1 systemd System and Service Manager
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 three 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. %E will expand to either
152 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
154 * The ExecStart= lines of unit files are no longer required to
155 reference absolute paths. If non-absolute paths are specified the
156 specified binary name is searched within the service manager's
157 built-in $PATH, which may be queried with 'elogind-path
158 search-binaries-default'. It's generally recommended to continue to
159 use absolute paths for all binaries specified in unit files.
161 * Units gained a new load state "bad-setting", which is used when a
162 unit file was loaded, but contained fatal errors which prevent it
163 from being started (for example, a service unit has been defined
164 lacking both ExecStart= and ExecStop= lines).
166 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
167 support alternative debuggers, for example lldb. The old name
168 continues to be available however, for compatibility reasons. Use the
169 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
170 to pick an alternative debugger instead of the default gdb.
172 * systemctl and the other tools will now output escape sequences that
173 generate proper clickable hyperlinks in various terminal emulators
174 where useful (for example, in the "systemctl status" output you can
175 now click on the unit file name to quickly open it in the
176 editor/viewer of your choice). Note that not all terminal emulators
177 support this functionality yet, but many do. Unfortunately, the
178 "less" pager doesn't support this yet, hence this functionality is
179 currently automatically turned off when a pager is started (which
180 happens quite often due to auto-paging). We hope to remove this
181 limitation as soon as "less" learns these escape sequences. This new
182 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
183 environment variable. For details on these escape sequences see:
184 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
186 * networkd's .network files now support a new IPv6MTUBytes= option for
187 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
188 option in the [Route] section to configure the MTU to use for
189 specific routes. It also gained support for configuration of the DHCP
190 "UserClass" option through the new UserClass= setting. It gained
191 three new options in the new [CAN] section for configuring CAN
192 networks. The MULTICAST and ALLMULTI interface flags may now be
193 controlled explicitly with the new Multicast= and AllMulticast=
196 * networkd will now automatically make use of the kernel's route
197 expiration feature, if it is available.
199 * udevd's .link files now support setting the number of receive and
200 transmit channels, using the RxChannels=, TxChannels=,
201 OtherChannels=, CombinedChannels= settings.
203 * Support for UDPSegmentationOffload= has been removed, given its
204 limited support in hardware, and waning software support.
206 * networkd's .netdev files now support creating "netdevsim" interfaces.
208 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
209 to query the unit belonging to a specific kernel control group.
211 * elogind-analyze gained a new verb "cat-config", which may be used to
212 dump the contents of any configuration file, with all its matching
213 drop-in files added in, and honouring the usual search and masking
214 logic applied to elogind configuration files. For example use
215 "elogind-analyze cat-config elogind/system.conf" to get the complete
216 system configuration file of elogind how it would be loaded by PID 1
217 itself. Similar to this, various tools such as elogind-tmpfiles or
218 elogind-sysusers, gained a new option "--cat-config", which does the
219 corresponding operation for their own configuration settings. For
220 example, "elogind-tmpfiles --cat-config" will now output the full
221 list of tmpfiles.d/ lines in place.
223 * timedatectl gained three new verbs: "show" shows bus properties of
224 elogind-timedated, "timesync-status" shows the current NTP
225 synchronization state of elogind-timesyncd, and "show-timesync"
226 shows bus properties of elogind-timesyncd.
228 * elogind-timesyncd gained a bus interface on which it exposes details
231 * elogind-nspawn gained a new --rlimit= switch for setting initial
232 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
233 understood by elogind-timedated. It takes a colon-separated list of
234 unit names of NTP client services. The list is used by
235 "timedatectl set-ntp".
237 resource limits for the container payload. There's a new switch
238 --hostname= to explicitly override the container's hostname. A new
239 --no-new-privileges= switch may be used to control the
240 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
241 --oom-score-adjust= switch controls the OOM scoring adjustment value
242 for the payload. The new --cpu-affinity= switch controls the CPU
243 affinity of the container payload. The new --resolv-conf= switch
244 allows more detailed control of /etc/resolv.conf handling of the
245 container. Similarly, the new --timezone= switch allows more detailed
246 control of /etc/localtime handling of the container.
248 * elogind-detect-virt gained a new --list switch, which will print a
249 list of all currently known VM and container environments.
251 * Support for "Portable Services" has been added, see
252 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
253 experimental, but this is expected to change soon. Reflecting this
254 experimental state, the "portablectl" binary is not installed into
255 /usr/bin yet. The binary has to be called with the full path
256 /usr/lib/elogind/portablectl instead.
258 * journalctl's and systemctl's -o switch now knows a new log output
259 mode "with-unit". The output it generates is very similar to the
260 regular "short" mode, but displays the unit name instead of the
261 syslog tag for each log line. Also, the date is shown with timezone
262 information. This mode is probably more useful than the classic
263 "short" output mode for most purposes, except where pixel-perfect
264 compatibility with classic /var/log/messages formatting is required.
266 * A new --dump-bus-properties switch has been added to the elogind
267 binary, which may be used to dump all supported D-Bus properties.
268 (Options which are still supported, but are deprecated, are *not*
271 * sd-bus gained a set of new calls:
272 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
273 enable/disable the "floating" state of a bus slot object,
274 i.e. whether the slot object pins the bus it is allocated for into
275 memory or if the bus slot object gets disconnected when the bus goes
276 away. sd_bus_open_with_description(),
277 sd_bus_open_user_with_description(),
278 sd_bus_open_system_with_description() may be used to allocate bus
279 objects and set their description string already during allocation.
281 * sd-event gained support for watching inotify events from the event
282 loop, in an efficient way, sharing inotify handles between multiple
283 users. For this a new function sd_event_add_inotify() has been added.
285 * sd-event and sd-bus gained support for calling special user-supplied
286 destructor functions for userdata pointers associated with
287 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
288 functions sd_bus_slot_set_destroy_callback,
289 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
290 sd_bus_track_get_destroy_callback,
291 sd_event_source_set_destroy_callback,
292 sd_event_source_get_destroy_callback have been added.
294 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
296 * PID 1 will now automatically reschedule .timer units whenever the
297 local timezone changes. (They previously got rescheduled
298 automatically when the system clock changed.)
300 * New documentation has been added to document cgroups delegation,
301 portable services and the various code quality tools we have set up:
303 https://github.com/systemd/systemd/blob/master/doc/CGROUP_DELEGATION.md
304 https://github.com/systemd/systemd/blob/master/doc/PORTABLE_SERVICES.md
305 https://github.com/systemd/systemd/blob/master/doc/CODE_QUALITY.md
307 * The Boot Loader Specification has been added to the source tree.
309 https://github.com/systemd/systemd/blob/master/doc/BOOT_LOADER_SPECIFICATION.md
311 While moving it into our source tree we have updated it and further
312 changes are now accepted through the usual github PR workflow.
314 * pam_elogind will now look for PAM userdata fields elogind.memory_max,
315 elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
316 earlier PAM modules. The data in these fields is used to initialize
317 the session scope's resource properties. Thus external PAM modules
318 may now configure per-session limits, for example sourced from
319 external user databases.
321 * socket units with Accept=yes will now maintain a "refused" counter in
322 addition to the existing "accepted" counter, counting connections
323 refused due to the enforced limits.
325 * The "elogind-path search-binaries-default" command may now be use to
326 query the default, built-in $PATH PID 1 will pass to the services it
329 * A new unit file setting PrivateMounts= has been added. It's a boolean
330 option. If enabled the unit's processes are invoked in their own file
331 system namespace. Note that this behaviour is also implied if any
332 other file system namespacing options (such as PrivateTmp=,
333 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
334 primarily useful for services that do not use any of the other file
335 system namespacing options. One such service is elogind-udevd.service
336 wher this is now used by default.
338 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
339 when the system is booted in UEFI "secure mode".
341 * A new unit "system-update-pre.target" is added, which defines an
342 optional synchronization point for offline system updates, as
343 implemented by the pre-existing "system-update.target" unit. It
344 allows ordering services before the service that executes the actual
345 update process in a generic way.
347 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
348 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
349 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
350 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
351 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
352 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
353 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
354 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
355 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
356 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
357 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
358 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
359 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
360 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
361 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
362 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
363 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
364 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
365 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
366 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
367 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
368 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
369 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
370 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
371 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
372 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
373 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
374 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
375 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
381 * The MemoryAccounting= unit property now defaults to on. After
382 discussions with the upstream control group maintainers we learnt
383 that the negative impact of cgroup memory accounting on current
384 kernels is finally relatively minimal, so that it should be safe to
385 enable this by default without affecting system performance. Besides
386 memory accounting only task accounting is turned on by default, all
387 other forms of resource accounting (CPU, IO, IP) remain off for now,
388 because it's not clear yet that their impact is small enough to move
389 from opt-in to opt-out. We recommend downstreams to leave memory
390 accounting on by default if kernel 4.14 or higher is primarily
391 used. On very resource constrained systems or when support for old
392 kernels is a necessity, -Dmemory-accounting-default=false can be used
393 to revert this change.
395 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
396 %udev_rules_update) and the journal catalog (%journal_catalog_update)
397 from the upgrade scriptlets of individual packages now do nothing.
398 Transfiletriggers have been added which will perform those updates
399 once at the end of the transaction.
401 Similar transfiletriggers have been added to execute any sysctl.d
402 and binfmt.d rules. Thus, it should be unnecessary to provide any
403 scriptlets to execute this configuration from package installation
406 * elogind-sysusers gained a mode where the configuration to execute is
407 specified on the command line, but this configuration is not executed
408 directly, but instead it is merged with the configuration on disk,
409 and the result is executed. This is useful for package installation
410 scripts which want to create the user before installing any files on
411 disk (in case some of those files are owned by that user), while
412 still allowing local admin overrides.
414 This functionality is exposed to rpm scriptlets through a new
415 %sysusers_create_package macro. Old %sysusers_create and
416 %sysusers_create_inline macros are deprecated.
418 A transfiletrigger for sysusers.d configuration is now installed,
419 which means that it should be unnecessary to call elogind-sysusers from
420 package installation scripts, unless the package installs any files
421 owned by those newly-created users, in which case
422 %sysusers_create_package should be used.
424 * Analogous change has been done for elogind-tmpfiles: it gained a mode
425 where the command-line configuration is merged with the configuration
426 on disk. This is exposed as the new %tmpfiles_create_package macro,
427 and %tmpfiles_create is deprecated. A transfiletrigger is installed
428 for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
429 from package installation scripts.
431 * sysusers.d configuration for a user may now also specify the group
432 number, in addition to the user number ("u username 123:456"), or
433 without the user number ("u username -:456").
435 * Configution items for elogind-sysusers can now be specified as
436 positional arguments when the new --inline switch is used.
438 * The login shell of users created through sysusers.d may now be
439 specified (previously, it was always /bin/sh for root and
440 /sbin/nologin for other users).
442 * elogind-analyze gained a new --global switch to look at global user
443 configuration. It also gained a unit-paths verb to list the unit load
444 paths that are compiled into elogind (which can be used with
445 --elogind, --user, or --global).
447 * udevadm trigger gained a new --settle/-w option to wait for any
448 triggered events to finish (but just those, and not any other events
449 which are triggered meanwhile).
451 * The action that elogind takes when the lid is closed and the
452 machine is connected to external power can now be configured using
453 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
454 was determined by HandleLidSwitch=, and, for backwards compatibility,
455 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
457 * journalctl will periodically call sd_journal_process() to make it
458 resilient against inotify queue overruns when journal files are
459 rotated very quickly.
461 * Two new functions in libelogind — sd_bus_get_n_queued_read and
462 sd_bus_get_n_queued_write — may be used to check the number of
463 pending bus messages.
465 * elogind gained a new
466 org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
467 which can be used to migrate foreign processes to scope and service
468 units. The primary user for this new API is elogind itself: the
469 elogind --user instance uses this call of the elogind --system
470 instance to migrate processes if it itself gets the request to
471 migrate processes and the kernel refuses this due to access
472 restrictions. Thanks to this "elogind-run --scope --user …" works
473 again in pure cgroups v2 environments when invoked from the user
476 * A new TemporaryFileSystem= setting can be used to mask out part of
477 the real file system tree with tmpfs mounts. This may be combined
478 with BindPaths= and BindReadOnlyPaths= to hide files or directories
479 not relevant to the unit, while still allowing some paths lower in
480 the tree to be accessed.
482 ProtectHome=tmpfs may now be used to hide user home and runtime
483 directories from units, in a way that is mostly equivalent to
484 "TemporaryFileSystem=/home /run/user /root".
486 * Non-service units are now started with KeyringMode=shared by default.
487 This means that mount and swapon and other mount tools have access
488 to keys in the main keyring.
490 * /sys/fs/bpf is now mounted automatically.
492 * QNX virtualization is now detected by elogind-detect-virt and may
493 be used in ConditionVirtualization=.
495 * IPAccounting= may now be enabled also for slice units.
497 * A new -Dsplit-bin= build configuration switch may be used to specify
498 whether bin and sbin directories are merged, or if they should be
499 included separately in $PATH and various listings of executable
500 directories. The build configuration scripts will try to autodetect
501 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
502 system, but distributions are encouraged to configure this
505 * A new -Dok-color= build configuration switch may be used to change
506 the colour of "OK" status messages.
508 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
509 PrivateNetwork=yes was buggy in previous versions of elogind. This
510 means that after the upgrade and daemon-reexec, any such units must
513 * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
514 will not exclude read-only files owned by root from cleanup.
516 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
517 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
518 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
519 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
520 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
521 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
522 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
523 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
524 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
525 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
526 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
527 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
528 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
529 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
530 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
531 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
537 * Some keyboards come with a zoom see-saw or rocker which until now got
538 mapped to the Linux "zoomin/out" keys in hwdb. However, these
539 keycodes are not recognized by any major desktop. They now produce
540 Up/Down key events so that they can be used for scrolling.
542 * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
543 slightly: previously, if an argument was specified for lines of this
544 type (i.e. the right-most column was set) this string was appended to
545 existing files each time elogind-tmpfiles was run. This behaviour was
546 different from what the documentation said, and not particularly
547 useful, as repeated elogind-tmpfiles invocations would not be
548 idempotent and grow such files without bounds. With this release
549 behaviour has been altered slightly, to match what the documentation
550 says: lines of this type only have an effect if the indicated files
551 don't exist yet, and only then the argument string is written to the
554 * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
555 elogind-tmpfiles behaviour: previously, read-only files owned by root
556 were always excluded from the file "aging" algorithm (i.e. the
557 automatic clean-up of directories like /tmp based on
558 atime/mtime/ctime). We intend to drop this restriction, and age files
559 by default even when owned by root and read-only. This behaviour was
560 inherited from older tools, but there have been requests to remove
561 it, and it's not obvious why this restriction was made in the first
562 place. Please speak up now, if you are aware of software that reqires
563 this behaviour, otherwise we'll remove the restriction in v238.
565 * A new environment variable $SYSTEMD_OFFLINE is now understood by
566 systemctl. It takes a boolean argument. If on, systemctl assumes it
567 operates on an "offline" OS tree, and will not attempt to talk to the
568 service manager. Previously, this mode was implicitly enabled if a
569 chroot() environment was detected, and this new environment variable
570 now provides explicit control.
572 * .path and .socket units may now be created transiently, too.
573 Previously only service, mount, automount and timer units were
574 supported as transient units. The elogind-run tool has been updated
575 to expose this new functionality, you may hence use it now to bind
576 arbitrary commands to path or socket activation on-the-fly from the
577 command line. Moreover, almost all properties are now exposed for the
578 unit types that already supported transient operation.
580 * The elogind-mount command gained support for a new --owner= parameter
581 which takes a user name, which is then resolved and included in uid=
582 and gid= mount options string of the file system to mount.
584 * A new unit condition ConditionControlGroupController= has been added
585 that checks whether a specific cgroup controller is available.
587 * Unit files, udev's .link files, and elogind-networkd's .netdev and
588 .network files all gained support for a new condition
589 ConditionKernelVersion= for checking against specific kernel
592 * In elogind-networkd, the [IPVLAN] section in .netdev files gained
593 support for configuring device flags in the Flags= setting. In the
594 same files, the [Tunnel] section gained support for configuring
595 AllowLocalRemote=. The [Route] section in .network files gained
596 support for configuring InitialCongestionWindow=,
597 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
598 understands RapidCommit=.
600 * elogind-networkd's DHCPv6 support gained support for Prefix
603 * sd-bus gained support for a new "watch-bind" feature. When this
604 feature is enabled, an sd_bus connection may be set up to connect to
605 an AF_UNIX socket in the file system as soon as it is created. This
606 functionality is useful for writing early-boot services that
607 automatically connect to the system bus as soon as it is started,
608 without ugly time-based polling. elogind-networkd and
609 elogind-resolved have been updated to make use of this
610 functionality. busctl exposes this functionality in a new
611 --watch-bind= command line switch.
613 * sd-bus will now optionally synthesize a local "Connected" signal as
614 soon as a D-Bus connection is set up fully. This message mirrors the
615 already existing "Disconnected" signal which is synthesized when the
616 connection is terminated. This signal is generally useful but
617 particularly handy in combination with the "watch-bind" feature
618 described above. Synthesizing of this message has to be requested
619 explicitly through the new API call sd_bus_set_connected_signal(). In
620 addition a new call sd_bus_is_ready() has been added that checks
621 whether a connection is fully set up (i.e. between the "Connected" and
622 "Disconnected" signals).
624 * sd-bus gained two new calls sd_bus_request_name_async() and
625 sd_bus_release_name_async() for asynchronously registering bus
626 names. Similar, there is now sd_bus_add_match_async() for installing
627 a signal match asynchronously. All of elogind's own services have
628 been updated to make use of these calls. Doing these operations
629 asynchronously has two benefits: it reduces the risk of deadlocks in
630 case of cyclic dependencies between bus services, and it speeds up
631 service initialization since synchronization points for bus
632 round-trips are removed.
634 * sd-bus gained two new calls sd_bus_match_signal() and
635 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
636 and sd_bus_add_match_async() but instead of taking a D-Bus match
637 string take match fields as normal function parameters.
639 * sd-bus gained two new calls sd_bus_set_sender() and
640 sd_bus_message_set_sender() for setting the sender name of outgoing
641 messages (either for all outgoing messages or for just one specific
642 one). These calls are only useful in direct connections as on
643 brokered connections the broker fills in the sender anyway,
644 overwriting whatever the client filled in.
646 * sd-event gained a new pseudo-handle that may be specified on all API
647 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
648 used this refers to the default event loop object of the calling
649 thread. Note however that this does not implicitly allocate one —
650 which has to be done prior by using sd_event_default(). Similarly
651 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
652 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
653 to the default bus of the specified type of the calling thread. Here
654 too this does not implicitly allocate bus connection objects, this
655 has to be done prior with sd_bus_default() and friends.
657 * sd-event gained a new call pair
658 sd_event_source_{get|set}_io_fd_own(). This may be used to request
659 automatic closure of the file descriptor an IO event source watches
660 when the event source is destroyed.
662 * elogind-networkd gained support for natively configuring WireGuard
665 * In previous versions elogind synthesized user records both for the
666 "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
667 internally. In order to simplify distribution-wide renames of the
668 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
669 new transitional flag file has been added: if
670 /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
671 user and group record within the elogind codebase is disabled.
673 * elogind-notify gained a new --uid= option for selecting the source
674 user/UID to use for notification messages sent to the service
677 * journalctl gained a new --grep= option to list only entries in which
678 the message matches a certain pattern. By default matching is case
679 insensitive if the pattern is lowercase, and case sensitive
680 otherwise. Option --case-sensitive=yes|no can be used to override
681 this an specify case sensitivity or case insensitivity.
683 * There's now a "elogind-analyze service-watchdogs" command for printing
684 the current state of the service runtime watchdog, and optionally
685 enabling or disabling the per-service watchdogs system-wide if given a
686 boolean argument (i.e. the concept you configure in WatchdogSec=), for
687 debugging purposes. There's also a kernel command line option
688 elogind.service_watchdogs= for controlling the same.
690 * Two new "log-level" and "log-target" options for elogind-analyze were
691 added that merge the now deprecated get-log-level, set-log-level and
692 get-log-target, set-log-target pairs. The deprecated options are still
693 understood for backwards compatibility. The two new options print the
694 current value when no arguments are given, and set them when a
695 level/target is given as an argument.
697 * sysusers.d's "u" lines now optionally accept both a UID and a GID
698 specification, separated by a ":" character, in order to create users
699 where UID and GID do not match.
701 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
702 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
703 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
704 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
705 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
706 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
707 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
708 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
709 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
710 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
711 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
712 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
713 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
714 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
715 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
716 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
717 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
718 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
719 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
720 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
727 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
728 in v235 has been extended to also set the dummy.ko module option
729 numdummies=0, preventing the kernel from automatically creating
730 dummy0. All dummy interfaces must now be explicitly created.
732 * Unknown '%' specifiers in configuration files are now rejected. This
733 applies to units and tmpfiles.d configuration. Any percent characters
734 that are followed by a letter or digit that are not supposed to be
735 interpreted as the beginning of a specifier should be escaped by
736 doubling ("%%"). (So "size=5%" is still accepted, as well as
737 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
738 valid specifiers today.)
740 * systemd-resolved now maintains a new dynamic
741 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
742 recommended to make /etc/resolv.conf a symlink to it. This file
743 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
744 includes dynamically acquired search domains, achieving more correct
745 DNS resolution by software that bypasses local DNS APIs such as NSS.
747 * The "uaccess" udev tag has been dropped from /dev/kvm and
748 /dev/dri/renderD*. These devices now have the 0666 permissions by
749 default (but this may be changed at build-time). /dev/dri/renderD*
750 will now be owned by the "render" group along with /dev/kfd.
752 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
753 systemd-journal-gatewayd.service and
754 systemd-journal-upload.service. This means "nss-systemd" must be
755 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
756 services are resolved properly.
758 * In /etc/fstab two new mount options are now understood:
759 x-systemd.makefs and x-systemd.growfs. The former has the effect that
760 the configured file system is formatted before it is mounted, the
761 latter that the file system is resized to the full block device size
762 after it is mounted (i.e. if the file system is smaller than the
763 partition it resides on, it's grown). This is similar to the fsck
764 logic in /etc/fstab, and pulls in systemd-makefs@.service and
765 systemd-growfs@.service as necessary, similar to
766 systemd-fsck@.service. Resizing is currently only supported on ext4
769 * In systemd-networkd, the IPv6 RA logic now optionally may announce
770 DNS server and domain information.
772 * Support for the LUKS2 on-disk format for encrypted partitions has
773 been added. This requires libcryptsetup2 during compilation and
776 * The systemd --user instance will now signal "readiness" when its
777 basic.target unit has been reached, instead of when the run queue ran
778 empty for the first time.
780 * Tmpfiles.d with user configuration are now also supported.
781 systemd-tmpfiles gained a new --user switch, and snippets placed in
782 ~/.config/user-tmpfiles.d/ and corresponding directories will be
783 executed by systemd-tmpfiles --user running in the new
784 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
785 running in the user session.
787 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
788 %S resolves to the top-level state directory (/var/lib for the system
789 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
790 top-level cache directory (/var/cache for the system instance,
791 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
792 logs directory (/var/log for the system instance,
793 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
794 existing %t specifier, that resolves to the top-level runtime
795 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
798 * journalctl learnt a new parameter --output-fields= for limiting the
799 set of journal fields to output in verbose and JSON output modes.
801 * systemd-timesyncd's configuration file gained a new option
802 RootDistanceMaxSec= for setting the maximum root distance of servers
803 it'll use, as well as the new options PollIntervalMinSec= and
804 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
806 * bootctl gained a new command "list" for listing all available boot
807 menu items on systems that follow the boot loader specification.
809 * systemctl gained a new --dry-run switch that shows what would be done
810 instead of doing it, and is currently supported by the shutdown and
813 * ConditionSecurity= can now detect the TOMOYO security module.
815 * Unit file [Install] sections are now also respected in unit drop-in
816 files. This is intended to be used by drop-ins under /usr/lib/.
818 * systemd-firstboot may now also set the initial keyboard mapping.
820 * Udev "changed" events for devices which are exposed as systemd
821 .device units are now propagated to units specified in
822 ReloadPropagatedFrom= as reload requests.
824 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
825 unit template name (i.e. a name in the form of 'foobar@.service',
826 without the instance component between the '@' and - the '.'), then
827 the escaped sysfs path of the device is automatically used as the
830 * SystemCallFilter= in unit files has been extended so that an "errno"
831 can be specified individually for each system call. Example:
832 SystemCallFilter=~uname:EILSEQ.
834 * The cgroup delegation logic has been substantially updated. Delegate=
835 now optionally takes a list of controllers (instead of a boolean, as
836 before), which lists the controllers to delegate at least.
838 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
840 * A new LogLevelMax= setting configures the maximum log level any
841 process of the service may log at (i.e. anything with a lesser
842 priority than what is specified is automatically dropped). A new
843 LogExtraFields= setting allows configuration of additional journal
844 fields to attach to all log records generated by any of the unit's
847 * New StandardInputData= and StandardInputText= settings along with the
848 new option StandardInput=data may be used to configure textual or
849 binary data that shall be passed to the executed service process via
850 standard input, encoded in-line in the unit file.
852 * StandardInput=, StandardOutput= and StandardError= may now be used to
853 connect stdin/stdout/stderr of executed processes directly with a
854 file or AF_UNIX socket in the file system, using the new "file:" option.
856 * A new unit file option CollectMode= has been added, that allows
857 tweaking the garbage collection logic for units. It may be used to
858 tell systemd to garbage collect units that have failed automatically
859 (normally it only GCs units that exited successfully). systemd-run
860 and systemd-mount expose this new functionality with a new -G option.
862 * "machinectl bind" may now be used to bind mount non-directories
863 (i.e. regularfiles, devices, fifos, sockets).
865 * systemd-analyze gained a new verb "calendar" for validating and
866 testing calendar time specifications to use for OnCalendar= in timer
867 units. Besides validating the expression it will calculate the next
868 time the specified expression would elapse.
870 * In addition to the pre-existing FailureAction= unit file setting
871 there's now SuccessAction=, for configuring a shutdown action to
872 execute when a unit completes successfully. This is useful in
873 particular inside containers that shall terminate after some workload
874 has been completed. Also, both options are now supported for all unit
875 types, not just services.
877 * networkds's IP rule support gained two new options
878 IncomingInterface= and OutgoingInterface= for configuring the incoming
879 and outgoing interfaces of configured rules. systemd-networkd also
880 gained support for "vxcan" network devices.
882 * networkd gained a new setting RequiredForOnline=, taking a
883 boolean. If set, systemd-wait-online will take it into consideration
884 when determining that the system is up, otherwise it will ignore the
885 interface for this purpose.
887 * The sd_notify() protocol gained support for a new operation: with
888 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
889 store again, ahead of POLLHUP or POLLERR when they are removed
892 * A new document doc/UIDS-GIDS.md has been added to the source tree,
893 that documents the UID/GID range and assignment assumptions and
894 requirements of systemd.
896 * The watchdog device PID 1 will ping may now be configured through the
897 WatchdogDevice= configuration file setting, or by setting the
898 systemd.watchdog_service= kernel commandline option.
900 * systemd-resolved's gained support for registering DNS-SD services on
901 the local network using MulticastDNS. Services may either be
902 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
903 the same dir below /run, /usr/lib), or through its D-Bus API.
905 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
906 extend the effective start, runtime, and stop time. The service must
907 continue to send EXTEND_TIMEOUT_USEC within the period specified to
908 prevent the service manager from making the service as timedout.
910 * elogind-resolved's DNSSEC support gained support for RFC 8080
911 (Ed25519 keys and signatures).
913 * The elogind-resolve command line tool gained a new set of options
914 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
915 --set-nta= and --revert to configure per-interface DNS configuration
916 dynamically during runtime. It's useful for pushing DNS information
917 into elogind-resolved from DNS hook scripts that various interface
918 managing software supports (such as pppd).
920 * elogind-nspawn gained a new --network-namespace-path= command line
921 option, which may be used to make a container join an existing
922 network namespace, by specifying a path to a "netns" file.
924 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
925 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
926 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
927 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
928 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
929 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
930 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
931 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
932 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
933 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
934 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
935 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
936 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
937 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
938 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
939 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
940 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
941 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
942 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
943 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
944 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
945 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
946 Jędrzejewski-Szmek, Zeal Jagannatha
952 * INCOMPATIBILITY: systemd-logind.service and other long-running
953 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
954 communication with the outside. This generally improves security of
955 the system, and is in almost all cases a safe and good choice, as
956 these services do not and should not provide any network-facing
957 functionality. However, systemd-logind uses the glibc NSS API to
958 query the user database. This creates problems on systems where NSS
959 is set up to directly consult network services for user database
960 lookups. In particular, this creates incompatibilities with the
961 "nss-nis" module, which attempts to directly contact the NIS/YP
962 network servers it is configured for, and will now consistently
963 fail. In such cases, it is possible to turn off IP sandboxing for
964 systemd-logind.service (set IPAddressDeny= in its [Service] section
965 to the empty string, via a .d/ unit file drop-in). Downstream
966 distributions might want to update their nss-nis packaging to include
967 such a drop-in snippet, accordingly, to hide this incompatibility
968 from the user. Another option is to make use of glibc's nscd service
969 to proxy such network requests through a privilege-separated, minimal
970 local caching daemon, or to switch to more modern technologies such
971 sssd, whose NSS hook-ups generally do not involve direct network
972 access. In general, we think it's definitely time to question the
973 implementation choices of nss-nis, i.e. whether it's a good idea
974 today to embed a network-facing loadable module into all local
975 processes that need to query the user database, including the most
976 trivial and benign ones, such as "ls". For more details about
977 IPAddressDeny= see below.
979 * A new modprobe.d drop-in is now shipped by default that sets the
980 bonding module option max_bonds=0. This overrides the kernel default,
981 to avoid conflicts and ambiguity as to whether or not bond0 should be
982 managed by systemd-networkd or not. This resolves multiple issues
983 with bond0 properties not being applied, when bond0 is configured
984 with systemd-networkd. Distributors may choose to not package this,
985 however in that case users will be prevented from correctly managing
986 bond0 interface using systemd-networkd.
988 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
989 which print the logging level and target of the system manager. They
990 complement the existing "set-log-level" and "set-log-target" verbs
991 used to change those values.
993 * journald.conf gained a new boolean setting ReadKMsg= which defaults
994 to on. If turned off kernel log messages will not be read by
995 systemd-journald or included in the logs. It also gained a new
996 setting LineMax= for configuring the maximum line length in
997 STDOUT/STDERR log streams. The new default for this value is 48K, up
998 from the previous hardcoded 2048.
1000 * A new unit setting RuntimeDirectoryPreserve= has been added, which
1001 allows more detailed control of what to do with a runtime directory
1002 configured with RuntimeDirectory= (i.e. a directory below /run or
1003 $XDG_RUNTIME_DIR) after a unit is stopped.
1005 * The RuntimeDirectory= setting for units gained support for creating
1006 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
1007 one top-level directory.
1009 * Units gained new options StateDirectory=, CacheDirectory=,
1010 LogsDirectory= and ConfigurationDirectory= which are closely related
1011 to RuntimeDirectory= but manage per-service directories below
1012 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
1013 possible to write unit files which when activated automatically gain
1014 properly owned service specific directories in these locations, thus
1015 making unit files self-contained and increasing compatibility with
1016 stateless systems and factory reset where /etc or /var are
1017 unpopulated at boot. Matching these new settings there's also
1018 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1019 ConfigurationDirectoryMode= for configuring the access mode of these
1020 directories. These settings are particularly useful in combination
1021 with DynamicUser=yes as they provide secure, properly-owned,
1022 writable, and stateful locations for storage, excluded from the
1023 sandbox that such services live in otherwise.
1025 * Automake support has been removed from this release. systemd is now
1028 * systemd-journald will now aggressively cache client metadata during
1029 runtime, speeding up log write performance under pressure. This comes
1030 at a small price though: as much of the metadata is read
1031 asynchronously from /proc/ (and isn't implicitly attached to log
1032 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1033 metadata stored alongside a log entry might be slightly
1034 out-of-date. Previously it could only be slightly newer than the log
1035 message. The time window is small however, and given that the kernel
1036 is unlikely to be improved anytime soon in this regard, this appears
1039 * nss-myhostname/systemd-resolved will now by default synthesize an
1040 A/AAAA resource record for the "_gateway" hostname, pointing to the
1041 current default IP gateway. Previously it did that for the "gateway"
1042 name, hampering adoption, as some distributions wanted to leave that
1043 host name open for local use. The old behaviour may still be
1044 requested at build time.
1046 * systemd-networkd's [Address] section in .network files gained a new
1047 Scope= setting for configuring the IP address scope. The [Network]
1048 section gained a new boolean setting ConfigureWithoutCarrier= that
1049 tells systemd-networkd to ignore link sensing when configuring the
1050 device. The [DHCP] section gained a new Anonymize= boolean option for
1051 turning on a number of options suggested in RFC 7844. A new
1052 [RoutingPolicyRule] section has been added for configuring the IP
1053 routing policy. The [Route] section has gained support for a new
1054 Type= setting which permits configuring
1055 blackhole/unreachable/prohibit routes.
1057 * The [VRF] section in .netdev files gained a new Table= setting for
1058 configuring the routing table to use. The [Tunnel] section gained a
1059 new Independent= boolean field for configuring tunnels independent of
1060 an underlying network interface. The [Bridge] section gained a new
1061 GroupForwardMask= option for configuration of propagation of link
1062 local frames between bridge ports.
1064 * The WakeOnLan= setting in .link files gained support for a number of
1065 new modes. A new TCP6SegmentationOffload= setting has been added for
1066 configuring TCP/IPv6 hardware segmentation offload.
1068 * The IPv6 RA sender implementation may now optionally send out RDNSS
1069 and RDNSSL records to supply DNS configuration to peers.
1071 * systemd-nspawn gained support for a new --system-call-filter= command
1072 line option for adding and removing entries in the default system
1073 call filter it applies. Moreover systemd-nspawn has been changed to
1074 implement a system call whitelist instead of a blacklist.
1076 * systemd-run gained support for a new --pipe command line option. If
1077 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1078 are directly passed on to the activated transient service
1079 executable. This allows invoking arbitrary processes as systemd
1080 services (for example to take benefit of dependency management,
1081 accounting management, resource management or log management that is
1082 done automatically for services) — while still allowing them to be
1083 integrated in a classic UNIX shell pipeline.
1085 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1086 using ReloadPropagationTo= is configured, a reload is now propagated
1087 to configured units. (Previously this was only done on explicitly
1088 requested reloads, using "systemctl reload" or an equivalent
1091 * For each service unit a restart counter is now kept: it is increased
1092 each time the service is restarted due to Restart=, and may be
1093 queried using "systemctl show -p NRestarts …".
1095 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1096 @signal and @timer have been added, for usage with SystemCallFilter=
1097 in unit files and the new --system-call-filter= command line option
1098 of systemd-nspawn (see above).
1100 * ExecStart= lines in unit files gained two new modifiers: when a
1101 command line is prefixed with "!" the command will be executed as
1102 configured, except for the credentials applied by
1103 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1104 "+", but does still apply namespacing options unlike "+". There's
1105 also "!!" now, which is mostly identical, but becomes a NOP on
1106 systems that support ambient capabilities. This is useful to write
1107 unit files that work with ambient capabilities where possible but
1108 automatically fall back to traditional privilege dropping mechanisms
1109 on systems where this is not supported.
1111 * ListenNetlink= settings in socket units now support RDMA netlink
1114 * A new unit file setting LockPersonality= has been added which permits
1115 locking down the chosen execution domain ("personality") of a service
1118 * A new special target "getty-pre.target" has been added, which is
1119 ordered before all text logins, and may be used to order services
1120 before textual logins acquire access to the console.
1122 * systemd will now attempt to load the virtio-rng.ko kernel module very
1123 early on if a VM environment supporting this is detected. This should
1124 improve entropy during early boot in virtualized environments.
1126 * A _netdev option is now supported in /etc/crypttab that operates in a
1127 similar way as the same option in /etc/fstab: it permits configuring
1128 encrypted devices that need to be ordered after the network is up.
1129 Following this logic, two new special targets
1130 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1131 added that are to cryptsetup.target what remote-fs.target and
1132 remote-fs-pre.target are to local-fs.target.
1134 * Service units gained a new UnsetEnvironment= setting which permits
1135 unsetting specific environment variables for services that are
1136 normally passed to it (for example in order to mask out locale
1137 settings for specific services that can't deal with it).
1139 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1140 traffic accounting (packet count as well as byte count) is done for
1141 the service, and shown as part of "systemctl status" or "systemd-run
1144 * Service units acquired two new options IPAddressAllow= and
1145 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1146 for configuring a simple IP access control list for all sockets of
1147 the unit. These options are available also on .slice and .socket
1148 units, permitting flexible access list configuration for individual
1149 services as well as groups of services (as defined by a slice unit),
1150 including system-wide. Note that IP ACLs configured this way are
1151 enforced on every single IPv4 and IPv6 socket created by any process
1152 of the service unit, and apply to ingress as well as egress traffic.
1154 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
1155 structured log message is generated each time the unit is stopped,
1156 containing information about the consumed resources of this
1159 * A new setting KeyringMode= has been added to unit files, which may be
1160 used to control how the kernel keyring is set up for executed
1163 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1164 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1165 behaviour (that is: these commands return immediately after the
1166 operation was enqueued instead of waiting for the operation to
1167 complete). Previously, "systemctl poweroff" and "systemctl reboot"
1168 were asynchronous on systems using systemd-logind (i.e. almost
1169 always, and like they were on sysvinit), and the other three commands
1170 were unconditionally synchronous. With this release this is cleaned
1171 up, and callers will see the same asynchronous behaviour on all
1172 systems for all five operations.
1174 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1177 * .timer units now accept calendar specifications in other timezones
1178 than UTC or the local timezone.
1180 * The tmpfiles snippet var.conf has been changed to create
1181 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1182 the "utmp" group already, and it appears to be generally understood
1183 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1184 databases. Previously this was implemented correctly for all these
1185 databases excepts btmp, which has been opened up like this now
1186 too. Note that while the other databases are world-readable
1187 (i.e. 0644), btmp is not and remains more restrictive.
1189 * The systemd-resolve tool gained a new --reset-server-features
1190 switch. When invoked like this systemd-resolved will forget
1191 everything it learnt about the features supported by the configured
1192 upstream DNS servers, and restarts the feature probing logic on the
1193 next resolver look-up for them at the highest feature level
1196 * The status dump systemd-resolved sends to the logs upon receiving
1197 SIGUSR1 now also includes information about all DNS servers it is
1198 configured to use, and the features levels it probed for them.
1200 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1201 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
1202 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1203 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1204 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1205 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1206 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1207 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1208 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1209 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1210 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1211 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1212 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1213 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1214 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1215 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1216 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1217 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1218 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1219 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1221 — Berlin, 2017-10-06
1225 * Meson is now supported as build system in addition to Automake. It is
1226 our plan to remove Automake in one of our next releases, so that
1227 Meson becomes our exclusive build system. Hence, please start using
1228 the Meson build system in your downstream packaging. There's plenty
1229 of documentation around how to use Meson, the extremely brief
1232 ./autogen.sh && ./configure && make && sudo make install
1236 meson build && ninja -C build && sudo ninja -C build install
1238 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1239 which permits configuring a timeout on the time a job is
1240 running. This is particularly useful for setting timeouts on jobs for
1243 * Unit files gained two new options ConditionUser= and ConditionGroup=
1244 for conditionalizing units based on the identity of the user/group
1245 running a systemd user instance.
1247 * systemd-networkd now understands a new FlowLabel= setting in the
1248 [VXLAN] section of .network files, as well as a Priority= in
1249 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1250 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1251 gained support for configuration of GENEVE links, and IPv6 address
1252 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1254 * .link files now understand a new Port= setting.
1256 * systemd-networkd's DHCP support gained support for DHCP option 119
1257 (domain search list).
1259 * systemd-networkd gained support for serving IPv6 address ranges using
1260 the Router Advertisement protocol. The new .network configuration
1261 section [IPv6Prefix] may be used to configure the ranges to
1262 serve. This is implemented based on a new, minimal, native server
1263 implementation of RA.
1265 * journalctl's --output= switch gained support for a new parameter
1266 "short-iso-precise" for a mode where timestamps are shown as precise
1269 * systemd-udevd's "net_id" builtin may now generate stable network
1270 interface names from IBM PowerVM VIO devices as well as ACPI platform
1273 * MulticastDNS support in systemd-resolved may now be explicitly
1274 enabled/disabled using the new MulticastDNS= configuration file
1277 * systemd-resolved may now optionally use libidn2 instead of the libidn
1278 for processing internationalized domain names. Support for libidn2
1279 should be considered experimental and should not be enabled by
1282 * "machinectl pull-tar" and related call may now do verification of
1283 downloaded images using SUSE-style .sha256 checksum files in addition
1284 to the already existing support for validating using Ubuntu-style
1287 * sd-bus gained support for a new sd_bus_message_appendv() call which
1288 is va_list equivalent of sd_bus_message_append().
1290 * sd-boot gained support for validating images using SHIM/MOK.
1292 * The SMACK code learnt support for "onlycap".
1294 * systemd-mount --umount is now much smarter in figuring out how to
1295 properly unmount a device given its mount or device path.
1297 * The code to call libnss_dns as a fallback from libnss_resolve when
1298 the communication with systemd-resolved fails was removed. This
1299 fallback was redundant and interfered with the [!UNAVAIL=return]
1300 suffix. See nss-resolve(8) for the recommended configuration.
1302 * systemd-logind may now be restarted without losing state. It stores
1303 the file descriptors for devices it manages in the system manager
1304 using the FDSTORE= mechanism. Please note that further changes in
1305 other components may be required to make use of this (for example
1306 Xorg has code to listen for stops of systemd-logind and terminate
1307 itself when logind is stopped or restarted, in order to avoid using
1308 stale file descriptors for graphical devices, which is now
1309 counterproductive and must be reverted in order for restarts of
1310 systemd-logind to be safe. See
1311 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1313 * All kernel install plugins are called with the environment variable
1314 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1315 /etc/machine-id. If the file is missing or empty, the variable is
1316 empty and BOOT_DIR_ABS is the path of a temporary directory which is
1317 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
1318 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1320 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
1321 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1322 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1323 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1324 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
1325 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1326 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1327 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1328 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
1329 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1330 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1331 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1332 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1333 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1334 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1335 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1336 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1337 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1338 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1339 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1340 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1341 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1342 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1343 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1344 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
1345 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1346 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
1347 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1348 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1351 — Berlin, 2017-07-12
1355 * The "hybrid" control group mode has been modified to improve
1356 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1357 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1358 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1359 cgroups-v1 hierarchy), the only externally visible change being that
1360 the cgroups-v2 hierarchy is also mounted, to
1361 /sys/fs/cgroup/unified. This should provide a large degree of
1362 compatibility with "legacy" cgroups-v1, while taking benefit of the
1363 better management capabilities of cgroups-v2.
1365 * The default control group setup mode may be selected both a boot-time
1366 via a set of kernel command line parameters (specifically:
1367 systemd.unified_cgroup_hierarchy= and
1368 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1369 default selected on the configure command line
1370 (--with-default-hierarchy=). The upstream default is "hybrid"
1371 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1372 this will change in a future systemd version to be "unified" (pure
1373 cgroups-v2 mode). The third option for the compile time option is
1374 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1375 distributions to default to "hybrid" mode for release distributions,
1376 starting with v233. We recommend "unified" for development
1377 distributions (specifically: distributions such as Fedora's rawhide)
1378 as that's where things are headed in the long run. Use "legacy" for
1379 greatest stability and compatibility only.
1381 * Note one current limitation of "unified" and "hybrid" control group
1382 setup modes: the kernel currently does not permit the systemd --user
1383 instance (i.e. unprivileged code) to migrate processes between two
1384 disconnected cgroup subtrees, even if both are managed and owned by
1385 the user. This effectively means "systemd-run --user --scope" doesn't
1386 work when invoked from outside of any "systemd --user" service or
1387 scope. Specifically, it is not supported from session scopes. We are
1388 working on fixing this in a future systemd version. (See #3388 for
1389 further details about this.)
1391 * DBus policy files are now installed into /usr rather than /etc. Make
1392 sure your system has dbus >= 1.9.18 running before upgrading to this
1393 version, or override the install path with --with-dbuspolicydir= .
1395 * All python scripts shipped with systemd (specifically: the various
1396 tests written in Python) now require Python 3.
1398 * systemd unit tests can now run standalone (without the source or
1399 build directories), and can be installed into /usr/lib/systemd/tests/
1400 with 'make install-tests'.
1402 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1403 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1406 * Support for the %c, %r, %R specifiers in unit files has been
1407 removed. Specifiers are not supposed to be dependent on configuration
1408 in the unit file itself (so that they resolve the same regardless
1409 where used in the unit files), but these specifiers were influenced
1410 by the Slice= option.
1412 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1413 all cases. If distributions want to use a different shell for this
1414 purpose (for example Fedora's /sbin/sushell) they need to specify
1415 this explicitly at configure time using --with-debug-shell=.
1417 * The confirmation spawn prompt has been reworked to offer the
1420 (c)ontinue, proceed without asking anymore
1421 (D)ump, show the state of the unit
1422 (f)ail, don't execute the command and pretend it failed
1424 (i)nfo, show a short summary of the unit
1425 (j)obs, show jobs that are in progress
1426 (s)kip, don't execute the command and pretend it succeeded
1427 (y)es, execute the command
1429 The 'n' choice for the confirmation spawn prompt has been removed,
1430 because its meaning was confusing.
1432 The prompt may now also be redirected to an alternative console by
1433 specifying the console as parameter to systemd.confirm_spawn=.
1435 * Services of Type=notify require a READY=1 notification to be sent
1436 during startup. If no such message is sent, the service now fails,
1437 even if the main process exited with a successful exit code.
1439 * Services that fail to start up correctly now always have their
1440 ExecStopPost= commands executed. Previously, they'd enter "failed"
1441 state directly, without executing these commands.
1443 * The option MulticastDNS= of network configuration files has acquired
1444 an actual implementation. With MulticastDNS=yes a host can resolve
1445 names of remote hosts and reply to mDNS A and AAAA requests.
1447 * When units are about to be started an additional check is now done to
1448 ensure that all dependencies of type BindsTo= (when used in
1449 combination with After=) have been started.
1451 * systemd-analyze gained a new verb "syscall-filter" which shows which
1452 system call groups are defined for the SystemCallFilter= unit file
1453 setting, and which system calls they contain.
1455 * A new system call filter group "@filesystem" has been added,
1456 consisting of various file system related system calls. Group
1457 "@reboot" has been added, covering reboot, kexec and shutdown related
1458 calls. Finally, group "@swap" has been added covering swap
1459 configuration related calls.
1461 * A new unit file option RestrictNamespaces= has been added that may be
1462 used to restrict access to the various process namespace types the
1463 Linux kernel provides. Specifically, it may be used to take away the
1464 right for a service unit to create additional file system, network,
1465 user, and other namespaces. This sandboxing option is particularly
1466 relevant due to the high amount of recently discovered namespacing
1467 related vulnerabilities in the kernel.
1469 * systemd-udev's .link files gained support for a new AutoNegotiation=
1470 setting for configuring Ethernet auto-negotiation.
1472 * systemd-networkd's .network files gained support for a new
1473 ListenPort= setting in the [DHCP] section to explicitly configure the
1474 UDP client port the DHCP client shall listen on.
1476 * .network files gained a new Unmanaged= boolean setting for explicitly
1477 excluding one or more interfaces from management by systemd-networkd.
1479 * The systemd-networkd ProxyARP= option has been renamed to
1480 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1481 renamed to ReduceARPProxy=. The old names continue to be available
1484 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1485 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1487 * systemd-networkd's bonding device support gained support for two new
1488 configuration options ActiveSlave= and PrimarySlave=.
1490 * The various options in the [Match] section of .network files gained
1491 support for negative matching.
1493 * New systemd-specific mount options are now understood in /etc/fstab:
1495 x-systemd.mount-timeout= may be used to configure the maximum
1496 permitted runtime of the mount command.
1498 x-systemd.device-bound may be set to bind a mount point to its
1499 backing device unit, in order to automatically remove a mount point
1500 if its backing device is unplugged. This option may also be
1501 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1502 on the block device, which is now automatically set for all CDROM
1503 drives, so that mounted CDs are automatically unmounted when they are
1504 removed from the drive.
1506 x-systemd.after= and x-systemd.before= may be used to explicitly
1507 order a mount after or before another unit or mount point.
1509 * Enqueued start jobs for device units are now automatically garbage
1510 collected if there are no jobs waiting for them anymore.
1512 * systemctl list-jobs gained two new switches: with --after, for every
1513 queued job the jobs it's waiting for are shown; with --before the
1514 jobs which it's blocking are shown.
1516 * systemd-nspawn gained support for ephemeral boots from disk images
1517 (or in other words: --ephemeral and --image= may now be
1518 combined). Moreover, ephemeral boots are now supported for normal
1519 directories, even if the backing file system is not btrfs. Of course,
1520 if the file system does not support file system snapshots or
1521 reflinks, the initial copy operation will be relatively expensive, but
1522 this should still be suitable for many use cases.
1524 * Calendar time specifications in .timer units now support
1525 specifications relative to the end of a month by using "~" instead of
1526 "-" as separator between month and day. For example, "*-02~03" means
1527 "the third last day in February". In addition a new syntax for
1528 repeated events has been added using the "/" character. For example,
1529 "9..17/2:00" means "every two hours from 9am to 5pm".
1531 * systemd-socket-proxyd gained a new parameter --connections-max= for
1532 configuring the maximum number of concurrent connections.
1534 * sd-id128 gained a new API for generating unique IDs for the host in a
1535 way that does not leak the machine ID. Specifically,
1536 sd_id128_get_machine_app_specific() derives an ID based on the
1537 machine ID a in well-defined, non-reversible, stable way. This is
1538 useful whenever an identifier for the host is needed but where the
1539 identifier shall not be useful to identify the system beyond the
1540 scope of the application itself. (Internally this uses HMAC-SHA256 as
1541 keyed hash function using the machine ID as input.)
1543 * NotifyAccess= gained a new supported value "exec". When set
1544 notifications are accepted from all processes systemd itself invoked,
1545 including all control processes.
1547 * .nspawn files gained support for defining overlay mounts using the
1548 Overlay= and OverlayReadOnly= options. Previously this functionality
1549 was only available on the systemd-nspawn command line.
1551 * systemd-nspawn's --bind= and --overlay= options gained support for
1552 bind/overlay mounts whose source lies within the container tree by
1553 prefixing the source path with "+".
1555 * systemd-nspawn's --bind= and --overlay= options gained support for
1556 automatically allocating a temporary source directory in /var/tmp
1557 that is removed when the container dies. Specifically, if the source
1558 directory is specified as empty string this mechanism is selected. An
1559 example usage is --overlay=+/var::/var, which creates an overlay
1560 mount based on the original /var contained in the image, overlayed
1561 with a temporary directory in the host's /var/tmp. This way changes
1562 to /var are automatically flushed when the container shuts down.
1564 * systemd-nspawn --image= option does now permit raw file system block
1565 devices (in addition to images containing partition tables, as
1568 * The disk image dissection logic in systemd-nspawn gained support for
1569 automatically setting up LUKS encrypted as well as Verity protected
1570 partitions. When a container is booted from an encrypted image the
1571 passphrase is queried at start-up time. When a container with Verity
1572 data is started, the root hash is search in a ".roothash" file
1573 accompanying the disk image (alternatively, pass the root hash via
1574 the new --root-hash= command line option).
1576 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1577 be used to dissect disk images the same way as systemd-nspawn does
1578 it, following the Bootable Partition Specification. It may even be
1579 used to mount disk images with complex partition setups (including
1580 LUKS and Verity partitions) to a local host directory, in order to
1581 inspect them. This tool is not considered public API (yet), and is
1582 thus not installed into /usr/bin. Please do not rely on its
1583 existence, since it might go away or be changed in later systemd
1586 * A new generator "systemd-verity-generator" has been added, similar in
1587 style to "systemd-cryptsetup-generator", permitting automatic setup of
1588 Verity root partitions when systemd boots up. In order to make use of
1589 this your partition setup should follow the Discoverable Partitions
1590 Specification, and the GPT partition ID of the root file system
1591 partition should be identical to the upper 128bit of the Verity root
1592 hash. The GPT partition ID of the Verity partition protecting it
1593 should be the lower 128bit of the Verity root hash. If the partition
1594 image follows this model it is sufficient to specify a single
1595 "roothash=" kernel command line argument to both configure which root
1596 image and verity partition to use as well as the root hash for
1597 it. Note that systemd-nspawn's Verity support follows the same
1598 semantics, meaning that disk images with proper Verity data in place
1599 may be booted in containers with systemd-nspawn as well as on
1600 physical systems via the verity generator. Also note that the "mkosi"
1601 tool available at https://github.com/systemd/mkosi has been updated
1602 to generate Verity protected disk images following this scheme. In
1603 fact, it has been updated to generate disk images that optionally
1604 implement a complete UEFI SecureBoot trust chain, involving a signed
1605 kernel and initrd image that incorporates such a root hash as well as
1606 a Verity-enabled root partition.
1608 * The hardware database (hwdb) udev supports has been updated to carry
1609 accelerometer quirks.
1611 * All system services are now run with a fresh kernel keyring set up
1612 for them. The invocation ID is stored by default in it, thus
1613 providing a safe, non-overridable way to determine the invocation
1616 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1617 options for bind mounting arbitrary paths in a service-specific
1618 way. When these options are used, arbitrary host or service files and
1619 directories may be mounted to arbitrary locations in the service's
1622 * Documentation has been added that lists all of systemd's low-level
1623 environment variables:
1625 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
1627 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1628 whether a specific socket file descriptor matches a specified socket
1631 * systemd-firstboot has been updated to check for the
1632 systemd.firstboot= kernel command line option. It accepts a boolean
1633 and when set to false the first boot questions are skipped.
1635 * systemd-fstab-generator has been updated to check for the
1636 systemd.volatile= kernel command line option, which either takes an
1637 optional boolean parameter or the special value "state". If used the
1638 system may be booted in a "volatile" boot mode. Specifically,
1639 "systemd.volatile" is used, the root directory will be mounted as
1640 tmpfs, and only /usr is mounted from the actual root file system. If
1641 "systemd.volatile=state" is used, the root directory will be mounted
1642 as usual, but /var is mounted as tmpfs. This concept provides similar
1643 functionality as systemd-nspawn's --volatile= option, but provides it
1644 on physical boots. Use this option for implementing stateless
1645 systems, or testing systems with all state and/or configuration reset
1646 to the defaults. (Note though that many distributions are not
1647 prepared to boot up without a populated /etc or /var, though.)
1649 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1650 partitions. Previously it only supported LUKS encrypted partitions
1651 for all other uses, except for the root partition itself.
1653 * Socket units gained support for listening on AF_VSOCK sockets for
1654 communication in virtualized QEMU environments.
1656 * The "configure" script gained a new option --with-fallback-hostname=
1657 for specifying the fallback hostname to use if none is configured in
1658 /etc/hostname. For example, by specifying
1659 --with-fallback-hostname=fedora it is possible to default to a
1660 hostname of "fedora" on pristine installations.
1662 * systemd-cgls gained support for a new --unit= switch for listing only
1663 the control groups of a specific unit. Similar --user-unit= has been
1664 added for listing only the control groups of a specific user unit.
1666 * systemd-mount gained a new --umount switch for unmounting a mount or
1667 automount point (and all mount/automount points below it).
1669 * systemd will now refuse full configuration reloads (via systemctl
1670 daemon-reload and related calls) unless at least 16MiB of free space
1671 are available in /run. This is a safety precaution in order to ensure
1672 that generators can safely operate after the reload completed.
1674 * A new unit file option RootImage= has been added, which has a similar
1675 effect as RootDirectory= but mounts the service's root directory from
1676 a disk image instead of plain directory. This logic reuses the same
1677 image dissection and mount logic that systemd-nspawn already uses,
1678 and hence supports any disk images systemd-nspawn supports, including
1679 those following the Discoverable Partition Specification, as well as
1680 Verity enabled images. This option enables systemd to run system
1681 services directly off disk images acting as resource bundles,
1682 possibly even including full integrity data.
1684 * A new MountAPIVFS= unit file option has been added, taking a boolean
1685 argument. If enabled /proc, /sys and /dev (collectively called the
1686 "API VFS") will be mounted for the service. This is only relevant if
1687 RootDirectory= or RootImage= is used for the service, as these mounts
1688 are of course in place in the host mount namespace anyway.
1690 * systemd-nspawn gained support for a new --pivot-root= switch. If
1691 specified the root directory within the container image is pivoted to
1692 the specified mount point, while the original root disk is moved to a
1693 different place. This option enables booting of ostree images
1694 directly with systemd-nspawn.
1696 * The systemd build scripts will no longer complain if the NTP server
1697 addresses are not changed from the defaults. Google now supports
1698 these NTP servers officially. We still recommend downstreams to
1699 properly register an NTP pool with the NTP pool project though.
1701 * coredumpctl gained a new "--reverse" option for printing the list
1702 of coredumps in reverse order.
1704 * coredumpctl will now show additional information about truncated and
1705 inaccessible coredumps, as well as coredumps that are still being
1706 processed. It also gained a new --quiet switch for suppressing
1707 additional informational message in its output.
1709 * coredumpctl gained support for only showing coredumps newer and/or
1710 older than specific timestamps, using the new --since= and --until=
1711 options, reminiscent of journalctl's options by the same name.
1713 * The systemd-coredump logic has been improved so that it may be reused
1714 to collect backtraces in non-compiled languages, for example in
1715 scripting languages such as Python.
1717 * machinectl will now show the UID shift of local containers, if user
1718 namespacing is enabled for them.
1720 * systemd will now optionally run "environment generator" binaries at
1721 configuration load time. They may be used to add environment
1722 variables to the environment block passed to services invoked. One
1723 user environment generator is shipped by default that sets up
1724 environment variables based on files dropped into /etc/environment.d
1725 and ~/.config/environment.d/.
1727 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1730 * hostnamed has been updated to report a new chassis type of
1731 "convertible" to cover "foldable" laptops that can both act as a
1732 tablet and as a laptop, such as various Lenovo Yoga devices.
1734 Contributions from: Adrián López, Alexander Galanin, Alexander
1735 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1736 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1737 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1738 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1739 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1740 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1741 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1742 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
1743 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1744 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1745 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1746 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1747 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1748 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1749 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1750 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1751 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1752 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1753 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1754 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1755 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1756 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1757 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1758 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1759 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1760 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1761 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1764 — Berlin, 2017-03-01
1768 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1769 RestrictAddressFamilies= enabled. These sandboxing options should
1770 generally be compatible with the various external udev call-out
1771 binaries we are aware of, however there may be exceptions, in
1772 particular when exotic languages for these call-outs are used. In
1773 this case, consider turning off these settings locally.
1775 * The new RemoveIPC= option can be used to remove IPC objects owned by
1776 the user or group of a service when that service exits.
1778 * The new ProtectKernelModules= option can be used to disable explicit
1779 load and unload operations of kernel modules by a service. In
1780 addition access to /usr/lib/modules is removed if this option is set.
1782 * ProtectSystem= option gained a new value "strict", which causes the
1783 whole file system tree with the exception of /dev, /proc, and /sys,
1784 to be remounted read-only for a service.
1786 * The new ProtectKernelTunables= option can be used to disable
1787 modification of configuration files in /sys and /proc by a service.
1788 Various directories and files are remounted read-only, so access is
1789 restricted even if the file permissions would allow it.
1791 * The new ProtectControlGroups= option can be used to disable write
1792 access by a service to /sys/fs/cgroup.
1794 * Various systemd services have been hardened with
1795 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1796 RestrictAddressFamilies=.
1798 * Support for dynamically creating users for the lifetime of a service
1799 has been added. If DynamicUser=yes is specified, user and group IDs
1800 will be allocated from the range 61184..65519 for the lifetime of the
1801 service. They can be resolved using the new nss-systemd.so NSS
1802 module. The module must be enabled in /etc/nsswitch.conf. Services
1803 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1804 any resources allocated by the service will be cleaned up when the
1805 service exits. They also have ProtectHome=read-only and
1806 ProtectSystem=strict enabled, so they are not able to make any
1807 permanent modifications to the system.
1809 * The nss-systemd module also always resolves root and nobody, making
1810 it possible to have no /etc/passwd or /etc/group files in minimal
1811 container or chroot environments.
1813 * Services may be started with their own user namespace using the new
1814 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1815 under which the service is running are mapped. All other users are
1818 * Support for the cgroup namespace has been added to systemd-nspawn. If
1819 supported by kernel, the container system started by systemd-nspawn
1820 will have its own view of the cgroup hierarchy. This new behaviour
1821 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1823 * The new MemorySwapMax= option can be used to limit the maximum swap
1824 usage under the unified cgroup hierarchy.
1826 * Support for the CPU controller in the unified cgroup hierarchy has
1827 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1828 options. This controller requires out-of-tree patches for the kernel
1829 and the support is provisional.
1831 * Mount and automount units may now be created transiently
1832 (i.e. dynamically at runtime via the bus API, instead of requiring
1833 unit files in the file system).
1835 * systemd-mount is a new tool which may mount file systems – much like
1836 mount(8), optionally pulling in additional dependencies through
1837 transient .mount and .automount units. For example, this tool
1838 automatically runs fsck on a backing block device before mounting,
1839 and allows the automount logic to be used dynamically from the
1840 command line for establishing mount points. This tool is particularly
1841 useful when dealing with removable media, as it will ensure fsck is
1842 run – if necessary – before the first access and that the file system
1843 is quickly unmounted after each access by utilizing the automount
1844 logic. This maximizes the chance that the file system on the
1845 removable media stays in a clean state, and if it isn't in a clean
1846 state is fixed automatically.
1848 * LazyUnmount=yes option for mount units has been added to expose the
1849 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1852 * /efi will be used as the mount point of the EFI boot partition, if
1853 the directory is present, and the mount point was not configured
1854 through other means (e.g. fstab). If /efi directory does not exist,
1855 /boot will be used as before. This makes it easier to automatically
1856 mount the EFI partition on systems where /boot is used for something
1859 * When operating on GPT disk images for containers, systemd-nspawn will
1860 now mount the ESP to /boot or /efi according to the same rules as PID
1861 1 running on a host. This allows tools like "bootctl" to operate
1862 correctly within such containers, in order to make container images
1863 bootable on physical systems.
1865 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
1867 * Two new user session targets have been added to support running
1868 graphical sessions under the systemd --user instance:
1869 graphical-session.target and graphical-session-pre.target. See
1870 systemd.special(7) for a description of how those targets should be
1873 * The vconsole initialization code has been significantly reworked to
1874 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
1875 support unicode keymaps. Font and keymap configuration will now be
1876 copied to all allocated virtual consoles.
1878 * FreeBSD's bhyve virtualization is now detected.
1880 * Information recorded in the journal for core dumps now includes the
1881 contents of /proc/mountinfo and the command line of the process at
1882 the top of the process hierarchy (which is usually the init process
1885 * systemd-journal-gatewayd learned the --directory= option to serve
1886 files from the specified location.
1888 * journalctl --root=… can be used to peruse the journal in the
1889 /var/log/ directories inside of a container tree. This is similar to
1890 the existing --machine= option, but does not require the container to
1893 * The hardware database has been extended to support
1894 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1897 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1898 specify the click rate for mice which include a horizontal wheel with
1899 a click rate that is different than the one for the vertical wheel.
1901 * systemd-run gained a new --wait option that makes service execution
1902 synchronous. (Specifically, the command will not return until the
1903 specified service binary exited.)
1905 * systemctl gained a new --wait option that causes the start command to
1906 wait until the units being started have terminated again.
1908 * A new journal output mode "short-full" has been added which displays
1909 timestamps with abbreviated English day names and adds a timezone
1910 suffix. Those timestamps include more information than the default
1911 "short" output mode, and can be passed directly to journalctl's
1912 --since= and --until= options.
1914 * /etc/resolv.conf will be bind-mounted into containers started by
1915 systemd-nspawn, if possible, so any changes to resolv.conf contents
1916 are automatically propagated to the container.
1918 * The number of instances for socket-activated services originating
1919 from a single IP address can be limited with
1920 MaxConnectionsPerSource=, extending the existing setting of
1923 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1926 * .netdev and .network configuration can now be extended through
1929 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1930 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1931 can be enabled and disabled using the new UDPSegmentationOffload=,
1932 TCPSegmentationOffload=, GenericSegmentationOffload=,
1933 GenericReceiveOffload=, LargeReceiveOffload= options in the
1934 [Link] section of .link files.
1936 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1937 Port VLAN ID can be configured for bridge devices using the new STP=,
1938 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1939 section of .netdev files.
1941 * The route table to which routes received over DHCP or RA should be
1942 added can be configured with the new RouteTable= option in the [DHCP]
1943 and [IPv6AcceptRA] sections of .network files.
1945 * The Address Resolution Protocol can be disabled on links managed by
1946 systemd-networkd using the ARP=no setting in the [Link] section of
1949 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1950 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1951 encode information about the result and exit codes of the current
1952 service runtime cycle.
1954 * systemd-sysctl will now configure kernel parameters in the order
1955 they occur in the configuration files. This matches what sysctl
1956 has been traditionally doing.
1958 * kernel-install "plugins" that are executed to perform various
1959 tasks after a new kernel is added and before an old one is removed
1960 can now return a special value to terminate the procedure and
1961 prevent any later plugins from running.
1963 * Journald's SplitMode=login setting has been deprecated. It has been
1964 removed from documentation, and its use is discouraged. In a future
1965 release it will be completely removed, and made equivalent to current
1966 default of SplitMode=uid.
1968 * Storage=both option setting in /etc/systemd/coredump.conf has been
1969 removed. With fast LZ4 compression storing the core dump twice is not
1972 * The --share-system systemd-nspawn option has been replaced with an
1973 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1974 this functionality is discouraged. In addition the variables
1975 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1976 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1977 individual namespaces.
1979 * "machinectl list" now shows the IP address of running containers in
1980 the output, as well as OS release information.
1982 * "loginctl list" now shows the TTY of each session in the output.
1984 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1985 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1986 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1987 tracking objects in a "recursive" mode, where a single client can be
1988 counted multiple times, if it takes multiple references.
1990 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
1991 sd_bus_get_exit_on_disconnect(). They may be used to make a
1992 process using sd-bus automatically exit if the bus connection is
1995 * Bus clients of the service manager may now "pin" loaded units into
1996 memory, by taking an explicit reference on them. This is useful to
1997 ensure the client can retrieve runtime data about the service even
1998 after the service completed execution. Taking such a reference is
1999 available only for privileged clients and should be helpful to watch
2000 running services in a race-free manner, and in particular collect
2001 information about exit statuses and results.
2003 * The nss-resolve module has been changed to strictly return UNAVAIL
2004 when communication via D-Bus with resolved failed, and NOTFOUND when
2005 a lookup completed but was negative. This means it is now possible to
2006 neatly configure fallbacks using nsswitch.conf result checking
2007 expressions. Taking benefit of this, the new recommended
2008 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
2010 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
2012 * A new setting CtrlAltDelBurstAction= has been added to
2013 /etc/systemd/system.conf which may be used to configure the precise
2014 behaviour if the user on the console presses Ctrl-Alt-Del more often
2015 than 7 times in 2s. Previously this would unconditionally result in
2016 an expedited, immediate reboot. With this new setting the precise
2017 operation may be configured in more detail, and also turned off
2020 * In .netdev files two new settings RemoteChecksumTx= and
2021 RemoteChecksumRx= are now understood that permit configuring the
2022 remote checksumming logic for VXLAN networks.
2024 * The service manager learnt a new "invocation ID" concept for invoked
2025 services. Each runtime cycle of a service will get a new invocation
2026 ID (a 128bit random UUID) assigned that identifies the current
2027 run of the service uniquely and globally. A new invocation ID
2028 is generated each time a service starts up. The journal will store
2029 the invocation ID of a service along with any logged messages, thus
2030 making the invocation ID useful for matching the online runtime of a
2031 service with the offline log data it generated in a safe way without
2032 relying on synchronized timestamps. In many ways this new service
2033 invocation ID concept is similar to the kernel's boot ID concept that
2034 uniquely and globally identifies the runtime of each boot. The
2035 invocation ID of a service is passed to the service itself via an
2036 environment variable ($INVOCATION_ID). A new bus call
2037 GetUnitByInvocationID() has been added that is similar to GetUnit()
2038 but instead of retrieving the bus path for a unit by its name
2039 retrieves it by its invocation ID. The returned path is valid only as
2040 long as the passed invocation ID is current.
2042 * systemd-resolved gained a new "DNSStubListener" setting in
2043 resolved.conf. It either takes a boolean value or the special values
2044 "udp" and "tcp", and configures whether to enable the stub DNS
2045 listener on 127.0.0.53:53.
2047 * IP addresses configured via networkd may now carry additional
2048 configuration settings supported by the kernel. New options include:
2049 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2050 PrefixRoute=, AutoJoin=.
2052 * The PAM configuration fragment file for "user@.service" shipped with
2053 systemd (i.e. the --user instance of systemd) has been stripped to
2054 the minimum necessary to make the system boot. Previously, it
2055 contained Fedora-specific stanzas that did not apply to other
2056 distributions. It is expected that downstream distributions add
2057 additional configuration lines, matching their needs to this file,
2058 using it only as rough template of what systemd itself needs. Note
2059 that this reduced fragment does not even include an invocation of
2060 pam_limits which most distributions probably want to add, even though
2061 systemd itself does not need it. (There's also the new build time
2062 option --with-pamconfdir=no to disable installation of the PAM
2065 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2066 capability is now also dropped from its set (in addition to
2067 CAP_SYS_MKNOD as before).
2069 * In service unit files it is now possible to connect a specific named
2070 file descriptor with stdin/stdout/stdout of an executed service. The
2071 name may be specified in matching .socket units using the
2072 FileDescriptorName= setting.
2074 * A number of journal settings may now be configured on the kernel
2075 command line. Specifically, the following options are now understood:
2076 systemd.journald.max_level_console=,
2077 systemd.journald.max_level_store=,
2078 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2079 systemd.journald.max_level_wall=.
2081 * "systemctl is-enabled --full" will now show by which symlinks a unit
2082 file is enabled in the unit dependency tree.
2084 * Support for VeraCrypt encrypted partitions has been added to the
2085 "cryptsetup" logic and /etc/crypttab.
2087 * systemd-detect-virt gained support for a new --private-users switch
2088 that checks whether the invoking processes are running inside a user
2089 namespace. Similar, a new special value "private-users" for the
2090 existing ConditionVirtualization= setting has been added, permitting
2091 skipping of specific units in user namespace environments.
2093 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2094 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2095 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2096 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2097 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2098 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2099 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2100 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2101 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2102 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2103 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2104 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2105 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2106 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2107 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2108 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2109 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2110 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2111 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2112 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2113 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2114 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2115 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2116 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2117 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2118 Jędrzejewski-Szmek, Zeal Jagannatha
2120 — Santa Fe, 2016-11-03
2124 * In service units the various ExecXYZ= settings have been extended
2125 with an additional special character as first argument of the
2126 assigned value: if the character '+' is used the specified command
2127 line it will be run with full privileges, regardless of User=,
2128 Group=, CapabilityBoundingSet= and similar options. The effect is
2129 similar to the existing PermissionsStartOnly= option, but allows
2130 configuration of this concept for each executed command line
2133 * Services may now alter the service watchdog timeout at runtime by
2134 sending a WATCHDOG_USEC= message via sd_notify().
2136 * MemoryLimit= and related unit settings now optionally take percentage
2137 specifications. The percentage is taken relative to the amount of
2138 physical memory in the system (or in case of containers, the assigned
2139 amount of memory). This allows scaling service resources neatly with
2140 the amount of RAM available on the system. Similarly, systemd-logind's
2141 RuntimeDirectorySize= option now also optionally takes percentage
2144 * In similar fashion TasksMax= takes percentage values now, too. The
2145 value is taken relative to the configured maximum number of processes
2146 on the system. The per-service task maximum has been changed to 15%
2147 using this functionality. (Effectively this is an increase of 512 →
2148 4915 for service units, given the kernel's default pid_max setting.)
2150 * Calendar time specifications in .timer units now understand a ".."
2151 syntax for time ranges. Example: "4..7:10" may now be used for
2152 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2155 * The InaccessableDirectories=, ReadOnlyDirectories= and
2156 ReadWriteDirectories= unit file settings have been renamed to
2157 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2158 applied to all kinds of file nodes, and not just directories, with
2159 the exception of symlinks. Specifically these settings may now be
2160 used on block and character device nodes, UNIX sockets and FIFOS as
2161 well as regular files. The old names of these settings remain
2162 available for compatibility.
2164 * systemd will now log about all service processes it kills forcibly
2165 (using SIGKILL) because they remained after the clean shutdown phase
2166 of the service completed. This should help identifying services that
2167 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2168 systemd-logind's configuration a similar log message is generated for
2169 processes killed at the end of each session due to this setting.
2171 * systemd will now set the $JOURNAL_STREAM environment variable for all
2172 services whose stdout/stderr are connected to the Journal (which
2173 effectively means by default: all services). The variable contains
2174 the device and inode number of the file descriptor used for
2175 stdout/stderr. This may be used by invoked programs to detect whether
2176 their stdout/stderr is connected to the Journal, in which case they
2177 can switch over to direct Journal communication, thus being able to
2178 pass extended, structured metadata along with their log messages. As
2179 one example, this is now used by glib's logging primitives.
2181 * When using systemd's default tmp.mount unit for /tmp, the mount point
2182 will now be established with the "nosuid" and "nodev" options. This
2183 avoids privilege escalation attacks that put traps and exploits into
2184 /tmp. However, this might cause problems if you e. g. put container
2185 images or overlays into /tmp; if you need this, override tmp.mount's
2186 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2189 * systemd now supports the "memory" cgroup controller also on
2192 * The systemd-cgtop tool now optionally takes a control group path as
2193 command line argument. If specified, the control group list shown is
2194 limited to subgroups of that group.
2196 * The SystemCallFilter= unit file setting gained support for
2197 pre-defined, named system call filter sets. For example
2198 SystemCallFilter=@clock is now an effective way to make all clock
2199 changing-related system calls unavailable to a service. A number of
2200 similar pre-defined groups are defined. Writing system call filters
2201 for system services is simplified substantially with this new
2202 concept. Accordingly, all of systemd's own, long-running services now
2203 enable system call filtering based on this, by default.
2205 * A new service setting MemoryDenyWriteExecute= has been added, taking
2206 a boolean value. If turned on, a service may no longer create memory
2207 mappings that are writable and executable at the same time. This
2208 enhances security for services where this is enabled as it becomes
2209 harder to dynamically write and then execute memory in exploited
2210 service processes. This option has been enabled for all of systemd's
2211 own long-running services.
2213 * A new RestrictRealtime= service setting has been added, taking a
2214 boolean argument. If set the service's processes may no longer
2215 acquire realtime scheduling. This improves security as realtime
2216 scheduling may otherwise be used to easily freeze the system.
2218 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2219 value. This may be used for requesting that the system manager inside
2220 of the container reports start-up completion to nspawn which then
2221 propagates this notification further to the service manager
2222 supervising nspawn itself. A related option NotifyReady= in .nspawn
2223 files has been added too. This functionality allows ordering of the
2224 start-up of multiple containers using the usual systemd ordering
2227 * machinectl gained a new command "stop" that is an alias for
2230 * systemd-resolved gained support for contacting DNS servers on
2231 link-local IPv6 addresses.
2233 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2234 its caches. A method call for requesting the same operation has been
2235 added to the bus API too, and is made available via "systemd-resolve
2238 * systemd-resolve gained a new --status switch. If passed a brief
2239 summary of the used DNS configuration with per-interface information
2242 * resolved.conf gained a new Cache= boolean option, defaulting to
2243 on. If turned off local DNS caching is disabled. This comes with a
2244 performance penalty in particular when DNSSEC is enabled. Note that
2245 resolved disables its internal caching implicitly anyway, when the
2246 configured DNS server is on a host-local IP address such as ::1 or
2247 127.0.0.1, thus automatically avoiding double local caching.
2249 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2250 for DNS requests. This improves compatibility with local programs
2251 that do not use the libc NSS or systemd-resolved's bus APIs for name
2252 resolution. This minimal DNS service is only available to local
2253 programs and does not implement the full DNS protocol, but enough to
2254 cover local DNS clients. A new, static resolv.conf file, listing just
2255 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2256 now recommended to make /etc/resolv.conf a symlink to this file in
2257 order to route all DNS lookups to systemd-resolved, regardless if
2258 done via NSS, the bus API or raw DNS packets. Note that this local
2259 DNS service is not as fully featured as the libc NSS or
2260 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2261 used to deliver link-local address information (as this implies
2262 sending a local interface index along), LLMNR/mDNS support via this
2263 interface is severely restricted. It is thus strongly recommended for
2264 all applications to use the libc NSS API or native systemd-resolved
2267 * systemd-networkd's bridge support learned a new setting
2268 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2269 in .network files has been added for configuring VLAN bridging in
2270 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2272 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2273 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2274 now be acquired without relying on DHCPv6. Two new options
2275 UseDomains= and UseDNS= have been added to configure this behaviour.
2277 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2278 renamed IPv6AcceptRA=, without altering its behaviour. The old
2279 setting name remains available for compatibility reasons.
2281 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2282 Key=, InputKey= and OutputKey=.
2284 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2285 interface configuration.
2287 * "systemctl edit" may now be used to create new unit files by
2288 specifying the --force switch.
2290 * sd-event gained a new function sd_event_get_iteration() for
2291 requesting the current iteration counter of the event loop. It starts
2292 at zero and is increased by one with each event loop iteration.
2294 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2295 file. It can be used in lieu of %systemd_requires in packages which
2296 don't use any systemd functionality and are intended to be installed
2297 in minimal containers without systemd present. This macro provides
2298 ordering dependencies to ensure that if the package is installed in
2299 the same rpm transaction as systemd, systemd will be installed before
2300 the scriptlets for the package are executed, allowing unit presets
2303 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2304 been added to simplify packaging of generators.
2306 * The os-release file gained VERSION_CODENAME field for the
2307 distribution nickname (e.g. VERSION_CODENAME=woody).
2309 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2310 can be set to disable parsing of metadata and the creation
2311 of persistent symlinks for that device.
2313 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2314 to make them available to logged-in users has been reverted.
2316 * Much of the common code of the various systemd components is now
2317 built into an internal shared library libsystemd-shared-231.so
2318 (incorporating the systemd version number in the name, to be updated
2319 with future releases) that the components link to. This should
2320 decrease systemd footprint both in memory during runtime and on
2321 disk. Note that the shared library is not for public use, and is
2322 neither API nor ABI stable, but is likely to change with every new
2323 released update. Packagers need to make sure that binaries
2324 linking to libsystemd-shared.so are updated in step with the
2327 * Configuration for "mkosi" is now part of the systemd
2328 repository. mkosi is a tool to easily build legacy-free OS images,
2329 and is available on github: https://github.com/systemd/mkosi. If
2330 "mkosi" is invoked in the build tree a new raw OS image is generated
2331 incorporating the systemd sources currently being worked on and a
2332 clean, fresh distribution installation. The generated OS image may be
2333 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
2334 UEFI PC. This functionality is particularly useful to easily test
2335 local changes made to systemd in a pristine, defined environment. See
2336 doc/HACKING for details.
2338 * configure learned the --with-support-url= option to specify the
2339 distribution's bugtracker.
2341 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2342 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2343 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2344 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2345 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2346 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2347 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2348 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2349 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2350 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2351 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2352 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2353 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2354 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2355 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2356 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
2357 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2358 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
2359 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
2361 — Berlin, 2016-07-25
2365 * DNSSEC is now turned on by default in systemd-resolved (in
2366 "allow-downgrade" mode), but may be turned off during compile time by
2367 passing "--with-default-dnssec=no" to "configure" (and of course,
2368 during runtime with DNSSEC= in resolved.conf). We recommend
2369 downstreams to leave this on at least during development cycles and
2370 report any issues with the DNSSEC logic upstream. We are very
2371 interested in collecting feedback about the DNSSEC validator and its
2372 limitations in the wild. Note however, that DNSSEC support is
2373 probably nothing downstreams should turn on in stable distros just
2374 yet, as it might create incompatibilities with a few DNS servers and
2375 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2376 automatically whenever we detect such incompatible setups, but there
2377 might be systems we do not cover yet. Hence: please help us testing
2378 the DNSSEC code, leave this on where you can, report back, but then
2379 again don't consider turning this on in your stable, LTS or
2380 production release just yet. (Note that you have to enable
2381 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2382 and its DNSSEC mode for host name resolution from local
2385 * systemd-resolve conveniently resolves DANE records with the --tlsa
2386 option and OPENPGPKEY records with the --openpgp option. It also
2387 supports dumping raw DNS record data via the new --raw= switch.
2389 * systemd-logind will now by default terminate user processes that are
2390 part of the user session scope unit (session-XX.scope) when the user
2391 logs out. This behavior is controlled by the KillUserProcesses=
2392 setting in logind.conf, and the previous default of "no" is now
2393 changed to "yes". This means that user sessions will be properly
2394 cleaned up after, but additional steps are necessary to allow
2395 intentionally long-running processes to survive logout.
2397 While the user is logged in at least once, user@.service is running,
2398 and any service that should survive the end of any individual login
2399 session can be started at a user service or scope using systemd-run.
2400 systemd-run(1) man page has been extended with an example which shows
2401 how to run screen in a scope unit underneath user@.service. The same
2402 command works for tmux.
2404 After the user logs out of all sessions, user@.service will be
2405 terminated too, by default, unless the user has "lingering" enabled.
2406 To effectively allow users to run long-term tasks even if they are
2407 logged out, lingering must be enabled for them. See loginctl(1) for
2408 details. The default polkit policy was modified to allow users to
2409 set lingering for themselves without authentication.
2411 Previous defaults can be restored at compile time by the
2412 --without-kill-user-processes option to "configure".
2414 * systemd-logind gained new configuration settings SessionsMax= and
2415 InhibitorsMax=, both with a default of 8192. It will not register new
2416 user sessions or inhibitors above this limit.
2418 * systemd-logind will now reload configuration on SIGHUP.
2420 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
2421 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
2422 enable. Also, support for the "io" cgroup controller in the unified
2423 hierarchy has been added, so that the "memory", "pids" and "io" are
2424 now the controllers that are supported on the unified hierarchy.
2426 WARNING: it is not possible to use previous systemd versions with
2427 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2428 is necessary to also update systemd in the initramfs if using the
2429 unified hierarchy. An updated SELinux policy is also required.
2431 * LLDP support has been extended, and both passive (receive-only) and
2432 active (sender) modes are supported. Passive mode ("routers-only") is
2433 enabled by default in systemd-networkd. Active LLDP mode is enabled
2434 by default for containers on the internal network. The "networkctl
2435 lldp" command may be used to list information gathered. "networkctl
2436 status" will also show basic LLDP information on connected peers now.
2438 * The IAID and DUID unique identifier sent in DHCP requests may now be
2439 configured for the system and each .network file managed by
2440 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2442 * systemd-networkd gained support for configuring proxy ARP support for
2443 each interface, via the ProxyArp= setting in .network files. It also
2444 gained support for configuring the multicast querier feature of
2445 bridge devices, via the new MulticastQuerier= setting in .netdev
2446 files. Similarly, snooping on the IGMP traffic can be controlled
2447 via the new setting MulticastSnooping=.
2449 A new setting PreferredLifetime= has been added for addresses
2450 configured in .network file to configure the lifetime intended for an
2453 The systemd-networkd DHCP server gained the option EmitRouter=, which
2454 defaults to yes, to configure whether the DHCP Option 3 (Router)
2457 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
2458 systemd-socket-activate and installed into /usr/bin. It is now fully
2461 * systemd-journald now uses separate threads to flush changes to disk
2462 when closing journal files, thus reducing impact of slow disk I/O on
2463 logging performance.
2465 * The sd-journal API gained two new calls
2466 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2467 can be used to open journal files using file descriptors instead of
2468 file or directory paths. sd_journal_open_container() has been
2469 deprecated, sd_journal_open_directory_fd() should be used instead
2470 with the flag SD_JOURNAL_OS_ROOT.
2472 * journalctl learned a new output mode "-o short-unix" that outputs log
2473 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2474 UTC). It also gained support for a new --no-hostname setting to
2475 suppress the hostname column in the family of "short" output modes.
2477 * systemd-ask-password now optionally skips printing of the password to
2478 stdout with --no-output which can be useful in scripts.
2480 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2481 (devices tagged with ID_MAKER_TOOL) are now tagged with
2482 "uaccess" and are available to logged in users.
2484 * The DeviceAllow= unit setting now supports specifiers (with "%").
2486 * "systemctl show" gained a new --value switch, which allows print a
2487 only the contents of a specific unit property, without also printing
2488 the property's name. Similar support was added to "show*" verbs
2489 of loginctl and machinectl that output "key=value" lists.
2491 * A new unit type "generated" was added for files dynamically generated
2492 by generator tools. Similarly, a new unit type "transient" is used
2493 for unit files created using the runtime API. "systemctl enable" will
2494 refuse to operate on such files.
2496 * A new command "systemctl revert" has been added that may be used to
2497 revert to the vendor version of a unit file, in case local changes
2498 have been made by adding drop-ins or overriding the unit file.
2500 * "machinectl clean" gained a new verb to automatically remove all or
2501 just hidden container images.
2503 * systemd-tmpfiles gained support for a new line type "e" for emptying
2504 directories, if they exist, without creating them if they don't.
2506 * systemd-nspawn gained support for automatically patching the UID/GIDs
2507 of the owners and the ACLs of all files and directories in a
2508 container tree to match the UID/GID user namespacing range selected
2509 for the container invocation. This mode is enabled via the new
2510 --private-users-chown switch. It also gained support for
2511 automatically choosing a free, previously unused UID/GID range when
2512 starting a container, via the new --private-users=pick setting (which
2513 implies --private-users-chown). Together, these options for the first
2514 time make user namespacing for nspawn containers fully automatic and
2515 thus deployable. The systemd-nspawn@.service template unit file has
2516 been changed to use this functionality by default.
2518 * systemd-nspawn gained a new --network-zone= switch, that allows
2519 creating ad-hoc virtual Ethernet links between multiple containers,
2520 that only exist as long as at least one container referencing them is
2521 running. This allows easy connecting of multiple containers with a
2522 common link that implements an Ethernet broadcast domain. Each of
2523 these network "zones" may be named relatively freely by the user, and
2524 may be referenced by any number of containers, but each container may
2525 only reference one of these "zones". On the lower level, this is
2526 implemented by an automatically managed bridge network interface for
2527 each zone, that is created when the first container referencing its
2528 zone is created and removed when the last one referencing its zone
2531 * The default start timeout may now be configured on the kernel command
2532 line via systemd.default_timeout_start_sec=. It was already
2533 configurable via the DefaultTimeoutStartSec= option in
2534 /etc/systemd/system.conf.
2536 * Socket units gained a new TriggerLimitIntervalSec= and
2537 TriggerLimitBurst= setting to configure a limit on the activation
2538 rate of the socket unit.
2540 * The LimitNICE= setting now optionally takes normal UNIX nice values
2541 in addition to the raw integer limit value. If the specified
2542 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2543 value is understood as UNIX nice value. If not prefixed like this it
2544 is understood as raw RLIMIT_NICE limit.
2546 * Note that the effect of the PrivateDevices= unit file setting changed
2547 slightly with this release: the per-device /dev file system will be
2548 mounted read-only from this version on, and will have "noexec"
2549 set. This (minor) change of behavior might cause some (exceptional)
2550 legacy software to break, when PrivateDevices=yes is set for its
2551 service. Please leave PrivateDevices= off if you run into problems
2554 * systemd-bootchart has been split out to a separate repository:
2555 https://github.com/systemd/systemd-bootchart
2557 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2558 merged into the kernel in its current form.
2560 * The compatibility libraries libsystemd-daemon.so,
2561 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2562 which have been deprecated since systemd-209 have been removed along
2563 with the corresponding pkg-config files. All symbols provided by
2564 those libraries are provided by libsystemd.so.
2566 * The Capabilities= unit file setting has been removed (it is ignored
2567 for backwards compatibility). AmbientCapabilities= and
2568 CapabilityBoundingSet= should be used instead.
2570 * A new special target has been added, initrd-root-device.target,
2571 which creates a synchronization point for dependencies of the root
2572 device in early userspace. Initramfs builders must ensure that this
2573 target is now included in early userspace.
2575 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2576 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2577 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2578 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2579 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2580 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2581 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2582 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
2583 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2584 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
2585 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2586 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2587 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
2588 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2589 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2590 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
2591 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2592 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2593 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2594 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2595 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2596 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
2597 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2598 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2599 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2602 — Fairfax, 2016-05-21
2606 * The systemd-resolved DNS resolver service has gained a substantial
2607 set of new features, most prominently it may now act as a DNSSEC
2608 validating stub resolver. DNSSEC mode is currently turned off by
2609 default, but is expected to be turned on by default in one of the
2610 next releases. For now, we invite everybody to test the DNSSEC logic
2611 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2612 service also gained a full set of D-Bus interfaces, including calls
2613 to configure DNS and DNSSEC settings per link (for use by external
2614 network management software). systemd-resolved and systemd-networkd
2615 now distinguish between "search" and "routing" domains. The former
2616 are used to qualify single-label names, the latter are used purely
2617 for routing lookups within certain domains to specific links.
2618 resolved now also synthesizes RRs for all entries from /etc/hosts.
2620 * The systemd-resolve tool (which is a client utility for
2621 systemd-resolved) has been improved considerably and is now fully
2622 supported and documented. Hence it has moved from /usr/lib/systemd to
2625 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2628 * The coredump collection logic has been reworked: when a coredump is
2629 collected it is now written to disk, compressed and processed
2630 (including stacktrace extraction) from a new instantiated service
2631 systemd-coredump@.service, instead of directly from the
2632 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2633 processing large coredumps can take up a substantial amount of
2634 resources and time, and this previously happened entirely outside of
2635 systemd's service supervision. With the new logic the core_pattern
2636 hook only does minimal metadata collection before passing off control
2637 to the new instantiated service, which is configured with a time
2638 limit, a nice level and other settings to minimize negative impact on
2639 the rest of the system. Also note that the new logic will honour the
2640 RLIMIT_CORE setting of the crashed process, which now allows users
2641 and processes to turn off coredumping for their processes by setting
2644 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2645 and all forked processes by default. Previously, PID 1 would leave
2646 the setting at "0" for all processes, as set by the kernel. Note that
2647 the resource limit traditionally has no effect on the generated
2648 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2649 logic is used. Since the limit is now honoured (see above) its
2650 default has been changed so that the coredumping logic is enabled by
2651 default for all processes, while allowing specific opt-out.
2653 * When the stacktrace is extracted from processes of system users, this
2654 is now done as "systemd-coredump" user, in order to sandbox this
2655 potentially security sensitive parsing operation. (Note that when
2656 processing coredumps of normal users this is done under the user ID
2657 of process that crashed, as before.) Packagers should take notice
2658 that it is now necessary to create the "systemd-coredump" system user
2659 and group at package installation time.
2661 * The systemd-activate socket activation testing tool gained support
2662 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2663 and --seqpacket switches. It also has been extended to support both
2664 new-style and inetd-style file descriptor passing. Use the new
2665 --inetd switch to request inetd-style file descriptor passing.
2667 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2668 variable, which takes a boolean value. If set to false, ANSI color
2669 output is disabled in the tools even when run on a terminal that
2672 * The VXLAN support in networkd now supports two new settings
2673 DestinationPort= and PortRange=.
2675 * A new systemd.machine_id= kernel command line switch has been added,
2676 that may be used to set the machine ID in /etc/machine-id if it is
2677 not initialized yet. This command line option has no effect if the
2678 file is already initialized.
2680 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2681 specified command line as PID 2 rather than PID 1 in the
2682 container. In this mode PID 1 is a minimal stub init process that
2683 implements the special POSIX and Linux semantics of PID 1 regarding
2684 signal and child process management. Note that this stub init process
2685 is implemented in nspawn itself and requires no support from the
2686 container image. This new logic is useful to support running
2687 arbitrary commands in the container, as normal processes are
2688 generally not prepared to run as PID 1.
2690 * systemd-nspawn gained a new --chdir= switch for setting the current
2691 working directory for the process started in the container.
2693 * "journalctl /dev/sda" will now output all kernel log messages for
2694 specified device from the current boot, in addition to all devices
2695 that are parents of it. This should make log output about devices
2696 pretty useful, as long as kernel drivers attach enough metadata to
2697 the log messages. (The usual SATA drivers do.)
2699 * The sd-journal API gained two new calls
2700 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2701 that report whether log data from /run or /var has been found.
2703 * journalctl gained a new switch "--fields" that prints all journal
2704 record field names currently in use in the journal. This is backed
2705 by two new sd-journal API calls sd_journal_enumerate_fields() and
2706 sd_journal_restart_fields().
2708 * Most configurable timeouts in systemd now expect an argument of
2709 "infinity" to turn them off, instead of "0" as before. The semantics
2710 from now on is that a timeout of "0" means "now", and "infinity"
2711 means "never". To maintain backwards compatibility, "0" continues to
2712 turn off previously existing timeout settings.
2714 * "systemctl reload-or-try-restart" has been renamed to "systemctl
2715 try-reload-or-restart" to clarify what it actually does: the "try"
2716 logic applies to both reloading and restarting, not just restarting.
2717 The old name continues to be accepted for compatibility.
2719 * On boot-up, when PID 1 detects that the system clock is behind the
2720 release date of the systemd version in use, the clock is now set
2721 to the latter. Previously, this was already done in timesyncd, in order
2722 to avoid running with clocks set to the various clock epochs such as
2723 1902, 1938 or 1970. With this change the logic is now done in PID 1
2724 in addition to timesyncd during early boot-up, so that it is enforced
2725 before the first process is spawned by systemd. Note that the logic
2726 in timesyncd remains, as it is more comprehensive and ensures
2727 clock monotonicity by maintaining a persistent timestamp file in
2728 /var. Since /var is generally not available in earliest boot or the
2729 initrd, this part of the logic remains in timesyncd, and is not done
2732 * Support for tweaking details in net_cls.class_id through the
2733 NetClass= configuration directive has been removed, as the kernel
2734 people have decided to deprecate that controller in cgroup v2.
2735 Userspace tools such as nftables are moving over to setting rules
2736 that are specific to the full cgroup path of a task, which obsoletes
2737 these controllers anyway. The NetClass= directive is kept around for
2738 legacy compatibility reasons. For a more in-depth description of the
2739 kernel change, please refer to the respective upstream commit:
2741 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2743 * A new service setting RuntimeMaxSec= has been added that may be used
2744 to specify a maximum runtime for a service. If the timeout is hit, the
2745 service is terminated and put into a failure state.
2747 * A new service setting AmbientCapabilities= has been added. It allows
2748 configuration of additional Linux process capabilities that are
2749 passed to the activated processes. This is only available on very
2752 * The process resource limit settings in service units may now be used
2753 to configure hard and soft limits individually.
2755 * The various libsystemd APIs such as sd-bus or sd-event now publicly
2756 expose support for gcc's __attribute__((cleanup())) C extension.
2757 Specifically, for many object destructor functions alternative
2758 versions have been added that have names suffixed with "p" and take a
2759 pointer to a pointer to the object to destroy, instead of just a
2760 pointer to the object itself. This is useful because these destructor
2761 functions may be used directly as parameters to the cleanup
2762 construct. Internally, systemd has been a heavy user of this GCC
2763 extension for a long time, and with this change similar support is
2764 now available to consumers of the library outside of systemd. Note
2765 that by using this extension in your sources compatibility with old
2766 and strictly ANSI compatible C compilers is lost. However, all gcc or
2767 LLVM versions of recent years support this extension.
2769 * Timer units gained support for a new setting RandomizedDelaySec= that
2770 allows configuring some additional randomized delay to the configured
2771 time. This is useful to spread out timer events to avoid load peaks in
2772 clusters or larger setups.
2774 * Calendar time specifications now support sub-second accuracy.
2776 * Socket units now support listening on SCTP and UDP-lite protocol
2779 * The sd-event API now comes with a full set of man pages.
2781 * Older versions of systemd contained experimental support for
2782 compressing journal files and coredumps with the LZ4 compressor that
2783 was not compatible with the lz4 binary (due to API limitations of the
2784 lz4 library). This support has been removed; only support for files
2785 compatible with the lz4 binary remains. This LZ4 logic is now
2786 officially supported and no longer considered experimental.
2788 * The dkr image import logic has been removed again from importd. dkr's
2789 micro-services focus doesn't fit into the machine image focus of
2790 importd, and quickly got out of date with the upstream dkr API.
2792 * Creation of the /run/lock/lockdev/ directory was dropped from
2793 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
2794 been available for many years. If you still need this, you need to
2795 create your own tmpfiles.d config file with:
2797 d /run/lock/lockdev 0775 root lock -
2799 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2800 and RebootArgument= have been moved from the [Service] section of
2801 unit files to [Unit], and they are now supported on all unit types,
2802 not just service units. Of course, elogind will continue to
2803 understand these settings also at the old location, in order to
2804 maintain compatibility.
2806 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2807 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2808 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2809 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2810 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2811 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2812 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2813 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2814 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2815 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2816 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2817 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2818 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2819 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2820 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2821 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2822 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2823 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2824 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2826 — Berlin, 2016-02-11
2830 * A number of properties previously only settable in unit
2831 files are now also available as properties to set when
2832 creating transient units programmatically via the bus, as it
2833 is exposed with systemd-run's --property=
2834 setting. Specifically, these are: SyslogIdentifier=,
2835 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2836 EnvironmentFile=, ReadWriteDirectories=,
2837 ReadOnlyDirectories=, InaccessibleDirectories=,
2838 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2840 * When creating transient services via the bus API it is now
2841 possible to pass in a set of file descriptors to use as
2842 STDIN/STDOUT/STDERR for the invoked process.
2844 * Slice units may now be created transiently via the bus APIs,
2845 similar to the way service and scope units may already be
2846 created transiently.
2848 * Wherever systemd expects a calendar timestamp specification
2849 (like in journalctl's --since= and --until= switches) UTC
2850 timestamps are now supported. Timestamps suffixed with "UTC"
2851 are now considered to be in Universal Time Coordinated
2852 instead of the local timezone. Also, timestamps may now
2853 optionally be specified with sub-second accuracy. Both of
2854 these additions also apply to recurring calendar event
2855 specification, such as OnCalendar= in timer units.
2857 * journalctl gained a new "--sync" switch that asks the
2858 journal daemon to write all so far unwritten log messages to
2859 disk and sync the files, before returning.
2861 * systemd-tmpfiles learned two new line types "q" and "Q" that
2862 operate like "v", but also set up a basic btrfs quota
2863 hierarchy when used on a btrfs file system with quota
2866 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2867 instead of a subvolume (even on a btrfs file system) if the
2868 root directory is a plain directory, and not a
2869 subvolume. This should simplify things with certain chroot()
2870 environments which are not aware of the concept of btrfs
2873 * systemd-detect-virt gained a new --chroot switch to detect
2874 whether execution takes place in a chroot() environment.
2876 * CPUAffinity= now takes CPU index ranges in addition to
2879 * The various memory-related resource limit settings (such as
2880 LimitAS=) now understand the usual K, M, G, ... suffixes to
2881 the base of 1024 (IEC). Similar, the time-related resource
2882 limit settings understand the usual min, h, day, ...
2885 * There's a new system.conf setting DefaultTasksMax= to
2886 control the default TasksMax= setting for services and
2887 scopes running on the system. (TasksMax= is the primary
2888 setting that exposes the "pids" cgroup controller on systemd
2889 and was introduced in the previous systemd release.) The
2890 setting now defaults to 512, which means services that are
2891 not explicitly configured otherwise will only be able to
2892 create 512 processes or threads at maximum, from this
2893 version on. Note that this means that thread- or
2894 process-heavy services might need to be reconfigured to set
2895 TasksMax= to a higher value. It is sufficient to set
2896 TasksMax= in these specific unit files to a higher value, or
2897 even "infinity". Similar, there's now a logind.conf setting
2898 UserTasksMax= that defaults to 4096 and limits the total
2899 number of processes or tasks each user may own
2900 concurrently. nspawn containers also have the TasksMax=
2901 value set by default now, to 8192. Note that all of this
2902 only has an effect if the "pids" cgroup controller is
2903 enabled in the kernel. The general benefit of these changes
2904 should be a more robust and safer system, that provides a
2905 certain amount of per-service fork() bomb protection.
2907 * systemd-nspawn gained the new --network-veth-extra= switch
2908 to define additional and arbitrarily-named virtual Ethernet
2909 links between the host and the container.
2911 * A new service execution setting PassEnvironment= has been
2912 added that allows importing select environment variables
2913 from PID1's environment block into the environment block of
2916 * Timer units gained support for a new RemainAfterElapse=
2917 setting which takes a boolean argument. It defaults to on,
2918 exposing behaviour unchanged to previous releases. If set to
2919 off, timer units are unloaded after they elapsed if they
2920 cannot elapse again. This is particularly useful for
2921 transient timer units, which shall not stay around longer
2922 than until they first elapse.
2924 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
2925 default now (the kernel default is 16). This is beneficial
2926 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
2927 allows substantially larger numbers of queued
2928 datagrams. This should increase the capability of systemd to
2929 parallelize boot-up, as logging and sd_notify() are unlikely
2930 to stall execution anymore. If you need to change the value
2931 from the new defaults, use the usual sysctl.d/ snippets.
2933 * The compression framing format used by the journal or
2934 coredump processing has changed to be in line with what the
2935 official LZ4 tools generate. LZ4 compression support in
2936 systemd was considered unsupported previously, as the format
2937 was not compatible with the normal tools. With this release
2938 this has changed now, and it is hence safe for downstream
2939 distributions to turn it on. While not compressing as well
2940 as the XZ, LZ4 is substantially faster, which makes
2941 it a good default choice for the compression logic in the
2942 journal and in coredump handling.
2944 * Any reference to /etc/mtab has been dropped from
2945 systemd. The file has been obsolete since a while, but
2946 systemd refused to work on systems where it was incorrectly
2947 set up (it should be a symlink or non-existent). Please make
2948 sure to update to util-linux 2.27.1 or newer in conjunction
2949 with this systemd release, which also drops any reference to
2950 /etc/mtab. If you maintain a distribution make sure that no
2951 software you package still references it, as this is a
2952 likely source of bugs. There's also a glibc bug pending,
2953 asking for removal of any reference to this obsolete file:
2955 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
2957 Note that only util-linux versions built with
2958 --enable-libmount-force-mountinfo are supported.
2960 * Support for the ".snapshot" unit type has been removed. This
2961 feature turned out to be little useful and little used, and
2962 has now been removed from the core and from systemctl.
2964 * The dependency types RequiresOverridable= and
2965 RequisiteOverridable= have been removed from systemd. They
2966 have been used only very sparingly to our knowledge and
2967 other options that provide a similar effect (such as
2968 systemctl --mode=ignore-dependencies) are much more useful
2969 and commonly used. Moreover, they were only half-way
2970 implemented as the option to control behaviour regarding
2971 these dependencies was never added to systemctl. By removing
2972 these dependency types the execution engine becomes a bit
2973 simpler. Unit files that use these dependencies should be
2974 changed to use the non-Overridable dependency types
2975 instead. In fact, when parsing unit files with these
2976 options, that's what systemd will automatically convert them
2977 too, but it will also warn, asking users to fix the unit
2978 files accordingly. Removal of these dependency types should
2979 only affect a negligible number of unit files in the wild.
2981 * Behaviour of networkd's IPForward= option changed
2982 (again). It will no longer maintain a per-interface setting,
2983 but propagate one way from interfaces where this is enabled
2984 to the global kernel setting. The global setting will be
2985 enabled when requested by a network that is set up, but
2986 never be disabled again. This change was made to make sure
2987 IPv4 and IPv6 behaviour regarding packet forwarding is
2988 similar (as the Linux IPv6 stack does not support
2989 per-interface control of this setting) and to minimize
2992 * In unit files the behaviour of %u, %U, %h, %s has
2993 changed. These specifiers will now unconditionally resolve
2994 to the various user database fields of the user that the
2995 systemd instance is running as, instead of the user
2996 configured in the specific unit via User=. Note that this
2997 effectively doesn't change much, as resolving of these
2998 specifiers was already turned off in the --system instance
2999 of systemd, as we cannot do NSS lookups from PID 1. In the
3000 --user instance of systemd these specifiers where correctly
3001 resolved, but hardly made any sense, since the user instance
3002 lacks privileges to do user switches anyway, and User= is
3003 hence useless. Moreover, even in the --user instance of
3004 systemd behaviour was awkward as it would only take settings
3005 from User= assignment placed before the specifier into
3006 account. In order to unify and simplify the logic around
3007 this the specifiers will now always resolve to the
3008 credentials of the user invoking the manager (which in case
3009 of PID 1 is the root user).
3011 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
3012 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
3013 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
3014 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3015 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3016 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3017 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3018 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3019 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3020 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3021 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3022 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3023 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3024 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3027 — Berlin, 2015-11-18
3031 * systemd now depends on util-linux v2.27. More specifically,
3032 the newly added mount monitor feature in libmount now
3033 replaces systemd's former own implementation.
3035 * libmount mandates /etc/mtab not to be regular file, and
3036 systemd now enforces this condition at early boot.
3037 /etc/mtab has been deprecated and warned about for a very
3038 long time, so systems running systemd should already have
3039 stopped having this file around as anything else than a
3040 symlink to /proc/self/mounts.
3042 * Support for the "pids" cgroup controller has been added. It
3043 allows accounting the number of tasks in a cgroup and
3044 enforcing limits on it. This adds two new setting
3045 TasksAccounting= and TasksMax= to each unit, as well as a
3046 global option DefaultTasksAccounting=.
3048 * Support for the "net_cls" cgroup controller has been added.
3049 It allows assigning a net class ID to each task in the
3050 cgroup, which can then be used in firewall rules and traffic
3051 shaping configurations. Note that the kernel netfilter net
3052 class code does not currently work reliably for ingress
3053 packets on unestablished sockets.
3055 This adds a new config directive called NetClass= to CGroup
3056 enabled units. Allowed values are positive numbers for fixed
3057 assignments and "auto" for picking a free value
3060 * 'systemctl is-system-running' now returns 'offline' if the
3061 system is not booted with systemd. This command can now be
3062 used as a substitute for 'systemd-notify --booted'.
3064 * Watchdog timeouts have been increased to 3 minutes for all
3065 in-tree service files. Apparently, disk IO issues are more
3066 frequent than we hoped, and user reported >1 minute waiting
3069 * 'machine-id-commit' functionality has been merged into
3070 'machine-id-setup --commit'. The separate binary has been
3073 * The WorkingDirectory= directive in unit files may now be set
3074 to the special value '~'. In this case, the working
3075 directory is set to the home directory of the user
3076 configured in User=.
3078 * "machinectl shell" will now open the shell in the home
3079 directory of the selected user by default.
3081 * The CrashChVT= configuration file setting is renamed to
3082 CrashChangeVT=, following our usual logic of not
3083 abbreviating unnecessarily. The old directive is still
3084 supported for compat reasons. Also, this directive now takes
3085 an integer value between 1 and 63, or a boolean value. The
3086 formerly supported '-1' value for disabling stays around for
3089 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3090 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3091 RootDirectory= properties can now be set for transient
3094 * The systemd-analyze tool gained a new "set-log-target" verb
3095 to change the logging target the system manager logs to
3096 dynamically during runtime. This is similar to how
3097 "systemd-analyze set-log-level" already changes the log
3100 * In nspawn /sys is now mounted as tmpfs, with only a selected
3101 set of subdirectories mounted in from the real sysfs. This
3102 enhances security slightly, and is useful for ensuring user
3103 namespaces work correctly.
3105 * Support for USB FunctionFS activation has been added. This
3106 allows implementation of USB gadget services that are
3107 activated as soon as they are requested, so that they don't
3108 have to run continuously, similar to classic socket
3111 * The "systemctl exit" command now optionally takes an
3112 additional parameter that sets the exit code to return from
3113 the systemd manager when exiting. This is only relevant when
3114 running the systemd user instance, or when running the
3115 system instance in a container.
3117 * sd-bus gained the new API calls sd_bus_path_encode_many()
3118 and sd_bus_path_decode_many() that allow easy encoding and
3119 decoding of multiple identifier strings inside a D-Bus
3120 object path. Another new call sd_bus_default_flush_close()
3121 has been added to flush and close per-thread default
3124 * systemd-cgtop gained support for a -M/--machine= switch to
3125 show the control groups within a certain container only.
3127 * "systemctl kill" gained support for an optional --fail
3128 switch. If specified the requested operation will fail of no
3129 processes have been killed, because the unit had no
3130 processes attached, or similar.
3132 * A new systemd.crash_reboot=1 kernel command line option has
3133 been added that triggers a reboot after crashing. This can
3134 also be set through CrashReboot= in systemd.conf.
3136 * The RuntimeDirectory= setting now understands unit
3137 specifiers like %i or %f.
3139 * A new (still internal) library API sd-ipv4acd has been added,
3140 that implements address conflict detection for IPv4. It's
3141 based on code from sd-ipv4ll, and will be useful for
3142 detecting DHCP address conflicts.
3144 * File descriptors passed during socket activation may now be
3145 named. A new API sd_listen_fds_with_names() is added to
3146 access the names. The default names may be overridden,
3147 either in the .socket file using the FileDescriptorName=
3148 parameter, or by passing FDNAME= when storing the file
3149 descriptors using sd_notify().
3151 * systemd-networkd gained support for:
3153 - Setting the IPv6 Router Advertisement settings via
3154 IPv6AcceptRouterAdvertisements= in .network files.
3156 - Configuring the HelloTimeSec=, MaxAgeSec= and
3157 ForwardDelaySec= bridge parameters in .netdev files.
3159 - Configuring PreferredSource= for static routes in
3162 * The "ask-password" framework used to query for LUKS harddisk
3163 passwords or SSL passwords during boot gained support for
3164 caching passwords in the kernel keyring, if it is
3165 available. This makes sure that the user only has to type in
3166 a passphrase once if there are multiple objects to unlock
3167 with the same one. Previously, such password caching was
3168 available only when Plymouth was used; this moves the
3169 caching logic into the systemd codebase itself. The
3170 "systemd-ask-password" utility gained a new --keyname=
3171 switch to control which kernel keyring key to use for
3172 caching a password in. This functionality is also useful for
3173 enabling display managers such as gdm to automatically
3174 unlock the user's GNOME keyring if its passphrase, the
3175 user's password and the harddisk password are the same, if
3176 gdm-autologin is used.
3178 * When downloading tar or raw images using "machinectl
3179 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3180 file is now also downloaded, if it is available and stored
3181 next to the image file.
3183 * Units of type ".socket" gained a new boolean setting
3184 Writable= which is only useful in conjunction with
3185 ListenSpecial=. If true, enables opening the specified
3186 special file in O_RDWR mode rather than O_RDONLY mode.
3188 * systemd-rfkill has been reworked to become a singleton
3189 service that is activated through /dev/rfkill on each rfkill
3190 state change and saves the settings to disk. This way,
3191 systemd-rfkill is now compatible with devices that exist
3192 only intermittendly, and even restores state if the previous
3193 system shutdown was abrupt rather than clean.
3195 * The journal daemon gained support for vacuuming old journal
3196 files controlled by the number of files that shall remain,
3197 in addition to the already existing control by size and by
3198 date. This is useful as journal interleaving performance
3199 degrades with too many separate journal files, and allows
3200 putting an effective limit on them. The new setting defaults
3201 to 100, but this may be changed by setting SystemMaxFiles=
3202 and RuntimeMaxFiles= in journald.conf. Also, the
3203 "journalctl" tool gained the new --vacuum-files= switch to
3204 manually vacuum journal files to leave only the specified
3205 number of files in place.
3207 * udev will now create /dev/disk/by-path links for ATA devices
3208 on kernels where that is supported.
3210 * Galician, Serbian, Turkish and Korean translations were added.
3212 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3213 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3214 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3215 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3216 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3217 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3218 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3219 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3220 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3221 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3222 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3223 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3224 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3225 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3226 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3227 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3228 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3229 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3231 — Berlin, 2015-10-07
3235 * The DHCP implementation of systemd-networkd gained a set of
3238 - The DHCP server now supports emitting DNS and NTP
3239 information. It may be enabled and configured via
3240 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3241 and NTP information is enabled, but no servers are
3242 configured, the corresponding uplink information (if there
3243 is any) is propagated.
3245 - Server and client now support transmission and reception
3246 of timezone information. It can be configured via the
3247 newly introduced network options UseTimezone=,
3248 EmitTimezone=, and Timezone=. Transmission of timezone
3249 information is enabled between host and containers by
3250 default now: the container will change its local timezone
3251 to what the host has set.
3253 - Lease timeouts can now be configured via
3254 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3256 - The DHCP server improved on the stability of
3257 leases. Clients are more likely to get the same lease
3258 information back, even if the server loses state.
3260 - The DHCP server supports two new configuration options to
3261 control the lease address pool metrics, PoolOffset= and
3264 * The encapsulation limit of tunnels in systemd-networkd may
3265 now be configured via 'EncapsulationLimit='. It allows
3266 modifying the maximum additional levels of encapsulation
3267 that are permitted to be prepended to a packet.
3269 * systemd now supports the concept of user buses replacing
3270 session buses, if used with dbus-1.10 (and enabled via dbus
3271 --enable-user-session). It previously only supported this on
3272 kdbus-enabled systems, and this release expands this to
3273 'dbus-daemon' systems.
3275 * systemd-networkd now supports predictable interface names
3278 * systemd now optionally supports the new Linux kernel
3279 "unified" control group hierarchy. If enabled via the kernel
3280 command-line option 'systemd.unified_cgroup_hierarchy=1',
3281 systemd will try to mount the unified cgroup hierarchy
3282 directly on /sys/fs/cgroup. If not enabled, or not
3283 available, systemd will fall back to the legacy cgroup
3284 hierarchy setup, as before. Host system and containers can
3285 mix and match legacy and unified hierarchies as they
3286 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
3287 environment variable to individually select the hierarchy to
3288 use for executed containers. By default, nspawn will use the
3289 unified hierarchy for the containers if the host uses the
3290 unified hierarchy, and the legacy hierarchy otherwise.
3291 Please note that at this point the unified hierarchy is an
3292 experimental kernel feature and is likely to change in one
3293 of the next kernel releases. Therefore, it should not be
3294 enabled by default in downstream distributions yet. The
3295 minimum required kernel version for the unified hierarchy to
3296 work is 4.2. Note that when the unified hierarchy is used
3297 for the first time delegated access to controllers is
3298 safe. Because of this systemd-nspawn containers will get
3299 access to controllers now, as will systemd user
3300 sessions. This means containers and user sessions may now
3301 manage their own resources, partitioning up what the system
3304 * A new special scope unit "init.scope" has been introduced
3305 that encapsulates PID 1 of the system. It may be used to
3306 determine resource usage and enforce resource limits on PID
3307 1 itself. PID 1 hence moved out of the root of the control
3310 * The cgtop tool gained support for filtering out kernel
3311 threads when counting tasks in a control group. Also, the
3312 count of processes is now recursively summed up by
3313 default. Two options -k and --recursive= have been added to
3314 revert to old behaviour. The tool has also been updated to
3315 work correctly in containers now.
3317 * systemd-nspawn's --bind= and --bind-ro= options have been
3318 extended to allow creation of non-recursive bind mounts.
3320 * libsystemd gained two new calls sd_pid_get_cgroup() and
3321 sd_peer_get_cgroup() which return the control group path of
3322 a process or peer of a connected AF_UNIX socket. This
3323 function call is particularly useful when implementing
3324 delegated subtrees support in the control group hierarchy.
3326 * The "sd-event" event loop API of libsystemd now supports
3327 correct dequeuing of real-time signals, without losing
3330 * When systemd requests a PolicyKit decision when managing
3331 units it will now add additional fields to the request,
3332 including unit name and desired operation. This enables more
3333 powerful PolicyKit policies, that make decisions depending
3334 on these parameters.
3336 * nspawn learnt support for .nspawn settings files, that may
3337 accompany the image files or directories of containers, and
3338 may contain additional settings for the container. This is
3339 an alternative to configuring container parameters via the
3340 nspawn command line.
3342 Contributions from: Cristian Rodríguez, Daniel Mack, David
3343 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3344 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3345 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3346 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3347 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3348 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
3349 Andersen, Tom Gundersen, Torstein Husebø
3351 — Berlin, 2015-09-08
3355 * machinectl gained a new verb 'shell' which opens a fresh
3356 shell on the target container or the host. It is similar to
3357 the existing 'login' command of machinectl, but spawns the
3358 shell directly without prompting for username or
3359 password. The pseudo machine '.host' now refers to the local
3360 host and is used by default. Hence, 'machinectl shell' can
3361 be used as replacement for 'su -' which spawns a session as
3362 a fresh systemd unit in a way that is fully isolated from
3363 the originating session.
3365 * systemd-networkd learned to cope with private-zone DHCP
3366 options and allows other programs to query the values.
3368 * SELinux access control when enabling/disabling units is no
3369 longer enforced with this release. The previous
3370 implementation was incorrect, and a new corrected
3371 implementation is not yet available. As unit file operations
3372 are still protected via PolicyKit and D-Bus policy this is
3373 not a security problem. Yet, distributions which care about
3374 optimal SELinux support should probably not stabilize on
3377 * sd-bus gained support for matches of type "arg0has=", that
3378 test for membership of strings in string arrays sent in bus
3381 * systemd-resolved now dumps the contents of its DNS and LLMNR
3382 caches to the logs on reception of the SIGUSR1 signal. This
3383 is useful to debug DNS behaviour.
3385 * The coredumpctl tool gained a new --directory= option to
3386 operate on journal files in a specific directory.
3388 * "systemctl reboot" and related commands gained a new
3389 "--message=" option which may be used to set a free-text
3390 wall message when shutting down or rebooting the
3391 system. This message is also logged, which is useful for
3392 figuring out the reason for a reboot or shutdown a
3395 * The "systemd-resolve-host" tool's -i switch now takes
3396 network interface numbers as alternative to interface names.
3398 * A new unit file setting for services has been introduced:
3399 UtmpMode= allows configuration of how precisely systemd
3400 handles utmp and wtmp entries for the service if this is
3401 enabled. This allows writing services that appear similar to
3402 user sessions in the output of the "w", "who", "last" and
3405 * systemd-resolved will now locally synthesize DNS resource
3406 records for the "localhost" and "gateway" domains as well as
3407 the local hostname. This should ensure that clients querying
3408 RRs via resolved will get similar results as those going via
3409 NSS, if nss-myhostname is enabled.
3411 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3412 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3413 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3414 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3415 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3416 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3417 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3418 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3419 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3420 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3421 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3422 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3424 — Berlin, 2015-08-27
3428 * The systemd-efi-boot-generator functionality was merged into
3429 systemd-gpt-auto-generator.
3431 * systemd-networkd now supports Group Policy for vxlan
3432 devices. It can be enabled via the new boolean configuration
3433 option called 'GroupPolicyExtension='.
3435 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3436 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3437 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3439 — Berlin, 2015-07-31
3443 * The python-systemd code has been removed from the systemd repository.
3444 A new repository has been created which accommodates the code from
3445 now on, and we kindly ask distributions to create a separate package
3446 for this: https://github.com/systemd/python-systemd
3448 * The systemd daemon will now reload its main configuration
3449 (/etc/systemd/system.conf) on daemon-reload.
3451 * sd-dhcp now exposes vendor specific extensions via
3452 sd_dhcp_lease_get_vendor_specific().
3454 * systemd-networkd gained a number of new configuration options.
3456 - A new boolean configuration option for TAP devices called
3457 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
3458 device, thus allowing to send and receive GSO packets.
3460 - A new tunnel configuration option called 'CopyDSCP='.
3461 If enabled, the DSCP field of ip6 tunnels is copied into the
3462 decapsulated packet.
3464 - A set of boolean bridge configuration options were added.
3465 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3466 and 'UnicastFlood=' are now parsed by networkd and applied to the
3467 respective bridge link device via the respective IFLA_BRPORT_*
3470 - A new string configuration option to override the hostname sent
3471 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3472 is true, networkd will use the configured hostname instead of the
3473 system hostname when sending DHCP requests.
3475 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3476 networkd will configure the IPv6 flow-label of the tunnel device
3477 according to RFC2460.
3479 - The 'macvtap' virtual network devices are now supported, similar to
3480 the already supported 'macvlan' devices.
3482 * systemd-resolved now implements RFC5452 to improve resilience against
3483 cache poisoning. Additionally, source port randomization is enabled
3484 by default to further protect against DNS spoofing attacks.
3486 * nss-mymachines now supports translating UIDs and GIDs of running
3487 containers with user-namespaces enabled. If a container 'foo'
3488 translates a host uid 'UID' to the container uid 'TUID', then
3489 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3490 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3491 mapped as 'vg-foo-TGID'.
3493 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
3494 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3495 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3496 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3497 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3498 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3499 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3500 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3501 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3502 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3504 — Berlin, 2015-07-29
3508 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3509 There are no known issues with current sysfs, and udev does not need
3510 or should be used to work around such bugs.
3512 * udev does no longer enable USB HID power management. Several reports
3513 indicate, that some devices cannot handle that setting.
3515 * The udev accelerometer helper was removed. The functionality
3516 is now fully included in iio-sensor-proxy. But this means,
3517 older iio-sensor-proxy versions will no longer provide
3518 accelerometer/orientation data with this systemd version.
3519 Please upgrade iio-sensor-proxy to version 1.0.
3521 * networkd gained a new configuration option IPv6PrivacyExtensions=
3522 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3523 for Stateless Address") on selected networks.
3525 * For the sake of fewer build-time dependencies and less code in the
3526 main repository, the python bindings are about to be removed in the
3527 next release. A new repository has been created which accommodates
3528 the code from now on, and we kindly ask distributions to create a
3529 separate package for this. The removal will take place in v223.
3531 https://github.com/systemd/python-systemd
3533 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3534 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3535 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3536 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
3537 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3538 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
3539 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3540 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
3541 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3542 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3544 — Berlin, 2015-07-07
3548 * The sd-bus.h and sd-event.h APIs have now been declared
3549 stable and have been added to the official interface of
3550 libsystemd.so. sd-bus implements an alternative D-Bus client
3551 library, that is relatively easy to use, very efficient and
3552 supports both classic D-Bus as well as kdbus as transport
3553 backend. sd-event is a generic event loop abstraction that
3554 is built around Linux epoll, but adds features such as event
3555 prioritization or efficient timer handling. Both APIs are good
3556 choices for C programs looking for a bus and/or event loop
3557 implementation that is minimal and does not have to be
3558 portable to other kernels.
3560 * kdbus support is no longer compile-time optional. It is now
3561 always built-in. However, it can still be disabled at
3562 runtime using the kdbus=0 kernel command line setting, and
3563 that setting may be changed to default to off, by specifying
3564 --disable-kdbus at build-time. Note though that the kernel
3565 command line setting has no effect if the kdbus.ko kernel
3566 module is not installed, in which case kdbus is (obviously)
3567 also disabled. We encourage all downstream distributions to
3568 begin testing kdbus by adding it to the kernel images in the
3569 development distributions, and leaving kdbus support in
3572 * The minimal required util-linux version has been bumped to
3575 * Support for chkconfig (--enable-chkconfig) was removed in
3576 favor of calling an abstraction tool
3577 /lib/systemd/systemd-sysv-install. This needs to be
3578 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3579 in README for details.
3581 * If there's a systemd unit and a SysV init script for the
3582 same service name, and the user executes "systemctl enable"
3583 for it (or a related call), then this will now enable both
3584 (or execute the related operation on both), not just the
3587 * The libudev API documentation has been converted from gtkdoc
3590 * gudev has been removed from the systemd tree, it is now an
3593 * The systemd-cgtop tool learnt a new --raw switch to generate
3594 "raw" (machine parsable) output.
3596 * networkd's IPForwarding= .network file setting learnt the
3597 new setting "kernel", which ensures that networkd does not
3598 change the IP forwarding sysctl from the default kernel
3601 * The systemd-logind bus API now exposes a new boolean
3602 property "Docked" that reports whether logind considers the
3603 system "docked", i.e. connected to a docking station or not.
3605 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3606 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3607 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3608 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3609 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3610 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3611 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3612 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3613 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3614 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3615 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
3616 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3617 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3618 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3619 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3620 Fink, Zbigniew Jędrzejewski-Szmek
3622 — Berlin, 2015-06-19
3626 * The gudev library has been extracted into a separate repository
3627 available at: https://git.gnome.org/browse/libgudev/
3628 It is now managed as part of the Gnome project. Distributions
3629 are recommended to pass --disable-gudev to systemd and use
3630 gudev from the Gnome project instead. gudev is still included
3631 in systemd, for now. It will be removed soon, though. Please
3632 also see the announcement-thread on systemd-devel:
3633 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
3635 * systemd now exposes a CPUUsageNSec= property for each
3636 service unit on the bus, that contains the overall consumed
3637 CPU time of a service (the sum of what each process of the
3638 service consumed). This value is only available if
3639 CPUAccounting= is turned on for a service, and is then shown
3640 in the "systemctl status" output.
3642 * Support for configuring alternative mappings of the old SysV
3643 runlevels to systemd targets has been removed. They are now
3644 hardcoded in a way that runlevels 2, 3, 4 all map to
3645 multi-user.target and 5 to graphical.target (which
3646 previously was already the default behaviour).
3648 * The auto-mounter logic gained support for mount point
3649 expiry, using a new TimeoutIdleSec= setting in .automount
3650 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3652 * The EFI System Partition (ESP) as mounted to /boot by
3653 systemd-efi-boot-generator will now be unmounted
3654 automatically after 2 minutes of not being used. This should
3655 minimize the risk of ESP corruptions.
3657 * New /etc/fstab options x-systemd.requires= and
3658 x-systemd.requires-mounts-for= are now supported to express
3659 additional dependencies for mounts. This is useful for
3660 journalling file systems that support external journal
3661 devices or overlay file systems that require underlying file
3662 systems to be mounted.
3664 * systemd does not support direct live-upgrades (via systemctl
3665 daemon-reexec) from versions older than v44 anymore. As no
3666 distribution we are aware of shipped such old versions in a
3667 stable release this should not be problematic.
3669 * When systemd forks off a new per-connection service instance
3670 it will now set the $REMOTE_ADDR environment variable to the
3671 remote IP address, and $REMOTE_PORT environment variable to
3672 the remote IP port. This behaviour is similar to the
3673 corresponding environment variables defined by CGI.
3675 * systemd-networkd gained support for uplink failure
3676 detection. The BindCarrier= option allows binding interface
3677 configuration dynamically to the link sense of other
3678 interfaces. This is useful to achieve behaviour like in
3681 * systemd-networkd gained support for configuring the DHCP
3682 client identifier to use when requesting leases.
3684 * systemd-networkd now has a per-network UseNTP= option to
3685 configure whether NTP server information acquired via DHCP
3686 is passed on to services like systemd-timesyncd.
3688 * systemd-networkd gained support for vti6 tunnels.
3690 * Note that systemd-networkd manages the sysctl variable
3691 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3692 it is configured for since v219. The variable controls IP
3693 forwarding, and is a per-interface alternative to the global
3694 /proc/sys/net/ipv[46]/ip_forward. This setting is
3695 configurable in the IPForward= option, which defaults to
3696 "no". This means if networkd is used for an interface it is
3697 no longer sufficient to set the global sysctl option to turn
3698 on IP forwarding! Instead, the .network file option
3699 IPForward= needs to be turned on! Note that the
3700 implementation of this behaviour was broken in v219 and has
3703 * Many bonding and vxlan options are now configurable in
3706 * systemd-nspawn gained a new --property= setting to set unit
3707 properties for the container scope. This is useful for
3708 setting resource parameters (e.g. "CPUShares=500") on
3709 containers started from the command line.
3711 * systemd-nspawn gained a new --private-users= switch to make
3712 use of user namespacing available on recent Linux kernels.
3714 * systemd-nspawn may now be called as part of a shell pipeline
3715 in which case the pipes used for stdin and stdout are passed
3716 directly to the process invoked in the container, without
3717 indirection via a pseudo tty.
3719 * systemd-nspawn gained a new switch to control the UNIX
3720 signal to use when killing the init process of the container
3723 * systemd-nspawn gained a new --overlay= switch for mounting
3724 overlay file systems into the container using the new kernel
3727 * When a container image is imported via systemd-importd and
3728 the host file system is not btrfs, a loopback block device
3729 file is created in /var/lib/machines.raw with a btrfs file
3730 system inside. It is then mounted to /var/lib/machines to
3731 enable btrfs features for container management. The loopback
3732 file and btrfs file system is grown as needed when container
3733 images are imported via systemd-importd.
3735 * systemd-machined/systemd-importd gained support for btrfs
3736 quota, to enforce container disk space limits on disk. This
3737 is exposed in "machinectl set-limit".
3739 * systemd-importd now can import containers from local .tar,
3740 .raw and .qcow2 images, and export them to .tar and .raw. It
3741 can also import dkr v2 images now from the network (on top
3744 * systemd-importd gained support for verifying downloaded
3745 images with gpg2 (previously only gpg1 was supported).
3747 * systemd-machined, systemd-logind, systemd: most bus calls
3748 are now accessible to unprivileged processes via
3749 PolicyKit. Also, systemd-logind will now allow users to kill
3750 their own sessions without further privileges or
3753 * systemd-shutdownd has been removed. This service was
3754 previously responsible for implementing scheduled shutdowns
3755 as exposed in /usr/bin/shutdown's time parameter. This
3756 functionality has now been moved into systemd-logind and is
3757 accessible via a bus interface.
3759 * "systemctl reboot" gained a new switch --firmware-setup that
3760 can be used to reboot into the EFI firmware setup, if that
3761 is available. systemd-logind now exposes an API on the bus
3762 to trigger such reboots, in case graphical desktop UIs want
3763 to cover this functionality.
3765 * "systemctl enable", "systemctl disable" and "systemctl mask"
3766 now support a new "--now" switch. If specified the units
3767 that are enabled will also be started, and the ones
3768 disabled/masked also stopped.
3770 * The Gummiboot EFI boot loader tool has been merged into
3771 systemd, and renamed to "systemd-boot". The bootctl tool has been
3772 updated to support systemd-boot.
3774 * An EFI kernel stub has been added that may be used to create
3775 kernel EFI binaries that contain not only the actual kernel,
3776 but also an initrd, boot splash, command line and OS release
3777 information. This combined binary can then be signed as a
3778 single image, so that the firmware can verify it all in one
3779 step. systemd-boot has special support for EFI binaries created
3780 like this and can extract OS release information from them
3781 and show them in the boot menu. This functionality is useful
3782 to implement cryptographically verified boot schemes.
3784 * Optional support has been added to systemd-fsck to pass
3785 fsck's progress report to an AF_UNIX socket in the file
3788 * udev will no longer create device symlinks for all block
3789 devices by default. A blacklist for excluding special block
3790 devices from this logic has been turned into a whitelist
3791 that requires picking block devices explicitly that require
3794 * A new (currently still internal) API sd-device.h has been
3795 added to libsystemd. This modernized API is supposed to
3796 replace libudev eventually. In fact, already much of libudev
3797 is now just a wrapper around sd-device.h.
3799 * A new hwdb database for storing metadata about pointing
3800 stick devices has been added.
3802 * systemd-tmpfiles gained support for setting file attributes
3803 similar to the "chattr" tool with new 'h' and 'H' lines.
3805 * systemd-journald will no longer unconditionally set the
3806 btrfs NOCOW flag on new journal files. This is instead done
3807 with tmpfiles snippet using the new 'h' line type. This
3808 allows easy disabling of this logic, by masking the
3809 journal-nocow.conf tmpfiles file.
3811 * systemd-journald will now translate audit message types to
3812 human readable identifiers when writing them to the
3813 journal. This should improve readability of audit messages.
3815 * The LUKS logic gained support for the offset= and skip=
3816 options in /etc/crypttab, as previously implemented by
3819 * /usr/lib/os-release gained a new optional field VARIANT= for
3820 distributions that support multiple variants (such as a
3821 desktop edition, a server edition, ...)
3823 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3824 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3825 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3826 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3827 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3828 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3829 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3830 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3831 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3832 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3833 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3834 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3835 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3836 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3837 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3838 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3839 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3840 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3841 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3842 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3843 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3844 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3845 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3846 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3847 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3848 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3849 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3851 — Berlin, 2015-05-22
3855 * Introduce a new API "sd-hwdb.h" for querying the hardware
3856 metadata database. With this minimal interface one can query
3857 and enumerate the udev hwdb, decoupled from the old libudev
3858 library. libudev's interface for this is now only a wrapper
3859 around sd-hwdb. A new tool systemd-hwdb has been added to
3860 interface with and update the database.
3862 * When any of systemd's tools copies files (for example due to
3863 tmpfiles' C lines) a btrfs reflink will attempted first,
3864 before bytewise copying is done.
3866 * systemd-nspawn gained a new --ephemeral switch. When
3867 specified a btrfs snapshot is taken of the container's root
3868 directory, and immediately removed when the container
3869 terminates again. Thus, a container can be started whose
3870 changes never alter the container's root directory, and are
3871 lost on container termination. This switch can also be used
3872 for starting a container off the root file system of the
3873 host without affecting the host OS. This switch is only
3874 available on btrfs file systems.
3876 * systemd-nspawn gained a new --template= switch. It takes the
3877 path to a container tree to use as template for the tree
3878 specified via --directory=, should that directory be
3879 missing. This allows instantiating containers dynamically,
3880 on first run. This switch is only available on btrfs file
3883 * When a .mount unit refers to a mount point on which multiple
3884 mounts are stacked, and the .mount unit is stopped all of
3885 the stacked mount points will now be unmounted until no
3886 mount point remains.
3888 * systemd now has an explicit notion of supported and
3889 unsupported unit types. Jobs enqueued for unsupported unit
3890 types will now fail with an "unsupported" error code. More
3891 specifically .swap, .automount and .device units are not
3892 supported in containers, .busname units are not supported on
3893 non-kdbus systems. .swap and .automount are also not
3894 supported if their respective kernel compile time options
3897 * machinectl gained support for two new "copy-from" and
3898 "copy-to" commands for copying files from a running
3899 container to the host or vice versa.
3901 * machinectl gained support for a new "bind" command to bind
3902 mount host directories into local containers. This is
3903 currently only supported for nspawn containers.
3905 * networkd gained support for configuring bridge forwarding
3906 database entries (fdb) from .network files.
3908 * A new tiny daemon "systemd-importd" has been added that can
3909 download container images in tar, raw, qcow2 or dkr formats,
3910 and make them available locally in /var/lib/machines, so
3911 that they can run as nspawn containers. The daemon can GPG
3912 verify the downloads (not supported for dkr, since it has no
3913 provisions for verifying downloads). It will transparently
3914 decompress bz2, xz, gzip compressed downloads if necessary,
3915 and restore sparse files on disk. The daemon uses privilege
3916 separation to ensure the actual download logic runs with
3917 fewer privileges than the daemon itself. machinectl has
3918 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3919 make the functionality of importd available to the
3920 user. With this in place the Fedora and Ubuntu "Cloud"
3921 images can be downloaded and booted as containers unmodified
3922 (the Fedora images lack the appropriate GPG signature files
3923 currently, so they cannot be verified, but this will change
3924 soon, hopefully). Note that downloading images is currently
3925 only fully supported on btrfs.
3927 * machinectl is now able to list container images found in
3928 /var/lib/machines, along with some metadata about sizes of
3929 disk and similar. If the directory is located on btrfs and
3930 quota is enabled, this includes quota display. A new command
3931 "image-status" has been added that shows additional
3932 information about images.
3934 * machinectl is now able to clone container images
3935 efficiently, if the underlying file system (btrfs) supports
3936 it, with the new "machinectl clone" command. It also
3937 gained commands for renaming and removing images, as well as
3938 marking them read-only or read-write (supported also on
3939 legacy file systems).
3941 * networkd gained support for collecting LLDP network
3942 announcements, from hardware that supports this. This is
3943 shown in networkctl output.
3945 * systemd-run gained support for a new -t (--pty) switch for
3946 invoking a binary on a pty whose input and output is
3947 connected to the invoking terminal. This allows executing
3948 processes as system services while interactively
3949 communicating with them via the terminal. Most interestingly
3950 this is supported across container boundaries. Invoking
3951 "systemd-run -t /bin/bash" is an alternative to running a
3952 full login session, the difference being that the former
3953 will not register a session, nor go through the PAM session
3956 * tmpfiles gained support for a new "v" line type for creating
3957 btrfs subvolumes. If the underlying file system is a legacy
3958 file system, this automatically degrades to creating a
3959 normal directory. Among others /var/lib/machines is now
3960 created like this at boot, should it be missing.
3962 * The directory /var/lib/containers/ has been deprecated and
3963 been replaced by /var/lib/machines. The term "machines" has
3964 been used in the systemd context as generic term for both
3965 VMs and containers, and hence appears more appropriate for
3966 this, as the directory can also contain raw images bootable
3969 * systemd-nspawn when invoked with -M but without --directory=
3970 or --image= is now capable of searching for the container
3971 root directory, subvolume or disk image automatically, in
3972 /var/lib/machines. systemd-nspawn@.service has been updated
3973 to make use of this, thus allowing it to be used for raw
3976 * A new machines.target unit has been introduced that is
3977 supposed to group all containers/VMs invoked as services on
3978 the system. systemd-nspawn@.service has been updated to
3979 integrate with that.
3981 * machinectl gained a new "start" command, for invoking a
3982 container as a service. "machinectl start foo" is mostly
3983 equivalent to "systemctl start systemd-nspawn@foo.service",
3984 but handles escaping in a nicer way.
3986 * systemd-nspawn will now mount most of the cgroupfs tree
3987 read-only into each container, with the exception of the
3988 container's own subtree in the name=systemd hierarchy.
3990 * journald now sets the special FS_NOCOW file flag for its
3991 journal files. This should improve performance on btrfs, by
3992 avoiding heavy fragmentation when journald's write-pattern
3993 is used on COW file systems. It degrades btrfs' data
3994 integrity guarantees for the files to the same levels as for
3995 ext3/ext4 however. This should be OK though as journald does
3996 its own data integrity checks and all its objects are
3997 checksummed on disk. Also, journald should handle btrfs disk
3998 full events a lot more gracefully now, by processing SIGBUS
3999 errors, and not relying on fallocate() anymore.
4001 * When journald detects that journal files it is writing to
4002 have been deleted it will immediately start new journal
4005 * systemd now provides a way to store file descriptors
4006 per-service in PID 1. This is useful for daemons to ensure
4007 that fds they require are not lost during a daemon
4008 restart. The fds are passed to the daemon on the next
4009 invocation in the same way socket activation fds are
4010 passed. This is now used by journald to ensure that the
4011 various sockets connected to all the system's stdout/stderr
4012 are not lost when journald is restarted. File descriptors
4013 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
4014 an extension to sd_notify(). Note that a limit is enforced
4015 on the number of fds a service can store in PID 1, and it
4016 defaults to 0, so that no fds may be stored, unless this is
4017 explicitly turned on.
4019 * The default TERM variable to use for units connected to a
4020 terminal, when no other value is explicitly is set is now
4021 vt220 rather than vt102. This should be fairly safe still,
4022 but allows PgUp/PgDn work.
4024 * The /etc/crypttab option header= as known from Debian is now
4027 * "loginctl user-status" and "loginctl session-status" will
4028 now show the last 10 lines of log messages of the
4029 user/session following the status output. Similar,
4030 "machinectl status" will show the last 10 log lines
4031 associated with a virtual machine or container
4032 service. (Note that this is usually not the log messages
4033 done in the VM/container itself, but simply what the
4034 container manager logs. For nspawn this includes all console
4037 * "loginctl session-status" without further argument will now
4038 show the status of the session of the caller. Similar,
4039 "lock-session", "unlock-session", "activate",
4040 "enable-linger", "disable-linger" may now be called without
4041 session/user parameter in which case they apply to the
4042 caller's session/user.
4044 * An X11 session scriptlet is now shipped that uploads
4045 $DISPLAY and $XAUTHORITY into the environment of the systemd
4046 --user daemon if a session begins. This should improve
4047 compatibility with X11 enabled applications run as systemd
4050 * Generators are now subject to masking via /etc and /run, the
4051 same way as unit files.
4053 * networkd .network files gained support for configuring
4054 per-link IPv4/IPv6 packet forwarding as well as IPv4
4055 masquerading. This is by default turned on for veth links to
4056 containers, as registered by systemd-nspawn. This means that
4057 nspawn containers run with --network-veth will now get
4058 automatic routed access to the host's networks without any
4059 further configuration or setup, as long as networkd runs on
4062 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4063 or UDP posts of a container on the host. With this in place
4064 it is possible to run containers with private veth links
4065 (--network-veth), and have their functionality exposed on
4066 the host as if their services were running directly on the
4069 * systemd-nspawn's --network-veth switch now gained a short
4070 version "-n", since with the changes above it is now truly
4071 useful out-of-the-box. The systemd-nspawn@.service has been
4072 updated to make use of it too by default.
4074 * systemd-nspawn will now maintain a per-image R/W lock, to
4075 ensure that the same image is not started more than once
4076 writable. (It's OK to run an image multiple times
4077 simultaneously in read-only mode.)
4079 * systemd-nspawn's --image= option is now capable of
4080 dissecting and booting MBR and GPT disk images that contain
4081 only a single active Linux partition. Previously it
4082 supported only GPT disk images with proper GPT type
4083 IDs. This allows running cloud images from major
4084 distributions directly with systemd-nspawn, without
4087 * In addition to collecting mouse dpi data in the udev
4088 hardware database, there's now support for collecting angle
4089 information for mouse scroll wheels. The database is
4090 supposed to guarantee similar scrolling behavior on mice
4091 that it knows about. There's also support for collecting
4092 information about Touchpad types.
4094 * udev's input_id built-in will now also collect touch screen
4095 dimension data and attach it to probed devices.
4097 * /etc/os-release gained support for a Distribution Privacy
4100 * networkd gained support for creating "ipvlan", "gretap",
4101 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4103 * systemd-tmpfiles gained support for "a" lines for setting
4106 * systemd-nspawn will now mount /tmp in the container to
4107 tmpfs, automatically.
4109 * systemd now exposes the memory.usage_in_bytes cgroup
4110 attribute and shows it for each service in the "systemctl
4111 status" output, if available.
4113 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4114 immediate reboot is triggered. This useful if shutdown is
4115 hung and is unable to complete, to expedite the
4116 operation. Note that this kind of reboot will still unmount
4117 all file systems, and hence should not result in fsck being
4120 * A .device unit for an optical block device will now be
4121 considered active only when a medium is in the drive. Also,
4122 mount units are now bound to their backing devices thus
4123 triggering automatic unmounting when devices become
4124 unavailable. With this in place systemd will now
4125 automatically unmount left-over mounts when a CD-ROM is
4126 ejected or an USB stick is yanked from the system.
4128 * networkd-wait-online now has support for waiting for
4129 specific interfaces only (with globbing), and for giving up
4130 after a configurable timeout.
4132 * networkd now exits when idle. It will be automatically
4133 restarted as soon as interfaces show up, are removed or
4134 change state. networkd will stay around as long as there is
4135 at least one DHCP state machine or similar around, that keep
4138 * networkd may now configure IPv6 link-local addressing in
4139 addition to IPv4 link-local addressing.
4141 * The IPv6 "token" for use in SLAAC may now be configured for
4142 each .network interface in networkd.
4144 * Routes configured with networkd may now be assigned a scope
4147 * networkd's [Match] sections now support globbing and lists
4148 of multiple space-separated matches per item.
4150 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4151 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4152 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4153 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4154 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4155 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4156 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4157 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4158 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4159 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4160 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4161 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4162 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4163 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4164 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
4165 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4166 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4167 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4168 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4169 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4170 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4171 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
4172 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4173 Hoffmann, Zbigniew Jędrzejewski-Szmek
4175 — Berlin, 2015-02-16
4179 * When querying unit file enablement status (for example via
4180 "systemctl is-enabled"), a new state "indirect" is now known
4181 which indicates that a unit might not be enabled itself, but
4182 another unit listed in its Also= setting might be.
4184 * Similar to the various existing ConditionXYZ= settings for
4185 units, there are now matching AssertXYZ= settings. While
4186 failing conditions cause a unit to be skipped, but its job
4187 to succeed, failing assertions declared like this will cause
4188 a unit start operation and its job to fail.
4190 * hostnamed now knows a new chassis type "embedded".
4192 * systemctl gained a new "edit" command. When used on a unit
4193 file, this allows extending unit files with .d/ drop-in
4194 configuration snippets or editing the full file (after
4195 copying it from /usr/lib to /etc). This will invoke the
4196 user's editor (as configured with $EDITOR), and reload the
4197 modified configuration after editing.
4199 * "systemctl status" now shows the suggested enablement state
4200 for a unit, as declared in the (usually vendor-supplied)
4201 system preset files.
4203 * nss-myhostname will now resolve the single-label host name
4204 "gateway" to the locally configured default IP routing
4205 gateways, ordered by their metrics. This assigns a stable
4206 name to the used gateways, regardless which ones are
4207 currently configured. Note that the name will only be
4208 resolved after all other name sources (if nss-myhostname is
4209 configured properly) and should hence not negatively impact
4210 systems that use the single-label host name "gateway" in
4213 * systemd-inhibit now allows filtering by mode when listing
4216 * Scope and service units gained a new "Delegate" boolean
4217 property, which, when set, allows processes running inside the
4218 unit to further partition resources. This is primarily
4219 useful for systemd user instances as well as container
4222 * journald will now pick up audit messages directly from
4223 the kernel, and log them like any other log message. The
4224 audit fields are split up and fully indexed. This means that
4225 journalctl in many ways is now a (nicer!) alternative to
4226 ausearch, the traditional audit client. Note that this
4227 implements only a minimal audit client. If you want the
4228 special audit modes like reboot-on-log-overflow, please use
4229 the traditional auditd instead, which can be used in
4230 parallel to journald.
4232 * The ConditionSecurity= unit file option now understands the
4233 special string "audit" to check whether auditing is
4236 * journalctl gained two new commands --vacuum-size= and
4237 --vacuum-time= to delete old journal files until the
4238 remaining ones take up no more than the specified size on disk,
4239 or are not older than the specified time.
4241 * A new, native PPPoE library has been added to sd-network,
4242 systemd's library of light-weight networking protocols. This
4243 library will be used in a future version of networkd to
4244 enable PPPoE communication without an external pppd daemon.
4246 * The busctl tool now understands a new "capture" verb that
4247 works similar to "monitor", but writes a packet capture
4248 trace to STDOUT that can be redirected to a file which is
4249 compatible with libcap's capture file format. This can then
4250 be loaded in Wireshark and similar tools to inspect bus
4253 * The busctl tool now understands a new "tree" verb that shows
4254 the object trees of a specific service on the bus, or of all
4257 * The busctl tool now understands a new "introspect" verb that
4258 shows all interfaces and members of objects on the bus,
4259 including their signature and values. This is particularly
4260 useful to get more information about bus objects shown by
4261 the new "busctl tree" command.
4263 * The busctl tool now understands new verbs "call",
4264 "set-property" and "get-property" for invoking bus method
4265 calls, setting and getting bus object properties in a
4268 * busctl gained a new --augment-creds= argument that controls
4269 whether the tool shall augment credential information it
4270 gets from the bus with data from /proc, in a possibly
4273 * nspawn's --link-journal= switch gained two new values
4274 "try-guest" and "try-host" that work like "guest" and
4275 "host", but do not fail if the host has no persistent
4276 journalling enabled. -j is now equivalent to
4277 --link-journal=try-guest.
4279 * macvlan network devices created by nspawn will now have
4280 stable MAC addresses.
4282 * A new SmackProcessLabel= unit setting has been added, which
4283 controls the SMACK security label processes forked off by
4284 the respective unit shall use.
4286 * If compiled with --enable-xkbcommon, systemd-localed will
4287 verify x11 keymap settings by compiling the given keymap. It
4288 will spew out warnings if the compilation fails. This
4289 requires libxkbcommon to be installed.
4291 * When a coredump is collected, a larger number of metadata
4292 fields is now collected and included in the journal records
4293 created for it. More specifically, control group membership,
4294 environment variables, memory maps, working directory,
4295 chroot directory, /proc/$PID/status, and a list of open file
4296 descriptors is now stored in the log entry.
4298 * The udev hwdb now contains DPI information for mice. For
4301 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4303 * All systemd programs that read standalone configuration
4304 files in /etc now also support a corresponding series of
4305 .conf.d configuration directories in /etc/, /run/,
4306 /usr/local/lib/, /usr/lib/, and (if configured with
4307 --enable-split-usr) /lib/. In particular, the following
4308 configuration files now have corresponding configuration
4309 directories: system.conf user.conf, logind.conf,
4310 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4311 resolved.conf, timesyncd.conf, journal-remote.conf, and
4312 journal-upload.conf. Note that distributions should use the
4313 configuration directories in /usr/lib/; the directories in
4314 /etc/ are reserved for the system administrator.
4316 * systemd-rfkill will no longer take the rfkill device name
4317 into account when storing rfkill state on disk, as the name
4318 might be dynamically assigned and not stable. Instead, the
4319 ID_PATH udev variable combined with the rfkill type (wlan,
4320 bluetooth, ...) is used.
4322 * A new service systemd-machine-id-commit.service has been
4323 added. When used on systems where /etc is read-only during
4324 boot, and /etc/machine-id is not initialized (but an empty
4325 file), this service will copy the temporary machine ID
4326 created as replacement into /etc after the system is fully
4327 booted up. This is useful for systems that are freshly
4328 installed with a non-initialized machine ID, but should get
4329 a fixed machine ID for subsequent boots.
4331 * networkd's .netdev files now provide a large set of
4332 configuration parameters for VXLAN devices. Similarly, the
4333 bridge port cost parameter is now configurable in .network
4334 files. There's also new support for configuring IP source
4335 routing. networkd .link files gained support for a new
4336 OriginalName= match that is useful to match against the
4337 original interface name the kernel assigned. .network files
4338 may include MTU= and MACAddress= fields for altering the MTU
4339 and MAC address while being connected to a specific network
4342 * The LUKS logic gained supported for configuring
4343 UUID-specific key files. There's also new support for naming
4344 LUKS device from the kernel command line, using the new
4345 luks.name= argument.
4347 * Timer units may now be transiently created via the bus API
4348 (this was previously already available for scope and service
4349 units). In addition it is now possible to create multiple
4350 transient units at the same time with a single bus call. The
4351 "systemd-run" tool has been updated to make use of this for
4352 running commands on a specified time, in at(1)-style.
4354 * tmpfiles gained support for "t" lines, for assigning
4355 extended attributes to files. Among other uses this may be
4356 used to assign SMACK labels to files.
4358 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4359 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4360 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4361 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4362 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4363 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4364 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4365 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4366 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4367 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4368 Michael Biebl, Michael Chapman, Michael Marineau, Michal
4369 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4370 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4371 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4372 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4373 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4374 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4375 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
4377 — Berlin, 2014-12-10
4381 * journalctl gained the new options -t/--identifier= to match
4382 on the syslog identifier (aka "tag"), as well as --utc to
4383 show log timestamps in the UTC timezone. journalctl now also
4384 accepts -n/--lines=all to disable line capping in a pager.
4386 * journalctl gained a new switch, --flush, that synchronously
4387 flushes logs from /run/log/journal to /var/log/journal if
4388 persistent storage is enabled. systemd-journal-flush.service
4389 now waits until the operation is complete.
4391 * Services can notify the manager before they start a reload
4392 (by sending RELOADING=1) or shutdown (by sending
4393 STOPPING=1). This allows the manager to track and show the
4394 internal state of daemons and closes a race condition when
4395 the process is still running but has closed its D-Bus
4398 * Services with Type=oneshot do not have to have any ExecStart
4401 * User units are now loaded also from
4402 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4403 /run/systemd/user directory that was already previously
4404 supported, but is under the control of the user.
4406 * Job timeouts (i.e. time-outs on the time a job that is
4407 queued stays in the run queue) can now optionally result in
4408 immediate reboot or power-off actions (JobTimeoutAction= and
4409 JobTimeoutRebootArgument=). This is useful on ".target"
4410 units, to limit the maximum time a target remains
4411 undispatched in the run queue, and to trigger an emergency
4412 operation in such a case. This is now used by default to
4413 turn off the system if boot-up (as defined by everything in
4414 basic.target) hangs and does not complete for at least
4415 15min. Also, if power-off or reboot hang for at least 30min
4416 an immediate power-off/reboot operation is triggered. This
4417 functionality is particularly useful to increase reliability
4418 on embedded devices, but also on laptops which might
4419 accidentally get powered on when carried in a backpack and
4420 whose boot stays stuck in a hard disk encryption passphrase
4423 * systemd-logind can be configured to also handle lid switch
4424 events even when the machine is docked or multiple displays
4425 are attached (HandleLidSwitchDocked= option).
4427 * A helper binary and a service have been added which can be
4428 used to resume from hibernation in the initramfs. A
4429 generator will parse the resume= option on the kernel
4430 command line to trigger resume.
4432 * A user console daemon systemd-consoled has been
4433 added. Currently, it is a preview, and will so far open a
4434 single terminal on each session of the user marked as
4435 Desktop=systemd-console.
4437 * Route metrics can be specified for DHCP routes added by
4440 * The SELinux context of socket-activated services can be set
4441 from the information provided by the networking stack
4442 (SELinuxContextFromNet= option).
4444 * Userspace firmware loading support has been removed and
4445 the minimum supported kernel version is thus bumped to 3.7.
4447 * Timeout for udev workers has been increased from 1 to 3
4448 minutes, but a warning will be printed after 1 minute to
4449 help diagnose kernel modules that take a long time to load.
4451 * Udev rules can now remove tags on devices with TAG-="foobar".
4453 * systemd's readahead implementation has been removed. In many
4454 circumstances it didn't give expected benefits even for
4455 rotational disk drives and was becoming less relevant in the
4456 age of SSDs. As none of the developers has been using
4457 rotating media anymore, and nobody stepped up to actively
4458 maintain this component of systemd it has now been removed.
4460 * Swap units can use Options= to specify discard options.
4461 Discard options specified for swaps in /etc/fstab are now
4464 * Docker containers are now detected as a separate type of
4467 * The Password Agent protocol gained support for queries where
4468 the user input is shown, useful e.g. for user names.
4469 systemd-ask-password gained a new --echo option to turn that
4472 * The default sysctl.d/ snippets will now set:
4474 net.core.default_qdisc = fq_codel
4476 This selects Fair Queuing Controlled Delay as the default
4477 queuing discipline for network interfaces. fq_codel helps
4478 fight the network bufferbloat problem. It is believed to be
4479 a good default with no tuning required for most workloads.
4480 Downstream distributions may override this choice. On 10Gbit
4481 servers that do not do forwarding, "fq" may perform better.
4482 Systems without a good clocksource should use "pfifo_fast".
4484 * If kdbus is enabled during build a new option BusPolicy= is
4485 available for service units, that allows locking all service
4486 processes into a stricter bus policy, in order to limit
4487 access to various bus services, or even hide most of them
4488 from the service's view entirely.
4490 * networkctl will now show the .network and .link file
4491 networkd has applied to a specific interface.
4493 * sd-login gained a new API call sd_session_get_desktop() to
4494 query which desktop environment has been selected for a
4497 * UNIX utmp support is now compile-time optional to support
4498 legacy-free systems.
4500 * systemctl gained two new commands "add-wants" and
4501 "add-requires" for pulling in units from specific targets
4504 * If the word "rescue" is specified on the kernel command line
4505 the system will now boot into rescue mode (aka
4506 rescue.target), which was previously available only by
4507 specifying "1" or "systemd.unit=rescue.target" on the kernel
4508 command line. This new kernel command line option nicely
4509 mirrors the already existing "emergency" kernel command line
4512 * New kernel command line options mount.usr=, mount.usrflags=,
4513 mount.usrfstype= have been added that match root=, rootflags=,
4514 rootfstype= but allow mounting a specific file system to
4517 * The $NOTIFY_SOCKET is now also passed to control processes of
4518 services, not only the main process.
4520 * This version reenables support for fsck's -l switch. This
4521 means at least version v2.25 of util-linux is required for
4522 operation, otherwise dead-locks on device nodes may
4523 occur. Again: you need to update util-linux to at least
4524 v2.25 when updating systemd to v217.
4526 * The "multi-seat-x" tool has been removed from systemd, as
4527 its functionality has been integrated into X servers 1.16,
4528 and the tool is hence redundant. It is recommended to update
4529 display managers invoking this tool to simply invoke X
4530 directly from now on, again.
4532 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4533 message flag has been added for all of systemd's PolicyKit
4534 authenticated method calls has been added. In particular
4535 this now allows optional interactive authorization via
4536 PolicyKit for many of PID1's privileged operations such as
4537 unit file enabling and disabling.
4539 * "udevadm hwdb --update" learnt a new switch "--usr" for
4540 placing the rebuilt hardware database in /usr instead of
4541 /etc. When used only hardware database entries stored in
4542 /usr will be used, and any user database entries in /etc are
4543 ignored. This functionality is useful for vendors to ship a
4544 pre-built database on systems where local configuration is
4545 unnecessary or unlikely.
4547 * Calendar time specifications in .timer units now also
4548 understand the strings "semi-annually", "quarterly" and
4549 "minutely" as shortcuts (in addition to the preexisting
4550 "anually", "hourly", ...).
4552 * systemd-tmpfiles will now correctly create files in /dev
4553 at boot which are marked for creation only at boot. It is
4554 recommended to always create static device nodes with 'c!'
4555 and 'b!', so that they are created only at boot and not
4556 overwritten at runtime.
4558 * When the watchdog logic is used for a service (WatchdogSec=)
4559 and the watchdog timeout is hit the service will now be
4560 terminated with SIGABRT (instead of just SIGTERM), in order
4561 to make sure a proper coredump and backtrace is
4562 generated. This ensures that hanging services will result in
4563 similar coredump/backtrace behaviour as services that hit a
4566 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4567 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4568 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4569 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4570 Herrmann, David Sommerseth, David Strauss, Emil Renner
4571 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4572 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4573 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4574 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4575 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4576 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4577 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4578 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4579 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4580 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4581 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4582 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4583 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4584 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4585 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4586 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
4589 — Berlin, 2014-10-28
4593 * timedated no longer reads NTP implementation unit names from
4594 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
4595 implementations should add a
4597 Conflicts=systemd-timesyncd.service
4599 to their unit files to take over and replace systemd's NTP
4600 default functionality.
4602 * systemd-sysusers gained a new line type "r" for configuring
4603 which UID/GID ranges to allocate system users/groups
4604 from. Lines of type "u" may now add an additional column
4605 that specifies the home directory for the system user to be
4606 created. Also, systemd-sysusers may now optionally read user
4607 information from STDIN instead of a file. This is useful for
4608 invoking it from RPM preinst scriptlets that need to create
4609 users before the first RPM file is installed since these
4610 files might need to be owned by them. A new
4611 %sysusers_create_inline RPM macro has been introduced to do
4612 just that. systemd-sysusers now updates the shadow files as
4613 well as the user/group databases, which should enhance
4614 compatibility with certain tools like grpck.
4616 * A number of bus APIs of PID 1 now optionally consult
4617 PolicyKit to permit access for otherwise unprivileged
4618 clients under certain conditions. Note that this currently
4619 doesn't support interactive authentication yet, but this is
4620 expected to be added eventually, too.
4622 * /etc/machine-info now has new fields for configuring the
4623 deployment environment of the machine, as well as the
4624 location of the machine. hostnamectl has been updated with
4625 new command to update these fields.
4627 * systemd-timesyncd has been updated to automatically acquire
4628 NTP server information from systemd-networkd, which might
4629 have been discovered via DHCP.
4631 * systemd-resolved now includes a caching DNS stub resolver
4632 and a complete LLMNR name resolution implementation. A new
4633 NSS module "nss-resolve" has been added which can be used
4634 instead of glibc's own "nss-dns" to resolve hostnames via
4635 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4636 be resolved via systemd-resolved D-Bus APIs. In contrast to
4637 the glibc internal resolver systemd-resolved is aware of
4638 multi-homed system, and keeps DNS server and caches separate
4639 and per-interface. Queries are sent simultaneously on all
4640 interfaces that have DNS servers configured, in order to
4641 properly handle VPNs and local LANs which might resolve
4642 separate sets of domain names. systemd-resolved may acquire
4643 DNS server information from systemd-networkd automatically,
4644 which in turn might have discovered them via DHCP. A tool
4645 "systemd-resolve-host" has been added that may be used to
4646 query the DNS logic in resolved. systemd-resolved implements
4647 IDNA and automatically uses IDNA or UTF-8 encoding depending
4648 on whether classic DNS or LLMNR is used as transport. In the
4649 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4650 implementation to systemd-resolved.
4652 * A new NSS module nss-mymachines has been added, that
4653 automatically resolves the names of all local registered
4654 containers to their respective IP addresses.
4656 * A new client tool "networkctl" for systemd-networkd has been
4657 added. It currently is entirely passive and will query
4658 networking configuration from udev, rtnetlink and networkd,
4659 and present it to the user in a very friendly
4660 way. Eventually, we hope to extend it to become a full
4661 control utility for networkd.
4663 * .socket units gained a new DeferAcceptSec= setting that
4664 controls the kernels' TCP_DEFER_ACCEPT sockopt for
4665 TCP. Similarly, support for controlling TCP keep-alive
4666 settings has been added (KeepAliveTimeSec=,
4667 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4668 turning off Nagle's algorithm on TCP has been added
4671 * logind learned a new session type "web", for use in projects
4672 like Cockpit which register web clients as PAM sessions.
4674 * timer units with at least one OnCalendar= setting will now
4675 be started only after time-sync.target has been
4676 reached. This way they will not elapse before the system
4677 clock has been corrected by a local NTP client or
4678 similar. This is particular useful on RTC-less embedded
4679 machines, that come up with an invalid system clock.
4681 * systemd-nspawn's --network-veth= switch should now result in
4682 stable MAC addresses for both the outer and the inner side
4685 * systemd-nspawn gained a new --volatile= switch for running
4686 container instances with /etc or /var unpopulated.
4688 * The kdbus client code has been updated to use the new Linux
4689 3.17 memfd subsystem instead of the old kdbus-specific one.
4691 * systemd-networkd's DHCP client and server now support
4692 FORCERENEW. There are also new configuration options to
4693 configure the vendor client identifier and broadcast mode
4696 * systemd will no longer inform the kernel about the current
4697 timezone, as this is necessarily incorrect and racy as the
4698 kernel has no understanding of DST and similar
4699 concepts. This hence means FAT timestamps will be always
4700 considered UTC, similar to what Android is already
4701 doing. Also, when the RTC is configured to the local time
4702 (rather than UTC) systemd will never synchronize back to it,
4703 as this might confuse Windows at a later boot.
4705 * systemd-analyze gained a new command "verify" for offline
4706 validation of unit files.
4708 * systemd-networkd gained support for a couple of additional
4709 settings for bonding networking setups. Also, the metric for
4710 statically configured routes may now be configured. For
4711 network interfaces where this is appropriate the peer IP
4712 address may now be configured.
4714 * systemd-networkd's DHCP client will no longer request
4715 broadcasting by default, as this tripped up some networks.
4716 For hardware where broadcast is required the feature should
4717 be switched back on using RequestBroadcast=yes.
4719 * systemd-networkd will now set up IPv4LL addresses (when
4720 enabled) even if DHCP is configured successfully.
4722 * udev will now default to respect network device names given
4723 by the kernel when the kernel indicates that these are
4724 predictable. This behavior can be tweaked by changing
4725 NamePolicy= in the relevant .link file.
4727 * A new library systemd-terminal has been added that
4728 implements full TTY stream parsing and rendering. This
4729 library is supposed to be used later on for implementing a
4730 full userspace VT subsystem, replacing the current kernel
4733 * A new tool systemd-journal-upload has been added to push
4734 journal data to a remote system running
4735 systemd-journal-remote.
4737 * journald will no longer forward all local data to another
4738 running syslog daemon. This change has been made because
4739 rsyslog (which appears to be the most commonly used syslog
4740 implementation these days) no longer makes use of this, and
4741 instead pulls the data out of the journal on its own. Since
4742 forwarding the messages to a non-existent syslog server is
4743 more expensive than we assumed we have now turned this
4744 off. If you run a syslog server that is not a recent rsyslog
4745 version, you have to turn this option on again
4746 (ForwardToSyslog= in journald.conf).
4748 * journald now optionally supports the LZ4 compressor for
4749 larger journal fields. This compressor should perform much
4750 better than XZ which was the previous default.
4752 * machinectl now shows the IP addresses of local containers,
4753 if it knows them, plus the interface name of the container.
4755 * A new tool "systemd-escape" has been added that makes it
4756 easy to escape strings to build unit names and similar.
4758 * sd_notify() messages may now include a new ERRNO= field
4759 which is parsed and collected by systemd and shown among the
4760 "systemctl status" output for a service.
4762 * A new component "systemd-firstboot" has been added that
4763 queries the most basic systemd information (timezone,
4764 hostname, root password) interactively on first
4765 boot. Alternatively it may also be used to provision these
4766 things offline on OS images installed into directories.
4768 * The default sysctl.d/ snippets will now set
4770 net.ipv4.conf.default.promote_secondaries=1
4772 This has the benefit of no flushing secondary IP addresses
4773 when primary addresses are removed.
4775 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4776 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4777 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4778 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4779 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4780 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4781 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4782 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4783 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4784 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4785 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4786 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4787 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4788 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4789 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4791 — Berlin, 2014-08-19
4795 * A new tool systemd-sysusers has been added. This tool
4796 creates system users and groups in /etc/passwd and
4797 /etc/group, based on static declarative system user/group
4798 definitions in /usr/lib/sysusers.d/. This is useful to
4799 enable factory resets and volatile systems that boot up with
4800 an empty /etc directory, and thus need system users and
4801 groups created during early boot. systemd now also ships
4802 with two default sysusers.d/ files for the most basic
4803 users and groups systemd and the core operating system
4806 * A new tmpfiles snippet has been added that rebuilds the
4807 essential files in /etc on boot, should they be missing.
4809 * A directive for ensuring automatic clean-up of
4810 /var/cache/man/ has been removed from the default
4811 configuration. This line should now be shipped by the man
4812 implementation. The necessary change has been made to the
4813 man-db implementation. Note that you need to update your man
4814 implementation to one that ships this line, otherwise no
4815 automatic clean-up of /var/cache/man will take place.
4817 * A new condition ConditionNeedsUpdate= has been added that
4818 may conditionalize services to only run when /etc or /var
4819 are "older" than the vendor operating system resources in
4820 /usr. This is useful for reconstructing or updating /etc
4821 after an offline update of /usr or a factory reset, on the
4822 next reboot. Services that want to run once after such an
4823 update or reset should use this condition and order
4824 themselves before the new systemd-update-done.service, which
4825 will mark the two directories as fully updated. A number of
4826 service files have been added making use of this, to rebuild
4827 the udev hardware database, the journald message catalog and
4828 dynamic loader cache (ldconfig). The systemd-sysusers tool
4829 described above also makes use of this now. With this in
4830 place it is now possible to start up a minimal operating
4831 system with /etc empty cleanly. For more information on the
4832 concepts involved see this recent blog story:
4834 http://0pointer.de/blog/projects/stateless.html
4836 * A new system group "input" has been introduced, and all
4837 input device nodes get this group assigned. This is useful
4838 for system-level software to get access to input devices. It
4839 complements what is already done for "audio" and "video".
4841 * systemd-networkd learnt minimal DHCPv4 server support in
4842 addition to the existing DHCPv4 client support. It also
4843 learnt DHCPv6 client and IPv6 Router Solicitation client
4844 support. The DHCPv4 client gained support for static routes
4845 passed in from the server. Note that the [DHCPv4] section
4846 known in older systemd-networkd versions has been renamed to
4847 [DHCP] and is now also used by the DHCPv6 client. Existing
4848 .network files using settings of this section should be
4849 updated, though compatibility is maintained. Optionally, the
4850 client hostname may now be sent to the DHCP server.
4852 * networkd gained support for vxlan virtual networks as well
4853 as tun/tap and dummy devices.
4855 * networkd gained support for automatic allocation of address
4856 ranges for interfaces from a system-wide pool of
4857 addresses. This is useful for dynamically managing a large
4858 number of interfaces with a single network configuration
4859 file. In particular this is useful to easily assign
4860 appropriate IP addresses to the veth links of a large number
4861 of nspawn instances.
4863 * RPM macros for processing sysusers, sysctl and binfmt
4864 drop-in snippets at package installation time have been
4867 * The /etc/os-release file should now be placed in
4868 /usr/lib/os-release. The old location is automatically
4869 created as symlink. /usr/lib is the more appropriate
4870 location of this file, since it shall actually describe the
4871 vendor operating system shipped in /usr, and not the
4872 configuration stored in /etc.
4874 * .mount units gained a new boolean SloppyOptions= setting
4875 that maps to mount(8)'s -s option which enables permissive
4876 parsing of unknown mount options.
4878 * tmpfiles learnt a new "L+" directive which creates a symlink
4879 but (unlike "L") deletes a pre-existing file first, should
4880 it already exist and not already be the correct
4881 symlink. Similarly, "b+", "c+" and "p+" directives have been
4882 added as well, which create block and character devices, as
4883 well as fifos in the filesystem, possibly removing any
4884 pre-existing files of different types.
4886 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4887 'argument' field (which so far specified the source to
4888 symlink/copy the files from) is now optional. If omitted the
4889 same file os copied from /usr/share/factory/ suffixed by the
4890 full destination path. This is useful for populating /etc
4891 with essential files, by copying them from vendor defaults
4892 shipped in /usr/share/factory/etc.
4894 * A new command "systemctl preset-all" has been added that
4895 applies the service preset settings to all installed unit
4896 files. A new switch --preset-mode= has been added that
4897 controls whether only enable or only disable operations
4900 * A new command "systemctl is-system-running" has been added
4901 that allows checking the overall state of the system, for
4902 example whether it is fully up and running.
4904 * When the system boots up with an empty /etc, the equivalent
4905 to "systemctl preset-all" is executed during early boot, to
4906 make sure all default services are enabled after a factory
4909 * systemd now contains a minimal preset file that enables the
4910 most basic services systemd ships by default.
4912 * Unit files' [Install] section gained a new DefaultInstance=
4913 field for defining the default instance to create if a
4914 template unit is enabled with no instance specified.
4916 * A new passive target cryptsetup-pre.target has been added
4917 that may be used by services that need to make they run and
4918 finish before the first LUKS cryptographic device is set up.
4920 * The /dev/loop-control and /dev/btrfs-control device nodes
4921 are now owned by the "disk" group by default, opening up
4922 access to this group.
4924 * systemd-coredump will now automatically generate a
4925 stack trace of all core dumps taking place on the system,
4926 based on elfutils' libdw library. This stack trace is logged
4929 * systemd-coredump may now optionally store coredumps directly
4930 on disk (in /var/lib/systemd/coredump, possibly compressed),
4931 instead of storing them unconditionally in the journal. This
4932 mode is the new default. A new configuration file
4933 /etc/systemd/coredump.conf has been added to configure this
4934 and other parameters of systemd-coredump.
4936 * coredumpctl gained a new "info" verb to show details about a
4937 specific coredump. A new switch "-1" has also been added
4938 that makes sure to only show information about the most
4939 recent entry instead of all entries. Also, as the tool is
4940 generally useful now the "systemd-" prefix of the binary
4941 name has been removed. Distributions that want to maintain
4942 compatibility with the old name should add a symlink from
4943 the old name to the new name.
4945 * journald's SplitMode= now defaults to "uid". This makes sure
4946 that unprivileged users can access their own coredumps with
4947 coredumpctl without restrictions.
4949 * New kernel command line options "systemd.wants=" (for
4950 pulling an additional unit during boot), "systemd.mask="
4951 (for masking a specific unit for the boot), and
4952 "systemd.debug-shell" (for enabling the debug shell on tty9)
4953 have been added. This is implemented in the new generator
4954 "systemd-debug-generator".
4956 * systemd-nspawn will now by default filter a couple of
4957 syscalls for containers, among them those required for
4958 kernel module loading, direct x86 IO port access, swap
4959 management, and kexec. Most importantly though
4960 open_by_handle_at() is now prohibited for containers,
4961 closing a hole similar to a recently discussed vulnerability
4962 in docker regarding access to files on file hierarchies the
4963 container should normally not have access to. Note that, for
4964 nspawn, we generally make no security claims anyway (and
4965 this is explicitly documented in the man page), so this is
4966 just a fix for one of the most obvious problems.
4968 * A new man page file-hierarchy(7) has been added that
4969 contains a minimized, modernized version of the file system
4970 layout systemd expects, similar in style to the FHS
4971 specification or hier(5). A new tool systemd-path(1) has
4972 been added to query many of these paths for the local
4975 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4976 longer done. Since the directory now has a per-user size
4977 limit, and is cleaned on logout this appears unnecessary,
4978 in particular since this now brings the lifecycle of this
4979 directory closer in line with how IPC objects are handled.
4981 * systemd.pc now exports a number of additional directories,
4982 including $libdir (which is useful to identify the library
4983 path for the primary architecture of the system), and a
4984 couple of drop-in directories.
4986 * udev's predictable network interface names now use the dev_port
4987 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4988 distinguish between ports of the same PCI function. dev_id should
4989 only be used for ports using the same HW address, hence the need
4992 * machined has been updated to export the OS version of a
4993 container (read from /etc/os-release and
4994 /usr/lib/os-release) on the bus. This is now shown in
4995 "machinectl status" for a machine.
4997 * A new service setting RestartForceExitStatus= has been
4998 added. If configured to a set of exit signals or process
4999 return values, the service will be restarted when the main
5000 daemon process exits with any of them, regardless of the
5003 * systemctl's -H switch for connecting to remote systemd
5004 machines has been extended so that it may be used to
5005 directly connect to a specific container on the
5006 host. "systemctl -H root@foobar:waldi" will now connect as
5007 user "root" to host "foobar", and then proceed directly to
5008 the container named "waldi". Note that currently you have to
5009 authenticate as user "root" for this to work, as entering
5010 containers is a privileged operation.
5012 Contributions from: Andreas Henriksson, Benjamin Steinwender,
5013 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
5014 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5015 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5016 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5017 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5018 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5019 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5020 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5021 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5022 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5023 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5025 — Berlin, 2014-07-03
5029 * As an experimental feature, udev now tries to lock the
5030 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5031 executes events for the disk or any of its partitions.
5032 Applications like partitioning programs can lock the
5033 disk device node (flock(LOCK_EX)) and claim temporary
5034 device ownership that way; udev will entirely skip all event
5035 handling for this disk and its partitions. If the disk
5036 was opened for writing, the close will trigger a partition
5037 table rescan in udev's "watch" facility, and if needed
5038 synthesize "change" events for the disk and all its partitions.
5039 This is now unconditionally enabled, and if it turns out to
5040 cause major problems, we might turn it on only for specific
5041 devices, or might need to disable it entirely. Device Mapper
5042 devices are excluded from this logic.
5044 * We temporarily dropped the "-l" switch for fsck invocations,
5045 since they collide with the flock() logic above. util-linux
5046 upstream has been changed already to avoid this conflict,
5047 and we will readd "-l" as soon as util-linux with this
5048 change has been released.
5050 * The dependency on libattr has been removed. Since a long
5051 time, the extended attribute calls have moved to glibc, and
5052 libattr is thus unnecessary.
5054 * Virtualization detection works without privileges now. This
5055 means the systemd-detect-virt binary no longer requires
5056 CAP_SYS_PTRACE file capabilities, and our daemons can run
5057 with fewer privileges.
5059 * systemd-networkd now runs under its own "systemd-network"
5060 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5061 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5062 loses the ability to write to files owned by root this way.
5064 * Similarly, systemd-resolved now runs under its own
5065 "systemd-resolve" user with no capabilities remaining.
5067 * Similarly, systemd-bus-proxyd now runs under its own
5068 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5070 * systemd-networkd gained support for setting up "veth"
5071 virtual Ethernet devices for container connectivity, as well
5072 as GRE and VTI tunnels.
5074 * systemd-networkd will no longer automatically attempt to
5075 manually load kernel modules necessary for certain tunnel
5076 transports. Instead, it is assumed the kernel loads them
5077 automatically when required. This only works correctly on
5078 very new kernels. On older kernels, please consider adding
5079 the kernel modules to /etc/modules-load.d/ as a work-around.
5081 * The resolv.conf file systemd-resolved generates has been
5082 moved to /run/systemd/resolve/. If you have a symlink from
5083 /etc/resolv.conf, it might be necessary to correct it.
5085 * Two new service settings, ProtectHome= and ProtectSystem=,
5086 have been added. When enabled, they will make the user data
5087 (such as /home) inaccessible or read-only and the system
5088 (such as /usr) read-only, for specific services. This allows
5089 very light-weight per-service sandboxing to avoid
5090 modifications of user data or system files from
5091 services. These two new switches have been enabled for all
5092 of systemd's long-running services, where appropriate.
5094 * Socket units gained new SocketUser= and SocketGroup=
5095 settings to set the owner user and group of AF_UNIX sockets
5096 and FIFOs in the file system.
5098 * Socket units gained a new RemoveOnStop= setting. If enabled,
5099 all FIFOS and sockets in the file system will be removed
5100 when the specific socket unit is stopped.
5102 * Socket units gained a new Symlinks= setting. It takes a list
5103 of symlinks to create to file system sockets or FIFOs
5104 created by the specific Unix sockets. This is useful to
5105 manage symlinks to socket nodes with the same life-cycle as
5108 * The /dev/log socket and /dev/initctl FIFO have been moved to
5109 /run, and have been replaced by symlinks. This allows
5110 connecting to these facilities even if PrivateDevices=yes is
5111 used for a service (which makes /dev/log itself unavailable,
5112 but /run is left). This also has the benefit of ensuring
5113 that /dev only contains device nodes, directories and
5114 symlinks, and nothing else.
5116 * sd-daemon gained two new calls sd_pid_notify() and
5117 sd_pid_notifyf(). They are similar to sd_notify() and
5118 sd_notifyf(), but allow overriding of the source PID of
5119 notification messages if permissions permit this. This is
5120 useful to send notify messages on behalf of a different
5121 process (for example, the parent process). The
5122 systemd-notify tool has been updated to make use of this
5123 when sending messages (so that notification messages now
5124 originate from the shell script invoking systemd-notify and
5125 not the systemd-notify process itself. This should minimize
5126 a race where systemd fails to associate notification
5127 messages to services when the originating process already
5130 * A new "on-abnormal" setting for Restart= has been added. If
5131 set, it will result in automatic restarts on all "abnormal"
5132 reasons for a process to exit, which includes unclean
5133 signals, core dumps, timeouts and watchdog timeouts, but
5134 does not include clean and unclean exit codes or clean
5135 signals. Restart=on-abnormal is an alternative for
5136 Restart=on-failure for services that shall be able to
5137 terminate and avoid restarts on certain errors, by
5138 indicating so with an unclean exit code. Restart=on-failure
5139 or Restart=on-abnormal is now the recommended setting for
5140 all long-running services.
5142 * If the InaccessibleDirectories= service setting points to a
5143 mount point (or if there are any submounts contained within
5144 it), it is now attempted to completely unmount it, to make
5145 the file systems truly unavailable for the respective
5148 * The ReadOnlyDirectories= service setting and
5149 systemd-nspawn's --read-only parameter are now recursively
5150 applied to all submounts, too.
5152 * Mount units may now be created transiently via the bus APIs.
5154 * The support for SysV and LSB init scripts has been removed
5155 from the systemd daemon itself. Instead, it is now
5156 implemented as a generator that creates native systemd units
5157 from these scripts when needed. This enables us to remove a
5158 substantial amount of legacy code from PID 1, following the
5159 fact that many distributions only ship a very small number
5160 of LSB/SysV init scripts nowadays.
5162 * Privileged Xen (dom0) domains are not considered
5163 virtualization anymore by the virtualization detection
5164 logic. After all, they generally have unrestricted access to
5165 the hardware and usually are used to manage the unprivileged
5168 * systemd-tmpfiles gained a new "C" line type, for copying
5169 files or entire directories.
5171 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
5172 lines. So far, they have been non-globbing versions of the
5173 latter, and have thus been redundant. In future, it is
5174 recommended to only use "z". "m" has hence been removed
5175 from the documentation, even though it stays supported.
5177 * A tmpfiles snippet to recreate the most basic structure in
5178 /var has been added. This is enough to create the /var/run →
5179 /run symlink and create a couple of structural
5180 directories. This allows systems to boot up with an empty or
5181 volatile /var. Of course, while with this change, the core OS
5182 now is capable with dealing with a volatile /var, not all
5183 user services are ready for it. However, we hope that sooner
5184 or later, many service daemons will be changed upstream so
5185 that they are able to automatically create their necessary
5186 directories in /var at boot, should they be missing. This is
5187 the first step to allow state-less systems that only require
5188 the vendor image for /usr to boot.
5190 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5191 empty tmpfs instance to a specific directory. This is
5192 particularly useful for making use of the automatic
5193 reconstruction of /var (see above), by passing --tmpfs=/var.
5195 * Access modes specified in tmpfiles snippets may now be
5196 prefixed with "~", which indicates that they shall be masked
5197 by whether the existing file or directory is currently
5198 writable, readable or executable at all. Also, if specified,
5199 the sgid/suid/sticky bits will be masked for all
5202 * A new passive target unit "network-pre.target" has been
5203 added which is useful for services that shall run before any
5204 network is configured, for example firewall scripts.
5206 * The "floppy" group that previously owned the /dev/fd*
5207 devices is no longer used. The "disk" group is now used
5208 instead. Distributions should probably deprecate usage of
5211 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5212 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5213 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5214 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5215 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5216 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5219 — Berlin, 2014-06-11
5223 * A new "systemd-timesyncd" daemon has been added for
5224 synchronizing the system clock across the network. It
5225 implements an SNTP client. In contrast to NTP
5226 implementations such as chrony or the NTP reference server,
5227 this only implements a client side, and does not bother with
5228 the full NTP complexity, focusing only on querying time from
5229 one remote server and synchronizing the local clock to
5230 it. Unless you intend to serve NTP to networked clients or
5231 want to connect to local hardware clocks, this simple NTP
5232 client should be more than appropriate for most
5233 installations. The daemon runs with minimal privileges, and
5234 has been hooked up with networkd to only operate when
5235 network connectivity is available. The daemon saves the
5236 current clock to disk every time a new NTP sync has been
5237 acquired, and uses this to possibly correct the system clock
5238 early at bootup, in order to accommodate for systems that
5239 lack an RTC such as the Raspberry Pi and embedded devices,
5240 and to make sure that time monotonically progresses on these
5241 systems, even if it is not always correct. To make use of
5242 this daemon, a new system user and group "systemd-timesync"
5243 needs to be created on installation of systemd.
5245 * The queue "seqnum" interface of libudev has been disabled, as
5246 it was generally incompatible with device namespacing as
5247 sequence numbers of devices go "missing" if the devices are
5248 part of a different namespace.
5250 * "systemctl list-timers" and "systemctl list-sockets" gained
5251 a --recursive switch for showing units of these types also
5252 for all local containers, similar in style to the already
5253 supported --recursive switch for "systemctl list-units".
5255 * A new RebootArgument= setting has been added for service
5256 units, which may be used to specify a kernel reboot argument
5257 to use when triggering reboots with StartLimitAction=.
5259 * A new FailureAction= setting has been added for service
5260 units which may be used to specify an operation to trigger
5261 when a service fails. This works similarly to
5262 StartLimitAction=, but unlike it, controls what is done
5263 immediately rather than only after several attempts to
5264 restart the service in question.
5266 * hostnamed got updated to also expose the kernel name,
5267 release, and version on the bus. This is useful for
5268 executing commands like hostnamectl with the -H switch.
5269 systemd-analyze makes use of this to properly display
5270 details when running non-locally.
5272 * The bootchart tool can now show cgroup information in the
5273 graphs it generates.
5275 * The CFS CPU quota cgroup attribute is now exposed for
5276 services. The new CPUQuota= switch has been added for this
5277 which takes a percentage value. Setting this will have the
5278 result that a service may never get more CPU time than the
5279 specified percentage, even if the machine is otherwise idle.
5281 * systemd-networkd learned IPIP and SIT tunnel support.
5283 * LSB init scripts exposing a dependency on $network will now
5284 get a dependency on network-online.target rather than simply
5285 network.target. This should bring LSB handling closer to
5286 what it was on SysV systems.
5288 * A new fsck.repair= kernel option has been added to control
5289 how fsck shall deal with unclean file systems at boot.
5291 * The (.ini) configuration file parser will now silently
5292 ignore sections whose name begins with "X-". This may be
5293 used to maintain application-specific extension sections in unit
5296 * machined gained a new API to query the IP addresses of
5297 registered containers. "machinectl status" has been updated
5298 to show these addresses in its output.
5300 * A new call sd_uid_get_display() has been added to the
5301 sd-login APIs for querying the "primary" session of a
5302 user. The "primary" session of the user is elected from the
5303 user's sessions and generally a graphical session is
5304 preferred over a text one.
5306 * A minimal systemd-resolved daemon has been added. It
5307 currently simply acts as a companion to systemd-networkd and
5308 manages resolv.conf based on per-interface DNS
5309 configuration, possibly supplied via DHCP. In the long run
5310 we hope to extend this into a local DNSSEC enabled DNS and
5313 * The systemd-networkd-wait-online tool is now enabled by
5314 default. It will delay network-online.target until a network
5315 connection has been configured. The tool primarily integrates
5316 with networkd, but will also make a best effort to make sense
5317 of network configuration performed in some other way.
5319 * Two new service options StartupCPUShares= and
5320 StartupBlockIOWeight= have been added that work similarly to
5321 CPUShares= and BlockIOWeight= however only apply during
5322 system startup. This is useful to prioritize certain services
5323 differently during bootup than during normal runtime.
5325 * hostnamed has been changed to prefer the statically
5326 configured hostname in /etc/hostname (unless set to
5327 'localhost' or empty) over any dynamic one supplied by
5328 dhcp. With this change, the rules for picking the hostname
5329 match more closely the rules of other configuration settings
5330 where the local administrator's configuration in /etc always
5331 overrides any other settings.
5333 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
5334 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5335 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5336 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5337 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5338 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5339 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5340 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5341 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
5342 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5343 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5344 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5345 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5346 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5347 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5348 Lindskog, WaLyong Cho, Will Woods, Zbigniew
5351 — Beijing, 2014-05-28
5355 * When restoring the screen brightness at boot, stay away from
5356 the darkest setting or from the lowest 5% of the available
5357 range, depending on which is the larger value of both. This
5358 should effectively protect the user from rebooting into a
5359 black screen, should the brightness have been set to minimum
5362 * sd-login gained a new sd_machine_get_class() call to
5363 determine the class ("vm" or "container") of a machine
5364 registered with machined.
5366 * sd-login gained new calls
5367 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5368 to query the identity of the peer of a local AF_UNIX
5369 connection. They operate similarly to their sd_pid_get_xyz()
5372 * PID 1 will now maintain a system-wide system state engine
5373 with the states "starting", "running", "degraded",
5374 "maintenance", "stopping". These states are bound to system
5375 startup, normal runtime, runtime with at least one failed
5376 service, rescue/emergency mode and system shutdown. This
5377 state is shown in the "systemctl status" output when no unit
5378 name is passed. It is useful to determine system state, in
5379 particularly when doing so for many systems or containers at
5382 * A new command "list-machines" has been added to "systemctl"
5383 that lists all local OS containers and shows their system
5384 state (see above), if systemd runs inside of them.
5386 * systemctl gained a new "-r" switch to recursively enumerate
5387 units on all local containers, when used with the
5388 "list-unit" command (which is the default one that is
5389 executed when no parameters are specified).
5391 * The GPT automatic partition discovery logic will now honour
5392 two GPT partition flags: one may be set on a partition to
5393 cause it to be mounted read-only, and the other may be set
5394 on a partition to ignore it during automatic discovery.
5396 * Two new GPT type UUIDs have been added for automatic root
5397 partition discovery, for 32-bit and 64-bit ARM. This is not
5398 particularly useful for discovering the root directory on
5399 these architectures during bare-metal boots (since UEFI is
5400 not common there), but still very useful to allow booting of
5401 ARM disk images in nspawn with the -i option.
5403 * MAC addresses of interfaces created with nspawn's
5404 --network-interface= switch will now be generated from the
5405 machine name, and thus be stable between multiple invocations
5408 * logind will now automatically remove all IPC objects owned
5409 by a user if she or he fully logs out. This makes sure that
5410 users who are logged out cannot continue to consume IPC
5411 resources. This covers SysV memory, semaphores and message
5412 queues as well as POSIX shared memory and message
5413 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5414 limits. With this functionality, that is corrected. This may
5415 be turned off by using the RemoveIPC= switch of logind.conf.
5417 * The systemd-machine-id-setup and tmpfiles tools gained a
5418 --root= switch to operate on a specific root directory,
5421 * journald can now forward logged messages to the TTYs of all
5422 logged in users ("wall"). This is the default for all
5423 emergency messages now.
5425 * A new tool systemd-journal-remote has been added to stream
5426 journal log messages across the network.
5428 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5429 controller trees are mounted into it. Note that the
5430 directories mounted beneath it are not read-only. This is a
5431 security measure and is particularly useful because glibc
5432 actually includes a search logic to pick any tmpfs it can
5433 find to implement shm_open() if /dev/shm is not available
5434 (which it might very well be in namespaced setups).
5436 * machinectl gained a new "poweroff" command to cleanly power
5437 down a local OS container.
5439 * The PrivateDevices= unit file setting will now also drop the
5440 CAP_MKNOD capability from the capability bound set, and
5441 imply DevicePolicy=closed.
5443 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5444 comprehensively on all long-running systemd services where
5445 this is appropriate.
5447 * systemd-udevd will now run in a disassociated mount
5448 namespace. To mount directories from udev rules, make sure to
5449 pull in mount units via SYSTEMD_WANTS properties.
5451 * The kdbus support gained support for uploading policy into
5452 the kernel. sd-bus gained support for creating "monitoring"
5453 connections that can eavesdrop into all bus communication
5454 for debugging purposes.
5456 * Timestamps may now be specified in seconds since the UNIX
5457 epoch Jan 1st, 1970 by specifying "@" followed by the value
5460 * Native tcpwrap support in systemd has been removed. tcpwrap
5461 is old code, not really maintained anymore and has serious
5462 shortcomings, and better options such as firewalls
5463 exist. For setups that require tcpwrap usage, please
5464 consider invoking your socket-activated service via tcpd,
5465 like on traditional inetd.
5467 * A new system.conf configuration option
5468 DefaultTimerAccuracySec= has been added that controls the
5469 default AccuracySec= setting of .timer units.
5471 * Timer units gained a new WakeSystem= switch. If enabled,
5472 timers configured this way will cause the system to resume
5473 from system suspend (if the system supports that, which most
5476 * Timer units gained a new Persistent= switch. If enabled,
5477 timers configured this way will save to disk when they have
5478 been last triggered. This information is then used on next
5479 reboot to possible execute overdue timer events, that
5480 could not take place because the system was powered off.
5481 This enables simple anacron-like behaviour for timer units.
5483 * systemctl's "list-timers" will now also list the time a
5484 timer unit was last triggered in addition to the next time
5485 it will be triggered.
5487 * systemd-networkd will now assign predictable IPv4LL
5488 addresses to its local interfaces.
5490 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5491 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5492 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5493 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5494 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5495 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5496 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5497 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5500 — Berlin, 2014-03-25
5504 * A new unit file setting RestrictAddressFamilies= has been
5505 added to restrict which socket address families unit
5506 processes gain access to. This takes address family names
5507 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5508 attack surface of services via exotic protocol stacks. This
5509 is built on seccomp system call filters.
5511 * Two new unit file settings RuntimeDirectory= and
5512 RuntimeDirectoryMode= have been added that may be used to
5513 manage a per-daemon runtime directories below /run. This is
5514 an alternative for setting up directory permissions with
5515 tmpfiles snippets, and has the advantage that the runtime
5516 directory's lifetime is bound to the daemon runtime and that
5517 the daemon starts up with an empty directory each time. This
5518 is particularly useful when writing services that drop
5519 privileges using the User= or Group= setting.
5521 * The DeviceAllow= unit setting now supports globbing for
5522 matching against device group names.
5524 * The systemd configuration file system.conf gained new
5525 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5526 DefaultMemoryAccounting= to globally turn on/off accounting
5527 for specific resources (cgroups) for all units. These
5528 settings may still be overridden individually in each unit
5531 * systemd-gpt-auto-generator is now able to discover /srv and
5532 root partitions in addition to /home and swap partitions. It
5533 also supports LUKS-encrypted partitions now. With this in
5534 place, automatic discovery of partitions to mount following
5535 the Discoverable Partitions Specification
5536 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
5537 is now a lot more complete. This allows booting without
5538 /etc/fstab and without root= on the kernel command line on
5539 systems prepared appropriately.
5541 * systemd-nspawn gained a new --image= switch which allows
5542 booting up disk images and Linux installations on any block
5543 device that follow the Discoverable Partitions Specification
5544 (see above). This means that installations made with
5545 appropriately updated installers may now be started and
5546 deployed using container managers, completely
5547 unmodified. (We hope that libvirt-lxc will add support for
5548 this feature soon, too.)
5550 * systemd-nspawn gained a new --network-macvlan= setting to
5551 set up a private macvlan interface for the
5552 container. Similarly, systemd-networkd gained a new
5553 Kind=macvlan setting in .netdev files.
5555 * systemd-networkd now supports configuring local addresses
5558 * A new tool systemd-network-wait-online has been added to
5559 synchronously wait for network connectivity using
5562 * The sd-bus.h bus API gained a new sd_bus_track object for
5563 tracking the life-cycle of bus peers. Note that sd-bus.h is
5564 still not a public API though (unless you specify
5565 --enable-kdbus on the configure command line, which however
5566 voids your warranty and you get no API stability guarantee).
5568 * The $XDG_RUNTIME_DIR runtime directories for each user are
5569 now individual tmpfs instances, which has the benefit of
5570 introducing separate pools for each user, with individual
5571 size limits, and thus making sure that unprivileged clients
5572 can no longer negatively impact the system or other users by
5573 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5574 RuntimeDirectorySize= has been introduced that allows
5575 controlling the default size limit for all users. It
5576 defaults to 10% of the available physical memory. This is no
5577 replacement for quotas on tmpfs though (which the kernel
5578 still does not support), as /dev/shm and /tmp are still
5579 shared resources used by both the system and unprivileged
5582 * logind will now automatically turn off automatic suspending
5583 on laptop lid close when more than one display is
5584 connected. This was previously expected to be implemented
5585 individually in desktop environments (such as GNOME),
5586 however has been added to logind now, in order to fix a
5587 boot-time race where a desktop environment might not have
5588 been started yet and thus not been able to take an inhibitor
5589 lock at the time where logind already suspends the system
5590 due to a closed lid.
5592 * logind will now wait at least 30s after each system
5593 suspend/resume cycle, and 3min after system boot before
5594 suspending the system due to a closed laptop lid. This
5595 should give USB docking stations and similar enough time to
5596 be probed and configured after system resume and boot in
5597 order to then act as suspend blocker.
5599 * systemd-run gained a new --property= setting which allows
5600 initialization of resource control properties (and others)
5601 for the created scope or service unit. Example: "systemd-run
5602 --property=BlockIOWeight=10 updatedb" may be used to run
5603 updatedb at a low block IO scheduling weight.
5605 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5606 now also work in --scope mode.
5608 * When systemd is compiled with kdbus support, basic support
5609 for enforced policies is now in place. (Note that enabling
5610 kdbus still voids your warranty and no API compatibility
5613 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5614 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5615 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5616 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5617 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5618 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5619 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5620 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5621 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5622 Zbigniew Jędrzejewski-Szmek
5624 — Berlin, 2014-03-12
5628 * systemd will now relabel /dev after loading the SMACK policy
5629 according to SMACK rules.
5631 * A new unit file option AppArmorProfile= has been added to
5632 set the AppArmor profile for the processes of a unit.
5634 * A new condition check ConditionArchitecture= has been added
5635 to conditionalize units based on the system architecture, as
5636 reported by uname()'s "machine" field.
5638 * systemd-networkd now supports matching on the system
5639 virtualization, architecture, kernel command line, host name
5642 * logind is now a lot more aggressive when suspending the
5643 machine due to a closed laptop lid. Instead of acting only
5644 on the lid close action, it will continuously watch the lid
5645 status and act on it. This is useful for laptops where the
5646 power button is on the outside of the chassis so that it can
5647 be reached without opening the lid (such as the Lenovo
5648 Yoga). On those machines, logind will now immediately
5649 re-suspend the machine if the power button has been
5650 accidentally pressed while the laptop was suspended and in a
5651 backpack or similar.
5653 * logind will now watch SW_DOCK switches and inhibit reaction
5654 to the lid switch if it is pressed. This means that logind
5655 will not suspend the machine anymore if the lid is closed
5656 and the system is docked, if the laptop supports SW_DOCK
5657 notifications via the input layer. Note that ACPI docking
5658 stations do not generate this currently. Also note that this
5659 logic is usually not fully sufficient and Desktop
5660 Environments should take a lid switch inhibitor lock when an
5661 external display is connected, as systemd will not watch
5664 * nspawn will now make use of the devices cgroup controller by
5665 default, and only permit creation of and access to the usual
5666 API device nodes like /dev/null or /dev/random, as well as
5667 access to (but not creation of) the pty devices.
5669 * We will now ship a default .network file for
5670 systemd-networkd that automatically configures DHCP for
5671 network interfaces created by nspawn's --network-veth or
5672 --network-bridge= switches.
5674 * systemd will now understand the usual M, K, G, T suffixes
5675 according to SI conventions (i.e. to the base 1000) when
5676 referring to throughput and hardware metrics. It will stay
5677 with IEC conventions (i.e. to the base 1024) for software
5678 metrics, according to what is customary according to
5679 Wikipedia. We explicitly document which base applies for
5680 each configuration option.
5682 * The DeviceAllow= setting in unit files now supports a syntax
5683 to whitelist an entire group of devices node majors at once,
5684 based on the /proc/devices listing. For example, with the
5685 string "char-pts", it is now possible to whitelist all
5686 current and future pseudo-TTYs at once.
5688 * sd-event learned a new "post" event source. Event sources of
5689 this type are triggered by the dispatching of any event
5690 source of a type that is not "post". This is useful for
5691 implementing clean-up and check event sources that are
5692 triggered by other work being done in the program.
5694 * systemd-networkd is no longer statically enabled, but uses
5695 the usual [Install] sections so that it can be
5696 enabled/disabled using systemctl. It still is enabled by
5699 * When creating a veth interface pair with systemd-nspawn, the
5700 host side will now be prefixed with "vb-" if
5701 --network-bridge= is used, and with "ve-" if --network-veth
5702 is used. This way, it is easy to distinguish these cases on
5703 the host, for example to apply different configuration to
5704 them with systemd-networkd.
5706 * The compatibility libraries for libsystemd-journal.so,
5707 libsystem-id128.so, libsystemd-login.so and
5708 libsystemd-daemon.so do not make use of IFUNC
5709 anymore. Instead, we now build libsystemd.so multiple times
5710 under these alternative names. This means that the footprint
5711 is drastically increased, but given that these are
5712 transitional compatibility libraries, this should not matter
5713 much. This change has been made necessary to support the ARM
5714 platform for these compatibility libraries, as the ARM
5715 toolchain is not really at the same level as the toolchain
5716 for other architectures like x86 and does not support
5717 IFUNC. Please make sure to use --enable-compat-libs only
5718 during a transitional period!
5720 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
5721 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5722 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5723 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5724 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5725 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5726 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5727 Zbigniew Jędrzejewski-Szmek
5729 — Berlin, 2014-02-24
5733 * A new component "systemd-networkd" has been added that can
5734 be used to configure local network interfaces statically or
5735 via DHCP. It is capable of bringing up bridges, VLANs, and
5736 bonding. Currently, no hook-ups for interactive network
5737 configuration are provided. Use this for your initrd,
5738 container, embedded, or server setup if you need a simple,
5739 yet powerful, network configuration solution. This
5740 configuration subsystem is quite nifty, as it allows wildcard
5741 hotplug matching in interfaces. For example, with a single
5742 configuration snippet, you can configure that all Ethernet
5743 interfaces showing up are automatically added to a bridge,
5744 or similar. It supports link-sensing and more.
5746 * A new tool "systemd-socket-proxyd" has been added which can
5747 act as a bidirectional proxy for TCP sockets. This is
5748 useful for adding socket activation support to services that
5749 do not actually support socket activation, including virtual
5750 machines and the like.
5752 * Add a new tool to save/restore rfkill state on
5755 * Save/restore state of keyboard backlights in addition to
5756 display backlights on shutdown/boot.
5758 * udev learned a new SECLABEL{} construct to label device
5759 nodes with a specific security label when they appear. For
5760 now, only SECLABEL{selinux} is supported, but the syntax is
5761 prepared for additional security frameworks.
5763 * udev gained a new scheme to configure link-level attributes
5764 from files in /etc/systemd/network/*.link. These files can
5765 match against MAC address, device path, driver name and type,
5766 and will apply attributes like the naming policy, link speed,
5767 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
5768 address assignment policy (randomized, ...).
5770 * The configuration of network interface naming rules for
5771 "permanent interface names" has changed: a new NamePolicy=
5772 setting in the [Link] section of .link files determines the
5773 priority of possible naming schemes (onboard, slot, MAC,
5774 path). The default value of this setting is determined by
5775 /usr/lib/net/links/99-default.link. Old
5776 80-net-name-slot.rules udev configuration file has been
5777 removed, so local configuration overriding this file should
5778 be adapted to override 99-default.link instead.
5780 * When the User= switch is used in a unit file, also
5781 initialize $SHELL= based on the user database entry.
5783 * systemd no longer depends on libdbus. All communication is
5784 now done with sd-bus, systemd's low-level bus library
5787 * kdbus support has been added to PID 1 itself. When kdbus is
5788 enabled, this causes PID 1 to set up the system bus and
5789 enable support for a new ".busname" unit type that
5790 encapsulates bus name activation on kdbus. It works a little
5791 bit like ".socket" units, except for bus names. A new
5792 generator has been added that converts classic dbus1 service
5793 activation files automatically into native systemd .busname
5796 * sd-bus: add a light-weight vtable implementation that allows
5797 defining objects on the bus with a simple static const
5798 vtable array of its methods, signals and properties.
5800 * systemd will not generate or install static dbus
5801 introspection data anymore to /usr/share/dbus-1/interfaces,
5802 as the precise format of these files is unclear, and
5803 nothing makes use of it.
5805 * A proxy daemon is now provided to proxy clients connecting
5806 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5807 compatibility with classic D-Bus.
5809 * A bus driver implementation has been added that supports the
5810 classic D-Bus bus driver calls on kdbus, also for
5811 compatibility purposes.
5813 * A new API "sd-event.h" has been added that implements a
5814 minimal event loop API built around epoll. It provides a
5815 couple of features that direct epoll usage is lacking:
5816 prioritization of events, scales to large numbers of timer
5817 events, per-event timer slack (accuracy), system-wide
5818 coalescing of timer events, exit handlers, watchdog
5819 supervision support using systemd's sd_notify() API, child
5822 * A new API "sd-rntl.h" has been added that provides an API
5823 around the route netlink interface of the kernel, similar in
5824 style to "sd-bus.h".
5826 * A new API "sd-dhcp-client.h" has been added that provides a
5827 small DHCPv4 client-side implementation. This is used by
5830 * There is a new kernel command line option
5831 "systemd.restore_state=0|1". When set to "0", none of the
5832 systemd tools will restore saved runtime state to hardware
5833 devices. More specifically, the rfkill and backlight states
5836 * The FsckPassNo= compatibility option in mount/service units
5837 has been removed. The fstab generator will now add the
5838 necessary dependencies automatically, and does not require
5839 PID1's support for that anymore.
5841 * journalctl gained a new switch, --list-boots, that lists
5842 recent boots with their times and boot IDs.
5844 * The various tools like systemctl, loginctl, timedatectl,
5845 busctl, systemd-run, ... have gained a new switch "-M" to
5846 connect to a specific, local OS container (as direct
5847 connection, without requiring SSH). This works on any
5848 container that is registered with machined, such as those
5849 created by libvirt-lxc or nspawn.
5851 * systemd-run and systemd-analyze also gained support for "-H"
5852 to connect to remote hosts via SSH. This is particularly
5853 useful for systemd-run because it enables queuing of jobs
5854 onto remote systems.
5856 * machinectl gained a new command "login" to open a getty
5857 login in any local container. This works with any container
5858 that is registered with machined (such as those created by
5859 libvirt-lxc or nspawn), and which runs systemd inside.
5861 * machinectl gained a new "reboot" command that may be used to
5862 trigger a reboot on a specific container that is registered
5863 with machined. This works on any container that runs an init
5864 system of some kind.
5866 * systemctl gained a new "list-timers" command to print a nice
5867 listing of installed timer units with the times they elapse
5870 * Alternative reboot() parameters may now be specified on the
5871 "systemctl reboot" command line and are passed to the
5872 reboot() system call.
5874 * systemctl gained a new --job-mode= switch to configure the
5875 mode to queue a job with. This is a more generic version of
5876 --fail, --irreversible, and --ignore-dependencies, which are
5877 still available but not advertised anymore.
5879 * /etc/systemd/system.conf gained new settings to configure
5880 various default timeouts of units, as well as the default
5881 start limit interval and burst. These may still be overridden
5884 * PID1 will now export on the bus profile data of the security
5885 policy upload process (such as the SELinux policy upload to
5888 * journald: when forwarding logs to the console, include
5889 timestamps (following the setting in
5890 /sys/module/printk/parameters/time).
5892 * OnCalendar= in timer units now understands the special
5893 strings "yearly" and "annually". (Both are equivalent)
5895 * The accuracy of timer units is now configurable with the new
5896 AccuracySec= setting. It defaults to 1min.
5898 * A new dependency type JoinsNamespaceOf= has been added that
5899 allows running two services within the same /tmp and network
5900 namespace, if PrivateNetwork= or PrivateTmp= are used.
5902 * A new command "cat" has been added to systemctl. It outputs
5903 the original unit file of a unit, and concatenates the
5904 contents of additional "drop-in" unit file snippets, so that
5905 the full configuration is shown.
5907 * systemctl now supports globbing on the various "list-xyz"
5908 commands, like "list-units" or "list-sockets", as well as on
5909 those commands which take multiple unit names.
5911 * journalctl's --unit= switch gained support for globbing.
5913 * All systemd daemons now make use of the watchdog logic so
5914 that systemd automatically notices when they hang.
5916 * If the $container_ttys environment variable is set,
5917 getty-generator will automatically spawn a getty for each
5918 listed tty. This is useful for container managers to request
5919 login gettys to be spawned on as many ttys as needed.
5921 * %h, %s, %U specifier support is not available anymore when
5922 used in unit files for PID 1. This is because NSS calls are
5923 not safe from PID 1. They stay available for --user
5924 instances of systemd, and as special case for the root user.
5926 * loginctl gained a new "--no-legend" switch to turn off output
5929 * The "sd-login.h" API gained three new calls:
5930 sd_session_is_remote(), sd_session_get_remote_user(),
5931 sd_session_get_remote_host() to query information about
5934 * The udev hardware database now also carries vendor/product
5935 information of SDIO devices.
5937 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5938 determine whether watchdog notifications are requested by
5941 * Socket-activated per-connection services now include a
5942 short description of the connection parameters in the
5945 * tmpfiles gained a new "--boot" option. When this is not used,
5946 only lines where the command character is not suffixed with
5947 "!" are executed. When this option is specified, those
5948 options are executed too. This partitions tmpfiles
5949 directives into those that can be safely executed at any
5950 time, and those which should be run only at boot (for
5951 example, a line that creates /run/nologin).
5953 * A new API "sd-resolve.h" has been added which provides a simple
5954 asynchronous wrapper around glibc NSS host name resolution
5955 calls, such as getaddrinfo(). In contrast to glibc's
5956 getaddrinfo_a(), it does not use signals. In contrast to most
5957 other asynchronous name resolution libraries, this one does
5958 not reimplement DNS, but reuses NSS, so that alternate
5959 host name resolution systems continue to work, such as mDNS,
5960 LDAP, etc. This API is based on libasyncns, but it has been
5961 cleaned up for inclusion in systemd.
5963 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5964 "sd-daemon.h" are no longer found in individual libraries
5965 libsystemd-journal.so, libsystemd-login.so,
5966 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
5967 merged them into a single library, libsystemd.so, which
5968 provides all symbols. The reason for this is cyclic
5969 dependencies, as these libraries tend to use each other's
5970 symbols. So far, we have managed to workaround that by linking
5971 a copy of a good part of our code into each of these
5972 libraries again and again, which, however, makes certain
5973 things hard to do, like sharing static variables. Also, it
5974 substantially increases footprint. With this change, there
5975 is only one library for the basic APIs systemd
5976 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5977 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5978 library as well, however are subject to the --enable-kdbus
5979 switch (see below). Note that "sd-dhcp-client.h" is not part
5980 of this library (this is because it only consumes, never
5981 provides, services of/to other APIs). To make the transition
5982 easy from the separate libraries to the unified one, we
5983 provide the --enable-compat-libs compile-time switch which
5984 will generate stub libraries that are compatible with the
5985 old ones but redirect all calls to the new one.
5987 * All of the kdbus logic and the new APIs "sd-bus.h",
5988 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
5989 and "sd-utf8.h" are compile-time optional via the
5990 "--enable-kdbus" switch, and they are not compiled in by
5991 default. To make use of kdbus, you have to explicitly enable
5992 the switch. Note however, that neither the kernel nor the
5993 userspace API for all of this is considered stable yet. We
5994 want to maintain the freedom to still change the APIs for
5995 now. By specifying this build-time switch, you acknowledge
5996 that you are aware of the instability of the current
5999 * Also, note that while kdbus is pretty much complete,
6000 it lacks one thing: proper policy support. This means you
6001 can build a fully working system with all features; however,
6002 it will be highly insecure. Policy support will be added in
6003 one of the next releases, at the same time that we will
6004 declare the APIs stable.
6006 * When the kernel command line argument "kdbus" is specified,
6007 systemd will automatically load the kdbus.ko kernel module. At
6008 this stage of development, it is only useful for testing kdbus
6009 and should not be used in production. Note: if "--enable-kdbus"
6010 is specified, and the kdbus.ko kernel module is available, and
6011 "kdbus" is added to the kernel command line, the entire system
6012 runs with kdbus instead of dbus-daemon, with the above mentioned
6013 problem of missing the system policy enforcement. Also a future
6014 version of kdbus.ko or a newer systemd will not be compatible with
6015 each other, and will unlikely be able to boot the machine if only
6016 one of them is updated.
6018 * systemctl gained a new "import-environment" command which
6019 uploads the caller's environment (or parts thereof) into the
6020 service manager so that it is inherited by services started
6021 by the manager. This is useful to upload variables like
6022 $DISPLAY into the user service manager.
6024 * A new PrivateDevices= switch has been added to service units
6025 which allows running a service with a namespaced /dev
6026 directory that does not contain any device nodes for
6027 physical devices. More specifically, it only includes devices
6028 such as /dev/null, /dev/urandom, and /dev/zero which are API
6031 * logind has been extended to support behaviour like VT
6032 switching on seats that do not support a VT. This makes
6033 multi-session available on seats that are not the first seat
6034 (seat0), and on systems where kernel support for VTs has
6035 been disabled at compile-time.
6037 * If a process holds a delay lock for system sleep or shutdown
6038 and fails to release it in time, we will now log its
6039 identity. This makes it easier to identify processes that
6040 cause slow suspends or power-offs.
6042 * When parsing /etc/crypttab, support for a new key-slot=
6043 option as supported by Debian is added. It allows indicating
6044 which LUKS slot to use on disk, speeding up key loading.
6046 * The sd_journal_sendv() API call has been checked and
6047 officially declared to be async-signal-safe so that it may
6048 be invoked from signal handlers for logging purposes.
6050 * Boot-time status output is now enabled automatically after a
6051 short timeout if boot does not progress, in order to give
6052 the user an indication what she or he is waiting for.
6054 * The boot-time output has been improved to show how much time
6055 remains until jobs expire.
6057 * The KillMode= switch in service units gained a new possible
6058 value "mixed". If set, and the unit is shut down, then the
6059 initial SIGTERM signal is sent only to the main daemon
6060 process, while the following SIGKILL signal is sent to
6061 all remaining processes of the service.
6063 * When a scope unit is registered, a new property "Controller"
6064 may be set. If set to a valid bus name, systemd will send a
6065 RequestStop() signal to this name when it would like to shut
6066 down the scope. This may be used to hook manager logic into
6067 the shutdown logic of scope units. Also, scope units may now
6068 be put in a special "abandoned" state, in which case the
6069 manager process which created them takes no further
6070 responsibilities for it.
6072 * When reading unit files, systemd will now verify
6073 the access mode of these files, and warn about certain
6074 suspicious combinations. This has been added to make it
6075 easier to track down packaging bugs where unit files are
6076 marked executable or world-writable.
6078 * systemd-nspawn gained a new "--setenv=" switch to set
6079 container-wide environment variables. The similar option in
6080 systemd-activate was renamed from "--environment=" to
6081 "--setenv=" for consistency.
6083 * systemd-nspawn has been updated to create a new kdbus domain
6084 for each container that is invoked, thus allowing each
6085 container to have its own set of system and user buses,
6086 independent of the host.
6088 * systemd-nspawn gained a new --drop-capability= switch to run
6089 the container with less capabilities than the default. Both
6090 --drop-capability= and --capability= now take the special
6091 string "all" for dropping or keeping all capabilities.
6093 * systemd-nspawn gained new switches for executing containers
6094 with specific SELinux labels set.
6096 * systemd-nspawn gained a new --quiet switch to not generate
6097 any additional output but the container's own console
6100 * systemd-nspawn gained a new --share-system switch to run a
6101 container without PID namespacing enabled.
6103 * systemd-nspawn gained a new --register= switch to control
6104 whether the container is registered with systemd-machined or
6105 not. This is useful for containers that do not run full
6106 OS images, but only specific apps.
6108 * systemd-nspawn gained a new --keep-unit which may be used
6109 when invoked as the only program from a service unit, and
6110 results in registration of the unit service itself in
6111 systemd-machined, instead of a newly opened scope unit.
6113 * systemd-nspawn gained a new --network-interface= switch for
6114 moving arbitrary interfaces to the container. The new
6115 --network-veth switch creates a virtual Ethernet connection
6116 between host and container. The new --network-bridge=
6117 switch then allows assigning the host side of this virtual
6118 Ethernet connection to a bridge device.
6120 * systemd-nspawn gained a new --personality= switch for
6121 setting the kernel personality for the container. This is
6122 useful when running a 32-bit container on a 64-bit host. A
6123 similar option Personality= is now also available for service
6126 * logind will now also track a "Desktop" identifier for each
6127 session which encodes the desktop environment of it. This is
6128 useful for desktop environments that want to identify
6129 multiple running sessions of itself easily.
6131 * A new SELinuxContext= setting for service units has been
6132 added that allows setting a specific SELinux execution
6133 context for a service.
6135 * Most systemd client tools will now honour $SYSTEMD_LESS for
6136 settings of the "less" pager. By default, these tools will
6137 override $LESS to allow certain operations to work, such as
6138 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6139 influence this logic.
6141 * systemd's "seccomp" hook-up has been changed to make use of
6142 the libseccomp library instead of using its own
6143 implementation. This has benefits for portability among
6146 * For usage together with SystemCallFilter=, a new
6147 SystemCallErrorNumber= setting has been introduced that
6148 allows configuration of a system error number to be returned
6149 on filtered system calls, instead of immediately killing the
6150 process. Also, SystemCallArchitectures= has been added to
6151 limit access to system calls of a particular architecture
6152 (in order to turn off support for unused secondary
6153 architectures). There is also a global
6154 SystemCallArchitectures= setting in system.conf now to turn
6155 off support for non-native system calls system-wide.
6157 * systemd requires a kernel with a working name_to_handle_at(),
6158 please see the kernel config requirements in the README file.
6160 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6161 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6162 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6163 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6164 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6165 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6166 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6167 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6168 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6169 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6170 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6171 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6172 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6173 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6174 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6175 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6176 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6177 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6178 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6179 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6180 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6181 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6182 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6183 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6185 — Berlin, 2014-02-20
6189 * logind has gained support for facilitating privileged input
6190 and drm device access for unprivileged clients. This work is
6191 useful to allow Wayland display servers (and similar
6192 programs, such as kmscon) to run under the user's ID and
6193 access input and drm devices which are normally
6194 protected. When this is used (and the kernel is new enough)
6195 logind will "mute" IO on the file descriptors passed to
6196 Wayland as long as it is in the background and "unmute" it
6197 if it returns into the foreground. This allows secure
6198 session switching without allowing background sessions to
6199 eavesdrop on input and display data. This also introduces
6200 session switching support if VT support is turned off in the
6201 kernel, and on seats that are not seat0.
6203 * A new kernel command line option luks.options= is understood
6204 now which allows specifying LUKS options for usage for LUKS
6205 encrypted partitions specified with luks.uuid=.
6207 * tmpfiles.d(5) snippets may now use specifier expansion in
6208 path names. More specifically %m, %b, %H, %v, are now
6209 replaced by the local machine id, boot id, hostname, and
6210 kernel version number.
6212 * A new tmpfiles.d(5) command "m" has been introduced which
6213 may be used to change the owner/group/access mode of a file
6214 or directory if it exists, but do nothing if it does not.
6216 * This release removes high-level support for the
6217 MemorySoftLimit= cgroup setting. The underlying kernel
6218 cgroup attribute memory.soft_limit= is currently badly
6219 designed and likely to be removed from the kernel API in its
6220 current form, hence we should not expose it for now.
6222 * The memory.use_hierarchy cgroup attribute is now enabled for
6223 all cgroups systemd creates in the memory cgroup
6224 hierarchy. This option is likely to be come the built-in
6225 default in the kernel anyway, and the non-hierarchical mode
6226 never made much sense in the intrinsically hierarchical
6229 * A new field _SYSTEMD_SLICE= is logged along with all journal
6230 messages containing the slice a message was generated
6231 from. This is useful to allow easy per-customer filtering of
6232 logs among other things.
6234 * systemd-journald will no longer adjust the group of journal
6235 files it creates to the "systemd-journal" group. Instead we
6236 rely on the journal directory to be owned by the
6237 "systemd-journal" group, and its setgid bit set, so that the
6238 kernel file system layer will automatically enforce that
6239 journal files inherit this group assignment. The reason for
6240 this change is that we cannot allow NSS look-ups from
6241 journald which would be necessary to resolve
6242 "systemd-journal" to a numeric GID, because this might
6243 create deadlocks if NSS involves synchronous queries to
6244 other daemons (such as nscd, or sssd) which in turn are
6245 logging clients of journald and might block on it, which
6246 would then dead lock. A tmpfiles.d(5) snippet included in
6247 systemd will make sure the setgid bit and group are
6248 properly set on the journal directory if it exists on every
6249 boot. However, we recommend adjusting it manually after
6250 upgrades too (or from RPM scriptlets), so that the change is
6251 not delayed until next reboot.
6253 * Backlight and random seed files in /var/lib/ have moved into
6254 the /var/lib/systemd/ directory, in order to centralize all
6255 systemd generated files in one directory.
6257 * Boot time performance measurements (as displayed by
6258 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6259 performance information if that's available to determine how
6260 much time BIOS and boot loader initialization required. With
6261 a sufficiently new BIOS you hence no longer need to boot
6262 with Gummiboot to get access to such information.
6264 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6265 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6266 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6267 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6268 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6269 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6270 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6272 — Berlin, 2013-10-02
6276 * The Restart= option for services now understands a new
6277 on-watchdog setting, which will restart the service
6278 automatically if the service stops sending out watchdog keep
6279 alive messages (as configured with WatchdogSec=).
6281 * The getty generator (which is responsible for bringing up a
6282 getty on configured serial consoles) will no longer only
6283 start a getty on the primary kernel console but on all
6284 others, too. This makes the order in which console= is
6285 specified on the kernel command line less important.
6287 * libsystemd-logind gained a new sd_session_get_vt() call to
6288 retrieve the VT number of a session.
6290 * If the option "tries=0" is set for an entry of /etc/crypttab
6291 its passphrase is queried indefinitely instead of any
6292 maximum number of tries.
6294 * If a service with a configure PID file terminates its PID
6295 file will now be removed automatically if it still exists
6296 afterwards. This should put an end to stale PID files.
6298 * systemd-run will now also take relative binary path names
6299 for execution and no longer insists on absolute paths.
6301 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6302 paths that are optionally prefixed with "-" to indicate that
6303 it should not be considered a failure if they do not exist.
6305 * journalctl -o (and similar commands) now understands a new
6306 output mode "short-precise", it is similar to "short" but
6307 shows timestamps with usec accuracy.
6309 * The option "discard" (as known from Debian) is now
6310 synonymous to "allow-discards" in /etc/crypttab. In fact,
6311 "discard" is preferred now (since it is easier to remember
6314 * Some licensing clean-ups were made, so that more code is now
6315 LGPL-2.1 licensed than before.
6317 * A minimal tool to save/restore the display backlight
6318 brightness across reboots has been added. It will store the
6319 backlight setting as late as possible at shutdown, and
6320 restore it as early as possible during reboot.
6322 * A logic to automatically discover and enable home and swap
6323 partitions on GPT disks has been added. With this in place
6324 /etc/fstab becomes optional for many setups as systemd can
6325 discover certain partitions located on the root disk
6326 automatically. Home partitions are recognized under their
6327 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6328 partitions are recognized under their GPT type ID
6329 0657fd6da4ab43c484e50933c84b4f4f.
6331 * systemd will no longer pass any environment from the kernel
6332 or initrd to system services. If you want to set an
6333 environment for all services, do so via the kernel command
6334 line systemd.setenv= assignment.
6336 * The systemd-sysctl tool no longer natively reads the file
6337 /etc/sysctl.conf. If desired, the file should be symlinked
6338 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6339 legacy support by a symlink rather than built-in code, it
6340 also makes the otherwise hidden order of application of the
6341 different files visible. (Note that this partly reverts to a
6342 pre-198 application order of sysctl knobs!)
6344 * The "systemctl set-log-level" and "systemctl dump" commands
6345 have been moved to systemd-analyze.
6347 * systemd-run learned the new --remain-after-exit switch,
6348 which causes the scope unit not to be cleaned up
6349 automatically after the process terminated.
6351 * tmpfiles learned a new --exclude-prefix= switch to exclude
6352 certain paths from operation.
6354 * journald will now automatically flush all messages to disk
6355 as soon as a message at the log level CRIT, ALERT or EMERG
6358 Contributions from: Andrew Cook, Brandon Philips, Christian
6359 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6360 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6361 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6362 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6363 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6364 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6365 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6366 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6367 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6368 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6369 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6370 William Giokas, Zbigniew Jędrzejewski-Szmek
6372 — Berlin, 2013-09-13
6376 * The documentation has been updated to cover the various new
6377 concepts introduced with 205.
6379 * Unit files now understand the new %v specifier which
6380 resolves to the kernel version string as returned by "uname
6383 * systemctl now supports filtering the unit list output by
6384 load state, active state and sub state, using the new
6387 * "systemctl status" will now show the results of the
6388 condition checks (like ConditionPathExists= and similar) of
6389 the last start attempts of the unit. They are also logged to
6392 * "journalctl -b" may now be used to look for boot output of a
6393 specific boot. Try "journalctl -b -1" for the previous boot,
6394 but the syntax is substantially more powerful.
6396 * "journalctl --show-cursor" has been added which prints the
6397 cursor string the last shown log line. This may then be used
6398 with the new "journalctl --after-cursor=" switch to continue
6399 browsing logs from that point on.
6401 * "journalctl --force" may now be used to force regeneration
6404 * Creation of "dead" device nodes has been moved from udev
6405 into kmod and tmpfiles. Previously, udev would read the kmod
6406 databases to pre-generate dead device nodes based on meta
6407 information contained in kernel modules, so that these would
6408 be auto-loaded on access rather then at boot. As this
6409 does not really have much to do with the exposing actual
6410 kernel devices to userspace this has always been slightly
6411 alien in the udev codebase. Following the new scheme kmod
6412 will now generate a runtime snippet for tmpfiles from the
6413 module meta information and it now is tmpfiles' job to the
6414 create the nodes. This also allows overriding access and
6415 other parameters for the nodes using the usual tmpfiles
6416 facilities. As side effect this allows us to remove the
6417 CAP_SYS_MKNOD capability bit from udevd entirely.
6419 * logind's device ACLs may now be applied to these "dead"
6420 devices nodes too, thus finally allowing managed access to
6421 devices such as /dev/snd/sequencer without loading the
6422 backing module right-away.
6424 * A new RPM macro has been added that may be used to apply
6425 tmpfiles configuration during package installation.
6427 * systemd-detect-virt and ConditionVirtualization= now can
6428 detect User-Mode-Linux machines (UML).
6430 * journald will now implicitly log the effective capabilities
6431 set of processes in the message metadata.
6433 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6435 * The initrd interface has been simplified (more specifically,
6436 support for passing performance data via environment
6437 variables and fsck results via files in /run has been
6438 removed). These features were non-essential, and are
6439 nowadays available in a much nicer way by having systemd in
6440 the initrd serialize its state and have the hosts systemd
6441 deserialize it again.
6443 * The udev "keymap" data files and tools to apply keyboard
6444 specific mappings of scan to key codes, and force-release
6445 scan code lists have been entirely replaced by a udev
6446 "keyboard" builtin and a hwdb data file.
6448 * systemd will now honour the kernel's "quiet" command line
6449 argument also during late shutdown, resulting in a
6450 completely silent shutdown when used.
6452 * There's now an option to control the SO_REUSEPORT socket
6453 option in .socket units.
6455 * Instance units will now automatically get a per-template
6456 subslice of system.slice unless something else is explicitly
6457 configured. For example, instances of sshd@.service will now
6458 implicitly be placed in system-sshd.slice rather than
6459 system.slice as before.
6461 * Test coverage support may now be enabled at build time.
6463 Contributions from: Dave Reisner, Frederic Crozat, Harald
6464 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6465 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6466 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6467 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6468 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6469 Giokas, Zbigniew Jędrzejewski-Szmek
6471 — Berlin, 2013-07-23
6475 * Two new unit types have been introduced:
6477 Scope units are very similar to service units, however, are
6478 created out of pre-existing processes — instead of PID 1
6479 forking off the processes. By using scope units it is
6480 possible for system services and applications to group their
6481 own child processes (worker processes) in a powerful way
6482 which then maybe used to organize them, or kill them
6483 together, or apply resource limits on them.
6485 Slice units may be used to partition system resources in an
6486 hierarchical fashion and then assign other units to them. By
6487 default there are now three slices: system.slice (for all
6488 system services), user.slice (for all user sessions),
6489 machine.slice (for VMs and containers).
6491 Slices and scopes have been introduced primarily in
6492 context of the work to move cgroup handling to a
6493 single-writer scheme, where only PID 1
6494 creates/removes/manages cgroups.
6496 * There's a new concept of "transient" units. In contrast to
6497 normal units these units are created via an API at runtime,
6498 not from configuration from disk. More specifically this
6499 means it is now possible to run arbitrary programs as
6500 independent services, with all execution parameters passed
6501 in via bus APIs rather than read from disk. Transient units
6502 make systemd substantially more dynamic then it ever was,
6503 and useful as a general batch manager.
6505 * logind has been updated to make use of scope and slice units
6506 for managing user sessions. As a user logs in he will get
6507 his own private slice unit, to which all sessions are added
6508 as scope units. We also added support for automatically
6509 adding an instance of user@.service for the user into the
6510 slice. Effectively logind will no longer create cgroup
6511 hierarchies on its own now, it will defer entirely to PID 1
6512 for this by means of scope, service and slice units. Since
6513 user sessions this way become entities managed by PID 1
6514 the output of "systemctl" is now a lot more comprehensive.
6516 * A new mini-daemon "systemd-machined" has been added which
6517 may be used by virtualization managers to register local
6518 VMs/containers. nspawn has been updated accordingly, and
6519 libvirt will be updated shortly. machined will collect a bit
6520 of meta information about the VMs/containers, and assign
6521 them their own scope unit (see above). The collected
6522 meta-data is then made available via the "machinectl" tool,
6523 and exposed in "ps" and similar tools. machined/machinectl
6524 is compile-time optional.
6526 * As discussed earlier, the low-level cgroup configuration
6527 options ControlGroup=, ControlGroupModify=,
6528 ControlGroupPersistent=, ControlGroupAttribute= have been
6529 removed. Please use high-level attribute settings instead as
6530 well as slice units.
6532 * A new bus call SetUnitProperties() has been added to alter
6533 various runtime parameters of a unit. This is primarily
6534 useful to alter cgroup parameters dynamically in a nice way,
6535 but will be extended later on to make more properties
6536 modifiable at runtime. systemctl gained a new set-properties
6537 command that wraps this call.
6539 * A new tool "systemd-run" has been added which can be used to
6540 run arbitrary command lines as transient services or scopes,
6541 while configuring a number of settings via the command
6542 line. This tool is currently very basic, however already
6543 very useful. We plan to extend this tool to even allow
6544 queuing of execution jobs with time triggers from the
6545 command line, similar in fashion to "at".
6547 * nspawn will now inform the user explicitly that kernels with
6548 audit enabled break containers, and suggest the user to turn
6551 * Support for detecting the IMA and AppArmor security
6552 frameworks with ConditionSecurity= has been added.
6554 * journalctl gained a new "-k" switch for showing only kernel
6555 messages, mimicking dmesg output; in addition to "--user"
6556 and "--system" switches for showing only user's own logs
6559 * systemd-delta can now show information about drop-in
6560 snippets extending unit files.
6562 * libsystemd-bus has been substantially updated but is still
6563 not available as public API.
6565 * systemd will now look for the "debug" argument on the kernel
6566 command line and enable debug logging, similar to what
6567 "systemd.log_level=debug" already did before.
6569 * "systemctl set-default", "systemctl get-default" has been
6570 added to configure the default.target symlink, which
6571 controls what to boot into by default.
6573 * "systemctl set-log-level" has been added as a convenient
6574 way to raise and lower systemd logging threshold.
6576 * "systemd-analyze plot" will now show the time the various
6577 generators needed for execution, as well as information
6578 about the unit file loading.
6580 * libsystemd-journal gained a new sd_journal_open_files() call
6581 for opening specific journal files. journactl also gained a
6582 new switch to expose this new functionality. Previously we
6583 only supported opening all files from a directory, or all
6584 files from the system, as opening individual files only is
6585 racy due to journal file rotation.
6587 * systemd gained the new DefaultEnvironment= setting in
6588 /etc/systemd/system.conf to set environment variables for
6591 * If a privileged process logs a journal message with the
6592 OBJECT_PID= field set, then journald will automatically
6593 augment this with additional OBJECT_UID=, OBJECT_GID=,
6594 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6595 system services want to log events about specific client
6596 processes. journactl/systemctl has been updated to make use
6597 of this information if all log messages regarding a specific
6600 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6601 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6602 Reisner, David Coppa, David King, David Strauss, Eelco
6603 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6604 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6605 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6606 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6607 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6608 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6609 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6610 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6611 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6612 Łukasz Stelmach, 장동준
6616 * The Python bindings gained some minimal support for the APIs
6617 exposed by libsystemd-logind.
6619 * ConditionSecurity= gained support for detecting SMACK. Since
6620 this condition already supports SELinux and AppArmor we only
6621 miss IMA for this. Patches welcome!
6623 Contributions from: Karol Lewandowski, Lennart Poettering,
6624 Zbigniew Jędrzejewski-Szmek
6628 * systemd-nspawn will now create /etc/resolv.conf if
6629 necessary, before bind-mounting the host's file onto it.
6631 * systemd-nspawn will now store meta information about a
6632 container on the container's cgroup as extended attribute
6633 fields, including the root directory.
6635 * The cgroup hierarchy has been reworked in many ways. All
6636 objects any of the components systemd creates in the cgroup
6637 tree are now suffixed. More specifically, user sessions are
6638 now placed in cgroups suffixed with ".session", users in
6639 cgroups suffixed with ".user", and nspawn containers in
6640 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6641 names are now escaped in a simple scheme to avoid collision
6642 of userspace object names with kernel filenames. This work
6643 is preparation for making these objects relocatable in the
6644 cgroup tree, in order to allow easy resource partitioning of
6645 these objects without causing naming conflicts.
6647 * systemctl list-dependencies gained the new switches
6648 --plain, --reverse, --after and --before.
6650 * systemd-inhibit now shows the process name of processes that
6651 have taken an inhibitor lock.
6653 * nss-myhostname will now also resolve "localhost"
6654 implicitly. This makes /etc/hosts an optional file and
6655 nicely handles that on IPv6 ::1 maps to both "localhost" and
6658 * libsystemd-logind.so gained a new call
6659 sd_get_machine_names() to enumerate running containers and
6660 VMs (currently only supported by very new libvirt and
6661 nspawn). sd_login_monitor can now be used to watch
6662 VMs/containers coming and going.
6664 * .include is not allowed recursively anymore, and only in
6665 unit files. Usually it is better to use drop-in snippets in
6666 .d/*.conf anyway, as introduced with systemd 198.
6668 * systemd-analyze gained a new "critical-chain" command that
6669 determines the slowest chain of units run during system
6670 boot-up. It is very useful for tracking down where
6671 optimizing boot time is the most beneficial.
6673 * systemd will no longer allow manipulating service paths in
6674 the name=systemd:/system cgroup tree using ControlGroup= in
6675 units. (But is still fine with it in all other dirs.)
6677 * There's a new systemd-nspawn@.service service file that may
6678 be used to easily run nspawn containers as system
6679 services. With the container's root directory in
6680 /var/lib/container/foobar it is now sufficient to run
6681 "systemctl start systemd-nspawn@foobar.service" to boot it.
6683 * systemd-cgls gained a new parameter "--machine" to list only
6684 the processes within a certain container.
6686 * ConditionSecurity= now can check for "apparmor". We still
6687 are lacking checks for SMACK and IMA for this condition
6688 check though. Patches welcome!
6690 * A new configuration file /etc/systemd/sleep.conf has been
6691 added that may be used to configure which kernel operation
6692 systemd is supposed to execute when "suspend", "hibernate"
6693 or "hybrid-sleep" is requested. This makes the new kernel
6694 "freeze" state accessible to the user.
6696 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6697 the passed argument if applicable.
6699 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6700 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6701 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6702 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6703 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6704 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6705 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6710 * The output of 'systemctl list-jobs' got some polishing. The
6711 '--type=' argument may now be passed more than once. A new
6712 command 'systemctl list-sockets' has been added which shows
6713 a list of kernel sockets systemd is listening on with the
6714 socket units they belong to, plus the units these socket
6717 * The experimental libsystemd-bus library got substantial
6718 updates to work in conjunction with the (also experimental)
6719 kdbus kernel project. It works well enough to exchange
6720 messages with some sophistication. Note that kdbus is not
6721 ready yet, and the library is mostly an elaborate test case
6722 for now, and not installable.
6724 * systemd gained a new unit 'systemd-static-nodes.service'
6725 that generates static device nodes earlier during boot, and
6726 can run in conjunction with udev.
6728 * libsystemd-login gained a new call sd_pid_get_user_unit()
6729 to retrieve the user systemd unit a process is running
6730 in. This is useful for systems where systemd is used as
6733 * systemd-nspawn now places all containers in the new /machine
6734 top-level cgroup directory in the name=systemd
6735 hierarchy. libvirt will soon do the same, so that we get a
6736 uniform separation of /system, /user and /machine for system
6737 services, user processes and containers/virtual
6738 machines. This new cgroup hierarchy is also useful to stick
6739 stable names to specific container instances, which can be
6740 recognized later this way (this name may be controlled
6741 via systemd-nspawn's new -M switch). libsystemd-login also
6742 gained a new call sd_pid_get_machine_name() to retrieve the
6743 name of the container/VM a specific process belongs to.
6745 * bootchart can now store its data in the journal.
6747 * libsystemd-journal gained a new call
6748 sd_journal_add_conjunction() for AND expressions to the
6749 matching logic. This can be used to express more complex
6750 logical expressions.
6752 * journactl can now take multiple --unit= and --user-unit=
6755 * The cryptsetup logic now understands the "luks.key=" kernel
6756 command line switch for specifying a file to read the
6757 decryption key from. Also, if a configured key file is not
6758 found the tool will now automatically fall back to prompting
6761 * Python systemd.journal module was updated to wrap recently
6762 added functions from libsystemd-journal. The interface was
6763 changed to bring the low level interface in s.j._Reader
6764 closer to the C API, and the high level interface in
6765 s.j.Reader was updated to wrap and convert all data about
6768 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6769 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6770 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6771 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6772 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6773 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6777 * journalctl --update-catalog now understands a new --root=
6778 option to operate on catalogs found in a different root
6781 * During shutdown after systemd has terminated all running
6782 services a final killing loop kills all remaining left-over
6783 processes. We will now print the name of these processes
6784 when we send SIGKILL to them, since this usually indicates a
6787 * If /etc/crypttab refers to password files stored on
6788 configured mount points automatic dependencies will now be
6789 generated to ensure the specific mount is established first
6790 before the key file is attempted to be read.
6792 * 'systemctl status' will now show information about the
6793 network sockets a socket unit is listening on.
6795 * 'systemctl status' will also shown information about any
6796 drop-in configuration file for units. (Drop-In configuration
6797 files in this context are files such as
6798 /etc/systemd/systemd/foobar.service.d/*.conf)
6800 * systemd-cgtop now optionally shows summed up CPU times of
6801 cgroups. Press '%' while running cgtop to switch between
6802 percentage and absolute mode. This is useful to determine
6803 which cgroups use up the most CPU time over the entire
6804 runtime of the system. systemd-cgtop has also been updated
6805 to be 'pipeable' for processing with further shell tools.
6807 * 'hostnamectl set-hostname' will now allow setting of FQDN
6810 * The formatting and parsing of time span values has been
6811 changed. The parser now understands fractional expressions
6812 such as "5.5h". The formatter will now output fractional
6813 expressions for all time spans under 1min, i.e. "5.123456s"
6814 rather than "5s 123ms 456us". For time spans under 1s
6815 millisecond values are shown, for those under 1ms
6816 microsecond values are shown. This should greatly improve
6817 all time-related output of systemd.
6819 * libsystemd-login and libsystemd-journal gained new
6820 functions for querying the poll() events mask and poll()
6821 timeout value for integration into arbitrary event
6824 * localectl gained the ability to list available X11 keymaps
6825 (models, layouts, variants, options).
6827 * 'systemd-analyze dot' gained the ability to filter for
6828 specific units via shell-style globs, to create smaller,
6829 more useful graphs. I.e. it is now possible to create simple
6830 graphs of all the dependencies between only target units, or
6831 of all units that Avahi has dependencies with.
6833 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6834 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6835 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6836 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6837 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6838 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6839 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6843 * The boot-time readahead implementation for rotating media
6844 will now read the read-ahead data in multiple passes which
6845 consist of all read requests made in equidistant time
6846 intervals. This means instead of strictly reading read-ahead
6847 data in its physical order on disk we now try to find a
6848 middle ground between physical and access time order.
6850 * /etc/os-release files gained a new BUILD_ID= field for usage
6851 on operating systems that provide continuous builds of OS
6854 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6855 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6856 William Douglas, Zbigniew Jędrzejewski-Szmek
6860 * systemd-python gained an API exposing libsystemd-daemon.
6862 * The SMACK setup logic gained support for uploading CIPSO
6865 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6866 ReadOnlyDirectories= and InaccessibleDirectories= has
6867 changed. The private /tmp and /var/tmp directories are now
6868 shared by all processes of a service (which means
6869 ExecStartPre= may now leave data in /tmp that ExecStart= of
6870 the same service can still access). When a service is
6871 stopped its temporary directories are immediately deleted
6872 (normal clean-up with tmpfiles is still done in addition to
6875 * By default, systemd will now set a couple of sysctl
6876 variables in the kernel: the safe sysrq options are turned
6877 on, IP route verification is turned on, and source routing
6878 disabled. The recently added hardlink and softlink
6879 protection of the kernel is turned on. These settings should
6880 be reasonably safe, and good defaults for all new systems.
6882 * The predictable network naming logic may now be turned off
6883 with a new kernel command line switch: net.ifnames=0.
6885 * A new libsystemd-bus module has been added that implements a
6886 pretty complete D-Bus client library. For details see:
6888 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
6890 * journald will now explicitly flush the journal files to disk
6891 at the latest 5min after each write. The file will then also
6892 be marked offline until the next write. This should increase
6893 reliability in case of a crash. The synchronization delay
6894 can be configured via SyncIntervalSec= in journald.conf.
6896 * There's a new remote-fs-setup.target unit that can be used
6897 to pull in specific services when at least one remote file
6898 system is to be mounted.
6900 * There are new targets timers.target and paths.target as
6901 canonical targets to pull user timer and path units in
6902 from. This complements sockets.target with a similar
6903 purpose for socket units.
6905 * libudev gained a new call udev_device_set_attribute_value()
6906 to set sysfs attributes of a device.
6908 * The udev daemon now sets the default number of worker
6909 processes executed in parallel based on the number of available
6910 CPUs instead of the amount of available RAM. This is supposed
6911 to provide a more reliable default and limit a too aggressive
6912 parallelism for setups with 1000s of devices connected.
6914 Contributions from: Auke Kok, Colin Walters, Cristian
6915 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6916 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6917 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6918 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6919 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6920 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6921 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6922 Zbigniew Jędrzejewski-Szmek
6926 * Configuration of unit files may now be extended via drop-in
6927 files without having to edit/override the unit files
6928 themselves. More specifically, if the administrator wants to
6929 change one value for a service file foobar.service he can
6930 now do so by dropping in a configuration snippet into
6931 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
6932 will load all these snippets and apply them on top of the
6933 main unit configuration file, possibly extending or
6934 overriding its settings. Using these drop-in snippets is
6935 generally nicer than the two earlier options for changing
6936 unit files locally: copying the files from
6937 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6938 them there; or creating a new file in /etc/systemd/system/
6939 that incorporates the original one via ".include". Drop-in
6940 snippets into these .d/ directories can be placed in any
6941 directory systemd looks for units in, and the usual
6942 overriding semantics between /usr/lib, /etc and /run apply
6945 * Most unit file settings which take lists of items can now be
6946 reset by assigning the empty string to them. For example,
6947 normally, settings such as Environment=FOO=BAR append a new
6948 environment variable assignment to the environment block,
6949 each time they are used. By assigning Environment= the empty
6950 string the environment block can be reset to empty. This is
6951 particularly useful with the .d/*.conf drop-in snippets
6952 mentioned above, since this adds the ability to reset list
6953 settings from vendor unit files via these drop-ins.
6955 * systemctl gained a new "list-dependencies" command for
6956 listing the dependencies of a unit recursively.
6958 * Inhibitors are now honored and listed by "systemctl
6959 suspend", "systemctl poweroff" (and similar) too, not only
6960 GNOME. These commands will also list active sessions by
6963 * Resource limits (as exposed by the various control group
6964 controllers) can now be controlled dynamically at runtime
6965 for all units. More specifically, you can now use a command
6966 like "systemctl set-cgroup-attr foobar.service cpu.shares
6967 2000" to alter the CPU shares a specific service gets. These
6968 settings are stored persistently on disk, and thus allow the
6969 administrator to easily adjust the resource usage of
6970 services with a few simple commands. This dynamic resource
6971 management logic is also available to other programs via the
6972 bus. Almost any kernel cgroup attribute and controller is
6975 * systemd-vconsole-setup will now copy all font settings to
6976 all allocated VTs, where it previously applied them only to
6979 * libsystemd-login gained the new sd_session_get_tty() API
6982 * This release drops support for a few legacy or
6983 distribution-specific LSB facility names when parsing init
6984 scripts: $x-display-manager, $mail-transfer-agent,
6985 $mail-transport-agent, $mail-transfer-agent, $smtp,
6986 $null. Also, the mail-transfer-agent.target unit backing
6987 this has been removed. Distributions which want to retain
6988 compatibility with this should carry the burden for
6989 supporting this themselves and patch support for these back
6990 in, if they really need to. Also, the facilities $syslog and
6991 $local_fs are now ignored, since systemd does not support
6992 early-boot LSB init scripts anymore, and these facilities
6993 are implied anyway for normal services. syslog.target has
6996 * There are new bus calls on PID1's Manager object for
6997 cancelling jobs, and removing snapshot units. Previously,
6998 both calls were only available on the Job and Snapshot
7001 * systemd-journal-gatewayd gained SSL support.
7003 * The various "environment" files, such as /etc/locale.conf
7004 now support continuation lines with a backslash ("\") as
7005 last character in the line, similarly in style (but different)
7006 to how this is supported in shells.
7008 * For normal user processes the _SYSTEMD_USER_UNIT= field is
7009 now implicitly appended to every log entry logged. systemctl
7010 has been updated to filter by this field when operating on a
7011 user systemd instance.
7013 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
7014 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7015 the container. This makes it easier to boot unmodified
7016 Fedora systems in a container, which however still requires
7017 audit=0 to be passed on the kernel command line. Auditing in
7018 kernel and userspace is unfortunately still too broken in
7019 context of containers, hence we recommend compiling it out
7020 of the kernel or using audit=0. Hopefully this will be fixed
7021 one day for good in the kernel.
7023 * nspawn gained the new --bind= and --bind-ro= parameters to
7024 bind mount specific directories from the host into the
7027 * nspawn will now mount its own devpts file system instance
7028 into the container, in order not to leak pty devices from
7029 the host into the container.
7031 * systemd will now read the firmware boot time performance
7032 information from the EFI variables, if the used boot loader
7033 supports this, and takes it into account for boot performance
7034 analysis via "systemd-analyze". This is currently supported
7035 only in conjunction with Gummiboot, but could be supported
7036 by other boot loaders too. For details see:
7038 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
7040 * A new generator has been added that automatically mounts the
7041 EFI System Partition (ESP) to /boot, if that directory
7042 exists, is empty, and no other file system has been
7043 configured to be mounted there.
7045 * logind will now send out PrepareForSleep(false) out
7046 unconditionally, after coming back from suspend. This may be
7047 used by applications as asynchronous notification for
7048 system resume events.
7050 * "systemctl unlock-sessions" has been added, that allows
7051 unlocking the screens of all user sessions at once, similar
7052 to how "systemctl lock-sessions" already locked all users
7053 sessions. This is backed by a new D-Bus call UnlockSessions().
7055 * "loginctl seat-status" will now show the master device of a
7056 seat. (i.e. the device of a seat that needs to be around for
7057 the seat to be considered available, usually the graphics
7060 * tmpfiles gained a new "X" line type, that allows
7061 configuration of files and directories (with wildcards) that
7062 shall be excluded from automatic cleanup ("aging").
7064 * udev default rules set the device node permissions now only
7065 at "add" events, and do not change them any longer with a
7066 later "change" event.
7068 * The log messages for lid events and power/sleep keypresses
7069 now carry a message ID.
7071 * We now have a substantially larger unit test suite, but this
7072 continues to be work in progress.
7074 * udevadm hwdb gained a new --root= parameter to change the
7075 root directory to operate relative to.
7077 * logind will now issue a background sync() request to the kernel
7078 early at shutdown, so that dirty buffers are flushed to disk early
7079 instead of at the last moment, in order to optimize shutdown
7082 * A new bootctl tool has been added that is an interface for
7083 certain boot loader operations. This is currently a preview
7084 and is likely to be extended into a small mechanism daemon
7085 like timedated, localed, hostnamed, and can be used by
7086 graphical UIs to enumerate available boot options, and
7087 request boot into firmware operations.
7089 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7090 the rest of the package. It also has been updated to work
7091 correctly in initrds.
7093 * Policykit previously has been runtime optional, and is now
7094 also compile time optional via a configure switch.
7096 * systemd-analyze has been reimplemented in C. Also "systemctl
7097 dot" has moved into systemd-analyze.
7099 * "systemctl status" with no further parameters will now print
7100 the status of all active or failed units.
7102 * Operations such as "systemctl start" can now be executed
7103 with a new mode "--irreversible" which may be used to queue
7104 operations that cannot accidentally be reversed by a later
7105 job queuing. This is by default used to make shutdown
7106 requests more robust.
7108 * The Python API of systemd now gained a new module for
7109 reading journal files.
7111 * A new tool kernel-install has been added that can install
7112 kernel images according to the Boot Loader Specification:
7114 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7116 * Boot time console output has been improved to provide
7117 animated boot time output for hanging jobs.
7119 * A new tool systemd-activate has been added which can be used
7120 to test socket activation with, directly from the command
7121 line. This should make it much easier to test and debug
7122 socket activation in daemons.
7124 * journalctl gained a new "--reverse" (or -r) option to show
7125 journal output in reverse order (i.e. newest line first).
7127 * journalctl gained a new "--pager-end" (or -e) option to jump
7128 to immediately jump to the end of the journal in the
7129 pager. This is only supported in conjunction with "less".
7131 * journalctl gained a new "--user-unit=" option, that works
7132 similarly to "--unit=" but filters for user units rather than
7135 * A number of unit files to ease adoption of systemd in
7136 initrds has been added. This moves some minimal logic from
7137 the various initrd implementations into systemd proper.
7139 * The journal files are now owned by a new group
7140 "systemd-journal", which exists specifically to allow access
7141 to the journal, and nothing else. Previously, we used the
7142 "adm" group for that, which however possibly covers more
7143 than just journal/log file access. This new group is now
7144 already used by systemd-journal-gatewayd to ensure this
7145 daemon gets access to the journal files and as little else
7146 as possible. Note that "make install" will also set FS ACLs
7147 up for /var/log/journal to give "adm" and "wheel" read
7148 access to it, in addition to "systemd-journal" which owns
7149 the journal files. We recommend that packaging scripts also
7150 add read access to "adm" + "wheel" to /var/log/journal, and
7151 all existing/future journal files. To normal users and
7152 administrators little changes, however packagers need to
7153 ensure to create the "systemd-journal" system group at
7154 package installation time.
7156 * The systemd-journal-gatewayd now runs as unprivileged user
7157 systemd-journal-gateway:systemd-journal-gateway. Packaging
7158 scripts need to create these system user/group at
7161 * timedated now exposes a new boolean property CanNTP that
7162 indicates whether a local NTP service is available or not.
7164 * systemd-detect-virt will now also detect xen PVs
7166 * The pstore file system is now mounted by default, if it is
7169 * In addition to the SELinux and IMA policies we will now also
7170 load SMACK policies at early boot.
7172 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7173 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7174 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7175 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7176 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7177 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7178 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7179 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7180 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7181 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7182 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7183 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7184 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7185 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7189 * Timer units now support calendar time events in addition to
7190 monotonic time events. That means you can now trigger a unit
7191 based on a calendar time specification such as "Thu,Fri
7192 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7193 or fifth day of any month of the year 2013, given that it is
7194 a thursday or friday. This brings timer event support
7195 considerably closer to cron's capabilities. For details on
7196 the supported calendar time specification language see
7199 * udev now supports a number of different naming policies for
7200 network interfaces for predictable names, and a combination
7201 of these policies is now the default. Please see this wiki
7202 document for details:
7204 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
7206 * Auke Kok's bootchart implementation has been added to the
7207 systemd tree. It is an optional component that can graph the
7208 boot in quite some detail. It is one of the best bootchart
7209 implementations around and minimal in its code and
7212 * nss-myhostname has been integrated into the systemd source
7213 tree. nss-myhostname guarantees that the local hostname
7214 always stays resolvable via NSS. It has been a weak
7215 requirement of systemd-hostnamed since a long time, and
7216 since its code is actually trivial we decided to just
7217 include it in systemd's source tree. It can be turned off
7218 with a configure switch.
7220 * The read-ahead logic is now capable of properly detecting
7221 whether a btrfs file system is on SSD or rotating media, in
7222 order to optimize the read-ahead scheme. Previously, it was
7223 only capable of detecting this on traditional file systems
7226 * In udev, additional device properties are now read from the
7227 IAB in addition to the OUI database. Also, Bluetooth company
7228 identities are attached to the devices as well.
7230 * In service files %U may be used as specifier that is
7231 replaced by the configured user name of the service.
7233 * nspawn may now be invoked without a controlling TTY. This
7234 makes it suitable for invocation as its own service. This
7235 may be used to set up a simple containerized server system
7236 using only core OS tools.
7238 * systemd and nspawn can now accept socket file descriptors
7239 when they are started for socket activation. This enables
7240 implementation of socket activated nspawn
7241 containers. i.e. think about autospawning an entire OS image
7242 when the first SSH or HTTP connection is received. We expect
7243 that similar functionality will also be added to libvirt-lxc
7246 * journalctl will now suppress ANSI color codes when
7247 presenting log data.
7249 * systemctl will no longer show control group information for
7250 a unit if the control group is empty anyway.
7252 * logind can now automatically suspend/hibernate/shutdown the
7255 * /etc/machine-info and hostnamed now also expose the chassis
7256 type of the system. This can be used to determine whether
7257 the local system is a laptop, desktop, handset or
7258 tablet. This information may either be configured by the
7259 user/vendor or is automatically determined from ACPI and DMI
7260 information if possible.
7262 * A number of PolicyKit actions are now bound together with
7263 "imply" rules. This should simplify creating UIs because
7264 many actions will now authenticate similar ones as well.
7266 * Unit files learnt a new condition ConditionACPower= which
7267 may be used to conditionalize a unit depending on whether an
7268 AC power source is connected or not, of whether the system
7269 is running on battery power.
7271 * systemctl gained a new "is-failed" verb that may be used in
7272 shell scripts and suchlike to check whether a specific unit
7273 is in the "failed" state.
7275 * The EnvironmentFile= setting in unit files now supports file
7276 globbing, and can hence be used to easily read a number of
7277 environment files at once.
7279 * systemd will no longer detect and recognize specific
7280 distributions. All distribution-specific #ifdeffery has been
7281 removed, systemd is now fully generic and
7282 distribution-agnostic. Effectively, not too much is lost as
7283 a lot of the code is still accessible via explicit configure
7284 switches. However, support for some distribution specific
7285 legacy configuration file formats has been dropped. We
7286 recommend distributions to simply adopt the configuration
7287 files everybody else uses now and convert the old
7288 configuration from packaging scripts. Most distributions
7289 already did that. If that's not possible or desirable,
7290 distributions are welcome to forward port the specific
7291 pieces of code locally from the git history.
7293 * When logging a message about a unit systemd will now always
7294 log the unit name in the message meta data.
7296 * localectl will now also discover system locale data that is
7297 not stored in locale archives, but directly unpacked.
7299 * logind will no longer unconditionally use framebuffer
7300 devices as seat masters, i.e. as devices that are required
7301 to be existing before a seat is considered preset. Instead,
7302 it will now look for all devices that are tagged as
7303 "seat-master" in udev. By default, framebuffer devices will
7304 be marked as such, but depending on local systems, other
7305 devices might be marked as well. This may be used to
7306 integrate graphics cards using closed source drivers (such
7307 as NVidia ones) more nicely into logind. Note however, that
7308 we recommend using the open source NVidia drivers instead,
7309 and no udev rules for the closed-source drivers will be
7310 shipped from us upstream.
7312 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7313 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7314 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7315 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7316 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7317 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7318 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7319 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7320 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7321 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7322 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7323 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7328 * udev gained support for loading additional device properties
7329 from an indexed database that is keyed by vendor/product IDs
7330 and similar device identifiers. For the beginning this
7331 "hwdb" is populated with data from the well-known PCI and
7332 USB database, but also includes PNP, ACPI and OID data. In
7333 the longer run this indexed database shall grow into
7334 becoming the one central database for non-essential
7335 userspace device metadata. Previously, data from the PCI/USB
7336 database was only attached to select devices, since the
7337 lookup was a relatively expensive operation due to O(n) time
7338 complexity (with n being the number of entries in the
7339 database). Since this is now O(1), we decided to add in this
7340 data for all devices where this is available, by
7341 default. Note that the indexed database needs to be rebuilt
7342 when new data files are installed. To achieve this you need
7343 to update your packaging scripts to invoke "udevadm hwdb
7344 --update" after installation of hwdb data files. For
7345 RPM-based distributions we introduced the new
7346 %udev_hwdb_update macro for this purpose.
7348 * The Journal gained support for the "Message Catalog", an
7349 indexed database to link up additional information with
7350 journal entries. For further details please check:
7352 https://www.freedesktop.org/wiki/Software/systemd/catalog
7354 The indexed message catalog database also needs to be
7355 rebuilt after installation of message catalog files. Use
7356 "journalctl --update-catalog" for this. For RPM-based
7357 distributions we introduced the %journal_catalog_update
7358 macro for this purpose.
7360 * The Python Journal bindings gained support for the standard
7361 Python logging framework.
7363 * The Journal API gained new functions for checking whether
7364 the underlying file system of a journal file is capable of
7365 properly reporting file change notifications, or whether
7366 applications that want to reflect journal changes "live"
7367 need to recheck journal files continuously in appropriate
7370 * It is now possible to set the "age" field for tmpfiles
7371 entries to 0, indicating that files matching this entry
7372 shall always be removed when the directories are cleaned up.
7374 * coredumpctl gained a new "gdb" verb which invokes gdb
7375 right-away on the selected coredump.
7377 * There's now support for "hybrid sleep" on kernels that
7378 support this, in addition to "suspend" and "hibernate". Use
7379 "systemctl hybrid-sleep" to make use of this.
7381 * logind's HandleSuspendKey= setting (and related settings)
7382 now gained support for a new "lock" setting to simply
7383 request the screen lock on all local sessions, instead of
7384 actually executing a suspend or hibernation.
7386 * systemd will now mount the EFI variables file system by
7389 * Socket units now gained support for configuration of the
7390 SMACK security label.
7392 * timedatectl will now output the time of the last and next
7393 daylight saving change.
7395 * We dropped support for various legacy and distro-specific
7396 concepts, such as insserv, early-boot SysV services
7397 (i.e. those for non-standard runlevels such as 'b' or 'S')
7398 or ArchLinux /etc/rc.conf support. We recommend the
7399 distributions who still need support this to either continue
7400 to maintain the necessary patches downstream, or find a
7401 different solution. (Talk to us if you have questions!)
7403 * Various systemd components will now bypass PolicyKit checks
7404 for root and otherwise handle properly if PolicyKit is not
7405 found to be around. This should fix most issues for
7406 PolicyKit-less systems. Quite frankly this should have been
7407 this way since day one. It is absolutely our intention to
7408 make systemd work fine on PolicyKit-less systems, and we
7409 consider it a bug if something does not work as it should if
7410 PolicyKit is not around.
7412 * For embedded systems it is now possible to build udev and
7413 systemd without blkid and/or kmod support.
7415 * "systemctl switch-root" is now capable of switching root
7416 more than once. I.e. in addition to transitions from the
7417 initrd to the host OS it is now possible to transition to
7418 further OS images from the host. This is useful to implement
7419 offline updating tools.
7421 * Various other additions have been made to the RPM macros
7422 shipped with systemd. Use %udev_rules_update() after
7423 installing new udev rules files. %_udevhwdbdir,
7424 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7425 %_sysctldir are now available which resolve to the right
7426 directories for packages to place various data files in.
7428 * journalctl gained the new --full switch (in addition to
7429 --all, to disable ellipsation for long messages.
7431 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7432 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7433 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7434 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7435 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7436 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7437 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7438 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7439 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7443 * journalctl gained new --since= and --until= switches to
7444 filter by time. It also now supports nice filtering for
7445 units via --unit=/-u.
7447 * Type=oneshot services may use ExecReload= and do the
7450 * The journal daemon now supports time-based rotation and
7451 vacuuming, in addition to the usual disk-space based
7454 * The journal will now index the available field values for
7455 each field name. This enables clients to show pretty drop
7456 downs of available match values when filtering. The bash
7457 completion of journalctl has been updated
7458 accordingly. journalctl gained a new switch -F to list all
7459 values a certain field takes in the journal database.
7461 * More service events are now written as structured messages
7462 to the journal, and made recognizable via message IDs.
7464 * The timedated, localed and hostnamed mini-services which
7465 previously only provided support for changing time, locale
7466 and hostname settings from graphical DEs such as GNOME now
7467 also have a minimal (but very useful) text-based client
7468 utility each. This is probably the nicest way to changing
7469 these settings from the command line now, especially since
7470 it lists available options and is fully integrated with bash
7473 * There's now a new tool "systemd-coredumpctl" to list and
7474 extract coredumps from the journal.
7476 * We now install a README each in /var/log/ and
7477 /etc/rc.d/init.d explaining where the system logs and init
7478 scripts went. This hopefully should help folks who go to
7479 that dirs and look into the otherwise now empty void and
7480 scratch their heads.
7482 * When user-services are invoked (by systemd --user) the
7483 $MANAGERPID env var is set to the PID of systemd.
7485 * SIGRTMIN+24 when sent to a --user instance will now result
7486 in immediate termination of systemd.
7488 * gatewayd received numerous feature additions such as a
7489 "follow" mode, for live syncing and filtering.
7491 * browse.html now allows filtering and showing detailed
7492 information on specific entries. Keyboard navigation and
7493 mouse screen support has been added.
7495 * gatewayd/journalctl now supports HTML5/JSON
7496 Server-Sent-Events as output.
7498 * The SysV init script compatibility logic will now
7499 heuristically determine whether a script supports the
7500 "reload" verb, and only then make this available as
7503 * "systemctl status --follow" has been removed, use "journalctl
7506 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7507 have been removed since they are hardly useful to be
7510 * And I'd like to take the opportunity to specifically mention
7511 Zbigniew for his great contributions. Zbigniew, you rock!
7513 Contributions from: Andrew Eikum, Christian Hesse, Colin
7514 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
7515 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7516 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7517 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7518 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7519 Jędrzejewski-Szmek, Сковорода Никита Андреевич
7523 * If /etc/vconsole.conf is non-existent or empty we will no
7524 longer load any console font or key map at boot by
7525 default. Instead the kernel defaults will be left
7526 intact. This is definitely the right thing to do, as no
7527 configuration should mean no configuration, and hard-coding
7528 font names that are different on all archs is probably a bad
7529 idea. Also, the kernel default key map and font should be
7530 good enough for most cases anyway, and mostly identical to
7531 the userspace fonts/key maps we previously overloaded them
7532 with. If distributions want to continue to default to a
7533 non-kernel font or key map they should ship a default
7534 /etc/vconsole.conf with the appropriate contents.
7536 Contributions from: Colin Walters, Daniel J Walsh, Dave
7537 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7538 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7542 * journalctl gained a new --cursor= switch to show entries
7543 starting from the specified location in the journal.
7545 * We now enforce a size limit on journal entry fields exported
7546 with "-o json" in journalctl. Fields larger than 4K will be
7547 assigned null. This can be turned off with --all.
7549 * An (optional) journal gateway daemon is now available as
7550 "systemd-journal-gatewayd.service". This service provides
7551 access to the journal via HTTP and JSON. This functionality
7552 will be used to implement live log synchronization in both
7553 pull and push modes, but has various other users too, such
7554 as easy log access for debugging of embedded devices. Right
7555 now it is already useful to retrieve the journal via HTTP:
7557 # systemctl start systemd-journal-gatewayd.service
7558 # wget http://localhost:19531/entries
7560 This will download the journal contents in a
7561 /var/log/messages compatible format. The same as JSON:
7563 # curl -H"Accept: application/json" http://localhost:19531/entries
7565 This service is also accessible via a web browser where a
7566 single static HTML5 app is served that uses the JSON logic
7567 to enable the user to do some basic browsing of the
7568 journal. This will be extended later on. Here's an example
7569 screenshot of this app in its current state:
7571 http://0pointer.de/public/journal-gatewayd
7573 Contributions from: Kay Sievers, Lennart Poettering, Robert
7574 Milasan, Tom Gundersen
7578 * The bash completion logic is now available for journalctl
7581 * We do not mount the "cpuset" controller anymore together with
7582 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7583 started if no parameters are assigned to it. "cpuset" hence
7584 broke code that assumed it could create "cpu" groups and
7587 * journalctl -f will now subscribe to terminal size changes,
7588 and line break accordingly.
7590 Contributions from: Dave Reisner, Kay Sievers, Lennart
7591 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
7595 * nspawn will now create a symlink /etc/localtime in the
7596 container environment, copying the host's timezone
7597 setting. Previously this has been done via a bind mount, but
7598 since symlinks cannot be bind mounted this has now been
7599 changed to create/update the appropriate symlink.
7601 * journalctl -n's line number argument is now optional, and
7602 will default to 10 if omitted.
7604 * journald will now log the maximum size the journal files may
7605 take up on disk. This is particularly useful if the default
7606 built-in logic of determining this parameter from the file
7607 system size is used. Use "systemctl status
7608 systemd-journald.service" to see this information.
7610 * The multi-seat X wrapper tool has been stripped down. As X
7611 is now capable of enumerating graphics devices via udev in a
7612 seat-aware way the wrapper is not strictly necessary
7613 anymore. A stripped down temporary stop-gap is still shipped
7614 until the upstream display managers have been updated to
7615 fully support the new X logic. Expect this wrapper to be
7616 removed entirely in one of the next releases.
7618 * HandleSleepKey= in logind.conf has been split up into
7619 HandleSuspendKey= and HandleHibernateKey=. The old setting
7620 is not available anymore. X11 and the kernel are
7621 distinguishing between these keys and we should too. This
7622 also means the inhibition lock for these keys has been split
7625 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7626 Poettering, Lukas Nykryn, Václav Pavlín
7630 * Whenever a unit changes state we will now log this to the
7631 journal and show along the unit's own log output in
7634 * ConditionPathIsMountPoint= can now properly detect bind
7635 mount points too. (Previously, a bind mount of one file
7636 system to another place in the same file system could not be
7637 detected as mount, since they shared struct stat's st_dev
7640 * We will now mount the cgroup controllers cpu, cpuacct,
7641 cpuset and the controllers net_cls, net_prio together by
7644 * nspawn containers will now have a virtualized boot
7645 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7646 over with a randomized ID at container initialization). This
7647 has the effect of making "journalctl -b" do the right thing
7650 * The JSON output journal serialization has been updated not
7651 to generate "endless" list objects anymore, but rather one
7652 JSON object per line. This is more in line how most JSON
7653 parsers expect JSON objects. The new output mode
7654 "json-pretty" has been added to provide similar output, but
7655 neatly aligned for readability by humans.
7657 * We dropped all explicit sync() invocations in the shutdown
7658 code. The kernel does this implicitly anyway in the kernel
7659 reboot() syscall. halt(8)'s -n option is now a compatibility
7662 * We now support virtualized reboot() in containers, as
7663 supported by newer kernels. We will fall back to exit() if
7664 CAP_SYS_REBOOT is not available to the container. Also,
7665 nspawn makes use of this now and will actually reboot the
7666 container if the containerized OS asks for that.
7668 * journalctl will only show local log output by default
7669 now. Use --merge (-m) to show remote log output, too.
7671 * libsystemd-journal gained the new sd_journal_get_usage()
7672 call to determine the current disk usage of all journal
7673 files. This is exposed in the new "journalctl --disk-usage"
7676 * journald gained a new configuration setting SplitMode= in
7677 journald.conf which may be used to control how user journals
7678 are split off. See journald.conf(5) for details.
7680 * A new condition type ConditionFileNotEmpty= has been added.
7682 * tmpfiles' "w" lines now support file globbing, to write
7683 multiple files at once.
7685 * We added Python bindings for the journal submission
7686 APIs. More Python APIs for a number of selected APIs will
7687 likely follow. Note that we intend to add native bindings
7688 only for the Python language, as we consider it common
7689 enough to deserve bindings shipped within systemd. There are
7690 various projects outside of systemd that provide bindings
7691 for languages such as PHP or Lua.
7693 * Many conditions will now resolve specifiers such as %i. In
7694 addition, PathChanged= and related directives of .path units
7695 now support specifiers as well.
7697 * There's now a new RPM macro definition for the system preset
7700 * journald will now warn if it ca not forward a message to the
7701 syslog daemon because its socket is full.
7703 * timedated will no longer write or process /etc/timezone,
7704 except on Debian. As we do not support late mounted /usr
7705 anymore /etc/localtime always being a symlink is now safe,
7706 and hence the information in /etc/timezone is not necessary
7709 * logind will now always reserve one VT for a text getty (VT6
7710 by default). Previously if more than 6 X sessions where
7711 started they took up all the VTs with auto-spawned gettys,
7712 so that no text gettys were available anymore.
7714 * udev will now automatically inform the btrfs kernel logic
7715 about btrfs RAID components showing up. This should make
7716 simple hotplug based btrfs RAID assembly work.
7718 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7719 (but not for its children which will stay at the kernel
7720 default). This should allow setups with a lot more listening
7723 * systemd will now always pass the configured timezone to the
7724 kernel at boot. timedated will do the same when the timezone
7727 * logind's inhibition logic has been updated. By default,
7728 logind will now handle the lid switch, the power and sleep
7729 keys all the time, even in graphical sessions. If DEs want
7730 to handle these events on their own they should take the new
7731 handle-power-key, handle-sleep-key and handle-lid-switch
7732 inhibitors during their runtime. A simple way to achieve
7733 that is to invoke the DE wrapped in an invocation of:
7735 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7737 * Access to unit operations is now checked via SELinux taking
7738 the unit file label and client process label into account.
7740 * systemd will now notify the administrator in the journal
7741 when he over-mounts a non-empty directory.
7743 * There are new specifiers that are resolved in unit files,
7744 for the host name (%H), the machine ID (%m) and the boot ID
7747 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
7748 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7749 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7750 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7751 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7752 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7753 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7757 * Support for reading structured kernel messages from
7758 /dev/kmsg has now been added and is enabled by default.
7760 * Support for reading kernel messages from /proc/kmsg has now
7761 been removed. If you want kernel messages in the journal
7762 make sure to run a recent kernel (>= 3.5) that supports
7763 reading structured messages from /dev/kmsg (see
7764 above). /proc/kmsg is now exclusive property of classic
7765 syslog daemons again.
7767 * The libudev API gained the new
7768 udev_device_new_from_device_id() call.
7770 * The logic for file system namespace (ReadOnlyDirectory=,
7771 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7772 require pivot_root() anymore. This means fewer temporary
7773 directories are created below /tmp for this feature.
7775 * nspawn containers will now see and receive all submounts
7776 made on the host OS below the root file system of the
7779 * Forward Secure Sealing is now supported for Journal files,
7780 which provide cryptographical sealing of journal files so
7781 that attackers cannot alter log history anymore without this
7782 being detectable. Lennart will soon post a blog story about
7783 this explaining it in more detail.
7785 * There are two new service settings RestartPreventExitStatus=
7786 and SuccessExitStatus= which allow configuration of exit
7787 status (exit code or signal) which will be excepted from the
7788 restart logic, resp. consider successful.
7790 * journalctl gained the new --verify switch that can be used
7791 to check the integrity of the structure of journal files and
7792 (if Forward Secure Sealing is enabled) the contents of
7795 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7796 and similar symlinks pre-created. This makes running shells
7797 as container init process a lot more fun.
7799 * The fstab support can now handle PARTUUID= and PARTLABEL=
7802 * A new ConditionHost= condition has been added to match
7803 against the hostname (with globs) and machine ID. This is
7804 useful for clusters where a single OS image is used to
7805 provision a large number of hosts which shall run slightly
7806 different sets of services.
7808 * Services which hit the restart limit will now be placed in a
7811 Contributions from: Bertram Poettering, Dave Reisner, Huang
7812 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7813 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7817 * When running in --user mode systemd will now become a
7818 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7819 tree a lot more organized.
7821 * A new PartOf= unit dependency type has been introduced that
7822 may be used to group services in a natural way.
7824 * "systemctl enable" may now be used to enable instances of
7827 * journalctl now prints error log levels in red, and
7828 warning/notice log levels in bright white. It also supports
7829 filtering by log level now.
7831 * cgtop gained a new -n switch (similar to top), to configure
7832 the maximum number of iterations to run for. It also gained
7833 -b, to run in batch mode (accepting no input).
7835 * The suffix ".service" may now be omitted on most systemctl
7836 command lines involving service unit names.
7838 * There's a new bus call in logind to lock all sessions, as
7839 well as a loginctl verb for it "lock-sessions".
7841 * libsystemd-logind.so gained a new call sd_journal_perror()
7842 that works similar to libc perror() but logs to the journal
7843 and encodes structured information about the error number.
7845 * /etc/crypttab entries now understand the new keyfile-size=
7848 * shutdown(8) now can send a (configurable) wall message when
7849 a shutdown is cancelled.
7851 * The mount propagation mode for the root file system will now
7852 default to "shared", which is useful to make containers work
7853 nicely out-of-the-box so that they receive new mounts from
7854 the host. This can be undone locally by running "mount
7855 --make-rprivate /" if needed.
7857 * The prefdm.service file has been removed. Distributions
7858 should maintain this unit downstream if they intend to keep
7859 it around. However, we recommend writing normal unit files
7860 for display managers instead.
7862 * Since systemd is a crucial part of the OS we will now
7863 default to a number of compiler switches that improve
7864 security (hardening) such as read-only relocations, stack
7865 protection, and suchlike.
7867 * The TimeoutSec= setting for services is now split into
7868 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7869 of individual time outs for the start and the stop phase of
7872 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7873 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7874 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7875 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7876 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7877 Gundersen, Zbigniew Jędrzejewski-Szmek
7881 * The journal and id128 C APIs are now fully documented as man
7884 * Extra safety checks have been added when transitioning from
7885 the initial RAM disk to the main system to avoid accidental
7888 * /etc/crypttab entries now understand the new keyfile-offset=
7891 * systemctl -t can now be used to filter by unit load state.
7893 * The journal C API gained the new sd_journal_wait() call to
7894 make writing synchronous journal clients easier.
7896 * journalctl gained the new -D switch to show journals from a
7899 * journalctl now displays a special marker between log
7900 messages of two different boots.
7902 * The journal is now explicitly flushed to /var via a service
7903 systemd-journal-flush.service, rather than implicitly simply
7904 by seeing /var/log/journal to be writable.
7906 * journalctl (and the journal C APIs) can now match for much
7907 more complex expressions, with alternatives and
7910 * When transitioning from the initial RAM disk to the main
7911 system we will now kill all processes in a killing spree to
7912 ensure no processes stay around by accident.
7914 * Three new specifiers may be used in unit files: %u, %h, %s
7915 resolve to the user name, user home directory resp. user
7916 shell. This is useful for running systemd user instances.
7918 * We now automatically rotate journal files if their data
7919 object hash table gets a fill level > 75%. We also size the
7920 hash table based on the configured maximum file size. This
7921 together should lower hash collisions drastically and thus
7922 speed things up a bit.
7924 * journalctl gained the new "--header" switch to introspect
7925 header data of journal files.
7927 * A new setting SystemCallFilters= has been added to services
7928 which may be used to apply blacklists or whitelists to
7929 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7931 * nspawn gained a new --link-journal= switch (and quicker: -j)
7932 to link the container journal with the host. This makes it
7933 very easy to centralize log viewing on the host for all
7934 guests while still keeping the journal files separated.
7936 * Many bugfixes and optimizations
7938 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7939 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7940 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7945 * Several tools now understand kernel command line arguments,
7946 which are only read when run in an initial RAM disk. They
7947 usually follow closely their normal counterparts, but are
7950 * There's a new tool to analyze the readahead files that are
7951 automatically generated at boot. Use:
7953 /usr/lib/systemd/systemd-readahead analyze /.readahead
7955 * We now provide an early debug shell on tty9 if this enabled. Use:
7957 systemctl enable debug-shell.service
7959 * All plymouth related units have been moved into the Plymouth
7960 package. Please make sure to upgrade your Plymouth version
7963 * systemd-tmpfiles now supports getting passed the basename of
7964 a configuration file only, in which case it will look for it
7965 in all appropriate directories automatically.
7967 * udevadm info now takes a /dev or /sys path as argument, and
7968 does the right thing. Example:
7970 udevadm info /dev/sda
7971 udevadm info /sys/class/block/sda
7973 * systemctl now prints a warning if a unit is stopped but a
7974 unit that might trigger it continues to run. Example: a
7975 service is stopped but the socket that activates it is left
7978 * "systemctl status" will now mention if the log output was
7979 shortened due to rotation since a service has been started.
7981 * The journal API now exposes functions to determine the
7982 "cutoff" times due to rotation.
7984 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7985 immediately flushing of runtime logs to /var if possible,
7986 resp. for triggering immediate rotation of the journal
7989 * It is now considered an error if a service is attempted to
7990 be stopped that is not loaded.
7992 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7994 * systemd-analyze now supports Python 3
7996 * tmpfiles now supports cleaning up directories via aging
7997 where the first level dirs are always kept around but
7998 directories beneath it automatically aged. This is enabled
7999 by prefixing the age field with '~'.
8001 * Seat objects now expose CanGraphical, CanTTY properties
8002 which is required to deal with very fast bootups where the
8003 display manager might be running before the graphics drivers
8004 completed initialization.
8006 * Seat objects now expose a State property.
8008 * We now include RPM macros for service enabling/disabling
8009 based on the preset logic. We recommend RPM based
8010 distributions to make use of these macros if possible. This
8011 makes it simpler to reuse RPM spec files across
8014 * We now make sure that the collected systemd unit name is
8015 always valid when services log to the journal via
8018 * There's a new man page kernel-command-line(7) detailing all
8019 command line options we understand.
8021 * The fstab generator may now be disabled at boot by passing
8022 fstab=0 on the kernel command line.
8024 * A new kernel command line option modules-load= is now understood
8025 to load a specific kernel module statically, early at boot.
8027 * Unit names specified on the systemctl command line are now
8028 automatically escaped as needed. Also, if file system or
8029 device paths are specified they are automatically turned
8030 into the appropriate mount or device unit names. Example:
8032 systemctl status /home
8033 systemctl status /dev/sda
8035 * The SysVConsole= configuration option has been removed from
8036 system.conf parsing.
8038 * The SysV search path is no longer exported on the D-Bus
8041 * The Names= option has been removed from unit file parsing.
8043 * There's a new man page bootup(7) detailing the boot process.
8045 * Every unit and every generator we ship with systemd now
8046 comes with full documentation. The self-explanatory boot is
8049 * A couple of services gained "systemd-" prefixes in their
8050 name if they wrap systemd code, rather than only external
8051 code. Among them fsck@.service which is now
8052 systemd-fsck@.service.
8054 * The HaveWatchdog property has been removed from the D-Bus
8057 * systemd.confirm_spawn= on the kernel command line should now
8060 * There's a new man page crypttab(5) which details all options
8061 we actually understand.
8063 * systemd-nspawn gained a new --capability= switch to pass
8064 additional capabilities to the container.
8066 * timedated will now read known NTP implementation unit names
8067 from /usr/lib/systemd/ntp-units.d/*.list,
8068 systemd-timedated-ntp.target has been removed.
8070 * journalctl gained a new switch "-b" that lists log data of
8071 the current boot only.
8073 * The notify socket is in the abstract namespace again, in
8074 order to support daemons which chroot() at start-up.
8076 * There is a new Storage= configuration option for journald
8077 which allows configuration of where log data should go. This
8078 also provides a way to disable journal logging entirely, so
8079 that data collected is only forwarded to the console, the
8080 kernel log buffer or another syslog implementation.
8082 * Many bugfixes and optimizations
8084 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8085 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8086 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8087 Shawn Landden, Tom Gundersen
8091 * "systemctl help <unit>" now shows the man page if one is
8094 * Several new man pages have been added.
8096 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8097 MaxLevelConsole= can now be specified in
8098 journald.conf. These options allow reducing the amount of
8099 data stored on disk or forwarded by the log level.
8101 * TimerSlackNSec= can now be specified in system.conf for
8102 PID1. This allows system-wide power savings.
8104 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8105 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8110 * logind is now capable of (optionally) handling power and
8111 sleep keys as well as the lid switch.
8113 * journalctl now understands the syntax "journalctl
8114 /usr/bin/avahi-daemon" to get all log output of a specific
8117 * CapabilityBoundingSet= in system.conf now also influences
8118 the capability bound set of usermode helpers of the kernel.
8120 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8121 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8122 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8123 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8127 * Note that we skipped 139 releases here in order to set the
8128 new version to something that is greater than both udev's
8129 and systemd's most recent version number.
8131 * udev: all udev sources are merged into the systemd source tree now.
8132 All future udev development will happen in the systemd tree. It
8133 is still fully supported to use the udev daemon and tools without
8134 systemd running, like in initramfs or other init systems. Building
8135 udev though, will require the *build* of the systemd tree, but
8136 udev can be properly *run* without systemd.
8138 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8139 should be used to create dead device nodes as workarounds for broken
8142 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8143 no longer supported. udev_monitor_new_from_netlink() needs to be
8144 used to subscribe to events.
8146 * udev: when udevd is started by systemd, processes which are left
8147 behind by forking them off of udev rules, are unconditionally cleaned
8148 up and killed now after the event handling has finished. Services or
8149 daemons must be started as systemd services. Services can be
8150 pulled-in by udev to get started, but they can no longer be directly
8151 forked by udev rules.
8153 * udev: the daemon binary is called systemd-udevd now and installed
8154 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8155 to adapt to that, create symlink, or rename the binary after building
8158 * libudev no longer provides these symbols:
8159 udev_monitor_from_socket()
8160 udev_queue_get_failed_list_entry()
8161 udev_get_{dev,sys,run}_path()
8162 The versions number was bumped and symbol versioning introduced.
8164 * systemd-loginctl and systemd-journalctl have been renamed
8165 to loginctl and journalctl to match systemctl.
8167 * The config files: /etc/systemd/systemd-logind.conf and
8168 /etc/systemd/systemd-journald.conf have been renamed to
8169 logind.conf and journald.conf. Package updates should rename
8170 the files to the new names on upgrade.
8172 * For almost all files the license is now LGPL2.1+, changed
8173 from the previous GPL2.0+. Exceptions are some minor stuff
8174 of udev (which will be changed to LGPL2.1 eventually, too),
8175 and the MIT licensed sd-daemon.[ch] library that is suitable
8176 to be used as drop-in files.
8178 * systemd and logind now handle system sleep states, in
8179 particular suspending and hibernating.
8181 * logind now implements a sleep/shutdown/idle inhibiting logic
8182 suitable for a variety of uses. Soonishly Lennart will blog
8183 about this in more detail.
8185 * var-run.mount and var-lock.mount are no longer provided
8186 (which previously bind mounted these directories to their new
8187 places). Distributions which have not converted these
8188 directories to symlinks should consider stealing these files
8189 from git history and add them downstream.
8191 * We introduced the Documentation= field for units and added
8192 this to all our shipped units. This is useful to make it
8193 easier to explore the boot and the purpose of the various
8196 * All smaller setup units (such as
8197 systemd-vconsole-setup.service) now detect properly if they
8198 are run in a container and are skipped when
8199 appropriate. This guarantees an entirely noise-free boot in
8200 Linux container environments such as systemd-nspawn.
8202 * A framework for implementing offline system updates is now
8203 integrated, for details see:
8204 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
8206 * A new service type Type=idle is available now which helps us
8207 avoiding ugly interleaving of getty output and boot status
8210 * There's now a system-wide CapabilityBoundingSet= option to
8211 globally reduce the set of capabilities for the
8212 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8213 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8214 even CAP_NET_ADMIN system-wide for secure systems.
8216 * There are now system-wide DefaultLimitXXX= options to
8217 globally change the defaults of the various resource limits
8218 for all units started by PID 1.
8220 * Harald Hoyer's systemd test suite has been integrated into
8221 systemd which allows easy testing of systemd builds in qemu
8222 and nspawn. (This is really awesome! Ask us for details!)
8224 * The fstab parser is now implemented as generator, not inside
8227 * systemctl will now warn you if .mount units generated from
8228 /etc/fstab are out of date due to changes in fstab that
8229 have not been read by systemd yet.
8231 * systemd is now suitable for usage in initrds. Dracut has
8232 already been updated to make use of this. With this in place
8233 initrds get a slight bit faster but primarily are much
8234 easier to introspect and debug since "systemctl status" in
8235 the host system can be used to introspect initrd services,
8236 and the journal from the initrd is kept around too.
8238 * systemd-delta has been added, a tool to explore differences
8239 between user/admin configuration and vendor defaults.
8241 * PrivateTmp= now affects both /tmp and /var/tmp.
8243 * Boot time status messages are now much prettier and feature
8244 proper english language. Booting up systemd has never been
8247 * Read-ahead pack files now include the inode number of all
8248 files to pre-cache. When the inode changes the pre-caching
8249 is not attempted. This should be nicer to deal with updated
8250 packages which might result in changes of read-ahead
8253 * We now temporaritly lower the kernel's read_ahead_kb variable
8254 when collecting read-ahead data to ensure the kernel's
8255 built-in read-ahead does not add noise to our measurements
8256 of necessary blocks to pre-cache.
8258 * There's now RequiresMountsFor= to add automatic dependencies
8259 for all mounts necessary for a specific file system path.
8261 * MountAuto= and SwapAuto= have been removed from
8262 system.conf. Mounting file systems at boot has to take place
8265 * nspawn now learned a new switch --uuid= to set the machine
8266 ID on the command line.
8268 * nspawn now learned the -b switch to automatically search
8271 * vt102 is now the default TERM for serial TTYs, upgraded from
8274 * systemd-logind now works on VT-less systems.
8276 * The build tree has been reorganized. The individual
8277 components now have directories of their own.
8279 * A new condition type ConditionPathIsReadWrite= is now available.
8281 * nspawn learned the new -C switch to create cgroups for the
8282 container in other hierarchies.
8284 * We now have support for hardware watchdogs, configurable in
8287 * The scheduled shutdown logic now has a public API.
8289 * We now mount /tmp as tmpfs by default, but this can be
8290 masked and /etc/fstab can override it.
8292 * Since udisks does not make use of /media anymore we are not
8293 mounting a tmpfs on it anymore.
8295 * journalctl gained a new --local switch to only interleave
8296 locally generated journal files.
8298 * We can now load the IMA policy at boot automatically.
8300 * The GTK tools have been split off into a systemd-ui.
8302 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8303 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8304 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8305 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8306 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8307 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8308 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8309 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8310 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8315 * This is mostly a bugfix release
8317 * Support optional initialization of the machine ID from the
8318 KVM or container configured UUID.
8320 * Support immediate reboots with "systemctl reboot -ff"
8322 * Show /etc/os-release data in systemd-analyze output
8324 * Many bugfixes for the journal, including endianness fixes and
8325 ensuring that disk space enforcement works
8327 * sd-login.h is C++ compatible again
8329 * Extend the /etc/os-release format on request of the Debian
8332 * We now refuse non-UTF8 strings used in various configuration
8333 and unit files. This is done to ensure we do not pass invalid
8334 data over D-Bus or expose it elsewhere.
8336 * Register Mimo USB Screens as suitable for automatic seat
8339 * Read SELinux client context from journal clients in a race
8342 * Reorder configuration file lookup order. /etc now always
8343 overrides /run in order to allow the administrator to always
8344 and unconditionally override vendor-supplied or
8345 automatically generated data.
8347 * The various user visible bits of the journal now have man
8348 pages. We still lack man pages for the journal API calls
8351 * We now ship all man pages in HTML format again in the
8354 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8355 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8356 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8361 * This is mostly a bugfix release
8363 * systems lacking /etc/os-release are no longer supported.
8365 * Various functionality updates to libsystemd-login.so
8367 * Track class of PAM logins to distinguish greeters from
8370 Contributions from: Kay Sievers, Lennart Poettering, Michael
8375 * This is an important bugfix release for v41.
8377 * Building man pages is now optional which should be useful
8378 for those building systemd from git but unwilling to install
8381 * Watchdog support for supervising services is now usable. In
8382 a future release support for hardware watchdogs
8383 (i.e. /dev/watchdog) will be added building on this.
8385 * Service start rate limiting is now configurable and can be
8386 turned off per service. When a start rate limit is hit a
8387 reboot can automatically be triggered.
8389 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8391 Contributions from: Benjamin Franzke, Bill Nottingham,
8392 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8393 Schmidt, Michał Górny, Piotr Drąg
8397 * The systemd binary is installed /usr/lib/systemd/systemd now;
8398 An existing /sbin/init symlink needs to be adapted with the
8401 * The code that loads kernel modules has been ported to invoke
8402 libkmod directly, instead of modprobe. This means we do not
8403 support systems with module-init-tools anymore.
8405 * Watchdog support is now already useful, but still not
8408 * A new kernel command line option systemd.setenv= is
8409 understood to set system wide environment variables
8410 dynamically at boot.
8412 * We now limit the set of capabilities of systemd-journald.
8414 * We now set SIGPIPE to ignore by default, since it only is
8415 useful in shell pipelines, and has little use in general
8416 code. This can be disabled with IgnoreSIPIPE=no in unit
8419 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8420 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8425 * This is mostly a bugfix release
8427 * We now expose the reason why a service failed in the
8428 "Result" D-Bus property.
8430 * Rudimentary service watchdog support (will be completed over
8431 the next few releases.)
8433 * When systemd forks off in order execute some service we will
8434 now immediately changes its argv[0] to reflect which process
8435 it will execute. This is useful to minimize the time window
8436 with a generic argv[0], which makes bootcharts more useful
8438 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8439 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8440 Mike Kazantsev, Ray Strode
8444 * This is mostly a test release, but incorporates many
8447 * New systemd-cgtop tool to show control groups by their
8450 * Linking against libacl for ACLs is optional again. If
8451 disabled, support tracking device access for active logins
8452 goes becomes unavailable, and so does access to the user
8453 journals by the respective users.
8455 * If a group "adm" exists, journal files are automatically
8456 owned by them, thus allow members of this group full access
8457 to the system journal as well as all user journals.
8459 * The journal now stores the SELinux context of the logging
8460 client for all entries.
8462 * Add C++ inclusion guards to all public headers
8464 * New output mode "cat" in the journal to print only text
8465 messages, without any meta data like date or time.
8467 * Include tiny X server wrapper as a temporary stop-gap to
8468 teach XOrg udev display enumeration. This is used by display
8469 managers such as gdm, and will go away as soon as XOrg
8470 learned native udev hotplugging for display devices.
8472 * Add new systemd-cat tool for executing arbitrary programs
8473 with STDERR/STDOUT connected to the journal. Can also act as
8474 BSD logger replacement, and does so by default.
8476 * Optionally store all locally generated coredumps in the
8477 journal along with meta data.
8479 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8480 writing short strings to files (for usage for /sys), and for
8481 creating symlinks, character and block device nodes.
8483 * New unit file option ControlGroupPersistent= to make cgroups
8484 persistent, following the mechanisms outlined in
8485 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
8487 * Support multiple local RTCs in a sane way
8489 * No longer monopolize IO when replaying readahead data on
8490 rotating disks, since we might starve non-file-system IO to
8491 death, since fanotify() will not see accesses done by blkid,
8494 * Do not show kernel threads in systemd-cgls anymore, unless
8495 requested with new -k switch.
8497 Contributions from: Dan Horák, Kay Sievers, Lennart
8498 Poettering, Michal Schmidt
8502 * This is mostly a test release, but incorporates many
8505 * The git repository moved to:
8506 git://anongit.freedesktop.org/systemd/systemd
8507 ssh://git.freedesktop.org/git/systemd/systemd
8509 * First release with the journal
8510 http://0pointer.de/blog/projects/the-journal.html
8512 * The journal replaces both systemd-kmsg-syslogd and
8513 systemd-stdout-bridge.
8515 * New sd_pid_get_unit() API call in libsystemd-logind
8517 * Many systemadm clean-ups
8519 * Introduce remote-fs-pre.target which is ordered before all
8520 remote mounts and may be used to start services before all
8523 * Added Mageia support
8525 * Add bash completion for systemd-loginctl
8527 * Actively monitor PID file creation for daemons which exit in
8528 the parent process before having finished writing the PID
8529 file in the daemon process. Daemons which do this need to be
8530 fixed (i.e. PID file creation must have finished before the
8531 parent exits), but we now react a bit more gracefully to them.
8533 * Add colourful boot output, mimicking the well-known output
8534 of existing distributions.
8536 * New option PassCredentials= for socket units, for
8537 compatibility with a recent kernel ABI breakage.
8539 * /etc/rc.local is now hooked in via a generator binary, and
8540 thus will no longer act as synchronization point during
8543 * systemctl list-unit-files now supports --root=.
8545 * systemd-tmpfiles now understands two new commands: z, Z for
8546 relabelling files according to the SELinux database. This is
8547 useful to apply SELinux labels to specific files in /sys,
8550 * Output of SysV services is now forwarded to both the console
8551 and the journal by default, not only just the console.
8553 * New man pages for all APIs from libsystemd-login.
8555 * The build tree got reorganized and the build system is a
8556 lot more modular allowing embedded setups to specifically
8557 select the components of systemd they are interested in.
8559 * Support for Linux systems lacking the kernel VT subsystem is
8562 * configure's --with-rootdir= got renamed to
8563 --with-rootprefix= to follow the naming used by udev and
8566 * Unless specified otherwise we will now install to /usr instead
8567 of /usr/local by default.
8569 * Processes with '@' in argv[0][0] are now excluded from the
8570 final shut-down killing spree, following the logic explained
8572 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
8574 * All processes remaining in a service cgroup when we enter
8575 the START or START_PRE states are now killed with
8576 SIGKILL. That means it is no longer possible to spawn
8577 background processes from ExecStart= lines (which was never
8578 supported anyway, and bad style).
8580 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8581 reloading of units together.
8583 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
8584 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8585 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8586 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8587 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob