label.
7377 * timedatectl will now output the time of the last and next
7378 daylight saving change.
7380 * We dropped support for various legacy and distro-specific
7381 concepts, such as insserv, early-boot SysV services
7382 (i.e. those for non-standard runlevels such as 'b' or 'S')
7383 or ArchLinux /etc/rc.conf support. We recommend the
7384 distributions who still need support this to either continue
7385 to maintain the necessary patches downstream, or find a
7386 different solution. (Talk to us if you have questions!)
7388 * Various systemd components will now bypass PolicyKit checks
7389 for root and otherwise handle properly if PolicyKit is not
7390 found to be around. This should fix most issues for
7391 PolicyKit-less systems. Quite frankly this should have been
7392 this way since day one. It is absolutely our intention to
7393 make systemd work fine on PolicyKit-less systems, and we
7394 consider it a bug if something does not work as it should if
7395 PolicyKit is not around.
7397 * For embedded systems it is now possible to build udev and
7398 systemd without blkid and/or kmod support.
7400 * "systemctl switch-root" is now capable of switching root
7401 more than once. I.e. in addition to transitions from the
7402 initrd to the host OS it is now possible to transition to
7403 further OS images from the host. This is useful to implement
7404 offline updating tools.
7406 * Various other additions have been made to the RPM macros
7407 shipped with systemd. Use %udev_rules_update() after
7408 installing new udev rules files. %_udevhwdbdir,
7409 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7410 %_sysctldir are now available which resolve to the right
7411 directories for packages to place various data files in.
7413 * journalctl gained the new --full switch (in addition to
7414 --all, to disable ellipsation for long messages.
7416 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7417 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7418 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7419 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7420 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7421 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7422 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7423 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7424 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7428 * journalctl gained new --since= and --until= switches to
7429 filter by time. It also now supports nice filtering for
7430 units via --unit=/-u.
7432 * Type=oneshot services may use ExecReload= and do the
7435 * The journal daemon now supports time-based rotation and
7436 vacuuming, in addition to the usual disk-space based
7439 * The journal will now index the available field values for
7440 each field name. This enables clients to show pretty drop
7441 downs of available match values when filtering. The bash
7442 completion of journalctl has been updated
7443 accordingly. journalctl gained a new switch -F to list all
7444 values a certain field takes in the journal database.
7446 * More service events are now written as structured messages
7447 to the journal, and made recognizable via message IDs.
7449 * The timedated, localed and hostnamed mini-services which
7450 previously only provided support for changing time, locale
7451 and hostname settings from graphical DEs such as GNOME now
7452 also have a minimal (but very useful) text-based client
7453 utility each. This is probably the nicest way to changing
7454 these settings from the command line now, especially since
7455 it lists available options and is fully integrated with bash
7458 * There's now a new tool "systemd-coredumpctl" to list and
7459 extract coredumps from the journal.
7461 * We now install a README each in /var/log/ and
7462 /etc/rc.d/init.d explaining where the system logs and init
7463 scripts went. This hopefully should help folks who go to
7464 that dirs and look into the otherwise now empty void and
7465 scratch their heads.
7467 * When user-services are invoked (by systemd --user) the
7468 $MANAGERPID env var is set to the PID of systemd.
7470 * SIGRTMIN+24 when sent to a --user instance will now result
7471 in immediate termination of systemd.
7473 * gatewayd received numerous feature additions such as a
7474 "follow" mode, for live syncing and filtering.
7476 * browse.html now allows filtering and showing detailed
7477 information on specific entries. Keyboard navigation and
7478 mouse screen support has been added.
7480 * gatewayd/journalctl now supports HTML5/JSON
7481 Server-Sent-Events as output.
7483 * The SysV init script compatibility logic will now
7484 heuristically determine whether a script supports the
7485 "reload" verb, and only then make this available as
7488 * "systemctl status --follow" has been removed, use "journalctl
7491 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7492 have been removed since they are hardly useful to be
7495 * And I'd like to take the opportunity to specifically mention
7496 Zbigniew for his great contributions. Zbigniew, you rock!
7498 Contributions from: Andrew Eikum, Christian Hesse, Colin
7499 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
7500 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7501 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7502 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7503 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7504 Jędrzejewski-Szmek, Сковорода Никита Андреевич
7508 * If /etc/vconsole.conf is non-existent or empty we will no
7509 longer load any console font or key map at boot by
7510 default. Instead the kernel defaults will be left
7511 intact. This is definitely the right thing to do, as no
7512 configuration should mean no configuration, and hard-coding
7513 font names that are different on all archs is probably a bad
7514 idea. Also, the kernel default key map and font should be
7515 good enough for most cases anyway, and mostly identical to
7516 the userspace fonts/key maps we previously overloaded them
7517 with. If distributions want to continue to default to a
7518 non-kernel font or key map they should ship a default
7519 /etc/vconsole.conf with the appropriate contents.
7521 Contributions from: Colin Walters, Daniel J Walsh, Dave
7522 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7523 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7527 * journalctl gained a new --cursor= switch to show entries
7528 starting from the specified location in the journal.
7530 * We now enforce a size limit on journal entry fields exported
7531 with "-o json" in journalctl. Fields larger than 4K will be
7532 assigned null. This can be turned off with --all.
7534 * An (optional) journal gateway daemon is now available as
7535 "systemd-journal-gatewayd.service". This service provides
7536 access to the journal via HTTP and JSON. This functionality
7537 will be used to implement live log synchronization in both
7538 pull and push modes, but has various other users too, such
7539 as easy log access for debugging of embedded devices. Right
7540 now it is already useful to retrieve the journal via HTTP:
7542 # systemctl start systemd-journal-gatewayd.service
7543 # wget http://localhost:19531/entries
7545 This will download the journal contents in a
7546 /var/log/messages compatible format. The same as JSON:
7548 # curl -H"Accept: application/json" http://localhost:19531/entries
7550 This service is also accessible via a web browser where a
7551 single static HTML5 app is served that uses the JSON logic
7552 to enable the user to do some basic browsing of the
7553 journal. This will be extended later on. Here's an example
7554 screenshot of this app in its current state:
7556 http://0pointer.de/public/journal-gatewayd
7558 Contributions from: Kay Sievers, Lennart Poettering, Robert
7559 Milasan, Tom Gundersen
7563 * The bash completion logic is now available for journalctl
7566 * We do not mount the "cpuset" controller anymore together with
7567 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7568 started if no parameters are assigned to it. "cpuset" hence
7569 broke code that assumed it could create "cpu" groups and
7572 * journalctl -f will now subscribe to terminal size changes,
7573 and line break accordingly.
7575 Contributions from: Dave Reisner, Kay Sievers, Lennart
7576 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
7580 * nspawn will now create a symlink /etc/localtime in the
7581 container environment, copying the host's timezone
7582 setting. Previously this has been done via a bind mount, but
7583 since symlinks cannot be bind mounted this has now been
7584 changed to create/update the appropriate symlink.
7586 * journalctl -n's line number argument is now optional, and
7587 will default to 10 if omitted.
7589 * journald will now log the maximum size the journal files may
7590 take up on disk. This is particularly useful if the default
7591 built-in logic of determining this parameter from the file
7592 system size is used. Use "systemctl status
7593 systemd-journald.service" to see this information.
7595 * The multi-seat X wrapper tool has been stripped down. As X
7596 is now capable of enumerating graphics devices via udev in a
7597 seat-aware way the wrapper is not strictly necessary
7598 anymore. A stripped down temporary stop-gap is still shipped
7599 until the upstream display managers have been updated to
7600 fully support the new X logic. Expect this wrapper to be
7601 removed entirely in one of the next releases.
7603 * HandleSleepKey= in logind.conf has been split up into
7604 HandleSuspendKey= and HandleHibernateKey=. The old setting
7605 is not available anymore. X11 and the kernel are
7606 distinguishing between these keys and we should too. This
7607 also means the inhibition lock for these keys has been split
7610 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7611 Poettering, Lukas Nykryn, Václav Pavlín
7615 * Whenever a unit changes state we will now log this to the
7616 journal and show along the unit's own log output in
7619 * ConditionPathIsMountPoint= can now properly detect bind
7620 mount points too. (Previously, a bind mount of one file
7621 system to another place in the same file system could not be
7622 detected as mount, since they shared struct stat's st_dev
7625 * We will now mount the cgroup controllers cpu, cpuacct,
7626 cpuset and the controllers net_cls, net_prio together by
7629 * nspawn containers will now have a virtualized boot
7630 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7631 over with a randomized ID at container initialization). This
7632 has the effect of making "journalctl -b" do the right thing
7635 * The JSON output journal serialization has been updated not
7636 to generate "endless" list objects anymore, but rather one
7637 JSON object per line. This is more in line how most JSON
7638 parsers expect JSON objects. The new output mode
7639 "json-pretty" has been added to provide similar output, but
7640 neatly aligned for readability by humans.
7642 * We dropped all explicit sync() invocations in the shutdown
7643 code. The kernel does this implicitly anyway in the kernel
7644 reboot() syscall. halt(8)'s -n option is now a compatibility
7647 * We now support virtualized reboot() in containers, as
7648 supported by newer kernels. We will fall back to exit() if
7649 CAP_SYS_REBOOT is not available to the container. Also,
7650 nspawn makes use of this now and will actually reboot the
7651 container if the containerized OS asks for that.
7653 * journalctl will only show local log output by default
7654 now. Use --merge (-m) to show remote log output, too.
7656 * libsystemd-journal gained the new sd_journal_get_usage()
7657 call to determine the current disk usage of all journal
7658 files. This is exposed in the new "journalctl --disk-usage"
7661 * journald gained a new configuration setting SplitMode= in
7662 journald.conf which may be used to control how user journals
7663 are split off. See journald.conf(5) for details.
7665 * A new condition type ConditionFileNotEmpty= has been added.
7667 * tmpfiles' "w" lines now support file globbing, to write
7668 multiple files at once.
7670 * We added Python bindings for the journal submission
7671 APIs. More Python APIs for a number of selected APIs will
7672 likely follow. Note that we intend to add native bindings
7673 only for the Python language, as we consider it common
7674 enough to deserve bindings shipped within systemd. There are
7675 various projects outside of systemd that provide bindings
7676 for languages such as PHP or Lua.
7678 * Many conditions will now resolve specifiers such as %i. In
7679 addition, PathChanged= and related directives of .path units
7680 now support specifiers as well.
7682 * There's now a new RPM macro definition for the system preset
7685 * journald will now warn if it ca not forward a message to the
7686 syslog daemon because its socket is full.
7688 * timedated will no longer write or process /etc/timezone,
7689 except on Debian. As we do not support late mounted /usr
7690 anymore /etc/localtime always being a symlink is now safe,
7691 and hence the information in /etc/timezone is not necessary
7694 * logind will now always reserve one VT for a text getty (VT6
7695 by default). Previously if more than 6 X sessions where
7696 started they took up all the VTs with auto-spawned gettys,
7697 so that no text gettys were available anymore.
7699 * udev will now automatically inform the btrfs kernel logic
7700 about btrfs RAID components showing up. This should make
7701 simple hotplug based btrfs RAID assembly work.
7703 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7704 (but not for its children which will stay at the kernel
7705 default). This should allow setups with a lot more listening
7708 * systemd will now always pass the configured timezone to the
7709 kernel at boot. timedated will do the same when the timezone
7712 * logind's inhibition logic has been updated. By default,
7713 logind will now handle the lid switch, the power and sleep
7714 keys all the time, even in graphical sessions. If DEs want
7715 to handle these events on their own they should take the new
7716 handle-power-key, handle-sleep-key and handle-lid-switch
7717 inhibitors during their runtime. A simple way to achieve
7718 that is to invoke the DE wrapped in an invocation of:
7720 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7722 * Access to unit operations is now checked via SELinux taking
7723 the unit file label and client process label into account.
7725 * systemd will now notify the administrator in the journal
7726 when he over-mounts a non-empty directory.
7728 * There are new specifiers that are resolved in unit files,
7729 for the host name (%H), the machine ID (%m) and the boot ID
7732 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
7733 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7734 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7735 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7736 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7737 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7738 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7742 * Support for reading structured kernel messages from
7743 /dev/kmsg has now been added and is enabled by default.
7745 * Support for reading kernel messages from /proc/kmsg has now
7746 been removed. If you want kernel messages in the journal
7747 make sure to run a recent kernel (>= 3.5) that supports
7748 reading structured messages from /dev/kmsg (see
7749 above). /proc/kmsg is now exclusive property of classic
7750 syslog daemons again.
7752 * The libudev API gained the new
7753 udev_device_new_from_device_id() call.
7755 * The logic for file system namespace (ReadOnlyDirectory=,
7756 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7757 require pivot_root() anymore. This means fewer temporary
7758 directories are created below /tmp for this feature.
7760 * nspawn containers will now see and receive all submounts
7761 made on the host OS below the root file system of the
7764 * Forward Secure Sealing is now supported for Journal files,
7765 which provide cryptographical sealing of journal files so
7766 that attackers cannot alter log history anymore without this
7767 being detectable. Lennart will soon post a blog story about
7768 this explaining it in more detail.
7770 * There are two new service settings RestartPreventExitStatus=
7771 and SuccessExitStatus= which allow configuration of exit
7772 status (exit code or signal) which will be excepted from the
7773 restart logic, resp. consider successful.
7775 * journalctl gained the new --verify switch that can be used
7776 to check the integrity of the structure of journal files and
7777 (if Forward Secure Sealing is enabled) the contents of
7780 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7781 and similar symlinks pre-created. This makes running shells
7782 as container init process a lot more fun.
7784 * The fstab support can now handle PARTUUID= and PARTLABEL=
7787 * A new ConditionHost= condition has been added to match
7788 against the hostname (with globs) and machine ID. This is
7789 useful for clusters where a single OS image is used to
7790 provision a large number of hosts which shall run slightly
7791 different sets of services.
7793 * Services which hit the restart limit will now be placed in a
7796 Contributions from: Bertram Poettering, Dave Reisner, Huang
7797 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7798 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7802 * When running in --user mode systemd will now become a
7803 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7804 tree a lot more organized.
7806 * A new PartOf= unit dependency type has been introduced that
7807 may be used to group services in a natural way.
7809 * "systemctl enable" may now be used to enable instances of
7812 * journalctl now prints error log levels in red, and
7813 warning/notice log levels in bright white. It also supports
7814 filtering by log level now.
7816 * cgtop gained a new -n switch (similar to top), to configure
7817 the maximum number of iterations to run for. It also gained
7818 -b, to run in batch mode (accepting no input).
7820 * The suffix ".service" may now be omitted on most systemctl
7821 command lines involving service unit names.
7823 * There's a new bus call in logind to lock all sessions, as
7824 well as a loginctl verb for it "lock-sessions".
7826 * libsystemd-logind.so gained a new call sd_journal_perror()
7827 that works similar to libc perror() but logs to the journal
7828 and encodes structured information about the error number.
7830 * /etc/crypttab entries now understand the new keyfile-size=
7833 * shutdown(8) now can send a (configurable) wall message when
7834 a shutdown is cancelled.
7836 * The mount propagation mode for the root file system will now
7837 default to "shared", which is useful to make containers work
7838 nicely out-of-the-box so that they receive new mounts from
7839 the host. This can be undone locally by running "mount
7840 --make-rprivate /" if needed.
7842 * The prefdm.service file has been removed. Distributions
7843 should maintain this unit downstream if they intend to keep
7844 it around. However, we recommend writing normal unit files
7845 for display managers instead.
7847 * Since systemd is a crucial part of the OS we will now
7848 default to a number of compiler switches that improve
7849 security (hardening) such as read-only relocations, stack
7850 protection, and suchlike.
7852 * The TimeoutSec= setting for services is now split into
7853 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7854 of individual time outs for the start and the stop phase of
7857 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7858 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7859 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7860 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7861 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7862 Gundersen, Zbigniew Jędrzejewski-Szmek
7866 * The journal and id128 C APIs are now fully documented as man
7869 * Extra safety checks have been added when transitioning from
7870 the initial RAM disk to the main system to avoid accidental
7873 * /etc/crypttab entries now understand the new keyfile-offset=
7876 * systemctl -t can now be used to filter by unit load state.
7878 * The journal C API gained the new sd_journal_wait() call to
7879 make writing synchronous journal clients easier.
7881 * journalctl gained the new -D switch to show journals from a
7884 * journalctl now displays a special marker between log
7885 messages of two different boots.
7887 * The journal is now explicitly flushed to /var via a service
7888 systemd-journal-flush.service, rather than implicitly simply
7889 by seeing /var/log/journal to be writable.
7891 * journalctl (and the journal C APIs) can now match for much
7892 more complex expressions, with alternatives and
7895 * When transitioning from the initial RAM disk to the main
7896 system we will now kill all processes in a killing spree to
7897 ensure no processes stay around by accident.
7899 * Three new specifiers may be used in unit files: %u, %h, %s
7900 resolve to the user name, user home directory resp. user
7901 shell. This is useful for running systemd user instances.
7903 * We now automatically rotate journal files if their data
7904 object hash table gets a fill level > 75%. We also size the
7905 hash table based on the configured maximum file size. This
7906 together should lower hash collisions drastically and thus
7907 speed things up a bit.
7909 * journalctl gained the new "--header" switch to introspect
7910 header data of journal files.
7912 * A new setting SystemCallFilters= has been added to services
7913 which may be used to apply blacklists or whitelists to
7914 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7916 * nspawn gained a new --link-journal= switch (and quicker: -j)
7917 to link the container journal with the host. This makes it
7918 very easy to centralize log viewing on the host for all
7919 guests while still keeping the journal files separated.
7921 * Many bugfixes and optimizations
7923 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7924 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7925 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7930 * Several tools now understand kernel command line arguments,
7931 which are only read when run in an initial RAM disk. They
7932 usually follow closely their normal counterparts, but are
7935 * There's a new tool to analyze the readahead files that are
7936 automatically generated at boot. Use:
7938 /usr/lib/systemd/systemd-readahead analyze /.readahead
7940 * We now provide an early debug shell on tty9 if this enabled. Use:
7942 systemctl enable debug-shell.service
7944 * All plymouth related units have been moved into the Plymouth
7945 package. Please make sure to upgrade your Plymouth version
7948 * systemd-tmpfiles now supports getting passed the basename of
7949 a configuration file only, in which case it will look for it
7950 in all appropriate directories automatically.
7952 * udevadm info now takes a /dev or /sys path as argument, and
7953 does the right thing. Example:
7955 udevadm info /dev/sda
7956 udevadm info /sys/class/block/sda
7958 * systemctl now prints a warning if a unit is stopped but a
7959 unit that might trigger it continues to run. Example: a
7960 service is stopped but the socket that activates it is left
7963 * "systemctl status" will now mention if the log output was
7964 shortened due to rotation since a service has been started.
7966 * The journal API now exposes functions to determine the
7967 "cutoff" times due to rotation.
7969 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7970 immediately flushing of runtime logs to /var if possible,
7971 resp. for triggering immediate rotation of the journal
7974 * It is now considered an error if a service is attempted to
7975 be stopped that is not loaded.
7977 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7979 * systemd-analyze now supports Python 3
7981 * tmpfiles now supports cleaning up directories via aging
7982 where the first level dirs are always kept around but
7983 directories beneath it automatically aged. This is enabled
7984 by prefixing the age field with '~'.
7986 * Seat objects now expose CanGraphical, CanTTY properties
7987 which is required to deal with very fast bootups where the
7988 display manager might be running before the graphics drivers
7989 completed initialization.
7991 * Seat objects now expose a State property.
7993 * We now include RPM macros for service enabling/disabling
7994 based on the preset logic. We recommend RPM based
7995 distributions to make use of these macros if possible. This
7996 makes it simpler to reuse RPM spec files across
7999 * We now make sure that the collected systemd unit name is
8000 always valid when services log to the journal via
8003 * There's a new man page kernel-command-line(7) detailing all
8004 command line options we understand.
8006 * The fstab generator may now be disabled at boot by passing
8007 fstab=0 on the kernel command line.
8009 * A new kernel command line option modules-load= is now understood
8010 to load a specific kernel module statically, early at boot.
8012 * Unit names specified on the systemctl command line are now
8013 automatically escaped as needed. Also, if file system or
8014 device paths are specified they are automatically turned
8015 into the appropriate mount or device unit names. Example:
8017 systemctl status /home
8018 systemctl status /dev/sda
8020 * The SysVConsole= configuration option has been removed from
8021 system.conf parsing.
8023 * The SysV search path is no longer exported on the D-Bus
8026 * The Names= option has been removed from unit file parsing.
8028 * There's a new man page bootup(7) detailing the boot process.
8030 * Every unit and every generator we ship with systemd now
8031 comes with full documentation. The self-explanatory boot is
8034 * A couple of services gained "systemd-" prefixes in their
8035 name if they wrap systemd code, rather than only external
8036 code. Among them fsck@.service which is now
8037 systemd-fsck@.service.
8039 * The HaveWatchdog property has been removed from the D-Bus
8042 * systemd.confirm_spawn= on the kernel command line should now
8045 * There's a new man page crypttab(5) which details all options
8046 we actually understand.
8048 * systemd-nspawn gained a new --capability= switch to pass
8049 additional capabilities to the container.
8051 * timedated will now read known NTP implementation unit names
8052 from /usr/lib/systemd/ntp-units.d/*.list,
8053 systemd-timedated-ntp.target has been removed.
8055 * journalctl gained a new switch "-b" that lists log data of
8056 the current boot only.
8058 * The notify socket is in the abstract namespace again, in
8059 order to support daemons which chroot() at start-up.
8061 * There is a new Storage= configuration option for journald
8062 which allows configuration of where log data should go. This
8063 also provides a way to disable journal logging entirely, so
8064 that data collected is only forwarded to the console, the
8065 kernel log buffer or another syslog implementation.
8067 * Many bugfixes and optimizations
8069 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8070 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8071 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8072 Shawn Landden, Tom Gundersen
8076 * "systemctl help <unit>" now shows the man page if one is
8079 * Several new man pages have been added.
8081 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8082 MaxLevelConsole= can now be specified in
8083 journald.conf. These options allow reducing the amount of
8084 data stored on disk or forwarded by the log level.
8086 * TimerSlackNSec= can now be specified in system.conf for
8087 PID1. This allows system-wide power savings.
8089 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8090 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8095 * logind is now capable of (optionally) handling power and
8096 sleep keys as well as the lid switch.
8098 * journalctl now understands the syntax "journalctl
8099 /usr/bin/avahi-daemon" to get all log output of a specific
8102 * CapabilityBoundingSet= in system.conf now also influences
8103 the capability bound set of usermode helpers of the kernel.
8105 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8106 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8107 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8108 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8112 * Note that we skipped 139 releases here in order to set the
8113 new version to something that is greater than both udev's
8114 and systemd's most recent version number.
8116 * udev: all udev sources are merged into the systemd source tree now.
8117 All future udev development will happen in the systemd tree. It
8118 is still fully supported to use the udev daemon and tools without
8119 systemd running, like in initramfs or other init systems. Building
8120 udev though, will require the *build* of the systemd tree, but
8121 udev can be properly *run* without systemd.
8123 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8124 should be used to create dead device nodes as workarounds for broken
8127 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8128 no longer supported. udev_monitor_new_from_netlink() needs to be
8129 used to subscribe to events.
8131 * udev: when udevd is started by systemd, processes which are left
8132 behind by forking them off of udev rules, are unconditionally cleaned
8133 up and killed now after the event handling has finished. Services or
8134 daemons must be started as systemd services. Services can be
8135 pulled-in by udev to get started, but they can no longer be directly
8136 forked by udev rules.
8138 * udev: the daemon binary is called systemd-udevd now and installed
8139 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8140 to adapt to that, create symlink, or rename the binary after building
8143 * libudev no longer provides these symbols:
8144 udev_monitor_from_socket()
8145 udev_queue_get_failed_list_entry()
8146 udev_get_{dev,sys,run}_path()
8147 The versions number was bumped and symbol versioning introduced.
8149 * systemd-loginctl and systemd-journalctl have been renamed
8150 to loginctl and journalctl to match systemctl.
8152 * The config files: /etc/systemd/systemd-logind.conf and
8153 /etc/systemd/systemd-journald.conf have been renamed to
8154 logind.conf and journald.conf. Package updates should rename
8155 the files to the new names on upgrade.
8157 * For almost all files the license is now LGPL2.1+, changed
8158 from the previous GPL2.0+. Exceptions are some minor stuff
8159 of udev (which will be changed to LGPL2.1 eventually, too),
8160 and the MIT licensed sd-daemon.[ch] library that is suitable
8161 to be used as drop-in files.
8163 * systemd and logind now handle system sleep states, in
8164 particular suspending and hibernating.
8166 * logind now implements a sleep/shutdown/idle inhibiting logic
8167 suitable for a variety of uses. Soonishly Lennart will blog
8168 about this in more detail.
8170 * var-run.mount and var-lock.mount are no longer provided
8171 (which previously bind mounted these directories to their new
8172 places). Distributions which have not converted these
8173 directories to symlinks should consider stealing these files
8174 from git history and add them downstream.
8176 * We introduced the Documentation= field for units and added
8177 this to all our shipped units. This is useful to make it
8178 easier to explore the boot and the purpose of the various
8181 * All smaller setup units (such as
8182 systemd-vconsole-setup.service) now detect properly if they
8183 are run in a container and are skipped when
8184 appropriate. This guarantees an entirely noise-free boot in
8185 Linux container environments such as systemd-nspawn.
8187 * A framework for implementing offline system updates is now
8188 integrated, for details see:
8189 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
8191 * A new service type Type=idle is available now which helps us
8192 avoiding ugly interleaving of getty output and boot status
8195 * There's now a system-wide CapabilityBoundingSet= option to
8196 globally reduce the set of capabilities for the
8197 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8198 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8199 even CAP_NET_ADMIN system-wide for secure systems.
8201 * There are now system-wide DefaultLimitXXX= options to
8202 globally change the defaults of the various resource limits
8203 for all units started by PID 1.
8205 * Harald Hoyer's systemd test suite has been integrated into
8206 systemd which allows easy testing of systemd builds in qemu
8207 and nspawn. (This is really awesome! Ask us for details!)
8209 * The fstab parser is now implemented as generator, not inside
8212 * systemctl will now warn you if .mount units generated from
8213 /etc/fstab are out of date due to changes in fstab that
8214 have not been read by systemd yet.
8216 * systemd is now suitable for usage in initrds. Dracut has
8217 already been updated to make use of this. With this in place
8218 initrds get a slight bit faster but primarily are much
8219 easier to introspect and debug since "systemctl status" in
8220 the host system can be used to introspect initrd services,
8221 and the journal from the initrd is kept around too.
8223 * systemd-delta has been added, a tool to explore differences
8224 between user/admin configuration and vendor defaults.
8226 * PrivateTmp= now affects both /tmp and /var/tmp.
8228 * Boot time status messages are now much prettier and feature
8229 proper english language. Booting up systemd has never been
8232 * Read-ahead pack files now include the inode number of all
8233 files to pre-cache. When the inode changes the pre-caching
8234 is not attempted. This should be nicer to deal with updated
8235 packages which might result in changes of read-ahead
8238 * We now temporaritly lower the kernel's read_ahead_kb variable
8239 when collecting read-ahead data to ensure the kernel's
8240 built-in read-ahead does not add noise to our measurements
8241 of necessary blocks to pre-cache.
8243 * There's now RequiresMountsFor= to add automatic dependencies
8244 for all mounts necessary for a specific file system path.
8246 * MountAuto= and SwapAuto= have been removed from
8247 system.conf. Mounting file systems at boot has to take place
8250 * nspawn now learned a new switch --uuid= to set the machine
8251 ID on the command line.
8253 * nspawn now learned the -b switch to automatically search
8256 * vt102 is now the default TERM for serial TTYs, upgraded from
8259 * systemd-logind now works on VT-less systems.
8261 * The build tree has been reorganized. The individual
8262 components now have directories of their own.
8264 * A new condition type ConditionPathIsReadWrite= is now available.
8266 * nspawn learned the new -C switch to create cgroups for the
8267 container in other hierarchies.
8269 * We now have support for hardware watchdogs, configurable in
8272 * The scheduled shutdown logic now has a public API.
8274 * We now mount /tmp as tmpfs by default, but this can be
8275 masked and /etc/fstab can override it.
8277 * Since udisks does not make use of /media anymore we are not
8278 mounting a tmpfs on it anymore.
8280 * journalctl gained a new --local switch to only interleave
8281 locally generated journal files.
8283 * We can now load the IMA policy at boot automatically.
8285 * The GTK tools have been split off into a systemd-ui.
8287 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8288 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8289 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8290 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8291 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8292 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8293 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8294 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8295 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8300 * This is mostly a bugfix release
8302 * Support optional initialization of the machine ID from the
8303 KVM or container configured UUID.
8305 * Support immediate reboots with "systemctl reboot -ff"
8307 * Show /etc/os-release data in systemd-analyze output
8309 * Many bugfixes for the journal, including endianness fixes and
8310 ensuring that disk space enforcement works
8312 * sd-login.h is C++ compatible again
8314 * Extend the /etc/os-release format on request of the Debian
8317 * We now refuse non-UTF8 strings used in various configuration
8318 and unit files. This is done to ensure we do not pass invalid
8319 data over D-Bus or expose it elsewhere.
8321 * Register Mimo USB Screens as suitable for automatic seat
8324 * Read SELinux client context from journal clients in a race
8327 * Reorder configuration file lookup order. /etc now always
8328 overrides /run in order to allow the administrator to always
8329 and unconditionally override vendor-supplied or
8330 automatically generated data.
8332 * The various user visible bits of the journal now have man
8333 pages. We still lack man pages for the journal API calls
8336 * We now ship all man pages in HTML format again in the
8339 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8340 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8341 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8346 * This is mostly a bugfix release
8348 * systems lacking /etc/os-release are no longer supported.
8350 * Various functionality updates to libsystemd-login.so
8352 * Track class of PAM logins to distinguish greeters from
8355 Contributions from: Kay Sievers, Lennart Poettering, Michael
8360 * This is an important bugfix release for v41.
8362 * Building man pages is now optional which should be useful
8363 for those building systemd from git but unwilling to install
8366 * Watchdog support for supervising services is now usable. In
8367 a future release support for hardware watchdogs
8368 (i.e. /dev/watchdog) will be added building on this.
8370 * Service start rate limiting is now configurable and can be
8371 turned off per service. When a start rate limit is hit a
8372 reboot can automatically be triggered.
8374 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8376 Contributions from: Benjamin Franzke, Bill Nottingham,
8377 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8378 Schmidt, Michał Górny, Piotr Drąg
8382 * The systemd binary is installed /usr/lib/systemd/systemd now;
8383 An existing /sbin/init symlink needs to be adapted with the
8386 * The code that loads kernel modules has been ported to invoke
8387 libkmod directly, instead of modprobe. This means we do not
8388 support systems with module-init-tools anymore.
8390 * Watchdog support is now already useful, but still not
8393 * A new kernel command line option systemd.setenv= is
8394 understood to set system wide environment variables
8395 dynamically at boot.
8397 * We now limit the set of capabilities of systemd-journald.
8399 * We now set SIGPIPE to ignore by default, since it only is
8400 useful in shell pipelines, and has little use in general
8401 code. This can be disabled with IgnoreSIPIPE=no in unit
8404 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8405 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8410 * This is mostly a bugfix release
8412 * We now expose the reason why a service failed in the
8413 "Result" D-Bus property.
8415 * Rudimentary service watchdog support (will be completed over
8416 the next few releases.)
8418 * When systemd forks off in order execute some service we will
8419 now immediately changes its argv[0] to reflect which process
8420 it will execute. This is useful to minimize the time window
8421 with a generic argv[0], which makes bootcharts more useful
8423 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8424 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8425 Mike Kazantsev, Ray Strode
8429 * This is mostly a test release, but incorporates many
8432 * New systemd-cgtop tool to show control groups by their
8435 * Linking against libacl for ACLs is optional again. If
8436 disabled, support tracking device access for active logins
8437 goes becomes unavailable, and so does access to the user
8438 journals by the respective users.
8440 * If a group "adm" exists, journal files are automatically
8441 owned by them, thus allow members of this group full access
8442 to the system journal as well as all user journals.
8444 * The journal now stores the SELinux context of the logging
8445 client for all entries.
8447 * Add C++ inclusion guards to all public headers
8449 * New output mode "cat" in the journal to print only text
8450 messages, without any meta data like date or time.
8452 * Include tiny X server wrapper as a temporary stop-gap to
8453 teach XOrg udev display enumeration. This is used by display
8454 managers such as gdm, and will go away as soon as XOrg
8455 learned native udev hotplugging for display devices.
8457 * Add new systemd-cat tool for executing arbitrary programs
8458 with STDERR/STDOUT connected to the journal. Can also act as
8459 BSD logger replacement, and does so by default.
8461 * Optionally store all locally generated coredumps in the
8462 journal along with meta data.
8464 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8465 writing short strings to files (for usage for /sys), and for
8466 creating symlinks, character and block device nodes.
8468 * New unit file option ControlGroupPersistent= to make cgroups
8469 persistent, following the mechanisms outlined in
8470 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
8472 * Support multiple local RTCs in a sane way
8474 * No longer monopolize IO when replaying readahead data on
8475 rotating disks, since we might starve non-file-system IO to
8476 death, since fanotify() will not see accesses done by blkid,
8479 * Do not show kernel threads in systemd-cgls anymore, unless
8480 requested with new -k switch.
8482 Contributions from: Dan Horák, Kay Sievers, Lennart
8483 Poettering, Michal Schmidt
8487 * This is mostly a test release, but incorporates many
8490 * The git repository moved to:
8491 git://anongit.freedesktop.org/systemd/systemd
8492 ssh://git.freedesktop.org/git/systemd/systemd
8494 * First release with the journal
8495 http://0pointer.de/blog/projects/the-journal.html
8497 * The journal replaces both systemd-kmsg-syslogd and
8498 systemd-stdout-bridge.
8500 * New sd_pid_get_unit() API call in libsystemd-logind
8502 * Many systemadm clean-ups
8504 * Introduce remote-fs-pre.target which is ordered before all
8505 remote mounts and may be used to start services before all
8508 * Added Mageia support
8510 * Add bash completion for systemd-loginctl
8512 * Actively monitor PID file creation for daemons which exit in
8513 the parent process before having finished writing the PID
8514 file in the daemon process. Daemons which do this need to be
8515 fixed (i.e. PID file creation must have finished before the
8516 parent exits), but we now react a bit more gracefully to them.
8518 * Add colourful boot output, mimicking the well-known output
8519 of existing distributions.
8521 * New option PassCredentials= for socket units, for
8522 compatibility with a recent kernel ABI breakage.
8524 * /etc/rc.local is now hooked in via a generator binary, and
8525 thus will no longer act as synchronization point during
8528 * systemctl list-unit-files now supports --root=.
8530 * systemd-tmpfiles now understands two new commands: z, Z for
8531 relabelling files according to the SELinux database. This is
8532 useful to apply SELinux labels to specific files in /sys,
8535 * Output of SysV services is now forwarded to both the console
8536 and the journal by default, not only just the console.
8538 * New man pages for all APIs from libsystemd-login.
8540 * The build tree got reorganized and the build system is a
8541 lot more modular allowing embedded setups to specifically
8542 select the components of systemd they are interested in.
8544 * Support for Linux systems lacking the kernel VT subsystem is
8547 * configure's --with-rootdir= got renamed to
8548 --with-rootprefix= to follow the naming used by udev and
8551 * Unless specified otherwise we will now install to /usr instead
8552 of /usr/local by default.
8554 * Processes with '@' in argv[0][0] are now excluded from the
8555 final shut-down killing spree, following the logic explained
8557 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
8559 * All processes remaining in a service cgroup when we enter
8560 the START or START_PRE states are now killed with
8561 SIGKILL. That means it is no longer possible to spawn
8562 background processes from ExecStart= lines (which was never
8563 supported anyway, and bad style).
8565 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8566 reloading of units together.
8568 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
8569 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8570 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8571 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8572 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob