1 systemd System and Service Manager
3 CHANGES WITH 239 in spe:
5 * NETWORK INTERFACE DEVICE NAMING CHANGES: elogind-udevd's "net_id"
6 builtin will name network interfaces differently than in previous
7 versions for virtual network interfaces created with SR-IOV and NPAR
8 and for devices where the PCI network controller device does not have
9 a slot number associated.
11 SR-IOV virtual devices are now named based on the name of the parent
12 interface, with a suffix of "v<N>", where <N> is the virtual device
13 number. Previously those virtual devices were named as if completely
16 The ninth and later NPAR virtual devices will be named following the
17 scheme used for the first eight NPAR partitions. Previously those
18 devices were not renamed and the kernel default (eth<n>) was used.
20 "net_id" will also generate names for PCI devices where the PCI
21 network controller device does not have an associated slot number
22 itself, but one of its parents does. Previously those devices were
23 not renamed and the kernel default (eth<n>) was used.
25 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
26 elogind.service. Since v235, IPAddressDeny=any has been set to
27 the unit. So, it is expected that the default behavior of
28 elogind is not changed. However, if distribution packagers or
29 administrators disabled or modified IPAddressDeny= setting by a
30 drop-in config file, then it may be necessary to update the file to
31 re-enable AF_INET and AF_INET6 to support network user name services,
34 * When the RestrictNamespaces= unit property is specified multiple
35 times, then the specified types are merged now. Previously, only the
36 last assignment was used. So, if distribution packagers or
37 administrators modified the setting by a drop-in config file, then it
38 may be necessary to update the file.
40 * When OnFailure= is used in combination with Restart= on a service
41 unit, then the specified units will no longer be triggered on
42 failures that result in restarting. Previously, the specified units
43 would be activated each time the unit failed, even when the unit was
44 going to be restarted automatically. This behaviour contradicted the
45 documentation. With this release the code is adjusted to match the
48 * elogind-tmpfiles will now print a notice whenever it encounters
49 tmpfiles.d/ lines referencing the /var/run/ directory. It will
50 recommend reworking them to use the /run/ directory instead (for
51 which /var/run/ is simply a symlinked compatibility alias). This way
52 elogind-tmpfiles can properly detect line conflicts and merge lines
53 referencing the same file by two paths, without having to access
56 * systemctl disable/unmask/preset/preset-all cannot be used with
57 --runtime. Previously this was allowed, but resulted in unintuitive
58 behaviour that wasn't useful. systemctl disable/unmask will now undo
59 both runtime and persistent enablement/masking, i.e. it will remove
60 any relevant symlinks both in /run and /etc.
62 * Note that all long-running system services shipped with elogind will
63 now default to a system call whitelist (rather than a blacklist, as
64 before). In particular, elogind-udevd will now enforce one too. For
65 most cases this should be safe, however downstream distributions
66 which disabled sandboxing of elogind-udevd (specifically the
67 MountFlags= setting), might want to disable this security feature
68 too, as the default whitelisting will prohibit all mount, swap,
69 reboot and clock changing operations from udev rules.
71 * sd-boot acquired new loader configuration settings to optionally turn
72 off Windows and MacOS boot partition discovery as well as
73 reboot-into-firmware menu items. It is also able to pick a better
74 screen resolution for HiDPI systems, and now provides loader
75 configuration settings to change the resolution explicitly.
77 * The elogind-resolve tool has been renamed to resolvectl (it also
78 * elogind-resolved now supports DNS-over-TLS. It's still
79 turned off by default, use DNSOverTLS=opportunistic to turn it on in
80 resolved.conf. We intend to make this the default as soon as couple
81 of additional techniques for optimizing the initial latency caused by
82 establishing a TLS/TCP connection are implemented.
84 * elogind-resolved.service and elogind-networkd.service now set
85 DynamicUser=yes. The users elogind-resolve and elogind-network are
86 not created by elogind-sysusers.
88 remains available under the old name, for compatibility), and its
89 interface is now verb-based, similar in style to the other <xyz>ctl
90 tools, such as systemctl or loginctl.
92 * The resolvectl/elogind-resolve tool also provides 'resolvconf'
93 compatibility. It may be symlinked under the 'resolvconf' name, in
94 which case it will take arguments and input compatible with the
95 Debian and FreeBSD resolvconf tool.
97 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
98 where the system initially suspends, and after a time-out resumes and
101 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
102 set the client will only send a DUID as client identifier.
104 * The nss-elogind glibc NSS module will now enumerate dynamic users and
105 groups in effect. Previously, it could resolve UIDs/GIDs to user
106 names/groups and vice versa, but did not support enumeration.
108 * journald's Compress= configuration setting now optionally accepts a
109 byte threshold value. All journal objects larger than this threshold
110 will be compressed, smaller ones will not. Previously this threshold
111 was not configurable and set to 512.
113 * A new system.conf setting NoNewPrivileges= is now available which may
114 be used to turn off acquisition of new privileges system-wide
115 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
116 for all its children). Note that turning this option on means setuid
117 binaries and file system capabilities lose their special powers.
118 While turning on this option is a big step towards a more secure
119 system, doing so is likely to break numerous pre-existing UNIX tools,
120 in particular su and sudo.
122 * A new service elogind-time-sync-wait.service has been added. If
123 enabled it will delay the time-sync.target unit at boot until time
124 synchronization has been received from the network. This
125 functionality is useful on systems lacking a local RTC or where it is
126 acceptable that the boot process shall be delayed by external network
129 * When hibernating, elogind will now inform the kernel of the image
130 write offset, on kernels new enough to support this. This means swap
131 files should work for hibernation now.
133 * When loading unit files, elogind will now look for drop-in unit files
134 extensions in additional places. Previously, for a unit file name
135 "foo-bar-baz.service" it would look for dropin files in
136 "foo-bar-baz.service.d/*.conf". Now, it will also look in
137 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
138 service name truncated after all inner dashes. This scheme allows
139 writing drop-ins easily that apply to a whole set of unit files at
140 once. It's particularly useful for mount and slice units (as their
141 naming is prefix based), but is also useful for service and other
142 units, for packages that install multiple unit files at once,
143 following a strict naming regime of beginning the unit file name with
144 the package's name. Two new specifiers are now supported in unit
145 files to match this: %j and %J are replaced by the part of the unit
146 name following the last dash.
148 * Unit files and other configuration files that support specifier
149 expansion now understand another two new specifiers: %T and %V will
150 resolve to /tmp and /var/tmp respectively, or whatever temporary
151 directory has been set for the calling user.
153 * The ExecStart= lines of unit files are no longer required to
154 reference absolute paths. If non-absolute paths are specified the
155 specified binary name is searched within the service manager's
156 built-in $PATH, which may be queried with 'elogind-path
157 search-binaries-default'. It's generally recommended to continue to
158 use absolute paths for all binaries specified in unit files.
160 * Units gained a new load state "bad-setting", which is used when a
161 unit file was loaded, but contained fatal errors which prevent it
162 from being started (for example, a service unit has been defined
163 lacking both ExecStart= and ExecStop= lines).
165 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
166 support alternative debuggers, for example lldb. The old name
167 continues to be available however, for compatibility reasons. Use the
168 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
169 to pick an alternative debugger instead of the default gdb.
171 * systemctl and the other tools will now output escape sequences that
172 generate proper clickable hyperlinks in various terminal emulators
173 where useful (for example, in the "systemctl status" output you can
174 now click on the unit file name to quickly open it in the
175 editor/viewer of your choice). Note that not all terminal emulators
176 support this functionality yet, but many do. Unfortunately, the
177 "less" pager doesn't support this yet, hence this functionality is
178 currently automatically turned off when a pager is started (which
179 happens quite often due to auto-paging). We hope to remove this
180 limitation as soon as "less" learns these escape sequences. This new
181 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
182 environment variable. For details on these escape sequences see:
183 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
185 * networkd's .network files now support a new IPv6MTUBytes= option for
186 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
187 option in the [Route] section to configure the MTU to use for
188 specific routes. It also gained support for configuration of the DHCP
189 "UserClass" option through the new UserClass= setting. It gained
190 three new options in the new [CAN] section for configuring CAN
191 networks. The MULTICAST and ALLMULTI interface flags may now be
192 controlled explicitly with the new Multicast= and AllMulticast=
195 * networkd will now automatically make use of the kernel's route
196 expiration feature, if it is available.
198 * udevd's .link files now support setting the number of receive and
199 transmit channels, using the RxChannels=, TxChannels=,
200 OtherChannels=, CombinedChannels= settings.
202 * Support for UDPSegmentationOffload= has been removed, given its
203 limited support in hardware, and waning software support.
205 * networkd's .netdev files now support creating "netdevsim" interfaces.
207 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
208 to query the unit belonging to a specific kernel control group.
210 * elogind-analyze gained a new verb "cat-config", which may be used to
211 dump the contents of any configuration file, with all its matching
212 drop-in files added in, and honouring the usual search and masking
213 logic applied to elogind configuration files. For example use
214 "elogind-analyze cat-config elogind/system.conf" to get the complete
215 system configuration file of elogind how it would be loaded by PID 1
216 itself. Similar to this, various tools such as elogind-tmpfiles or
217 elogind-sysusers, gained a new option "--cat-config", which does the
218 corresponding operation for their own configuration settings. For
219 example, "elogind-tmpfiles --cat-config" will now output the full
220 list of tmpfiles.d/ lines in place.
222 * timedatectl gained two new verbs "timesync-status" (to show the
223 current NTP synchronization state of elogind-timesyncd) and
224 "show-timesync" (to show bus properties of elogind-timesyncd).
226 * elogind-timesyncd gained a bus interface on which it exposes details
229 * elogind-nspawn gained a new --rlimit= switch for setting initial
230 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
231 understood by elogind-timedated. It takes a colon-separated list of
232 unit names of NTP client services. The list is used by
233 "timedatectl set-ntp".
235 resource limits for the container payload. There's a new switch
236 --hostname= to explicitly override the container's hostname. A new
237 --no-new-privileges= switch may be used to control the
238 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
239 --oom-score-adjust= switch controls the OOM scoring adjustment value
240 for the payload. The new --cpu-affinity= switch controls the CPU
241 affinity of the container payload. The new --resolv-conf= switch
242 allows more detailed control of /etc/resolv.conf handling of the
243 container. Similarly, the new --timezone= switch allows more detailed
244 control of /etc/localtime handling of the container.
246 * elogind-detect-virt gained a new --list switch, which will print a
247 list of all currently known VM and container environments.
249 * Support for "Portable Services" has been added, see
250 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
251 experimental, but this is expected to change soon. Reflecting this
252 experimental state, the "portablectl" binary is not installed into
253 /usr/bin yet. The binary has to be called with the full path
254 /usr/lib/elogind/portablectl instead.
256 * journalctl's and systemctl's -o switch now knows a new log output
257 mode "with-unit". The output it generates is very similar to the
258 regular "short" mode, but displays the unit name instead of the
259 syslog tag for each log line. Also, the date is shown with timezone
260 information. This mode is probably more useful than the classic
261 "short" output mode for most purposes, except where pixel-perfect
262 compatibility with classic /var/log/messages formatting is required.
264 * A new --dump-bus-properties switch has been added to the elogind
265 binary, which may be used to dump all supported D-Bus properties.
266 (Options which are still supported, but are deprecated, are *not*
269 * sd-bus gained a set of new calls:
270 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
271 enable/disable the "floating" state of a bus slot object,
272 i.e. whether the slot object pins the bus it is allocated for into
273 memory or if the bus slot object gets disconnected when the bus goes
274 away. sd_bus_open_with_description(),
275 sd_bus_open_user_with_description(),
276 sd_bus_open_system_with_description() may be used to allocate bus
277 objects and set their description string already during allocation.
279 * sd-event gained support for watching inotify events from the event
280 loop, in an efficient way, sharing inotify handles between multiple
281 users. For this a new function sd_event_add_inotify() has been added.
283 * sd-event and sd-bus gained support for calling special user-supplied
284 destructor functions for userdata pointers associated with
285 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
286 functions sd_bus_slot_set_destroy_callback,
287 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
288 sd_bus_track_get_destroy_callback,
289 sd_event_source_set_destroy_callback,
290 sd_event_source_get_destroy_callback have been added.
292 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
294 * PID 1 will now automatically reschedule .timer units whenever the
295 local timezone changes. (They previously got rescheduled
296 automatically when the system clock changed.)
298 * New documentation has been added to document cgroups delegation,
299 portable services and the various code quality tools we have set up:
301 https://github.com/systemd/systemd/blob/master/doc/CGROUP_DELEGATION.md
302 https://github.com/systemd/systemd/blob/master/doc/PORTABLE_SERVICES.md
303 https://github.com/systemd/systemd/blob/master/doc/CODE_QUALITY.md
305 * pam_elogind will now look for PAM userdata fields elogind.memory_max,
306 elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
307 * 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 earlier PAM modules. The data in these fields is used to initialize
315 the session scope's resource properties. Thus external PAM modules
316 may now configure per-session limits, for example sourced from
317 external user databases.
319 * socket units with Accept=yes will now maintain a "refused" counter in
320 addition to the existing "accepted" counter, counting connections
321 refused due to the enforced limits.
323 * The "elogind-path search-binaries-default" command may now be use to
324 query the default, built-in $PATH PID 1 will pass to the services it
327 * A new unit file setting PrivateMounts= has been added. It's a boolean
328 option. If enabled the unit's processes are invoked in their own file
329 system namespace. Note that this behaviour is also implied if any
330 other file system namespacing options (such as PrivateTmp=,
331 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
332 primarily useful for services that do not use any of the other file
333 system namespacing options. One such service is elogind-udevd.service
334 wher this is now used by default.
336 * A new unit "system-update-pre.target" is added, which defines an
337 optional synchronization point for offline system updates, as
338 implemented by the pre-existing "system-update.target" unit. It
339 allows ordering services before the service that executes the actual
340 update process in a generic way.
342 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
343 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
344 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
345 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
346 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
347 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
348 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
349 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
350 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
351 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
352 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
353 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
354 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
355 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
356 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
357 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
358 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
359 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
360 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
361 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
362 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
363 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
364 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
365 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
366 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
367 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
368 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
369 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
370 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
376 * The MemoryAccounting= unit property now defaults to on. After
377 discussions with the upstream control group maintainers we learnt
378 that the negative impact of cgroup memory accounting on current
379 kernels is finally relatively minimal, so that it should be safe to
380 enable this by default without affecting system performance. Besides
381 memory accounting only task accounting is turned on by default, all
382 other forms of resource accounting (CPU, IO, IP) remain off for now,
383 because it's not clear yet that their impact is small enough to move
384 from opt-in to opt-out. We recommend downstreams to leave memory
385 accounting on by default if kernel 4.14 or higher is primarily
386 used. On very resource constrained systems or when support for old
387 kernels is a necessity, -Dmemory-accounting-default=false can be used
388 to revert this change.
390 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
391 %udev_rules_update) and the journal catalog (%journal_catalog_update)
392 from the upgrade scriptlets of individual packages now do nothing.
393 Transfiletriggers have been added which will perform those updates
394 once at the end of the transaction.
396 Similar transfiletriggers have been added to execute any sysctl.d
397 and binfmt.d rules. Thus, it should be unnecessary to provide any
398 scriptlets to execute this configuration from package installation
401 * elogind-sysusers gained a mode where the configuration to execute is
402 specified on the command line, but this configuration is not executed
403 directly, but instead it is merged with the configuration on disk,
404 and the result is executed. This is useful for package installation
405 scripts which want to create the user before installing any files on
406 disk (in case some of those files are owned by that user), while
407 still allowing local admin overrides.
409 This functionality is exposed to rpm scriptlets through a new
410 %sysusers_create_package macro. Old %sysusers_create and
411 %sysusers_create_inline macros are deprecated.
413 A transfiletrigger for sysusers.d configuration is now installed,
414 which means that it should be unnecessary to call elogind-sysusers from
415 package installation scripts, unless the package installs any files
416 owned by those newly-created users, in which case
417 %sysusers_create_package should be used.
419 * Analogous change has been done for elogind-tmpfiles: it gained a mode
420 where the command-line configuration is merged with the configuration
421 on disk. This is exposed as the new %tmpfiles_create_package macro,
422 and %tmpfiles_create is deprecated. A transfiletrigger is installed
423 for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
424 from package installation scripts.
426 * sysusers.d configuration for a user may now also specify the group
427 number, in addition to the user number ("u username 123:456"), or
428 without the user number ("u username -:456").
430 * Configution items for elogind-sysusers can now be specified as
431 positional arguments when the new --inline switch is used.
433 * The login shell of users created through sysusers.d may now be
434 specified (previously, it was always /bin/sh for root and
435 /sbin/nologin for other users).
437 * elogind-analyze gained a new --global switch to look at global user
438 configuration. It also gained a unit-paths verb to list the unit load
439 paths that are compiled into elogind (which can be used with
440 --elogind, --user, or --global).
442 * udevadm trigger gained a new --settle/-w option to wait for any
443 triggered events to finish (but just those, and not any other events
444 which are triggered meanwhile).
446 * The action that elogind takes when the lid is closed and the
447 machine is connected to external power can now be configured using
448 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
449 was determined by HandleLidSwitch=, and, for backwards compatibility,
450 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
452 * journalctl will periodically call sd_journal_process() to make it
453 resilient against inotify queue overruns when journal files are
454 rotated very quickly.
456 * Two new functions in libelogind — sd_bus_get_n_queued_read and
457 sd_bus_get_n_queued_write — may be used to check the number of
458 pending bus messages.
460 * elogind gained a new
461 org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
462 which can be used to migrate foreign processes to scope and service
463 units. The primary user for this new API is elogind itself: the
464 elogind --user instance uses this call of the elogind --system
465 instance to migrate processes if it itself gets the request to
466 migrate processes and the kernel refuses this due to access
467 restrictions. Thanks to this "elogind-run --scope --user …" works
468 again in pure cgroups v2 environments when invoked from the user
471 * A new TemporaryFileSystem= setting can be used to mask out part of
472 the real file system tree with tmpfs mounts. This may be combined
473 with BindPaths= and BindReadOnlyPaths= to hide files or directories
474 not relevant to the unit, while still allowing some paths lower in
475 the tree to be accessed.
477 ProtectHome=tmpfs may now be used to hide user home and runtime
478 directories from units, in a way that is mostly equivalent to
479 "TemporaryFileSystem=/home /run/user /root".
481 * Non-service units are now started with KeyringMode=shared by default.
482 This means that mount and swapon and other mount tools have access
483 to keys in the main keyring.
485 * /sys/fs/bpf is now mounted automatically.
487 * QNX virtualization is now detected by elogind-detect-virt and may
488 be used in ConditionVirtualization=.
490 * IPAccounting= may now be enabled also for slice units.
492 * A new -Dsplit-bin= build configuration switch may be used to specify
493 whether bin and sbin directories are merged, or if they should be
494 included separately in $PATH and various listings of executable
495 directories. The build configuration scripts will try to autodetect
496 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
497 system, but distributions are encouraged to configure this
500 * A new -Dok-color= build configuration switch may be used to change
501 the colour of "OK" status messages.
503 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
504 PrivateNetwork=yes was buggy in previous versions of elogind. This
505 means that after the upgrade and daemon-reexec, any such units must
508 * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
509 will not exclude read-only files owned by root from cleanup.
511 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
512 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
513 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
514 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
515 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
516 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
517 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
518 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
519 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
520 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
521 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
522 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
523 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
524 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
525 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
526 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
532 * Some keyboards come with a zoom see-saw or rocker which until now got
533 mapped to the Linux "zoomin/out" keys in hwdb. However, these
534 keycodes are not recognized by any major desktop. They now produce
535 Up/Down key events so that they can be used for scrolling.
537 * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
538 slightly: previously, if an argument was specified for lines of this
539 type (i.e. the right-most column was set) this string was appended to
540 existing files each time elogind-tmpfiles was run. This behaviour was
541 different from what the documentation said, and not particularly
542 useful, as repeated elogind-tmpfiles invocations would not be
543 idempotent and grow such files without bounds. With this release
544 behaviour has been altered slightly, to match what the documentation
545 says: lines of this type only have an effect if the indicated files
546 don't exist yet, and only then the argument string is written to the
549 * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
550 elogind-tmpfiles behaviour: previously, read-only files owned by root
551 were always excluded from the file "aging" algorithm (i.e. the
552 automatic clean-up of directories like /tmp based on
553 atime/mtime/ctime). We intend to drop this restriction, and age files
554 by default even when owned by root and read-only. This behaviour was
555 inherited from older tools, but there have been requests to remove
556 it, and it's not obvious why this restriction was made in the first
557 place. Please speak up now, if you are aware of software that reqires
558 this behaviour, otherwise we'll remove the restriction in v238.
560 * A new environment variable $SYSTEMD_OFFLINE is now understood by
561 systemctl. It takes a boolean argument. If on, systemctl assumes it
562 operates on an "offline" OS tree, and will not attempt to talk to the
563 service manager. Previously, this mode was implicitly enabled if a
564 chroot() environment was detected, and this new environment variable
565 now provides explicit control.
567 * .path and .socket units may now be created transiently, too.
568 Previously only service, mount, automount and timer units were
569 supported as transient units. The elogind-run tool has been updated
570 to expose this new functionality, you may hence use it now to bind
571 arbitrary commands to path or socket activation on-the-fly from the
572 command line. Moreover, almost all properties are now exposed for the
573 unit types that already supported transient operation.
575 * The elogind-mount command gained support for a new --owner= parameter
576 which takes a user name, which is then resolved and included in uid=
577 and gid= mount options string of the file system to mount.
579 * A new unit condition ConditionControlGroupController= has been added
580 that checks whether a specific cgroup controller is available.
582 * Unit files, udev's .link files, and elogind-networkd's .netdev and
583 .network files all gained support for a new condition
584 ConditionKernelVersion= for checking against specific kernel
587 * In elogind-networkd, the [IPVLAN] section in .netdev files gained
588 support for configuring device flags in the Flags= setting. In the
589 same files, the [Tunnel] section gained support for configuring
590 AllowLocalRemote=. The [Route] section in .network files gained
591 support for configuring InitialCongestionWindow=,
592 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
593 understands RapidCommit=.
595 * elogind-networkd's DHCPv6 support gained support for Prefix
598 * sd-bus gained support for a new "watch-bind" feature. When this
599 feature is enabled, an sd_bus connection may be set up to connect to
600 an AF_UNIX socket in the file system as soon as it is created. This
601 functionality is useful for writing early-boot services that
602 automatically connect to the system bus as soon as it is started,
603 without ugly time-based polling. elogind-networkd and
604 elogind-resolved have been updated to make use of this
605 functionality. busctl exposes this functionality in a new
606 --watch-bind= command line switch.
608 * sd-bus will now optionally synthesize a local "Connected" signal as
609 soon as a D-Bus connection is set up fully. This message mirrors the
610 already existing "Disconnected" signal which is synthesized when the
611 connection is terminated. This signal is generally useful but
612 particularly handy in combination with the "watch-bind" feature
613 described above. Synthesizing of this message has to be requested
614 explicitly through the new API call sd_bus_set_connected_signal(). In
615 addition a new call sd_bus_is_ready() has been added that checks
616 whether a connection is fully set up (i.e. between the "Connected" and
617 "Disconnected" signals).
619 * sd-bus gained two new calls sd_bus_request_name_async() and
620 sd_bus_release_name_async() for asynchronously registering bus
621 names. Similar, there is now sd_bus_add_match_async() for installing
622 a signal match asynchronously. All of elogind's own services have
623 been updated to make use of these calls. Doing these operations
624 asynchronously has two benefits: it reduces the risk of deadlocks in
625 case of cyclic dependencies between bus services, and it speeds up
626 service initialization since synchronization points for bus
627 round-trips are removed.
629 * sd-bus gained two new calls sd_bus_match_signal() and
630 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
631 and sd_bus_add_match_async() but instead of taking a D-Bus match
632 string take match fields as normal function parameters.
634 * sd-bus gained two new calls sd_bus_set_sender() and
635 sd_bus_message_set_sender() for setting the sender name of outgoing
636 messages (either for all outgoing messages or for just one specific
637 one). These calls are only useful in direct connections as on
638 brokered connections the broker fills in the sender anyway,
639 overwriting whatever the client filled in.
641 * sd-event gained a new pseudo-handle that may be specified on all API
642 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
643 used this refers to the default event loop object of the calling
644 thread. Note however that this does not implicitly allocate one —
645 which has to be done prior by using sd_event_default(). Similarly
646 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
647 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
648 to the default bus of the specified type of the calling thread. Here
649 too this does not implicitly allocate bus connection objects, this
650 has to be done prior with sd_bus_default() and friends.
652 * sd-event gained a new call pair
653 sd_event_source_{get|set}_io_fd_own(). This may be used to request
654 automatic closure of the file descriptor an IO event source watches
655 when the event source is destroyed.
657 * elogind-networkd gained support for natively configuring WireGuard
660 * In previous versions elogind synthesized user records both for the
661 "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
662 internally. In order to simplify distribution-wide renames of the
663 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
664 new transitional flag file has been added: if
665 /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
666 user and group record within the elogind codebase is disabled.
668 * elogind-notify gained a new --uid= option for selecting the source
669 user/UID to use for notification messages sent to the service
672 * journalctl gained a new --grep= option to list only entries in which
673 the message matches a certain pattern. By default matching is case
674 insensitive if the pattern is lowercase, and case sensitive
675 otherwise. Option --case-sensitive=yes|no can be used to override
676 this an specify case sensitivity or case insensitivity.
678 * There's now a "elogind-analyze service-watchdogs" command for printing
679 the current state of the service runtime watchdog, and optionally
680 enabling or disabling the per-service watchdogs system-wide if given a
681 boolean argument (i.e. the concept you configure in WatchdogSec=), for
682 debugging purposes. There's also a kernel command line option
683 elogind.service_watchdogs= for controlling the same.
685 * Two new "log-level" and "log-target" options for elogind-analyze were
686 added that merge the now deprecated get-log-level, set-log-level and
687 get-log-target, set-log-target pairs. The deprecated options are still
688 understood for backwards compatibility. The two new options print the
689 current value when no arguments are given, and set them when a
690 level/target is given as an argument.
692 * sysusers.d's "u" lines now optionally accept both a UID and a GID
693 specification, separated by a ":" character, in order to create users
694 where UID and GID do not match.
696 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
697 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
698 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
699 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
700 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
701 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
702 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
703 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
704 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
705 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
706 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
707 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
708 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
709 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
710 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
711 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
712 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
713 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
714 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
715 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
722 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
723 in v235 has been extended to also set the dummy.ko module option
724 numdummies=0, preventing the kernel from automatically creating
725 dummy0. All dummy interfaces must now be explicitly created.
727 * Unknown '%' specifiers in configuration files are now rejected. This
728 applies to units and tmpfiles.d configuration. Any percent characters
729 that are followed by a letter or digit that are not supposed to be
730 interpreted as the beginning of a specifier should be escaped by
731 doubling ("%%"). (So "size=5%" is still accepted, as well as
732 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
733 valid specifiers today.)
735 * systemd-resolved now maintains a new dynamic
736 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
737 recommended to make /etc/resolv.conf a symlink to it. This file
738 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
739 includes dynamically acquired search domains, achieving more correct
740 DNS resolution by software that bypasses local DNS APIs such as NSS.
742 * The "uaccess" udev tag has been dropped from /dev/kvm and
743 /dev/dri/renderD*. These devices now have the 0666 permissions by
744 default (but this may be changed at build-time). /dev/dri/renderD*
745 will now be owned by the "render" group along with /dev/kfd.
747 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
748 systemd-journal-gatewayd.service and
749 systemd-journal-upload.service. This means "nss-systemd" must be
750 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
751 services are resolved properly.
753 * In /etc/fstab two new mount options are now understood:
754 x-systemd.makefs and x-systemd.growfs. The former has the effect that
755 the configured file system is formatted before it is mounted, the
756 latter that the file system is resized to the full block device size
757 after it is mounted (i.e. if the file system is smaller than the
758 partition it resides on, it's grown). This is similar to the fsck
759 logic in /etc/fstab, and pulls in systemd-makefs@.service and
760 systemd-growfs@.service as necessary, similar to
761 systemd-fsck@.service. Resizing is currently only supported on ext4
764 * In systemd-networkd, the IPv6 RA logic now optionally may announce
765 DNS server and domain information.
767 * Support for the LUKS2 on-disk format for encrypted partitions has
768 been added. This requires libcryptsetup2 during compilation and
771 * The systemd --user instance will now signal "readiness" when its
772 basic.target unit has been reached, instead of when the run queue ran
773 empty for the first time.
775 * Tmpfiles.d with user configuration are now also supported.
776 systemd-tmpfiles gained a new --user switch, and snippets placed in
777 ~/.config/user-tmpfiles.d/ and corresponding directories will be
778 executed by systemd-tmpfiles --user running in the new
779 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
780 running in the user session.
782 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
783 %S resolves to the top-level state directory (/var/lib for the system
784 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
785 top-level cache directory (/var/cache for the system instance,
786 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
787 logs directory (/var/log for the system instance,
788 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
789 existing %t specifier, that resolves to the top-level runtime
790 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
793 * journalctl learnt a new parameter --output-fields= for limiting the
794 set of journal fields to output in verbose and JSON output modes.
796 * systemd-timesyncd's configuration file gained a new option
797 RootDistanceMaxSec= for setting the maximum root distance of servers
798 it'll use, as well as the new options PollIntervalMinSec= and
799 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
801 * bootctl gained a new command "list" for listing all available boot
802 menu items on systems that follow the boot loader specification.
804 * systemctl gained a new --dry-run switch that shows what would be done
805 instead of doing it, and is currently supported by the shutdown and
808 * ConditionSecurity= can now detect the TOMOYO security module.
810 * Unit file [Install] sections are now also respected in unit drop-in
811 files. This is intended to be used by drop-ins under /usr/lib/.
813 * systemd-firstboot may now also set the initial keyboard mapping.
815 * Udev "changed" events for devices which are exposed as systemd
816 .device units are now propagated to units specified in
817 ReloadPropagatedFrom= as reload requests.
819 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
820 unit template name (i.e. a name in the form of 'foobar@.service',
821 without the instance component between the '@' and - the '.'), then
822 the escaped sysfs path of the device is automatically used as the
825 * SystemCallFilter= in unit files has been extended so that an "errno"
826 can be specified individually for each system call. Example:
827 SystemCallFilter=~uname:EILSEQ.
829 * The cgroup delegation logic has been substantially updated. Delegate=
830 now optionally takes a list of controllers (instead of a boolean, as
831 before), which lists the controllers to delegate at least.
833 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
835 * A new LogLevelMax= setting configures the maximum log level any
836 process of the service may log at (i.e. anything with a lesser
837 priority than what is specified is automatically dropped). A new
838 LogExtraFields= setting allows configuration of additional journal
839 fields to attach to all log records generated by any of the unit's
842 * New StandardInputData= and StandardInputText= settings along with the
843 new option StandardInput=data may be used to configure textual or
844 binary data that shall be passed to the executed service process via
845 standard input, encoded in-line in the unit file.
847 * StandardInput=, StandardOutput= and StandardError= may now be used to
848 connect stdin/stdout/stderr of executed processes directly with a
849 file or AF_UNIX socket in the file system, using the new "file:" option.
851 * A new unit file option CollectMode= has been added, that allows
852 tweaking the garbage collection logic for units. It may be used to
853 tell systemd to garbage collect units that have failed automatically
854 (normally it only GCs units that exited successfully). systemd-run
855 and systemd-mount expose this new functionality with a new -G option.
857 * "machinectl bind" may now be used to bind mount non-directories
858 (i.e. regularfiles, devices, fifos, sockets).
860 * systemd-analyze gained a new verb "calendar" for validating and
861 testing calendar time specifications to use for OnCalendar= in timer
862 units. Besides validating the expression it will calculate the next
863 time the specified expression would elapse.
865 * In addition to the pre-existing FailureAction= unit file setting
866 there's now SuccessAction=, for configuring a shutdown action to
867 execute when a unit completes successfully. This is useful in
868 particular inside containers that shall terminate after some workload
869 has been completed. Also, both options are now supported for all unit
870 types, not just services.
872 * networkds's IP rule support gained two new options
873 IncomingInterface= and OutgoingInterface= for configuring the incoming
874 and outgoing interfaces of configured rules. systemd-networkd also
875 gained support for "vxcan" network devices.
877 * networkd gained a new setting RequiredForOnline=, taking a
878 boolean. If set, systemd-wait-online will take it into consideration
879 when determining that the system is up, otherwise it will ignore the
880 interface for this purpose.
882 * The sd_notify() protocol gained support for a new operation: with
883 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
884 store again, ahead of POLLHUP or POLLERR when they are removed
887 * A new document doc/UIDS-GIDS.md has been added to the source tree,
888 that documents the UID/GID range and assignment assumptions and
889 requirements of systemd.
891 * The watchdog device PID 1 will ping may now be configured through the
892 WatchdogDevice= configuration file setting, or by setting the
893 systemd.watchdog_service= kernel commandline option.
895 * systemd-resolved's gained support for registering DNS-SD services on
896 the local network using MulticastDNS. Services may either be
897 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
898 the same dir below /run, /usr/lib), or through its D-Bus API.
900 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
901 extend the effective start, runtime, and stop time. The service must
902 continue to send EXTEND_TIMEOUT_USEC within the period specified to
903 prevent the service manager from making the service as timedout.
905 * elogind-resolved's DNSSEC support gained support for RFC 8080
906 (Ed25519 keys and signatures).
908 * The elogind-resolve command line tool gained a new set of options
909 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
910 --set-nta= and --revert to configure per-interface DNS configuration
911 dynamically during runtime. It's useful for pushing DNS information
912 into elogind-resolved from DNS hook scripts that various interface
913 managing software supports (such as pppd).
915 * elogind-nspawn gained a new --network-namespace-path= command line
916 option, which may be used to make a container join an existing
917 network namespace, by specifying a path to a "netns" file.
919 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
920 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
921 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
922 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
923 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
924 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
925 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
926 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
927 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
928 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
929 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
930 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
931 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
932 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
933 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
934 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
935 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
936 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
937 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
938 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
939 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
940 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
941 Jędrzejewski-Szmek, Zeal Jagannatha
947 * INCOMPATIBILITY: systemd-logind.service and other long-running
948 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
949 communication with the outside. This generally improves security of
950 the system, and is in almost all cases a safe and good choice, as
951 these services do not and should not provide any network-facing
952 functionality. However, systemd-logind uses the glibc NSS API to
953 query the user database. This creates problems on systems where NSS
954 is set up to directly consult network services for user database
955 lookups. In particular, this creates incompatibilities with the
956 "nss-nis" module, which attempts to directly contact the NIS/YP
957 network servers it is configured for, and will now consistently
958 fail. In such cases, it is possible to turn off IP sandboxing for
959 systemd-logind.service (set IPAddressDeny= in its [Service] section
960 to the empty string, via a .d/ unit file drop-in). Downstream
961 distributions might want to update their nss-nis packaging to include
962 such a drop-in snippet, accordingly, to hide this incompatibility
963 from the user. Another option is to make use of glibc's nscd service
964 to proxy such network requests through a privilege-separated, minimal
965 local caching daemon, or to switch to more modern technologies such
966 sssd, whose NSS hook-ups generally do not involve direct network
967 access. In general, we think it's definitely time to question the
968 implementation choices of nss-nis, i.e. whether it's a good idea
969 today to embed a network-facing loadable module into all local
970 processes that need to query the user database, including the most
971 trivial and benign ones, such as "ls". For more details about
972 IPAddressDeny= see below.
974 * A new modprobe.d drop-in is now shipped by default that sets the
975 bonding module option max_bonds=0. This overrides the kernel default,
976 to avoid conflicts and ambiguity as to whether or not bond0 should be
977 managed by systemd-networkd or not. This resolves multiple issues
978 with bond0 properties not being applied, when bond0 is configured
979 with systemd-networkd. Distributors may choose to not package this,
980 however in that case users will be prevented from correctly managing
981 bond0 interface using systemd-networkd.
983 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
984 which print the logging level and target of the system manager. They
985 complement the existing "set-log-level" and "set-log-target" verbs
986 used to change those values.
988 * journald.conf gained a new boolean setting ReadKMsg= which defaults
989 to on. If turned off kernel log messages will not be read by
990 systemd-journald or included in the logs. It also gained a new
991 setting LineMax= for configuring the maximum line length in
992 STDOUT/STDERR log streams. The new default for this value is 48K, up
993 from the previous hardcoded 2048.
995 * A new unit setting RuntimeDirectoryPreserve= has been added, which
996 allows more detailed control of what to do with a runtime directory
997 configured with RuntimeDirectory= (i.e. a directory below /run or
998 $XDG_RUNTIME_DIR) after a unit is stopped.
1000 * The RuntimeDirectory= setting for units gained support for creating
1001 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
1002 one top-level directory.
1004 * Units gained new options StateDirectory=, CacheDirectory=,
1005 LogsDirectory= and ConfigurationDirectory= which are closely related
1006 to RuntimeDirectory= but manage per-service directories below
1007 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
1008 possible to write unit files which when activated automatically gain
1009 properly owned service specific directories in these locations, thus
1010 making unit files self-contained and increasing compatibility with
1011 stateless systems and factory reset where /etc or /var are
1012 unpopulated at boot. Matching these new settings there's also
1013 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1014 ConfigurationDirectoryMode= for configuring the access mode of these
1015 directories. These settings are particularly useful in combination
1016 with DynamicUser=yes as they provide secure, properly-owned,
1017 writable, and stateful locations for storage, excluded from the
1018 sandbox that such services live in otherwise.
1020 * Automake support has been removed from this release. systemd is now
1023 * systemd-journald will now aggressively cache client metadata during
1024 runtime, speeding up log write performance under pressure. This comes
1025 at a small price though: as much of the metadata is read
1026 asynchronously from /proc/ (and isn't implicitly attached to log
1027 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1028 metadata stored alongside a log entry might be slightly
1029 out-of-date. Previously it could only be slightly newer than the log
1030 message. The time window is small however, and given that the kernel
1031 is unlikely to be improved anytime soon in this regard, this appears
1034 * nss-myhostname/systemd-resolved will now by default synthesize an
1035 A/AAAA resource record for the "_gateway" hostname, pointing to the
1036 current default IP gateway. Previously it did that for the "gateway"
1037 name, hampering adoption, as some distributions wanted to leave that
1038 host name open for local use. The old behaviour may still be
1039 requested at build time.
1041 * systemd-networkd's [Address] section in .network files gained a new
1042 Scope= setting for configuring the IP address scope. The [Network]
1043 section gained a new boolean setting ConfigureWithoutCarrier= that
1044 tells systemd-networkd to ignore link sensing when configuring the
1045 device. The [DHCP] section gained a new Anonymize= boolean option for
1046 turning on a number of options suggested in RFC 7844. A new
1047 [RoutingPolicyRule] section has been added for configuring the IP
1048 routing policy. The [Route] section has gained support for a new
1049 Type= setting which permits configuring
1050 blackhole/unreachable/prohibit routes.
1052 * The [VRF] section in .netdev files gained a new Table= setting for
1053 configuring the routing table to use. The [Tunnel] section gained a
1054 new Independent= boolean field for configuring tunnels independent of
1055 an underlying network interface. The [Bridge] section gained a new
1056 GroupForwardMask= option for configuration of propagation of link
1057 local frames between bridge ports.
1059 * The WakeOnLan= setting in .link files gained support for a number of
1060 new modes. A new TCP6SegmentationOffload= setting has been added for
1061 configuring TCP/IPv6 hardware segmentation offload.
1063 * The IPv6 RA sender implementation may now optionally send out RDNSS
1064 and RDNSSL records to supply DNS configuration to peers.
1066 * systemd-nspawn gained support for a new --system-call-filter= command
1067 line option for adding and removing entries in the default system
1068 call filter it applies. Moreover systemd-nspawn has been changed to
1069 implement a system call whitelist instead of a blacklist.
1071 * systemd-run gained support for a new --pipe command line option. If
1072 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1073 are directly passed on to the activated transient service
1074 executable. This allows invoking arbitrary processes as systemd
1075 services (for example to take benefit of dependency management,
1076 accounting management, resource management or log management that is
1077 done automatically for services) — while still allowing them to be
1078 integrated in a classic UNIX shell pipeline.
1080 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1081 using ReloadPropagationTo= is configured, a reload is now propagated
1082 to configured units. (Previously this was only done on explicitly
1083 requested reloads, using "systemctl reload" or an equivalent
1086 * For each service unit a restart counter is now kept: it is increased
1087 each time the service is restarted due to Restart=, and may be
1088 queried using "systemctl show -p NRestarts …".
1090 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1091 @signal and @timer have been added, for usage with SystemCallFilter=
1092 in unit files and the new --system-call-filter= command line option
1093 of systemd-nspawn (see above).
1095 * ExecStart= lines in unit files gained two new modifiers: when a
1096 command line is prefixed with "!" the command will be executed as
1097 configured, except for the credentials applied by
1098 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1099 "+", but does still apply namespacing options unlike "+". There's
1100 also "!!" now, which is mostly identical, but becomes a NOP on
1101 systems that support ambient capabilities. This is useful to write
1102 unit files that work with ambient capabilities where possible but
1103 automatically fall back to traditional privilege dropping mechanisms
1104 on systems where this is not supported.
1106 * ListenNetlink= settings in socket units now support RDMA netlink
1109 * A new unit file setting LockPersonality= has been added which permits
1110 locking down the chosen execution domain ("personality") of a service
1113 * A new special target "getty-pre.target" has been added, which is
1114 ordered before all text logins, and may be used to order services
1115 before textual logins acquire access to the console.
1117 * systemd will now attempt to load the virtio-rng.ko kernel module very
1118 early on if a VM environment supporting this is detected. This should
1119 improve entropy during early boot in virtualized environments.
1121 * A _netdev option is now supported in /etc/crypttab that operates in a
1122 similar way as the same option in /etc/fstab: it permits configuring
1123 encrypted devices that need to be ordered after the network is up.
1124 Following this logic, two new special targets
1125 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1126 added that are to cryptsetup.target what remote-fs.target and
1127 remote-fs-pre.target are to local-fs.target.
1129 * Service units gained a new UnsetEnvironment= setting which permits
1130 unsetting specific environment variables for services that are
1131 normally passed to it (for example in order to mask out locale
1132 settings for specific services that can't deal with it).
1134 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1135 traffic accounting (packet count as well as byte count) is done for
1136 the service, and shown as part of "systemctl status" or "systemd-run
1139 * Service units acquired two new options IPAddressAllow= and
1140 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1141 for configuring a simple IP access control list for all sockets of
1142 the unit. These options are available also on .slice and .socket
1143 units, permitting flexible access list configuration for individual
1144 services as well as groups of services (as defined by a slice unit),
1145 including system-wide. Note that IP ACLs configured this way are
1146 enforced on every single IPv4 and IPv6 socket created by any process
1147 of the service unit, and apply to ingress as well as egress traffic.
1149 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
1150 structured log message is generated each time the unit is stopped,
1151 containing information about the consumed resources of this
1154 * A new setting KeyringMode= has been added to unit files, which may be
1155 used to control how the kernel keyring is set up for executed
1158 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1159 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1160 behaviour (that is: these commands return immediately after the
1161 operation was enqueued instead of waiting for the operation to
1162 complete). Previously, "systemctl poweroff" and "systemctl reboot"
1163 were asynchronous on systems using systemd-logind (i.e. almost
1164 always, and like they were on sysvinit), and the other three commands
1165 were unconditionally synchronous. With this release this is cleaned
1166 up, and callers will see the same asynchronous behaviour on all
1167 systems for all five operations.
1169 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1172 * .timer units now accept calendar specifications in other timezones
1173 than UTC or the local timezone.
1175 * The tmpfiles snippet var.conf has been changed to create
1176 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1177 the "utmp" group already, and it appears to be generally understood
1178 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1179 databases. Previously this was implemented correctly for all these
1180 databases excepts btmp, which has been opened up like this now
1181 too. Note that while the other databases are world-readable
1182 (i.e. 0644), btmp is not and remains more restrictive.
1184 * The systemd-resolve tool gained a new --reset-server-features
1185 switch. When invoked like this systemd-resolved will forget
1186 everything it learnt about the features supported by the configured
1187 upstream DNS servers, and restarts the feature probing logic on the
1188 next resolver look-up for them at the highest feature level
1191 * The status dump systemd-resolved sends to the logs upon receiving
1192 SIGUSR1 now also includes information about all DNS servers it is
1193 configured to use, and the features levels it probed for them.
1195 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1196 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
1197 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1198 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1199 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1200 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1201 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1202 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1203 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1204 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1205 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1206 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1207 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1208 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1209 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1210 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1211 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1212 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1213 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1214 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1216 — Berlin, 2017-10-06
1220 * Meson is now supported as build system in addition to Automake. It is
1221 our plan to remove Automake in one of our next releases, so that
1222 Meson becomes our exclusive build system. Hence, please start using
1223 the Meson build system in your downstream packaging. There's plenty
1224 of documentation around how to use Meson, the extremely brief
1227 ./autogen.sh && ./configure && make && sudo make install
1231 meson build && ninja -C build && sudo ninja -C build install
1233 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1234 which permits configuring a timeout on the time a job is
1235 running. This is particularly useful for setting timeouts on jobs for
1238 * Unit files gained two new options ConditionUser= and ConditionGroup=
1239 for conditionalizing units based on the identity of the user/group
1240 running a systemd user instance.
1242 * systemd-networkd now understands a new FlowLabel= setting in the
1243 [VXLAN] section of .network files, as well as a Priority= in
1244 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1245 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1246 gained support for configuration of GENEVE links, and IPv6 address
1247 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1249 * .link files now understand a new Port= setting.
1251 * systemd-networkd's DHCP support gained support for DHCP option 119
1252 (domain search list).
1254 * systemd-networkd gained support for serving IPv6 address ranges using
1255 the Router Advertisement protocol. The new .network configuration
1256 section [IPv6Prefix] may be used to configure the ranges to
1257 serve. This is implemented based on a new, minimal, native server
1258 implementation of RA.
1260 * journalctl's --output= switch gained support for a new parameter
1261 "short-iso-precise" for a mode where timestamps are shown as precise
1264 * systemd-udevd's "net_id" builtin may now generate stable network
1265 interface names from IBM PowerVM VIO devices as well as ACPI platform
1268 * MulticastDNS support in systemd-resolved may now be explicitly
1269 enabled/disabled using the new MulticastDNS= configuration file
1272 * systemd-resolved may now optionally use libidn2 instead of the libidn
1273 for processing internationalized domain names. Support for libidn2
1274 should be considered experimental and should not be enabled by
1277 * "machinectl pull-tar" and related call may now do verification of
1278 downloaded images using SUSE-style .sha256 checksum files in addition
1279 to the already existing support for validating using Ubuntu-style
1282 * sd-bus gained support for a new sd_bus_message_appendv() call which
1283 is va_list equivalent of sd_bus_message_append().
1285 * sd-boot gained support for validating images using SHIM/MOK.
1287 * The SMACK code learnt support for "onlycap".
1289 * systemd-mount --umount is now much smarter in figuring out how to
1290 properly unmount a device given its mount or device path.
1292 * The code to call libnss_dns as a fallback from libnss_resolve when
1293 the communication with systemd-resolved fails was removed. This
1294 fallback was redundant and interfered with the [!UNAVAIL=return]
1295 suffix. See nss-resolve(8) for the recommended configuration.
1297 * systemd-logind may now be restarted without losing state. It stores
1298 the file descriptors for devices it manages in the system manager
1299 using the FDSTORE= mechanism. Please note that further changes in
1300 other components may be required to make use of this (for example
1301 Xorg has code to listen for stops of systemd-logind and terminate
1302 itself when logind is stopped or restarted, in order to avoid using
1303 stale file descriptors for graphical devices, which is now
1304 counterproductive and must be reverted in order for restarts of
1305 systemd-logind to be safe. See
1306 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1308 * All kernel install plugins are called with the environment variable
1309 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1310 /etc/machine-id. If the file is missing or empty, the variable is
1311 empty and BOOT_DIR_ABS is the path of a temporary directory which is
1312 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
1313 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1315 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
1316 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1317 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1318 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1319 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
1320 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1321 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1322 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1323 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
1324 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1325 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1326 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1327 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1328 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1329 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1330 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1331 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1332 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1333 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1334 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1335 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1336 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1337 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1338 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1339 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
1340 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1341 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
1342 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1343 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1346 — Berlin, 2017-07-12
1350 * The "hybrid" control group mode has been modified to improve
1351 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1352 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1353 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1354 cgroups-v1 hierarchy), the only externally visible change being that
1355 the cgroups-v2 hierarchy is also mounted, to
1356 /sys/fs/cgroup/unified. This should provide a large degree of
1357 compatibility with "legacy" cgroups-v1, while taking benefit of the
1358 better management capabilities of cgroups-v2.
1360 * The default control group setup mode may be selected both a boot-time
1361 via a set of kernel command line parameters (specifically:
1362 systemd.unified_cgroup_hierarchy= and
1363 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1364 default selected on the configure command line
1365 (--with-default-hierarchy=). The upstream default is "hybrid"
1366 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1367 this will change in a future systemd version to be "unified" (pure
1368 cgroups-v2 mode). The third option for the compile time option is
1369 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1370 distributions to default to "hybrid" mode for release distributions,
1371 starting with v233. We recommend "unified" for development
1372 distributions (specifically: distributions such as Fedora's rawhide)
1373 as that's where things are headed in the long run. Use "legacy" for
1374 greatest stability and compatibility only.
1376 * Note one current limitation of "unified" and "hybrid" control group
1377 setup modes: the kernel currently does not permit the systemd --user
1378 instance (i.e. unprivileged code) to migrate processes between two
1379 disconnected cgroup subtrees, even if both are managed and owned by
1380 the user. This effectively means "systemd-run --user --scope" doesn't
1381 work when invoked from outside of any "systemd --user" service or
1382 scope. Specifically, it is not supported from session scopes. We are
1383 working on fixing this in a future systemd version. (See #3388 for
1384 further details about this.)
1386 * DBus policy files are now installed into /usr rather than /etc. Make
1387 sure your system has dbus >= 1.9.18 running before upgrading to this
1388 version, or override the install path with --with-dbuspolicydir= .
1390 * All python scripts shipped with systemd (specifically: the various
1391 tests written in Python) now require Python 3.
1393 * systemd unit tests can now run standalone (without the source or
1394 build directories), and can be installed into /usr/lib/systemd/tests/
1395 with 'make install-tests'.
1397 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1398 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1401 * Support for the %c, %r, %R specifiers in unit files has been
1402 removed. Specifiers are not supposed to be dependent on configuration
1403 in the unit file itself (so that they resolve the same regardless
1404 where used in the unit files), but these specifiers were influenced
1405 by the Slice= option.
1407 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1408 all cases. If distributions want to use a different shell for this
1409 purpose (for example Fedora's /sbin/sushell) they need to specify
1410 this explicitly at configure time using --with-debug-shell=.
1412 * The confirmation spawn prompt has been reworked to offer the
1415 (c)ontinue, proceed without asking anymore
1416 (D)ump, show the state of the unit
1417 (f)ail, don't execute the command and pretend it failed
1419 (i)nfo, show a short summary of the unit
1420 (j)obs, show jobs that are in progress
1421 (s)kip, don't execute the command and pretend it succeeded
1422 (y)es, execute the command
1424 The 'n' choice for the confirmation spawn prompt has been removed,
1425 because its meaning was confusing.
1427 The prompt may now also be redirected to an alternative console by
1428 specifying the console as parameter to systemd.confirm_spawn=.
1430 * Services of Type=notify require a READY=1 notification to be sent
1431 during startup. If no such message is sent, the service now fails,
1432 even if the main process exited with a successful exit code.
1434 * Services that fail to start up correctly now always have their
1435 ExecStopPost= commands executed. Previously, they'd enter "failed"
1436 state directly, without executing these commands.
1438 * The option MulticastDNS= of network configuration files has acquired
1439 an actual implementation. With MulticastDNS=yes a host can resolve
1440 names of remote hosts and reply to mDNS A and AAAA requests.
1442 * When units are about to be started an additional check is now done to
1443 ensure that all dependencies of type BindsTo= (when used in
1444 combination with After=) have been started.
1446 * systemd-analyze gained a new verb "syscall-filter" which shows which
1447 system call groups are defined for the SystemCallFilter= unit file
1448 setting, and which system calls they contain.
1450 * A new system call filter group "@filesystem" has been added,
1451 consisting of various file system related system calls. Group
1452 "@reboot" has been added, covering reboot, kexec and shutdown related
1453 calls. Finally, group "@swap" has been added covering swap
1454 configuration related calls.
1456 * A new unit file option RestrictNamespaces= has been added that may be
1457 used to restrict access to the various process namespace types the
1458 Linux kernel provides. Specifically, it may be used to take away the
1459 right for a service unit to create additional file system, network,
1460 user, and other namespaces. This sandboxing option is particularly
1461 relevant due to the high amount of recently discovered namespacing
1462 related vulnerabilities in the kernel.
1464 * systemd-udev's .link files gained support for a new AutoNegotiation=
1465 setting for configuring Ethernet auto-negotiation.
1467 * systemd-networkd's .network files gained support for a new
1468 ListenPort= setting in the [DHCP] section to explicitly configure the
1469 UDP client port the DHCP client shall listen on.
1471 * .network files gained a new Unmanaged= boolean setting for explicitly
1472 excluding one or more interfaces from management by systemd-networkd.
1474 * The systemd-networkd ProxyARP= option has been renamed to
1475 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1476 renamed to ReduceARPProxy=. The old names continue to be available
1479 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1480 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1482 * systemd-networkd's bonding device support gained support for two new
1483 configuration options ActiveSlave= and PrimarySlave=.
1485 * The various options in the [Match] section of .network files gained
1486 support for negative matching.
1488 * New systemd-specific mount options are now understood in /etc/fstab:
1490 x-systemd.mount-timeout= may be used to configure the maximum
1491 permitted runtime of the mount command.
1493 x-systemd.device-bound may be set to bind a mount point to its
1494 backing device unit, in order to automatically remove a mount point
1495 if its backing device is unplugged. This option may also be
1496 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1497 on the block device, which is now automatically set for all CDROM
1498 drives, so that mounted CDs are automatically unmounted when they are
1499 removed from the drive.
1501 x-systemd.after= and x-systemd.before= may be used to explicitly
1502 order a mount after or before another unit or mount point.
1504 * Enqueued start jobs for device units are now automatically garbage
1505 collected if there are no jobs waiting for them anymore.
1507 * systemctl list-jobs gained two new switches: with --after, for every
1508 queued job the jobs it's waiting for are shown; with --before the
1509 jobs which it's blocking are shown.
1511 * systemd-nspawn gained support for ephemeral boots from disk images
1512 (or in other words: --ephemeral and --image= may now be
1513 combined). Moreover, ephemeral boots are now supported for normal
1514 directories, even if the backing file system is not btrfs. Of course,
1515 if the file system does not support file system snapshots or
1516 reflinks, the initial copy operation will be relatively expensive, but
1517 this should still be suitable for many use cases.
1519 * Calendar time specifications in .timer units now support
1520 specifications relative to the end of a month by using "~" instead of
1521 "-" as separator between month and day. For example, "*-02~03" means
1522 "the third last day in February". In addition a new syntax for
1523 repeated events has been added using the "/" character. For example,
1524 "9..17/2:00" means "every two hours from 9am to 5pm".
1526 * systemd-socket-proxyd gained a new parameter --connections-max= for
1527 configuring the maximum number of concurrent connections.
1529 * sd-id128 gained a new API for generating unique IDs for the host in a
1530 way that does not leak the machine ID. Specifically,
1531 sd_id128_get_machine_app_specific() derives an ID based on the
1532 machine ID a in well-defined, non-reversible, stable way. This is
1533 useful whenever an identifier for the host is needed but where the
1534 identifier shall not be useful to identify the system beyond the
1535 scope of the application itself. (Internally this uses HMAC-SHA256 as
1536 keyed hash function using the machine ID as input.)
1538 * NotifyAccess= gained a new supported value "exec". When set
1539 notifications are accepted from all processes systemd itself invoked,
1540 including all control processes.
1542 * .nspawn files gained support for defining overlay mounts using the
1543 Overlay= and OverlayReadOnly= options. Previously this functionality
1544 was only available on the systemd-nspawn command line.
1546 * systemd-nspawn's --bind= and --overlay= options gained support for
1547 bind/overlay mounts whose source lies within the container tree by
1548 prefixing the source path with "+".
1550 * systemd-nspawn's --bind= and --overlay= options gained support for
1551 automatically allocating a temporary source directory in /var/tmp
1552 that is removed when the container dies. Specifically, if the source
1553 directory is specified as empty string this mechanism is selected. An
1554 example usage is --overlay=+/var::/var, which creates an overlay
1555 mount based on the original /var contained in the image, overlayed
1556 with a temporary directory in the host's /var/tmp. This way changes
1557 to /var are automatically flushed when the container shuts down.
1559 * systemd-nspawn --image= option does now permit raw file system block
1560 devices (in addition to images containing partition tables, as
1563 * The disk image dissection logic in systemd-nspawn gained support for
1564 automatically setting up LUKS encrypted as well as Verity protected
1565 partitions. When a container is booted from an encrypted image the
1566 passphrase is queried at start-up time. When a container with Verity
1567 data is started, the root hash is search in a ".roothash" file
1568 accompanying the disk image (alternatively, pass the root hash via
1569 the new --root-hash= command line option).
1571 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1572 be used to dissect disk images the same way as systemd-nspawn does
1573 it, following the Bootable Partition Specification. It may even be
1574 used to mount disk images with complex partition setups (including
1575 LUKS and Verity partitions) to a local host directory, in order to
1576 inspect them. This tool is not considered public API (yet), and is
1577 thus not installed into /usr/bin. Please do not rely on its
1578 existence, since it might go away or be changed in later systemd
1581 * A new generator "systemd-verity-generator" has been added, similar in
1582 style to "systemd-cryptsetup-generator", permitting automatic setup of
1583 Verity root partitions when systemd boots up. In order to make use of
1584 this your partition setup should follow the Discoverable Partitions
1585 Specification, and the GPT partition ID of the root file system
1586 partition should be identical to the upper 128bit of the Verity root
1587 hash. The GPT partition ID of the Verity partition protecting it
1588 should be the lower 128bit of the Verity root hash. If the partition
1589 image follows this model it is sufficient to specify a single
1590 "roothash=" kernel command line argument to both configure which root
1591 image and verity partition to use as well as the root hash for
1592 it. Note that systemd-nspawn's Verity support follows the same
1593 semantics, meaning that disk images with proper Verity data in place
1594 may be booted in containers with systemd-nspawn as well as on
1595 physical systems via the verity generator. Also note that the "mkosi"
1596 tool available at https://github.com/systemd/mkosi has been updated
1597 to generate Verity protected disk images following this scheme. In
1598 fact, it has been updated to generate disk images that optionally
1599 implement a complete UEFI SecureBoot trust chain, involving a signed
1600 kernel and initrd image that incorporates such a root hash as well as
1601 a Verity-enabled root partition.
1603 * The hardware database (hwdb) udev supports has been updated to carry
1604 accelerometer quirks.
1606 * All system services are now run with a fresh kernel keyring set up
1607 for them. The invocation ID is stored by default in it, thus
1608 providing a safe, non-overridable way to determine the invocation
1611 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1612 options for bind mounting arbitrary paths in a service-specific
1613 way. When these options are used, arbitrary host or service files and
1614 directories may be mounted to arbitrary locations in the service's
1617 * Documentation has been added that lists all of systemd's low-level
1618 environment variables:
1620 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
1622 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1623 whether a specific socket file descriptor matches a specified socket
1626 * systemd-firstboot has been updated to check for the
1627 systemd.firstboot= kernel command line option. It accepts a boolean
1628 and when set to false the first boot questions are skipped.
1630 * systemd-fstab-generator has been updated to check for the
1631 systemd.volatile= kernel command line option, which either takes an
1632 optional boolean parameter or the special value "state". If used the
1633 system may be booted in a "volatile" boot mode. Specifically,
1634 "systemd.volatile" is used, the root directory will be mounted as
1635 tmpfs, and only /usr is mounted from the actual root file system. If
1636 "systemd.volatile=state" is used, the root directory will be mounted
1637 as usual, but /var is mounted as tmpfs. This concept provides similar
1638 functionality as systemd-nspawn's --volatile= option, but provides it
1639 on physical boots. Use this option for implementing stateless
1640 systems, or testing systems with all state and/or configuration reset
1641 to the defaults. (Note though that many distributions are not
1642 prepared to boot up without a populated /etc or /var, though.)
1644 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1645 partitions. Previously it only supported LUKS encrypted partitions
1646 for all other uses, except for the root partition itself.
1648 * Socket units gained support for listening on AF_VSOCK sockets for
1649 communication in virtualized QEMU environments.
1651 * The "configure" script gained a new option --with-fallback-hostname=
1652 for specifying the fallback hostname to use if none is configured in
1653 /etc/hostname. For example, by specifying
1654 --with-fallback-hostname=fedora it is possible to default to a
1655 hostname of "fedora" on pristine installations.
1657 * systemd-cgls gained support for a new --unit= switch for listing only
1658 the control groups of a specific unit. Similar --user-unit= has been
1659 added for listing only the control groups of a specific user unit.
1661 * systemd-mount gained a new --umount switch for unmounting a mount or
1662 automount point (and all mount/automount points below it).
1664 * systemd will now refuse full configuration reloads (via systemctl
1665 daemon-reload and related calls) unless at least 16MiB of free space
1666 are available in /run. This is a safety precaution in order to ensure
1667 that generators can safely operate after the reload completed.
1669 * A new unit file option RootImage= has been added, which has a similar
1670 effect as RootDirectory= but mounts the service's root directory from
1671 a disk image instead of plain directory. This logic reuses the same
1672 image dissection and mount logic that systemd-nspawn already uses,
1673 and hence supports any disk images systemd-nspawn supports, including
1674 those following the Discoverable Partition Specification, as well as
1675 Verity enabled images. This option enables systemd to run system
1676 services directly off disk images acting as resource bundles,
1677 possibly even including full integrity data.
1679 * A new MountAPIVFS= unit file option has been added, taking a boolean
1680 argument. If enabled /proc, /sys and /dev (collectively called the
1681 "API VFS") will be mounted for the service. This is only relevant if
1682 RootDirectory= or RootImage= is used for the service, as these mounts
1683 are of course in place in the host mount namespace anyway.
1685 * systemd-nspawn gained support for a new --pivot-root= switch. If
1686 specified the root directory within the container image is pivoted to
1687 the specified mount point, while the original root disk is moved to a
1688 different place. This option enables booting of ostree images
1689 directly with systemd-nspawn.
1691 * The systemd build scripts will no longer complain if the NTP server
1692 addresses are not changed from the defaults. Google now supports
1693 these NTP servers officially. We still recommend downstreams to
1694 properly register an NTP pool with the NTP pool project though.
1696 * coredumpctl gained a new "--reverse" option for printing the list
1697 of coredumps in reverse order.
1699 * coredumpctl will now show additional information about truncated and
1700 inaccessible coredumps, as well as coredumps that are still being
1701 processed. It also gained a new --quiet switch for suppressing
1702 additional informational message in its output.
1704 * coredumpctl gained support for only showing coredumps newer and/or
1705 older than specific timestamps, using the new --since= and --until=
1706 options, reminiscent of journalctl's options by the same name.
1708 * The systemd-coredump logic has been improved so that it may be reused
1709 to collect backtraces in non-compiled languages, for example in
1710 scripting languages such as Python.
1712 * machinectl will now show the UID shift of local containers, if user
1713 namespacing is enabled for them.
1715 * systemd will now optionally run "environment generator" binaries at
1716 configuration load time. They may be used to add environment
1717 variables to the environment block passed to services invoked. One
1718 user environment generator is shipped by default that sets up
1719 environment variables based on files dropped into /etc/environment.d
1720 and ~/.config/environment.d/.
1722 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1725 * hostnamed has been updated to report a new chassis type of
1726 "convertible" to cover "foldable" laptops that can both act as a
1727 tablet and as a laptop, such as various Lenovo Yoga devices.
1729 Contributions from: Adrián López, Alexander Galanin, Alexander
1730 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1731 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1732 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1733 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1734 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1735 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1736 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1737 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
1738 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1739 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1740 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1741 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1742 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1743 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1744 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1745 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1746 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1747 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1748 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1749 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1750 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1751 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1752 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1753 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1754 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1755 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1756 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1759 — Berlin, 2017-03-01
1763 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1764 RestrictAddressFamilies= enabled. These sandboxing options should
1765 generally be compatible with the various external udev call-out
1766 binaries we are aware of, however there may be exceptions, in
1767 particular when exotic languages for these call-outs are used. In
1768 this case, consider turning off these settings locally.
1770 * The new RemoveIPC= option can be used to remove IPC objects owned by
1771 the user or group of a service when that service exits.
1773 * The new ProtectKernelModules= option can be used to disable explicit
1774 load and unload operations of kernel modules by a service. In
1775 addition access to /usr/lib/modules is removed if this option is set.
1777 * ProtectSystem= option gained a new value "strict", which causes the
1778 whole file system tree with the exception of /dev, /proc, and /sys,
1779 to be remounted read-only for a service.
1781 * The new ProtectKernelTunables= option can be used to disable
1782 modification of configuration files in /sys and /proc by a service.
1783 Various directories and files are remounted read-only, so access is
1784 restricted even if the file permissions would allow it.
1786 * The new ProtectControlGroups= option can be used to disable write
1787 access by a service to /sys/fs/cgroup.
1789 * Various systemd services have been hardened with
1790 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1791 RestrictAddressFamilies=.
1793 * Support for dynamically creating users for the lifetime of a service
1794 has been added. If DynamicUser=yes is specified, user and group IDs
1795 will be allocated from the range 61184..65519 for the lifetime of the
1796 service. They can be resolved using the new nss-systemd.so NSS
1797 module. The module must be enabled in /etc/nsswitch.conf. Services
1798 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1799 any resources allocated by the service will be cleaned up when the
1800 service exits. They also have ProtectHome=read-only and
1801 ProtectSystem=strict enabled, so they are not able to make any
1802 permanent modifications to the system.
1804 * The nss-systemd module also always resolves root and nobody, making
1805 it possible to have no /etc/passwd or /etc/group files in minimal
1806 container or chroot environments.
1808 * Services may be started with their own user namespace using the new
1809 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1810 under which the service is running are mapped. All other users are
1813 * Support for the cgroup namespace has been added to systemd-nspawn. If
1814 supported by kernel, the container system started by systemd-nspawn
1815 will have its own view of the cgroup hierarchy. This new behaviour
1816 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1818 * The new MemorySwapMax= option can be used to limit the maximum swap
1819 usage under the unified cgroup hierarchy.
1821 * Support for the CPU controller in the unified cgroup hierarchy has
1822 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1823 options. This controller requires out-of-tree patches for the kernel
1824 and the support is provisional.
1826 * Mount and automount units may now be created transiently
1827 (i.e. dynamically at runtime via the bus API, instead of requiring
1828 unit files in the file system).
1830 * systemd-mount is a new tool which may mount file systems – much like
1831 mount(8), optionally pulling in additional dependencies through
1832 transient .mount and .automount units. For example, this tool
1833 automatically runs fsck on a backing block device before mounting,
1834 and allows the automount logic to be used dynamically from the
1835 command line for establishing mount points. This tool is particularly
1836 useful when dealing with removable media, as it will ensure fsck is
1837 run – if necessary – before the first access and that the file system
1838 is quickly unmounted after each access by utilizing the automount
1839 logic. This maximizes the chance that the file system on the
1840 removable media stays in a clean state, and if it isn't in a clean
1841 state is fixed automatically.
1843 * LazyUnmount=yes option for mount units has been added to expose the
1844 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1847 * /efi will be used as the mount point of the EFI boot partition, if
1848 the directory is present, and the mount point was not configured
1849 through other means (e.g. fstab). If /efi directory does not exist,
1850 /boot will be used as before. This makes it easier to automatically
1851 mount the EFI partition on systems where /boot is used for something
1854 * When operating on GPT disk images for containers, systemd-nspawn will
1855 now mount the ESP to /boot or /efi according to the same rules as PID
1856 1 running on a host. This allows tools like "bootctl" to operate
1857 correctly within such containers, in order to make container images
1858 bootable on physical systems.
1860 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
1862 * Two new user session targets have been added to support running
1863 graphical sessions under the systemd --user instance:
1864 graphical-session.target and graphical-session-pre.target. See
1865 systemd.special(7) for a description of how those targets should be
1868 * The vconsole initialization code has been significantly reworked to
1869 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
1870 support unicode keymaps. Font and keymap configuration will now be
1871 copied to all allocated virtual consoles.
1873 * FreeBSD's bhyve virtualization is now detected.
1875 * Information recorded in the journal for core dumps now includes the
1876 contents of /proc/mountinfo and the command line of the process at
1877 the top of the process hierarchy (which is usually the init process
1880 * systemd-journal-gatewayd learned the --directory= option to serve
1881 files from the specified location.
1883 * journalctl --root=… can be used to peruse the journal in the
1884 /var/log/ directories inside of a container tree. This is similar to
1885 the existing --machine= option, but does not require the container to
1888 * The hardware database has been extended to support
1889 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1892 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1893 specify the click rate for mice which include a horizontal wheel with
1894 a click rate that is different than the one for the vertical wheel.
1896 * systemd-run gained a new --wait option that makes service execution
1897 synchronous. (Specifically, the command will not return until the
1898 specified service binary exited.)
1900 * systemctl gained a new --wait option that causes the start command to
1901 wait until the units being started have terminated again.
1903 * A new journal output mode "short-full" has been added which displays
1904 timestamps with abbreviated English day names and adds a timezone
1905 suffix. Those timestamps include more information than the default
1906 "short" output mode, and can be passed directly to journalctl's
1907 --since= and --until= options.
1909 * /etc/resolv.conf will be bind-mounted into containers started by
1910 systemd-nspawn, if possible, so any changes to resolv.conf contents
1911 are automatically propagated to the container.
1913 * The number of instances for socket-activated services originating
1914 from a single IP address can be limited with
1915 MaxConnectionsPerSource=, extending the existing setting of
1918 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1921 * .netdev and .network configuration can now be extended through
1924 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1925 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1926 can be enabled and disabled using the new UDPSegmentationOffload=,
1927 TCPSegmentationOffload=, GenericSegmentationOffload=,
1928 GenericReceiveOffload=, LargeReceiveOffload= options in the
1929 [Link] section of .link files.
1931 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1932 Port VLAN ID can be configured for bridge devices using the new STP=,
1933 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1934 section of .netdev files.
1936 * The route table to which routes received over DHCP or RA should be
1937 added can be configured with the new RouteTable= option in the [DHCP]
1938 and [IPv6AcceptRA] sections of .network files.
1940 * The Address Resolution Protocol can be disabled on links managed by
1941 systemd-networkd using the ARP=no setting in the [Link] section of
1944 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1945 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1946 encode information about the result and exit codes of the current
1947 service runtime cycle.
1949 * systemd-sysctl will now configure kernel parameters in the order
1950 they occur in the configuration files. This matches what sysctl
1951 has been traditionally doing.
1953 * kernel-install "plugins" that are executed to perform various
1954 tasks after a new kernel is added and before an old one is removed
1955 can now return a special value to terminate the procedure and
1956 prevent any later plugins from running.
1958 * Journald's SplitMode=login setting has been deprecated. It has been
1959 removed from documentation, and its use is discouraged. In a future
1960 release it will be completely removed, and made equivalent to current
1961 default of SplitMode=uid.
1963 * Storage=both option setting in /etc/systemd/coredump.conf has been
1964 removed. With fast LZ4 compression storing the core dump twice is not
1967 * The --share-system systemd-nspawn option has been replaced with an
1968 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1969 this functionality is discouraged. In addition the variables
1970 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1971 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1972 individual namespaces.
1974 * "machinectl list" now shows the IP address of running containers in
1975 the output, as well as OS release information.
1977 * "loginctl list" now shows the TTY of each session in the output.
1979 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1980 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1981 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1982 tracking objects in a "recursive" mode, where a single client can be
1983 counted multiple times, if it takes multiple references.
1985 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
1986 sd_bus_get_exit_on_disconnect(). They may be used to make a
1987 process using sd-bus automatically exit if the bus connection is
1990 * Bus clients of the service manager may now "pin" loaded units into
1991 memory, by taking an explicit reference on them. This is useful to
1992 ensure the client can retrieve runtime data about the service even
1993 after the service completed execution. Taking such a reference is
1994 available only for privileged clients and should be helpful to watch
1995 running services in a race-free manner, and in particular collect
1996 information about exit statuses and results.
1998 * The nss-resolve module has been changed to strictly return UNAVAIL
1999 when communication via D-Bus with resolved failed, and NOTFOUND when
2000 a lookup completed but was negative. This means it is now possible to
2001 neatly configure fallbacks using nsswitch.conf result checking
2002 expressions. Taking benefit of this, the new recommended
2003 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
2005 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
2007 * A new setting CtrlAltDelBurstAction= has been added to
2008 /etc/systemd/system.conf which may be used to configure the precise
2009 behaviour if the user on the console presses Ctrl-Alt-Del more often
2010 than 7 times in 2s. Previously this would unconditionally result in
2011 an expedited, immediate reboot. With this new setting the precise
2012 operation may be configured in more detail, and also turned off
2015 * In .netdev files two new settings RemoteChecksumTx= and
2016 RemoteChecksumRx= are now understood that permit configuring the
2017 remote checksumming logic for VXLAN networks.
2019 * The service manager learnt a new "invocation ID" concept for invoked
2020 services. Each runtime cycle of a service will get a new invocation
2021 ID (a 128bit random UUID) assigned that identifies the current
2022 run of the service uniquely and globally. A new invocation ID
2023 is generated each time a service starts up. The journal will store
2024 the invocation ID of a service along with any logged messages, thus
2025 making the invocation ID useful for matching the online runtime of a
2026 service with the offline log data it generated in a safe way without
2027 relying on synchronized timestamps. In many ways this new service
2028 invocation ID concept is similar to the kernel's boot ID concept that
2029 uniquely and globally identifies the runtime of each boot. The
2030 invocation ID of a service is passed to the service itself via an
2031 environment variable ($INVOCATION_ID). A new bus call
2032 GetUnitByInvocationID() has been added that is similar to GetUnit()
2033 but instead of retrieving the bus path for a unit by its name
2034 retrieves it by its invocation ID. The returned path is valid only as
2035 long as the passed invocation ID is current.
2037 * systemd-resolved gained a new "DNSStubListener" setting in
2038 resolved.conf. It either takes a boolean value or the special values
2039 "udp" and "tcp", and configures whether to enable the stub DNS
2040 listener on 127.0.0.53:53.
2042 * IP addresses configured via networkd may now carry additional
2043 configuration settings supported by the kernel. New options include:
2044 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2045 PrefixRoute=, AutoJoin=.
2047 * The PAM configuration fragment file for "user@.service" shipped with
2048 systemd (i.e. the --user instance of systemd) has been stripped to
2049 the minimum necessary to make the system boot. Previously, it
2050 contained Fedora-specific stanzas that did not apply to other
2051 distributions. It is expected that downstream distributions add
2052 additional configuration lines, matching their needs to this file,
2053 using it only as rough template of what systemd itself needs. Note
2054 that this reduced fragment does not even include an invocation of
2055 pam_limits which most distributions probably want to add, even though
2056 systemd itself does not need it. (There's also the new build time
2057 option --with-pamconfdir=no to disable installation of the PAM
2060 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2061 capability is now also dropped from its set (in addition to
2062 CAP_SYS_MKNOD as before).
2064 * In service unit files it is now possible to connect a specific named
2065 file descriptor with stdin/stdout/stdout of an executed service. The
2066 name may be specified in matching .socket units using the
2067 FileDescriptorName= setting.
2069 * A number of journal settings may now be configured on the kernel
2070 command line. Specifically, the following options are now understood:
2071 systemd.journald.max_level_console=,
2072 systemd.journald.max_level_store=,
2073 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2074 systemd.journald.max_level_wall=.
2076 * "systemctl is-enabled --full" will now show by which symlinks a unit
2077 file is enabled in the unit dependency tree.
2079 * Support for VeraCrypt encrypted partitions has been added to the
2080 "cryptsetup" logic and /etc/crypttab.
2082 * systemd-detect-virt gained support for a new --private-users switch
2083 that checks whether the invoking processes are running inside a user
2084 namespace. Similar, a new special value "private-users" for the
2085 existing ConditionVirtualization= setting has been added, permitting
2086 skipping of specific units in user namespace environments.
2088 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2089 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2090 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2091 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2092 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2093 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2094 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2095 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2096 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2097 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2098 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2099 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2100 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2101 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2102 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2103 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2104 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2105 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2106 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2107 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2108 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2109 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2110 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2111 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2112 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2113 Jędrzejewski-Szmek, Zeal Jagannatha
2115 — Santa Fe, 2016-11-03
2119 * In service units the various ExecXYZ= settings have been extended
2120 with an additional special character as first argument of the
2121 assigned value: if the character '+' is used the specified command
2122 line it will be run with full privileges, regardless of User=,
2123 Group=, CapabilityBoundingSet= and similar options. The effect is
2124 similar to the existing PermissionsStartOnly= option, but allows
2125 configuration of this concept for each executed command line
2128 * Services may now alter the service watchdog timeout at runtime by
2129 sending a WATCHDOG_USEC= message via sd_notify().
2131 * MemoryLimit= and related unit settings now optionally take percentage
2132 specifications. The percentage is taken relative to the amount of
2133 physical memory in the system (or in case of containers, the assigned
2134 amount of memory). This allows scaling service resources neatly with
2135 the amount of RAM available on the system. Similarly, systemd-logind's
2136 RuntimeDirectorySize= option now also optionally takes percentage
2139 * In similar fashion TasksMax= takes percentage values now, too. The
2140 value is taken relative to the configured maximum number of processes
2141 on the system. The per-service task maximum has been changed to 15%
2142 using this functionality. (Effectively this is an increase of 512 →
2143 4915 for service units, given the kernel's default pid_max setting.)
2145 * Calendar time specifications in .timer units now understand a ".."
2146 syntax for time ranges. Example: "4..7:10" may now be used for
2147 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2150 * The InaccessableDirectories=, ReadOnlyDirectories= and
2151 ReadWriteDirectories= unit file settings have been renamed to
2152 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2153 applied to all kinds of file nodes, and not just directories, with
2154 the exception of symlinks. Specifically these settings may now be
2155 used on block and character device nodes, UNIX sockets and FIFOS as
2156 well as regular files. The old names of these settings remain
2157 available for compatibility.
2159 * systemd will now log about all service processes it kills forcibly
2160 (using SIGKILL) because they remained after the clean shutdown phase
2161 of the service completed. This should help identifying services that
2162 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2163 systemd-logind's configuration a similar log message is generated for
2164 processes killed at the end of each session due to this setting.
2166 * systemd will now set the $JOURNAL_STREAM environment variable for all
2167 services whose stdout/stderr are connected to the Journal (which
2168 effectively means by default: all services). The variable contains
2169 the device and inode number of the file descriptor used for
2170 stdout/stderr. This may be used by invoked programs to detect whether
2171 their stdout/stderr is connected to the Journal, in which case they
2172 can switch over to direct Journal communication, thus being able to
2173 pass extended, structured metadata along with their log messages. As
2174 one example, this is now used by glib's logging primitives.
2176 * When using systemd's default tmp.mount unit for /tmp, the mount point
2177 will now be established with the "nosuid" and "nodev" options. This
2178 avoids privilege escalation attacks that put traps and exploits into
2179 /tmp. However, this might cause problems if you e. g. put container
2180 images or overlays into /tmp; if you need this, override tmp.mount's
2181 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2184 * systemd now supports the "memory" cgroup controller also on
2187 * The systemd-cgtop tool now optionally takes a control group path as
2188 command line argument. If specified, the control group list shown is
2189 limited to subgroups of that group.
2191 * The SystemCallFilter= unit file setting gained support for
2192 pre-defined, named system call filter sets. For example
2193 SystemCallFilter=@clock is now an effective way to make all clock
2194 changing-related system calls unavailable to a service. A number of
2195 similar pre-defined groups are defined. Writing system call filters
2196 for system services is simplified substantially with this new
2197 concept. Accordingly, all of systemd's own, long-running services now
2198 enable system call filtering based on this, by default.
2200 * A new service setting MemoryDenyWriteExecute= has been added, taking
2201 a boolean value. If turned on, a service may no longer create memory
2202 mappings that are writable and executable at the same time. This
2203 enhances security for services where this is enabled as it becomes
2204 harder to dynamically write and then execute memory in exploited
2205 service processes. This option has been enabled for all of systemd's
2206 own long-running services.
2208 * A new RestrictRealtime= service setting has been added, taking a
2209 boolean argument. If set the service's processes may no longer
2210 acquire realtime scheduling. This improves security as realtime
2211 scheduling may otherwise be used to easily freeze the system.
2213 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2214 value. This may be used for requesting that the system manager inside
2215 of the container reports start-up completion to nspawn which then
2216 propagates this notification further to the service manager
2217 supervising nspawn itself. A related option NotifyReady= in .nspawn
2218 files has been added too. This functionality allows ordering of the
2219 start-up of multiple containers using the usual systemd ordering
2222 * machinectl gained a new command "stop" that is an alias for
2225 * systemd-resolved gained support for contacting DNS servers on
2226 link-local IPv6 addresses.
2228 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2229 its caches. A method call for requesting the same operation has been
2230 added to the bus API too, and is made available via "systemd-resolve
2233 * systemd-resolve gained a new --status switch. If passed a brief
2234 summary of the used DNS configuration with per-interface information
2237 * resolved.conf gained a new Cache= boolean option, defaulting to
2238 on. If turned off local DNS caching is disabled. This comes with a
2239 performance penalty in particular when DNSSEC is enabled. Note that
2240 resolved disables its internal caching implicitly anyway, when the
2241 configured DNS server is on a host-local IP address such as ::1 or
2242 127.0.0.1, thus automatically avoiding double local caching.
2244 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2245 for DNS requests. This improves compatibility with local programs
2246 that do not use the libc NSS or systemd-resolved's bus APIs for name
2247 resolution. This minimal DNS service is only available to local
2248 programs and does not implement the full DNS protocol, but enough to
2249 cover local DNS clients. A new, static resolv.conf file, listing just
2250 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2251 now recommended to make /etc/resolv.conf a symlink to this file in
2252 order to route all DNS lookups to systemd-resolved, regardless if
2253 done via NSS, the bus API or raw DNS packets. Note that this local
2254 DNS service is not as fully featured as the libc NSS or
2255 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2256 used to deliver link-local address information (as this implies
2257 sending a local interface index along), LLMNR/mDNS support via this
2258 interface is severely restricted. It is thus strongly recommended for
2259 all applications to use the libc NSS API or native systemd-resolved
2262 * systemd-networkd's bridge support learned a new setting
2263 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2264 in .network files has been added for configuring VLAN bridging in
2265 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2267 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2268 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2269 now be acquired without relying on DHCPv6. Two new options
2270 UseDomains= and UseDNS= have been added to configure this behaviour.
2272 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2273 renamed IPv6AcceptRA=, without altering its behaviour. The old
2274 setting name remains available for compatibility reasons.
2276 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2277 Key=, InputKey= and OutputKey=.
2279 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2280 interface configuration.
2282 * "systemctl edit" may now be used to create new unit files by
2283 specifying the --force switch.
2285 * sd-event gained a new function sd_event_get_iteration() for
2286 requesting the current iteration counter of the event loop. It starts
2287 at zero and is increased by one with each event loop iteration.
2289 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2290 file. It can be used in lieu of %systemd_requires in packages which
2291 don't use any systemd functionality and are intended to be installed
2292 in minimal containers without systemd present. This macro provides
2293 ordering dependencies to ensure that if the package is installed in
2294 the same rpm transaction as systemd, systemd will be installed before
2295 the scriptlets for the package are executed, allowing unit presets
2298 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2299 been added to simplify packaging of generators.
2301 * The os-release file gained VERSION_CODENAME field for the
2302 distribution nickname (e.g. VERSION_CODENAME=woody).
2304 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2305 can be set to disable parsing of metadata and the creation
2306 of persistent symlinks for that device.
2308 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2309 to make them available to logged-in users has been reverted.
2311 * Much of the common code of the various systemd components is now
2312 built into an internal shared library libsystemd-shared-231.so
2313 (incorporating the systemd version number in the name, to be updated
2314 with future releases) that the components link to. This should
2315 decrease systemd footprint both in memory during runtime and on
2316 disk. Note that the shared library is not for public use, and is
2317 neither API nor ABI stable, but is likely to change with every new
2318 released update. Packagers need to make sure that binaries
2319 linking to libsystemd-shared.so are updated in step with the
2322 * Configuration for "mkosi" is now part of the systemd
2323 repository. mkosi is a tool to easily build legacy-free OS images,
2324 and is available on github: https://github.com/systemd/mkosi. If
2325 "mkosi" is invoked in the build tree a new raw OS image is generated
2326 incorporating the systemd sources currently being worked on and a
2327 clean, fresh distribution installation. The generated OS image may be
2328 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
2329 UEFI PC. This functionality is particularly useful to easily test
2330 local changes made to systemd in a pristine, defined environment. See
2331 doc/HACKING for details.
2333 * configure learned the --with-support-url= option to specify the
2334 distribution's bugtracker.
2336 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2337 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2338 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2339 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2340 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2341 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2342 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2343 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2344 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2345 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2346 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2347 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2348 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2349 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2350 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2351 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
2352 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2353 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
2354 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
2356 — Berlin, 2016-07-25
2360 * DNSSEC is now turned on by default in systemd-resolved (in
2361 "allow-downgrade" mode), but may be turned off during compile time by
2362 passing "--with-default-dnssec=no" to "configure" (and of course,
2363 during runtime with DNSSEC= in resolved.conf). We recommend
2364 downstreams to leave this on at least during development cycles and
2365 report any issues with the DNSSEC logic upstream. We are very
2366 interested in collecting feedback about the DNSSEC validator and its
2367 limitations in the wild. Note however, that DNSSEC support is
2368 probably nothing downstreams should turn on in stable distros just
2369 yet, as it might create incompatibilities with a few DNS servers and
2370 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2371 automatically whenever we detect such incompatible setups, but there
2372 might be systems we do not cover yet. Hence: please help us testing
2373 the DNSSEC code, leave this on where you can, report back, but then
2374 again don't consider turning this on in your stable, LTS or
2375 production release just yet. (Note that you have to enable
2376 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2377 and its DNSSEC mode for host name resolution from local
2380 * systemd-resolve conveniently resolves DANE records with the --tlsa
2381 option and OPENPGPKEY records with the --openpgp option. It also
2382 supports dumping raw DNS record data via the new --raw= switch.
2384 * systemd-logind will now by default terminate user processes that are
2385 part of the user session scope unit (session-XX.scope) when the user
2386 logs out. This behavior is controlled by the KillUserProcesses=
2387 setting in logind.conf, and the previous default of "no" is now
2388 changed to "yes". This means that user sessions will be properly
2389 cleaned up after, but additional steps are necessary to allow
2390 intentionally long-running processes to survive logout.
2392 While the user is logged in at least once, user@.service is running,
2393 and any service that should survive the end of any individual login
2394 session can be started at a user service or scope using systemd-run.
2395 systemd-run(1) man page has been extended with an example which shows
2396 how to run screen in a scope unit underneath user@.service. The same
2397 command works for tmux.
2399 After the user logs out of all sessions, user@.service will be
2400 terminated too, by default, unless the user has "lingering" enabled.
2401 To effectively allow users to run long-term tasks even if they are
2402 logged out, lingering must be enabled for them. See loginctl(1) for
2403 details. The default polkit policy was modified to allow users to
2404 set lingering for themselves without authentication.
2406 Previous defaults can be restored at compile time by the
2407 --without-kill-user-processes option to "configure".
2409 * systemd-logind gained new configuration settings SessionsMax= and
2410 InhibitorsMax=, both with a default of 8192. It will not register new
2411 user sessions or inhibitors above this limit.
2413 * systemd-logind will now reload configuration on SIGHUP.
2415 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
2416 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
2417 enable. Also, support for the "io" cgroup controller in the unified
2418 hierarchy has been added, so that the "memory", "pids" and "io" are
2419 now the controllers that are supported on the unified hierarchy.
2421 WARNING: it is not possible to use previous systemd versions with
2422 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2423 is necessary to also update systemd in the initramfs if using the
2424 unified hierarchy. An updated SELinux policy is also required.
2426 * LLDP support has been extended, and both passive (receive-only) and
2427 active (sender) modes are supported. Passive mode ("routers-only") is
2428 enabled by default in systemd-networkd. Active LLDP mode is enabled
2429 by default for containers on the internal network. The "networkctl
2430 lldp" command may be used to list information gathered. "networkctl
2431 status" will also show basic LLDP information on connected peers now.
2433 * The IAID and DUID unique identifier sent in DHCP requests may now be
2434 configured for the system and each .network file managed by
2435 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2437 * systemd-networkd gained support for configuring proxy ARP support for
2438 each interface, via the ProxyArp= setting in .network files. It also
2439 gained support for configuring the multicast querier feature of
2440 bridge devices, via the new MulticastQuerier= setting in .netdev
2441 files. Similarly, snooping on the IGMP traffic can be controlled
2442 via the new setting MulticastSnooping=.
2444 A new setting PreferredLifetime= has been added for addresses
2445 configured in .network file to configure the lifetime intended for an
2448 The systemd-networkd DHCP server gained the option EmitRouter=, which
2449 defaults to yes, to configure whether the DHCP Option 3 (Router)
2452 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
2453 systemd-socket-activate and installed into /usr/bin. It is now fully
2456 * systemd-journald now uses separate threads to flush changes to disk
2457 when closing journal files, thus reducing impact of slow disk I/O on
2458 logging performance.
2460 * The sd-journal API gained two new calls
2461 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2462 can be used to open journal files using file descriptors instead of
2463 file or directory paths. sd_journal_open_container() has been
2464 deprecated, sd_journal_open_directory_fd() should be used instead
2465 with the flag SD_JOURNAL_OS_ROOT.
2467 * journalctl learned a new output mode "-o short-unix" that outputs log
2468 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2469 UTC). It also gained support for a new --no-hostname setting to
2470 suppress the hostname column in the family of "short" output modes.
2472 * systemd-ask-password now optionally skips printing of the password to
2473 stdout with --no-output which can be useful in scripts.
2475 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2476 (devices tagged with ID_MAKER_TOOL) are now tagged with
2477 "uaccess" and are available to logged in users.
2479 * The DeviceAllow= unit setting now supports specifiers (with "%").
2481 * "systemctl show" gained a new --value switch, which allows print a
2482 only the contents of a specific unit property, without also printing
2483 the property's name. Similar support was added to "show*" verbs
2484 of loginctl and machinectl that output "key=value" lists.
2486 * A new unit type "generated" was added for files dynamically generated
2487 by generator tools. Similarly, a new unit type "transient" is used
2488 for unit files created using the runtime API. "systemctl enable" will
2489 refuse to operate on such files.
2491 * A new command "systemctl revert" has been added that may be used to
2492 revert to the vendor version of a unit file, in case local changes
2493 have been made by adding drop-ins or overriding the unit file.
2495 * "machinectl clean" gained a new verb to automatically remove all or
2496 just hidden container images.
2498 * systemd-tmpfiles gained support for a new line type "e" for emptying
2499 directories, if they exist, without creating them if they don't.
2501 * systemd-nspawn gained support for automatically patching the UID/GIDs
2502 of the owners and the ACLs of all files and directories in a
2503 container tree to match the UID/GID user namespacing range selected
2504 for the container invocation. This mode is enabled via the new
2505 --private-users-chown switch. It also gained support for
2506 automatically choosing a free, previously unused UID/GID range when
2507 starting a container, via the new --private-users=pick setting (which
2508 implies --private-users-chown). Together, these options for the first
2509 time make user namespacing for nspawn containers fully automatic and
2510 thus deployable. The systemd-nspawn@.service template unit file has
2511 been changed to use this functionality by default.
2513 * systemd-nspawn gained a new --network-zone= switch, that allows
2514 creating ad-hoc virtual Ethernet links between multiple containers,
2515 that only exist as long as at least one container referencing them is
2516 running. This allows easy connecting of multiple containers with a
2517 common link that implements an Ethernet broadcast domain. Each of
2518 these network "zones" may be named relatively freely by the user, and
2519 may be referenced by any number of containers, but each container may
2520 only reference one of these "zones". On the lower level, this is
2521 implemented by an automatically managed bridge network interface for
2522 each zone, that is created when the first container referencing its
2523 zone is created and removed when the last one referencing its zone
2526 * The default start timeout may now be configured on the kernel command
2527 line via systemd.default_timeout_start_sec=. It was already
2528 configurable via the DefaultTimeoutStartSec= option in
2529 /etc/systemd/system.conf.
2531 * Socket units gained a new TriggerLimitIntervalSec= and
2532 TriggerLimitBurst= setting to configure a limit on the activation
2533 rate of the socket unit.
2535 * The LimitNICE= setting now optionally takes normal UNIX nice values
2536 in addition to the raw integer limit value. If the specified
2537 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2538 value is understood as UNIX nice value. If not prefixed like this it
2539 is understood as raw RLIMIT_NICE limit.
2541 * Note that the effect of the PrivateDevices= unit file setting changed
2542 slightly with this release: the per-device /dev file system will be
2543 mounted read-only from this version on, and will have "noexec"
2544 set. This (minor) change of behavior might cause some (exceptional)
2545 legacy software to break, when PrivateDevices=yes is set for its
2546 service. Please leave PrivateDevices= off if you run into problems
2549 * systemd-bootchart has been split out to a separate repository:
2550 https://github.com/systemd/systemd-bootchart
2552 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2553 merged into the kernel in its current form.
2555 * The compatibility libraries libsystemd-daemon.so,
2556 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2557 which have been deprecated since systemd-209 have been removed along
2558 with the corresponding pkg-config files. All symbols provided by
2559 those libraries are provided by libsystemd.so.
2561 * The Capabilities= unit file setting has been removed (it is ignored
2562 for backwards compatibility). AmbientCapabilities= and
2563 CapabilityBoundingSet= should be used instead.
2565 * A new special target has been added, initrd-root-device.target,
2566 which creates a synchronization point for dependencies of the root
2567 device in early userspace. Initramfs builders must ensure that this
2568 target is now included in early userspace.
2570 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2571 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2572 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2573 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2574 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2575 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2576 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2577 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
2578 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2579 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
2580 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2581 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2582 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
2583 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2584 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2585 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
2586 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2587 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2588 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2589 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2590 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2591 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
2592 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2593 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2594 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2597 — Fairfax, 2016-05-21
2601 * The systemd-resolved DNS resolver service has gained a substantial
2602 set of new features, most prominently it may now act as a DNSSEC
2603 validating stub resolver. DNSSEC mode is currently turned off by
2604 default, but is expected to be turned on by default in one of the
2605 next releases. For now, we invite everybody to test the DNSSEC logic
2606 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2607 service also gained a full set of D-Bus interfaces, including calls
2608 to configure DNS and DNSSEC settings per link (for use by external
2609 network management software). systemd-resolved and systemd-networkd
2610 now distinguish between "search" and "routing" domains. The former
2611 are used to qualify single-label names, the latter are used purely
2612 for routing lookups within certain domains to specific links.
2613 resolved now also synthesizes RRs for all entries from /etc/hosts.
2615 * The systemd-resolve tool (which is a client utility for
2616 systemd-resolved) has been improved considerably and is now fully
2617 supported and documented. Hence it has moved from /usr/lib/systemd to
2620 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2623 * The coredump collection logic has been reworked: when a coredump is
2624 collected it is now written to disk, compressed and processed
2625 (including stacktrace extraction) from a new instantiated service
2626 systemd-coredump@.service, instead of directly from the
2627 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2628 processing large coredumps can take up a substantial amount of
2629 resources and time, and this previously happened entirely outside of
2630 systemd's service supervision. With the new logic the core_pattern
2631 hook only does minimal metadata collection before passing off control
2632 to the new instantiated service, which is configured with a time
2633 limit, a nice level and other settings to minimize negative impact on
2634 the rest of the system. Also note that the new logic will honour the
2635 RLIMIT_CORE setting of the crashed process, which now allows users
2636 and processes to turn off coredumping for their processes by setting
2639 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2640 and all forked processes by default. Previously, PID 1 would leave
2641 the setting at "0" for all processes, as set by the kernel. Note that
2642 the resource limit traditionally has no effect on the generated
2643 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2644 logic is used. Since the limit is now honoured (see above) its
2645 default has been changed so that the coredumping logic is enabled by
2646 default for all processes, while allowing specific opt-out.
2648 * When the stacktrace is extracted from processes of system users, this
2649 is now done as "systemd-coredump" user, in order to sandbox this
2650 potentially security sensitive parsing operation. (Note that when
2651 processing coredumps of normal users this is done under the user ID
2652 of process that crashed, as before.) Packagers should take notice
2653 that it is now necessary to create the "systemd-coredump" system user
2654 and group at package installation time.
2656 * The systemd-activate socket activation testing tool gained support
2657 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2658 and --seqpacket switches. It also has been extended to support both
2659 new-style and inetd-style file descriptor passing. Use the new
2660 --inetd switch to request inetd-style file descriptor passing.
2662 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2663 variable, which takes a boolean value. If set to false, ANSI color
2664 output is disabled in the tools even when run on a terminal that
2667 * The VXLAN support in networkd now supports two new settings
2668 DestinationPort= and PortRange=.
2670 * A new systemd.machine_id= kernel command line switch has been added,
2671 that may be used to set the machine ID in /etc/machine-id if it is
2672 not initialized yet. This command line option has no effect if the
2673 file is already initialized.
2675 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2676 specified command line as PID 2 rather than PID 1 in the
2677 container. In this mode PID 1 is a minimal stub init process that
2678 implements the special POSIX and Linux semantics of PID 1 regarding
2679 signal and child process management. Note that this stub init process
2680 is implemented in nspawn itself and requires no support from the
2681 container image. This new logic is useful to support running
2682 arbitrary commands in the container, as normal processes are
2683 generally not prepared to run as PID 1.
2685 * systemd-nspawn gained a new --chdir= switch for setting the current
2686 working directory for the process started in the container.
2688 * "journalctl /dev/sda" will now output all kernel log messages for
2689 specified device from the current boot, in addition to all devices
2690 that are parents of it. This should make log output about devices
2691 pretty useful, as long as kernel drivers attach enough metadata to
2692 the log messages. (The usual SATA drivers do.)
2694 * The sd-journal API gained two new calls
2695 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2696 that report whether log data from /run or /var has been found.
2698 * journalctl gained a new switch "--fields" that prints all journal
2699 record field names currently in use in the journal. This is backed
2700 by two new sd-journal API calls sd_journal_enumerate_fields() and
2701 sd_journal_restart_fields().
2703 * Most configurable timeouts in systemd now expect an argument of
2704 "infinity" to turn them off, instead of "0" as before. The semantics
2705 from now on is that a timeout of "0" means "now", and "infinity"
2706 means "never". To maintain backwards compatibility, "0" continues to
2707 turn off previously existing timeout settings.
2709 * "systemctl reload-or-try-restart" has been renamed to "systemctl
2710 try-reload-or-restart" to clarify what it actually does: the "try"
2711 logic applies to both reloading and restarting, not just restarting.
2712 The old name continues to be accepted for compatibility.
2714 * On boot-up, when PID 1 detects that the system clock is behind the
2715 release date of the systemd version in use, the clock is now set
2716 to the latter. Previously, this was already done in timesyncd, in order
2717 to avoid running with clocks set to the various clock epochs such as
2718 1902, 1938 or 1970. With this change the logic is now done in PID 1
2719 in addition to timesyncd during early boot-up, so that it is enforced
2720 before the first process is spawned by systemd. Note that the logic
2721 in timesyncd remains, as it is more comprehensive and ensures
2722 clock monotonicity by maintaining a persistent timestamp file in
2723 /var. Since /var is generally not available in earliest boot or the
2724 initrd, this part of the logic remains in timesyncd, and is not done
2727 * Support for tweaking details in net_cls.class_id through the
2728 NetClass= configuration directive has been removed, as the kernel
2729 people have decided to deprecate that controller in cgroup v2.
2730 Userspace tools such as nftables are moving over to setting rules
2731 that are specific to the full cgroup path of a task, which obsoletes
2732 these controllers anyway. The NetClass= directive is kept around for
2733 legacy compatibility reasons. For a more in-depth description of the
2734 kernel change, please refer to the respective upstream commit:
2736 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2738 * A new service setting RuntimeMaxSec= has been added that may be used
2739 to specify a maximum runtime for a service. If the timeout is hit, the
2740 service is terminated and put into a failure state.
2742 * A new service setting AmbientCapabilities= has been added. It allows
2743 configuration of additional Linux process capabilities that are
2744 passed to the activated processes. This is only available on very
2747 * The process resource limit settings in service units may now be used
2748 to configure hard and soft limits individually.
2750 * The various libsystemd APIs such as sd-bus or sd-event now publicly
2751 expose support for gcc's __attribute__((cleanup())) C extension.
2752 Specifically, for many object destructor functions alternative
2753 versions have been added that have names suffixed with "p" and take a
2754 pointer to a pointer to the object to destroy, instead of just a
2755 pointer to the object itself. This is useful because these destructor
2756 functions may be used directly as parameters to the cleanup
2757 construct. Internally, systemd has been a heavy user of this GCC
2758 extension for a long time, and with this change similar support is
2759 now available to consumers of the library outside of systemd. Note
2760 that by using this extension in your sources compatibility with old
2761 and strictly ANSI compatible C compilers is lost. However, all gcc or
2762 LLVM versions of recent years support this extension.
2764 * Timer units gained support for a new setting RandomizedDelaySec= that
2765 allows configuring some additional randomized delay to the configured
2766 time. This is useful to spread out timer events to avoid load peaks in
2767 clusters or larger setups.
2769 * Calendar time specifications now support sub-second accuracy.
2771 * Socket units now support listening on SCTP and UDP-lite protocol
2774 * The sd-event API now comes with a full set of man pages.
2776 * Older versions of systemd contained experimental support for
2777 compressing journal files and coredumps with the LZ4 compressor that
2778 was not compatible with the lz4 binary (due to API limitations of the
2779 lz4 library). This support has been removed; only support for files
2780 compatible with the lz4 binary remains. This LZ4 logic is now
2781 officially supported and no longer considered experimental.
2783 * The dkr image import logic has been removed again from importd. dkr's
2784 micro-services focus doesn't fit into the machine image focus of
2785 importd, and quickly got out of date with the upstream dkr API.
2787 * Creation of the /run/lock/lockdev/ directory was dropped from
2788 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
2789 been available for many years. If you still need this, you need to
2790 create your own tmpfiles.d config file with:
2792 d /run/lock/lockdev 0775 root lock -
2794 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2795 and RebootArgument= have been moved from the [Service] section of
2796 unit files to [Unit], and they are now supported on all unit types,
2797 not just service units. Of course, elogind will continue to
2798 understand these settings also at the old location, in order to
2799 maintain compatibility.
2801 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2802 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2803 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2804 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2805 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2806 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2807 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2808 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2809 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2810 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2811 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2812 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2813 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2814 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2815 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2816 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2817 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2818 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2819 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2821 — Berlin, 2016-02-11
2825 * A number of properties previously only settable in unit
2826 files are now also available as properties to set when
2827 creating transient units programmatically via the bus, as it
2828 is exposed with systemd-run's --property=
2829 setting. Specifically, these are: SyslogIdentifier=,
2830 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2831 EnvironmentFile=, ReadWriteDirectories=,
2832 ReadOnlyDirectories=, InaccessibleDirectories=,
2833 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2835 * When creating transient services via the bus API it is now
2836 possible to pass in a set of file descriptors to use as
2837 STDIN/STDOUT/STDERR for the invoked process.
2839 * Slice units may now be created transiently via the bus APIs,
2840 similar to the way service and scope units may already be
2841 created transiently.
2843 * Wherever systemd expects a calendar timestamp specification
2844 (like in journalctl's --since= and --until= switches) UTC
2845 timestamps are now supported. Timestamps suffixed with "UTC"
2846 are now considered to be in Universal Time Coordinated
2847 instead of the local timezone. Also, timestamps may now
2848 optionally be specified with sub-second accuracy. Both of
2849 these additions also apply to recurring calendar event
2850 specification, such as OnCalendar= in timer units.
2852 * journalctl gained a new "--sync" switch that asks the
2853 journal daemon to write all so far unwritten log messages to
2854 disk and sync the files, before returning.
2856 * systemd-tmpfiles learned two new line types "q" and "Q" that
2857 operate like "v", but also set up a basic btrfs quota
2858 hierarchy when used on a btrfs file system with quota
2861 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2862 instead of a subvolume (even on a btrfs file system) if the
2863 root directory is a plain directory, and not a
2864 subvolume. This should simplify things with certain chroot()
2865 environments which are not aware of the concept of btrfs
2868 * systemd-detect-virt gained a new --chroot switch to detect
2869 whether execution takes place in a chroot() environment.
2871 * CPUAffinity= now takes CPU index ranges in addition to
2874 * The various memory-related resource limit settings (such as
2875 LimitAS=) now understand the usual K, M, G, ... suffixes to
2876 the base of 1024 (IEC). Similar, the time-related resource
2877 limit settings understand the usual min, h, day, ...
2880 * There's a new system.conf setting DefaultTasksMax= to
2881 control the default TasksMax= setting for services and
2882 scopes running on the system. (TasksMax= is the primary
2883 setting that exposes the "pids" cgroup controller on systemd
2884 and was introduced in the previous systemd release.) The
2885 setting now defaults to 512, which means services that are
2886 not explicitly configured otherwise will only be able to
2887 create 512 processes or threads at maximum, from this
2888 version on. Note that this means that thread- or
2889 process-heavy services might need to be reconfigured to set
2890 TasksMax= to a higher value. It is sufficient to set
2891 TasksMax= in these specific unit files to a higher value, or
2892 even "infinity". Similar, there's now a logind.conf setting
2893 UserTasksMax= that defaults to 4096 and limits the total
2894 number of processes or tasks each user may own
2895 concurrently. nspawn containers also have the TasksMax=
2896 value set by default now, to 8192. Note that all of this
2897 only has an effect if the "pids" cgroup controller is
2898 enabled in the kernel. The general benefit of these changes
2899 should be a more robust and safer system, that provides a
2900 certain amount of per-service fork() bomb protection.
2902 * systemd-nspawn gained the new --network-veth-extra= switch
2903 to define additional and arbitrarily-named virtual Ethernet
2904 links between the host and the container.
2906 * A new service execution setting PassEnvironment= has been
2907 added that allows importing select environment variables
2908 from PID1's environment block into the environment block of
2911 * Timer units gained support for a new RemainAfterElapse=
2912 setting which takes a boolean argument. It defaults to on,
2913 exposing behaviour unchanged to previous releases. If set to
2914 off, timer units are unloaded after they elapsed if they
2915 cannot elapse again. This is particularly useful for
2916 transient timer units, which shall not stay around longer
2917 than until they first elapse.
2919 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
2920 default now (the kernel default is 16). This is beneficial
2921 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
2922 allows substantially larger numbers of queued
2923 datagrams. This should increase the capability of systemd to
2924 parallelize boot-up, as logging and sd_notify() are unlikely
2925 to stall execution anymore. If you need to change the value
2926 from the new defaults, use the usual sysctl.d/ snippets.
2928 * The compression framing format used by the journal or
2929 coredump processing has changed to be in line with what the
2930 official LZ4 tools generate. LZ4 compression support in
2931 systemd was considered unsupported previously, as the format
2932 was not compatible with the normal tools. With this release
2933 this has changed now, and it is hence safe for downstream
2934 distributions to turn it on. While not compressing as well
2935 as the XZ, LZ4 is substantially faster, which makes
2936 it a good default choice for the compression logic in the
2937 journal and in coredump handling.
2939 * Any reference to /etc/mtab has been dropped from
2940 systemd. The file has been obsolete since a while, but
2941 systemd refused to work on systems where it was incorrectly
2942 set up (it should be a symlink or non-existent). Please make
2943 sure to update to util-linux 2.27.1 or newer in conjunction
2944 with this systemd release, which also drops any reference to
2945 /etc/mtab. If you maintain a distribution make sure that no
2946 software you package still references it, as this is a
2947 likely source of bugs. There's also a glibc bug pending,
2948 asking for removal of any reference to this obsolete file:
2950 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
2952 Note that only util-linux versions built with
2953 --enable-libmount-force-mountinfo are supported.
2955 * Support for the ".snapshot" unit type has been removed. This
2956 feature turned out to be little useful and little used, and
2957 has now been removed from the core and from systemctl.
2959 * The dependency types RequiresOverridable= and
2960 RequisiteOverridable= have been removed from systemd. They
2961 have been used only very sparingly to our knowledge and
2962 other options that provide a similar effect (such as
2963 systemctl --mode=ignore-dependencies) are much more useful
2964 and commonly used. Moreover, they were only half-way
2965 implemented as the option to control behaviour regarding
2966 these dependencies was never added to systemctl. By removing
2967 these dependency types the execution engine becomes a bit
2968 simpler. Unit files that use these dependencies should be
2969 changed to use the non-Overridable dependency types
2970 instead. In fact, when parsing unit files with these
2971 options, that's what systemd will automatically convert them
2972 too, but it will also warn, asking users to fix the unit
2973 files accordingly. Removal of these dependency types should
2974 only affect a negligible number of unit files in the wild.
2976 * Behaviour of networkd's IPForward= option changed
2977 (again). It will no longer maintain a per-interface setting,
2978 but propagate one way from interfaces where this is enabled
2979 to the global kernel setting. The global setting will be
2980 enabled when requested by a network that is set up, but
2981 never be disabled again. This change was made to make sure
2982 IPv4 and IPv6 behaviour regarding packet forwarding is
2983 similar (as the Linux IPv6 stack does not support
2984 per-interface control of this setting) and to minimize
2987 * In unit files the behaviour of %u, %U, %h, %s has
2988 changed. These specifiers will now unconditionally resolve
2989 to the various user database fields of the user that the
2990 systemd instance is running as, instead of the user
2991 configured in the specific unit via User=. Note that this
2992 effectively doesn't change much, as resolving of these
2993 specifiers was already turned off in the --system instance
2994 of systemd, as we cannot do NSS lookups from PID 1. In the
2995 --user instance of systemd these specifiers where correctly
2996 resolved, but hardly made any sense, since the user instance
2997 lacks privileges to do user switches anyway, and User= is
2998 hence useless. Moreover, even in the --user instance of
2999 systemd behaviour was awkward as it would only take settings
3000 from User= assignment placed before the specifier into
3001 account. In order to unify and simplify the logic around
3002 this the specifiers will now always resolve to the
3003 credentials of the user invoking the manager (which in case
3004 of PID 1 is the root user).
3006 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
3007 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
3008 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
3009 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3010 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3011 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3012 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3013 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3014 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3015 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3016 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3017 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3018 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3019 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3022 — Berlin, 2015-11-18
3026 * systemd now depends on util-linux v2.27. More specifically,
3027 the newly added mount monitor feature in libmount now
3028 replaces systemd's former own implementation.
3030 * libmount mandates /etc/mtab not to be regular file, and
3031 systemd now enforces this condition at early boot.
3032 /etc/mtab has been deprecated and warned about for a very
3033 long time, so systems running systemd should already have
3034 stopped having this file around as anything else than a
3035 symlink to /proc/self/mounts.
3037 * Support for the "pids" cgroup controller has been added. It
3038 allows accounting the number of tasks in a cgroup and
3039 enforcing limits on it. This adds two new setting
3040 TasksAccounting= and TasksMax= to each unit, as well as a
3041 global option DefaultTasksAccounting=.
3043 * Support for the "net_cls" cgroup controller has been added.
3044 It allows assigning a net class ID to each task in the
3045 cgroup, which can then be used in firewall rules and traffic
3046 shaping configurations. Note that the kernel netfilter net
3047 class code does not currently work reliably for ingress
3048 packets on unestablished sockets.
3050 This adds a new config directive called NetClass= to CGroup
3051 enabled units. Allowed values are positive numbers for fixed
3052 assignments and "auto" for picking a free value
3055 * 'systemctl is-system-running' now returns 'offline' if the
3056 system is not booted with systemd. This command can now be
3057 used as a substitute for 'systemd-notify --booted'.
3059 * Watchdog timeouts have been increased to 3 minutes for all
3060 in-tree service files. Apparently, disk IO issues are more
3061 frequent than we hoped, and user reported >1 minute waiting
3064 * 'machine-id-commit' functionality has been merged into
3065 'machine-id-setup --commit'. The separate binary has been
3068 * The WorkingDirectory= directive in unit files may now be set
3069 to the special value '~'. In this case, the working
3070 directory is set to the home directory of the user
3071 configured in User=.
3073 * "machinectl shell" will now open the shell in the home
3074 directory of the selected user by default.
3076 * The CrashChVT= configuration file setting is renamed to
3077 CrashChangeVT=, following our usual logic of not
3078 abbreviating unnecessarily. The old directive is still
3079 supported for compat reasons. Also, this directive now takes
3080 an integer value between 1 and 63, or a boolean value. The
3081 formerly supported '-1' value for disabling stays around for
3084 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3085 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3086 RootDirectory= properties can now be set for transient
3089 * The systemd-analyze tool gained a new "set-log-target" verb
3090 to change the logging target the system manager logs to
3091 dynamically during runtime. This is similar to how
3092 "systemd-analyze set-log-level" already changes the log
3095 * In nspawn /sys is now mounted as tmpfs, with only a selected
3096 set of subdirectories mounted in from the real sysfs. This
3097 enhances security slightly, and is useful for ensuring user
3098 namespaces work correctly.
3100 * Support for USB FunctionFS activation has been added. This
3101 allows implementation of USB gadget services that are
3102 activated as soon as they are requested, so that they don't
3103 have to run continuously, similar to classic socket
3106 * The "systemctl exit" command now optionally takes an
3107 additional parameter that sets the exit code to return from
3108 the systemd manager when exiting. This is only relevant when
3109 running the systemd user instance, or when running the
3110 system instance in a container.
3112 * sd-bus gained the new API calls sd_bus_path_encode_many()
3113 and sd_bus_path_decode_many() that allow easy encoding and
3114 decoding of multiple identifier strings inside a D-Bus
3115 object path. Another new call sd_bus_default_flush_close()
3116 has been added to flush and close per-thread default
3119 * systemd-cgtop gained support for a -M/--machine= switch to
3120 show the control groups within a certain container only.
3122 * "systemctl kill" gained support for an optional --fail
3123 switch. If specified the requested operation will fail of no
3124 processes have been killed, because the unit had no
3125 processes attached, or similar.
3127 * A new systemd.crash_reboot=1 kernel command line option has
3128 been added that triggers a reboot after crashing. This can
3129 also be set through CrashReboot= in systemd.conf.
3131 * The RuntimeDirectory= setting now understands unit
3132 specifiers like %i or %f.
3134 * A new (still internal) library API sd-ipv4acd has been added,
3135 that implements address conflict detection for IPv4. It's
3136 based on code from sd-ipv4ll, and will be useful for
3137 detecting DHCP address conflicts.
3139 * File descriptors passed during socket activation may now be
3140 named. A new API sd_listen_fds_with_names() is added to
3141 access the names. The default names may be overridden,
3142 either in the .socket file using the FileDescriptorName=
3143 parameter, or by passing FDNAME= when storing the file
3144 descriptors using sd_notify().
3146 * systemd-networkd gained support for:
3148 - Setting the IPv6 Router Advertisement settings via
3149 IPv6AcceptRouterAdvertisements= in .network files.
3151 - Configuring the HelloTimeSec=, MaxAgeSec= and
3152 ForwardDelaySec= bridge parameters in .netdev files.
3154 - Configuring PreferredSource= for static routes in
3157 * The "ask-password" framework used to query for LUKS harddisk
3158 passwords or SSL passwords during boot gained support for
3159 caching passwords in the kernel keyring, if it is
3160 available. This makes sure that the user only has to type in
3161 a passphrase once if there are multiple objects to unlock
3162 with the same one. Previously, such password caching was
3163 available only when Plymouth was used; this moves the
3164 caching logic into the systemd codebase itself. The
3165 "systemd-ask-password" utility gained a new --keyname=
3166 switch to control which kernel keyring key to use for
3167 caching a password in. This functionality is also useful for
3168 enabling display managers such as gdm to automatically
3169 unlock the user's GNOME keyring if its passphrase, the
3170 user's password and the harddisk password are the same, if
3171 gdm-autologin is used.
3173 * When downloading tar or raw images using "machinectl
3174 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3175 file is now also downloaded, if it is available and stored
3176 next to the image file.
3178 * Units of type ".socket" gained a new boolean setting
3179 Writable= which is only useful in conjunction with
3180 ListenSpecial=. If true, enables opening the specified
3181 special file in O_RDWR mode rather than O_RDONLY mode.
3183 * systemd-rfkill has been reworked to become a singleton
3184 service that is activated through /dev/rfkill on each rfkill
3185 state change and saves the settings to disk. This way,
3186 systemd-rfkill is now compatible with devices that exist
3187 only intermittendly, and even restores state if the previous
3188 system shutdown was abrupt rather than clean.
3190 * The journal daemon gained support for vacuuming old journal
3191 files controlled by the number of files that shall remain,
3192 in addition to the already existing control by size and by
3193 date. This is useful as journal interleaving performance
3194 degrades with too many separate journal files, and allows
3195 putting an effective limit on them. The new setting defaults
3196 to 100, but this may be changed by setting SystemMaxFiles=
3197 and RuntimeMaxFiles= in journald.conf. Also, the
3198 "journalctl" tool gained the new --vacuum-files= switch to
3199 manually vacuum journal files to leave only the specified
3200 number of files in place.
3202 * udev will now create /dev/disk/by-path links for ATA devices
3203 on kernels where that is supported.
3205 * Galician, Serbian, Turkish and Korean translations were added.
3207 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3208 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3209 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3210 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3211 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3212 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3213 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3214 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3215 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3216 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3217 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3218 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3219 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3220 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3221 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3222 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3223 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3224 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3226 — Berlin, 2015-10-07
3230 * The DHCP implementation of systemd-networkd gained a set of
3233 - The DHCP server now supports emitting DNS and NTP
3234 information. It may be enabled and configured via
3235 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3236 and NTP information is enabled, but no servers are
3237 configured, the corresponding uplink information (if there
3238 is any) is propagated.
3240 - Server and client now support transmission and reception
3241 of timezone information. It can be configured via the
3242 newly introduced network options UseTimezone=,
3243 EmitTimezone=, and Timezone=. Transmission of timezone
3244 information is enabled between host and containers by
3245 default now: the container will change its local timezone
3246 to what the host has set.
3248 - Lease timeouts can now be configured via
3249 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3251 - The DHCP server improved on the stability of
3252 leases. Clients are more likely to get the same lease
3253 information back, even if the server loses state.
3255 - The DHCP server supports two new configuration options to
3256 control the lease address pool metrics, PoolOffset= and
3259 * The encapsulation limit of tunnels in systemd-networkd may
3260 now be configured via 'EncapsulationLimit='. It allows
3261 modifying the maximum additional levels of encapsulation
3262 that are permitted to be prepended to a packet.
3264 * systemd now supports the concept of user buses replacing
3265 session buses, if used with dbus-1.10 (and enabled via dbus
3266 --enable-user-session). It previously only supported this on
3267 kdbus-enabled systems, and this release expands this to
3268 'dbus-daemon' systems.
3270 * systemd-networkd now supports predictable interface names
3273 * systemd now optionally supports the new Linux kernel
3274 "unified" control group hierarchy. If enabled via the kernel
3275 command-line option 'systemd.unified_cgroup_hierarchy=1',
3276 systemd will try to mount the unified cgroup hierarchy
3277 directly on /sys/fs/cgroup. If not enabled, or not
3278 available, systemd will fall back to the legacy cgroup
3279 hierarchy setup, as before. Host system and containers can
3280 mix and match legacy and unified hierarchies as they
3281 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
3282 environment variable to individually select the hierarchy to
3283 use for executed containers. By default, nspawn will use the
3284 unified hierarchy for the containers if the host uses the
3285 unified hierarchy, and the legacy hierarchy otherwise.
3286 Please note that at this point the unified hierarchy is an
3287 experimental kernel feature and is likely to change in one
3288 of the next kernel releases. Therefore, it should not be
3289 enabled by default in downstream distributions yet. The
3290 minimum required kernel version for the unified hierarchy to
3291 work is 4.2. Note that when the unified hierarchy is used
3292 for the first time delegated access to controllers is
3293 safe. Because of this systemd-nspawn containers will get
3294 access to controllers now, as will systemd user
3295 sessions. This means containers and user sessions may now
3296 manage their own resources, partitioning up what the system
3299 * A new special scope unit "init.scope" has been introduced
3300 that encapsulates PID 1 of the system. It may be used to
3301 determine resource usage and enforce resource limits on PID
3302 1 itself. PID 1 hence moved out of the root of the control
3305 * The cgtop tool gained support for filtering out kernel
3306 threads when counting tasks in a control group. Also, the
3307 count of processes is now recursively summed up by
3308 default. Two options -k and --recursive= have been added to
3309 revert to old behaviour. The tool has also been updated to
3310 work correctly in containers now.
3312 * systemd-nspawn's --bind= and --bind-ro= options have been
3313 extended to allow creation of non-recursive bind mounts.
3315 * libsystemd gained two new calls sd_pid_get_cgroup() and
3316 sd_peer_get_cgroup() which return the control group path of
3317 a process or peer of a connected AF_UNIX socket. This
3318 function call is particularly useful when implementing
3319 delegated subtrees support in the control group hierarchy.
3321 * The "sd-event" event loop API of libsystemd now supports
3322 correct dequeuing of real-time signals, without losing
3325 * When systemd requests a PolicyKit decision when managing
3326 units it will now add additional fields to the request,
3327 including unit name and desired operation. This enables more
3328 powerful PolicyKit policies, that make decisions depending
3329 on these parameters.
3331 * nspawn learnt support for .nspawn settings files, that may
3332 accompany the image files or directories of containers, and
3333 may contain additional settings for the container. This is
3334 an alternative to configuring container parameters via the
3335 nspawn command line.
3337 Contributions from: Cristian Rodríguez, Daniel Mack, David
3338 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3339 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3340 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3341 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3342 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3343 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
3344 Andersen, Tom Gundersen, Torstein Husebø
3346 — Berlin, 2015-09-08
3350 * machinectl gained a new verb 'shell' which opens a fresh
3351 shell on the target container or the host. It is similar to
3352 the existing 'login' command of machinectl, but spawns the
3353 shell directly without prompting for username or
3354 password. The pseudo machine '.host' now refers to the local
3355 host and is used by default. Hence, 'machinectl shell' can
3356 be used as replacement for 'su -' which spawns a session as
3357 a fresh systemd unit in a way that is fully isolated from
3358 the originating session.
3360 * systemd-networkd learned to cope with private-zone DHCP
3361 options and allows other programs to query the values.
3363 * SELinux access control when enabling/disabling units is no
3364 longer enforced with this release. The previous
3365 implementation was incorrect, and a new corrected
3366 implementation is not yet available. As unit file operations
3367 are still protected via PolicyKit and D-Bus policy this is
3368 not a security problem. Yet, distributions which care about
3369 optimal SELinux support should probably not stabilize on
3372 * sd-bus gained support for matches of type "arg0has=", that
3373 test for membership of strings in string arrays sent in bus
3376 * systemd-resolved now dumps the contents of its DNS and LLMNR
3377 caches to the logs on reception of the SIGUSR1 signal. This
3378 is useful to debug DNS behaviour.
3380 * The coredumpctl tool gained a new --directory= option to
3381 operate on journal files in a specific directory.
3383 * "systemctl reboot" and related commands gained a new
3384 "--message=" option which may be used to set a free-text
3385 wall message when shutting down or rebooting the
3386 system. This message is also logged, which is useful for
3387 figuring out the reason for a reboot or shutdown a
3390 * The "systemd-resolve-host" tool's -i switch now takes
3391 network interface numbers as alternative to interface names.
3393 * A new unit file setting for services has been introduced:
3394 UtmpMode= allows configuration of how precisely systemd
3395 handles utmp and wtmp entries for the service if this is
3396 enabled. This allows writing services that appear similar to
3397 user sessions in the output of the "w", "who", "last" and
3400 * systemd-resolved will now locally synthesize DNS resource
3401 records for the "localhost" and "gateway" domains as well as
3402 the local hostname. This should ensure that clients querying
3403 RRs via resolved will get similar results as those going via
3404 NSS, if nss-myhostname is enabled.
3406 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3407 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3408 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3409 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3410 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3411 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3412 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3413 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3414 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3415 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3416 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3417 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3419 — Berlin, 2015-08-27
3423 * The systemd-efi-boot-generator functionality was merged into
3424 systemd-gpt-auto-generator.
3426 * systemd-networkd now supports Group Policy for vxlan
3427 devices. It can be enabled via the new boolean configuration
3428 option called 'GroupPolicyExtension='.
3430 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3431 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3432 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3434 — Berlin, 2015-07-31
3438 * The python-systemd code has been removed from the systemd repository.
3439 A new repository has been created which accommodates the code from
3440 now on, and we kindly ask distributions to create a separate package
3441 for this: https://github.com/systemd/python-systemd
3443 * The systemd daemon will now reload its main configuration
3444 (/etc/systemd/system.conf) on daemon-reload.
3446 * sd-dhcp now exposes vendor specific extensions via
3447 sd_dhcp_lease_get_vendor_specific().
3449 * systemd-networkd gained a number of new configuration options.
3451 - A new boolean configuration option for TAP devices called
3452 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
3453 device, thus allowing to send and receive GSO packets.
3455 - A new tunnel configuration option called 'CopyDSCP='.
3456 If enabled, the DSCP field of ip6 tunnels is copied into the
3457 decapsulated packet.
3459 - A set of boolean bridge configuration options were added.
3460 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3461 and 'UnicastFlood=' are now parsed by networkd and applied to the
3462 respective bridge link device via the respective IFLA_BRPORT_*
3465 - A new string configuration option to override the hostname sent
3466 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3467 is true, networkd will use the configured hostname instead of the
3468 system hostname when sending DHCP requests.
3470 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3471 networkd will configure the IPv6 flow-label of the tunnel device
3472 according to RFC2460.
3474 - The 'macvtap' virtual network devices are now supported, similar to
3475 the already supported 'macvlan' devices.
3477 * systemd-resolved now implements RFC5452 to improve resilience against
3478 cache poisoning. Additionally, source port randomization is enabled
3479 by default to further protect against DNS spoofing attacks.
3481 * nss-mymachines now supports translating UIDs and GIDs of running
3482 containers with user-namespaces enabled. If a container 'foo'
3483 translates a host uid 'UID' to the container uid 'TUID', then
3484 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3485 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3486 mapped as 'vg-foo-TGID'.
3488 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
3489 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3490 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3491 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3492 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3493 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3494 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3495 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3496 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3497 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3499 — Berlin, 2015-07-29
3503 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3504 There are no known issues with current sysfs, and udev does not need
3505 or should be used to work around such bugs.
3507 * udev does no longer enable USB HID power management. Several reports
3508 indicate, that some devices cannot handle that setting.
3510 * The udev accelerometer helper was removed. The functionality
3511 is now fully included in iio-sensor-proxy. But this means,
3512 older iio-sensor-proxy versions will no longer provide
3513 accelerometer/orientation data with this systemd version.
3514 Please upgrade iio-sensor-proxy to version 1.0.
3516 * networkd gained a new configuration option IPv6PrivacyExtensions=
3517 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3518 for Stateless Address") on selected networks.
3520 * For the sake of fewer build-time dependencies and less code in the
3521 main repository, the python bindings are about to be removed in the
3522 next release. A new repository has been created which accommodates
3523 the code from now on, and we kindly ask distributions to create a
3524 separate package for this. The removal will take place in v223.
3526 https://github.com/systemd/python-systemd
3528 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3529 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3530 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3531 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
3532 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3533 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
3534 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3535 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
3536 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3537 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3539 — Berlin, 2015-07-07
3543 * The sd-bus.h and sd-event.h APIs have now been declared
3544 stable and have been added to the official interface of
3545 libsystemd.so. sd-bus implements an alternative D-Bus client
3546 library, that is relatively easy to use, very efficient and
3547 supports both classic D-Bus as well as kdbus as transport
3548 backend. sd-event is a generic event loop abstraction that
3549 is built around Linux epoll, but adds features such as event
3550 prioritization or efficient timer handling. Both APIs are good
3551 choices for C programs looking for a bus and/or event loop
3552 implementation that is minimal and does not have to be
3553 portable to other kernels.
3555 * kdbus support is no longer compile-time optional. It is now
3556 always built-in. However, it can still be disabled at
3557 runtime using the kdbus=0 kernel command line setting, and
3558 that setting may be changed to default to off, by specifying
3559 --disable-kdbus at build-time. Note though that the kernel
3560 command line setting has no effect if the kdbus.ko kernel
3561 module is not installed, in which case kdbus is (obviously)
3562 also disabled. We encourage all downstream distributions to
3563 begin testing kdbus by adding it to the kernel images in the
3564 development distributions, and leaving kdbus support in
3567 * The minimal required util-linux version has been bumped to
3570 * Support for chkconfig (--enable-chkconfig) was removed in
3571 favor of calling an abstraction tool
3572 /lib/systemd/systemd-sysv-install. This needs to be
3573 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3574 in README for details.
3576 * If there's a systemd unit and a SysV init script for the
3577 same service name, and the user executes "systemctl enable"
3578 for it (or a related call), then this will now enable both
3579 (or execute the related operation on both), not just the
3582 * The libudev API documentation has been converted from gtkdoc
3585 * gudev has been removed from the systemd tree, it is now an
3588 * The systemd-cgtop tool learnt a new --raw switch to generate
3589 "raw" (machine parsable) output.
3591 * networkd's IPForwarding= .network file setting learnt the
3592 new setting "kernel", which ensures that networkd does not
3593 change the IP forwarding sysctl from the default kernel
3596 * The systemd-logind bus API now exposes a new boolean
3597 property "Docked" that reports whether logind considers the
3598 system "docked", i.e. connected to a docking station or not.
3600 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3601 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3602 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3603 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3604 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3605 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3606 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3607 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3608 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3609 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3610 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
3611 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3612 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3613 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3614 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3615 Fink, Zbigniew Jędrzejewski-Szmek
3617 — Berlin, 2015-06-19
3621 * The gudev library has been extracted into a separate repository
3622 available at: https://git.gnome.org/browse/libgudev/
3623 It is now managed as part of the Gnome project. Distributions
3624 are recommended to pass --disable-gudev to systemd and use
3625 gudev from the Gnome project instead. gudev is still included
3626 in systemd, for now. It will be removed soon, though. Please
3627 also see the announcement-thread on systemd-devel:
3628 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
3630 * systemd now exposes a CPUUsageNSec= property for each
3631 service unit on the bus, that contains the overall consumed
3632 CPU time of a service (the sum of what each process of the
3633 service consumed). This value is only available if
3634 CPUAccounting= is turned on for a service, and is then shown
3635 in the "systemctl status" output.
3637 * Support for configuring alternative mappings of the old SysV
3638 runlevels to systemd targets has been removed. They are now
3639 hardcoded in a way that runlevels 2, 3, 4 all map to
3640 multi-user.target and 5 to graphical.target (which
3641 previously was already the default behaviour).
3643 * The auto-mounter logic gained support for mount point
3644 expiry, using a new TimeoutIdleSec= setting in .automount
3645 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3647 * The EFI System Partition (ESP) as mounted to /boot by
3648 systemd-efi-boot-generator will now be unmounted
3649 automatically after 2 minutes of not being used. This should
3650 minimize the risk of ESP corruptions.
3652 * New /etc/fstab options x-systemd.requires= and
3653 x-systemd.requires-mounts-for= are now supported to express
3654 additional dependencies for mounts. This is useful for
3655 journalling file systems that support external journal
3656 devices or overlay file systems that require underlying file
3657 systems to be mounted.
3659 * systemd does not support direct live-upgrades (via systemctl
3660 daemon-reexec) from versions older than v44 anymore. As no
3661 distribution we are aware of shipped such old versions in a
3662 stable release this should not be problematic.
3664 * When systemd forks off a new per-connection service instance
3665 it will now set the $REMOTE_ADDR environment variable to the
3666 remote IP address, and $REMOTE_PORT environment variable to
3667 the remote IP port. This behaviour is similar to the
3668 corresponding environment variables defined by CGI.
3670 * systemd-networkd gained support for uplink failure
3671 detection. The BindCarrier= option allows binding interface
3672 configuration dynamically to the link sense of other
3673 interfaces. This is useful to achieve behaviour like in
3676 * systemd-networkd gained support for configuring the DHCP
3677 client identifier to use when requesting leases.
3679 * systemd-networkd now has a per-network UseNTP= option to
3680 configure whether NTP server information acquired via DHCP
3681 is passed on to services like systemd-timesyncd.
3683 * systemd-networkd gained support for vti6 tunnels.
3685 * Note that systemd-networkd manages the sysctl variable
3686 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3687 it is configured for since v219. The variable controls IP
3688 forwarding, and is a per-interface alternative to the global
3689 /proc/sys/net/ipv[46]/ip_forward. This setting is
3690 configurable in the IPForward= option, which defaults to
3691 "no". This means if networkd is used for an interface it is
3692 no longer sufficient to set the global sysctl option to turn
3693 on IP forwarding! Instead, the .network file option
3694 IPForward= needs to be turned on! Note that the
3695 implementation of this behaviour was broken in v219 and has
3698 * Many bonding and vxlan options are now configurable in
3701 * systemd-nspawn gained a new --property= setting to set unit
3702 properties for the container scope. This is useful for
3703 setting resource parameters (e.g. "CPUShares=500") on
3704 containers started from the command line.
3706 * systemd-nspawn gained a new --private-users= switch to make
3707 use of user namespacing available on recent Linux kernels.
3709 * systemd-nspawn may now be called as part of a shell pipeline
3710 in which case the pipes used for stdin and stdout are passed
3711 directly to the process invoked in the container, without
3712 indirection via a pseudo tty.
3714 * systemd-nspawn gained a new switch to control the UNIX
3715 signal to use when killing the init process of the container
3718 * systemd-nspawn gained a new --overlay= switch for mounting
3719 overlay file systems into the container using the new kernel
3722 * When a container image is imported via systemd-importd and
3723 the host file system is not btrfs, a loopback block device
3724 file is created in /var/lib/machines.raw with a btrfs file
3725 system inside. It is then mounted to /var/lib/machines to
3726 enable btrfs features for container management. The loopback
3727 file and btrfs file system is grown as needed when container
3728 images are imported via systemd-importd.
3730 * systemd-machined/systemd-importd gained support for btrfs
3731 quota, to enforce container disk space limits on disk. This
3732 is exposed in "machinectl set-limit".
3734 * systemd-importd now can import containers from local .tar,
3735 .raw and .qcow2 images, and export them to .tar and .raw. It
3736 can also import dkr v2 images now from the network (on top
3739 * systemd-importd gained support for verifying downloaded
3740 images with gpg2 (previously only gpg1 was supported).
3742 * systemd-machined, systemd-logind, systemd: most bus calls
3743 are now accessible to unprivileged processes via
3744 PolicyKit. Also, systemd-logind will now allow users to kill
3745 their own sessions without further privileges or
3748 * systemd-shutdownd has been removed. This service was
3749 previously responsible for implementing scheduled shutdowns
3750 as exposed in /usr/bin/shutdown's time parameter. This
3751 functionality has now been moved into systemd-logind and is
3752 accessible via a bus interface.
3754 * "systemctl reboot" gained a new switch --firmware-setup that
3755 can be used to reboot into the EFI firmware setup, if that
3756 is available. systemd-logind now exposes an API on the bus
3757 to trigger such reboots, in case graphical desktop UIs want
3758 to cover this functionality.
3760 * "systemctl enable", "systemctl disable" and "systemctl mask"
3761 now support a new "--now" switch. If specified the units
3762 that are enabled will also be started, and the ones
3763 disabled/masked also stopped.
3765 * The Gummiboot EFI boot loader tool has been merged into
3766 systemd, and renamed to "systemd-boot". The bootctl tool has been
3767 updated to support systemd-boot.
3769 * An EFI kernel stub has been added that may be used to create
3770 kernel EFI binaries that contain not only the actual kernel,
3771 but also an initrd, boot splash, command line and OS release
3772 information. This combined binary can then be signed as a
3773 single image, so that the firmware can verify it all in one
3774 step. systemd-boot has special support for EFI binaries created
3775 like this and can extract OS release information from them
3776 and show them in the boot menu. This functionality is useful
3777 to implement cryptographically verified boot schemes.
3779 * Optional support has been added to systemd-fsck to pass
3780 fsck's progress report to an AF_UNIX socket in the file
3783 * udev will no longer create device symlinks for all block
3784 devices by default. A blacklist for excluding special block
3785 devices from this logic has been turned into a whitelist
3786 that requires picking block devices explicitly that require
3789 * A new (currently still internal) API sd-device.h has been
3790 added to libsystemd. This modernized API is supposed to
3791 replace libudev eventually. In fact, already much of libudev
3792 is now just a wrapper around sd-device.h.
3794 * A new hwdb database for storing metadata about pointing
3795 stick devices has been added.
3797 * systemd-tmpfiles gained support for setting file attributes
3798 similar to the "chattr" tool with new 'h' and 'H' lines.
3800 * systemd-journald will no longer unconditionally set the
3801 btrfs NOCOW flag on new journal files. This is instead done
3802 with tmpfiles snippet using the new 'h' line type. This
3803 allows easy disabling of this logic, by masking the
3804 journal-nocow.conf tmpfiles file.
3806 * systemd-journald will now translate audit message types to
3807 human readable identifiers when writing them to the
3808 journal. This should improve readability of audit messages.
3810 * The LUKS logic gained support for the offset= and skip=
3811 options in /etc/crypttab, as previously implemented by
3814 * /usr/lib/os-release gained a new optional field VARIANT= for
3815 distributions that support multiple variants (such as a
3816 desktop edition, a server edition, ...)
3818 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3819 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3820 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3821 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3822 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3823 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3824 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3825 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3826 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3827 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3828 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3829 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3830 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3831 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3832 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3833 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3834 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3835 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3836 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3837 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3838 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3839 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3840 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3841 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3842 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3843 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3844 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3846 — Berlin, 2015-05-22
3850 * Introduce a new API "sd-hwdb.h" for querying the hardware
3851 metadata database. With this minimal interface one can query
3852 and enumerate the udev hwdb, decoupled from the old libudev
3853 library. libudev's interface for this is now only a wrapper
3854 around sd-hwdb. A new tool systemd-hwdb has been added to
3855 interface with and update the database.
3857 * When any of systemd's tools copies files (for example due to
3858 tmpfiles' C lines) a btrfs reflink will attempted first,
3859 before bytewise copying is done.
3861 * systemd-nspawn gained a new --ephemeral switch. When
3862 specified a btrfs snapshot is taken of the container's root
3863 directory, and immediately removed when the container
3864 terminates again. Thus, a container can be started whose
3865 changes never alter the container's root directory, and are
3866 lost on container termination. This switch can also be used
3867 for starting a container off the root file system of the
3868 host without affecting the host OS. This switch is only
3869 available on btrfs file systems.
3871 * systemd-nspawn gained a new --template= switch. It takes the
3872 path to a container tree to use as template for the tree
3873 specified via --directory=, should that directory be
3874 missing. This allows instantiating containers dynamically,
3875 on first run. This switch is only available on btrfs file
3878 * When a .mount unit refers to a mount point on which multiple
3879 mounts are stacked, and the .mount unit is stopped all of
3880 the stacked mount points will now be unmounted until no
3881 mount point remains.
3883 * systemd now has an explicit notion of supported and
3884 unsupported unit types. Jobs enqueued for unsupported unit
3885 types will now fail with an "unsupported" error code. More
3886 specifically .swap, .automount and .device units are not
3887 supported in containers, .busname units are not supported on
3888 non-kdbus systems. .swap and .automount are also not
3889 supported if their respective kernel compile time options
3892 * machinectl gained support for two new "copy-from" and
3893 "copy-to" commands for copying files from a running
3894 container to the host or vice versa.
3896 * machinectl gained support for a new "bind" command to bind
3897 mount host directories into local containers. This is
3898 currently only supported for nspawn containers.
3900 * networkd gained support for configuring bridge forwarding
3901 database entries (fdb) from .network files.
3903 * A new tiny daemon "systemd-importd" has been added that can
3904 download container images in tar, raw, qcow2 or dkr formats,
3905 and make them available locally in /var/lib/machines, so
3906 that they can run as nspawn containers. The daemon can GPG
3907 verify the downloads (not supported for dkr, since it has no
3908 provisions for verifying downloads). It will transparently
3909 decompress bz2, xz, gzip compressed downloads if necessary,
3910 and restore sparse files on disk. The daemon uses privilege
3911 separation to ensure the actual download logic runs with
3912 fewer privileges than the daemon itself. machinectl has
3913 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3914 make the functionality of importd available to the
3915 user. With this in place the Fedora and Ubuntu "Cloud"
3916 images can be downloaded and booted as containers unmodified
3917 (the Fedora images lack the appropriate GPG signature files
3918 currently, so they cannot be verified, but this will change
3919 soon, hopefully). Note that downloading images is currently
3920 only fully supported on btrfs.
3922 * machinectl is now able to list container images found in
3923 /var/lib/machines, along with some metadata about sizes of
3924 disk and similar. If the directory is located on btrfs and
3925 quota is enabled, this includes quota display. A new command
3926 "image-status" has been added that shows additional
3927 information about images.
3929 * machinectl is now able to clone container images
3930 efficiently, if the underlying file system (btrfs) supports
3931 it, with the new "machinectl clone" command. It also
3932 gained commands for renaming and removing images, as well as
3933 marking them read-only or read-write (supported also on
3934 legacy file systems).
3936 * networkd gained support for collecting LLDP network
3937 announcements, from hardware that supports this. This is
3938 shown in networkctl output.
3940 * systemd-run gained support for a new -t (--pty) switch for
3941 invoking a binary on a pty whose input and output is
3942 connected to the invoking terminal. This allows executing
3943 processes as system services while interactively
3944 communicating with them via the terminal. Most interestingly
3945 this is supported across container boundaries. Invoking
3946 "systemd-run -t /bin/bash" is an alternative to running a
3947 full login session, the difference being that the former
3948 will not register a session, nor go through the PAM session
3951 * tmpfiles gained support for a new "v" line type for creating
3952 btrfs subvolumes. If the underlying file system is a legacy
3953 file system, this automatically degrades to creating a
3954 normal directory. Among others /var/lib/machines is now
3955 created like this at boot, should it be missing.
3957 * The directory /var/lib/containers/ has been deprecated and
3958 been replaced by /var/lib/machines. The term "machines" has
3959 been used in the systemd context as generic term for both
3960 VMs and containers, and hence appears more appropriate for
3961 this, as the directory can also contain raw images bootable
3964 * systemd-nspawn when invoked with -M but without --directory=
3965 or --image= is now capable of searching for the container
3966 root directory, subvolume or disk image automatically, in
3967 /var/lib/machines. systemd-nspawn@.service has been updated
3968 to make use of this, thus allowing it to be used for raw
3971 * A new machines.target unit has been introduced that is
3972 supposed to group all containers/VMs invoked as services on
3973 the system. systemd-nspawn@.service has been updated to
3974 integrate with that.
3976 * machinectl gained a new "start" command, for invoking a
3977 container as a service. "machinectl start foo" is mostly
3978 equivalent to "systemctl start systemd-nspawn@foo.service",
3979 but handles escaping in a nicer way.
3981 * systemd-nspawn will now mount most of the cgroupfs tree
3982 read-only into each container, with the exception of the
3983 container's own subtree in the name=systemd hierarchy.
3985 * journald now sets the special FS_NOCOW file flag for its
3986 journal files. This should improve performance on btrfs, by
3987 avoiding heavy fragmentation when journald's write-pattern
3988 is used on COW file systems. It degrades btrfs' data
3989 integrity guarantees for the files to the same levels as for
3990 ext3/ext4 however. This should be OK though as journald does
3991 its own data integrity checks and all its objects are
3992 checksummed on disk. Also, journald should handle btrfs disk
3993 full events a lot more gracefully now, by processing SIGBUS
3994 errors, and not relying on fallocate() anymore.
3996 * When journald detects that journal files it is writing to
3997 have been deleted it will immediately start new journal
4000 * systemd now provides a way to store file descriptors
4001 per-service in PID 1. This is useful for daemons to ensure
4002 that fds they require are not lost during a daemon
4003 restart. The fds are passed to the daemon on the next
4004 invocation in the same way socket activation fds are
4005 passed. This is now used by journald to ensure that the
4006 various sockets connected to all the system's stdout/stderr
4007 are not lost when journald is restarted. File descriptors
4008 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
4009 an extension to sd_notify(). Note that a limit is enforced
4010 on the number of fds a service can store in PID 1, and it
4011 defaults to 0, so that no fds may be stored, unless this is
4012 explicitly turned on.
4014 * The default TERM variable to use for units connected to a
4015 terminal, when no other value is explicitly is set is now
4016 vt220 rather than vt102. This should be fairly safe still,
4017 but allows PgUp/PgDn work.
4019 * The /etc/crypttab option header= as known from Debian is now
4022 * "loginctl user-status" and "loginctl session-status" will
4023 now show the last 10 lines of log messages of the
4024 user/session following the status output. Similar,
4025 "machinectl status" will show the last 10 log lines
4026 associated with a virtual machine or container
4027 service. (Note that this is usually not the log messages
4028 done in the VM/container itself, but simply what the
4029 container manager logs. For nspawn this includes all console
4032 * "loginctl session-status" without further argument will now
4033 show the status of the session of the caller. Similar,
4034 "lock-session", "unlock-session", "activate",
4035 "enable-linger", "disable-linger" may now be called without
4036 session/user parameter in which case they apply to the
4037 caller's session/user.
4039 * An X11 session scriptlet is now shipped that uploads
4040 $DISPLAY and $XAUTHORITY into the environment of the systemd
4041 --user daemon if a session begins. This should improve
4042 compatibility with X11 enabled applications run as systemd
4045 * Generators are now subject to masking via /etc and /run, the
4046 same way as unit files.
4048 * networkd .network files gained support for configuring
4049 per-link IPv4/IPv6 packet forwarding as well as IPv4
4050 masquerading. This is by default turned on for veth links to
4051 containers, as registered by systemd-nspawn. This means that
4052 nspawn containers run with --network-veth will now get
4053 automatic routed access to the host's networks without any
4054 further configuration or setup, as long as networkd runs on
4057 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4058 or UDP posts of a container on the host. With this in place
4059 it is possible to run containers with private veth links
4060 (--network-veth), and have their functionality exposed on
4061 the host as if their services were running directly on the
4064 * systemd-nspawn's --network-veth switch now gained a short
4065 version "-n", since with the changes above it is now truly
4066 useful out-of-the-box. The systemd-nspawn@.service has been
4067 updated to make use of it too by default.
4069 * systemd-nspawn will now maintain a per-image R/W lock, to
4070 ensure that the same image is not started more than once
4071 writable. (It's OK to run an image multiple times
4072 simultaneously in read-only mode.)
4074 * systemd-nspawn's --image= option is now capable of
4075 dissecting and booting MBR and GPT disk images that contain
4076 only a single active Linux partition. Previously it
4077 supported only GPT disk images with proper GPT type
4078 IDs. This allows running cloud images from major
4079 distributions directly with systemd-nspawn, without
4082 * In addition to collecting mouse dpi data in the udev
4083 hardware database, there's now support for collecting angle
4084 information for mouse scroll wheels. The database is
4085 supposed to guarantee similar scrolling behavior on mice
4086 that it knows about. There's also support for collecting
4087 information about Touchpad types.
4089 * udev's input_id built-in will now also collect touch screen
4090 dimension data and attach it to probed devices.
4092 * /etc/os-release gained support for a Distribution Privacy
4095 * networkd gained support for creating "ipvlan", "gretap",
4096 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4098 * systemd-tmpfiles gained support for "a" lines for setting
4101 * systemd-nspawn will now mount /tmp in the container to
4102 tmpfs, automatically.
4104 * systemd now exposes the memory.usage_in_bytes cgroup
4105 attribute and shows it for each service in the "systemctl
4106 status" output, if available.
4108 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4109 immediate reboot is triggered. This useful if shutdown is
4110 hung and is unable to complete, to expedite the
4111 operation. Note that this kind of reboot will still unmount
4112 all file systems, and hence should not result in fsck being
4115 * A .device unit for an optical block device will now be
4116 considered active only when a medium is in the drive. Also,
4117 mount units are now bound to their backing devices thus
4118 triggering automatic unmounting when devices become
4119 unavailable. With this in place systemd will now
4120 automatically unmount left-over mounts when a CD-ROM is
4121 ejected or an USB stick is yanked from the system.
4123 * networkd-wait-online now has support for waiting for
4124 specific interfaces only (with globbing), and for giving up
4125 after a configurable timeout.
4127 * networkd now exits when idle. It will be automatically
4128 restarted as soon as interfaces show up, are removed or
4129 change state. networkd will stay around as long as there is
4130 at least one DHCP state machine or similar around, that keep
4133 * networkd may now configure IPv6 link-local addressing in
4134 addition to IPv4 link-local addressing.
4136 * The IPv6 "token" for use in SLAAC may now be configured for
4137 each .network interface in networkd.
4139 * Routes configured with networkd may now be assigned a scope
4142 * networkd's [Match] sections now support globbing and lists
4143 of multiple space-separated matches per item.
4145 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4146 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4147 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4148 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4149 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4150 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4151 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4152 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4153 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4154 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4155 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4156 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4157 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4158 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4159 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
4160 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4161 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4162 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4163 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4164 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4165 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4166 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
4167 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4168 Hoffmann, Zbigniew Jędrzejewski-Szmek
4170 — Berlin, 2015-02-16
4174 * When querying unit file enablement status (for example via
4175 "systemctl is-enabled"), a new state "indirect" is now known
4176 which indicates that a unit might not be enabled itself, but
4177 another unit listed in its Also= setting might be.
4179 * Similar to the various existing ConditionXYZ= settings for
4180 units, there are now matching AssertXYZ= settings. While
4181 failing conditions cause a unit to be skipped, but its job
4182 to succeed, failing assertions declared like this will cause
4183 a unit start operation and its job to fail.
4185 * hostnamed now knows a new chassis type "embedded".
4187 * systemctl gained a new "edit" command. When used on a unit
4188 file, this allows extending unit files with .d/ drop-in
4189 configuration snippets or editing the full file (after
4190 copying it from /usr/lib to /etc). This will invoke the
4191 user's editor (as configured with $EDITOR), and reload the
4192 modified configuration after editing.
4194 * "systemctl status" now shows the suggested enablement state
4195 for a unit, as declared in the (usually vendor-supplied)
4196 system preset files.
4198 * nss-myhostname will now resolve the single-label host name
4199 "gateway" to the locally configured default IP routing
4200 gateways, ordered by their metrics. This assigns a stable
4201 name to the used gateways, regardless which ones are
4202 currently configured. Note that the name will only be
4203 resolved after all other name sources (if nss-myhostname is
4204 configured properly) and should hence not negatively impact
4205 systems that use the single-label host name "gateway" in
4208 * systemd-inhibit now allows filtering by mode when listing
4211 * Scope and service units gained a new "Delegate" boolean
4212 property, which, when set, allows processes running inside the
4213 unit to further partition resources. This is primarily
4214 useful for systemd user instances as well as container
4217 * journald will now pick up audit messages directly from
4218 the kernel, and log them like any other log message. The
4219 audit fields are split up and fully indexed. This means that
4220 journalctl in many ways is now a (nicer!) alternative to
4221 ausearch, the traditional audit client. Note that this
4222 implements only a minimal audit client. If you want the
4223 special audit modes like reboot-on-log-overflow, please use
4224 the traditional auditd instead, which can be used in
4225 parallel to journald.
4227 * The ConditionSecurity= unit file option now understands the
4228 special string "audit" to check whether auditing is
4231 * journalctl gained two new commands --vacuum-size= and
4232 --vacuum-time= to delete old journal files until the
4233 remaining ones take up no more than the specified size on disk,
4234 or are not older than the specified time.
4236 * A new, native PPPoE library has been added to sd-network,
4237 systemd's library of light-weight networking protocols. This
4238 library will be used in a future version of networkd to
4239 enable PPPoE communication without an external pppd daemon.
4241 * The busctl tool now understands a new "capture" verb that
4242 works similar to "monitor", but writes a packet capture
4243 trace to STDOUT that can be redirected to a file which is
4244 compatible with libcap's capture file format. This can then
4245 be loaded in Wireshark and similar tools to inspect bus
4248 * The busctl tool now understands a new "tree" verb that shows
4249 the object trees of a specific service on the bus, or of all
4252 * The busctl tool now understands a new "introspect" verb that
4253 shows all interfaces and members of objects on the bus,
4254 including their signature and values. This is particularly
4255 useful to get more information about bus objects shown by
4256 the new "busctl tree" command.
4258 * The busctl tool now understands new verbs "call",
4259 "set-property" and "get-property" for invoking bus method
4260 calls, setting and getting bus object properties in a
4263 * busctl gained a new --augment-creds= argument that controls
4264 whether the tool shall augment credential information it
4265 gets from the bus with data from /proc, in a possibly
4268 * nspawn's --link-journal= switch gained two new values
4269 "try-guest" and "try-host" that work like "guest" and
4270 "host", but do not fail if the host has no persistent
4271 journalling enabled. -j is now equivalent to
4272 --link-journal=try-guest.
4274 * macvlan network devices created by nspawn will now have
4275 stable MAC addresses.
4277 * A new SmackProcessLabel= unit setting has been added, which
4278 controls the SMACK security label processes forked off by
4279 the respective unit shall use.
4281 * If compiled with --enable-xkbcommon, systemd-localed will
4282 verify x11 keymap settings by compiling the given keymap. It
4283 will spew out warnings if the compilation fails. This
4284 requires libxkbcommon to be installed.
4286 * When a coredump is collected, a larger number of metadata
4287 fields is now collected and included in the journal records
4288 created for it. More specifically, control group membership,
4289 environment variables, memory maps, working directory,
4290 chroot directory, /proc/$PID/status, and a list of open file
4291 descriptors is now stored in the log entry.
4293 * The udev hwdb now contains DPI information for mice. For
4296 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4298 * All systemd programs that read standalone configuration
4299 files in /etc now also support a corresponding series of
4300 .conf.d configuration directories in /etc/, /run/,
4301 /usr/local/lib/, /usr/lib/, and (if configured with
4302 --enable-split-usr) /lib/. In particular, the following
4303 configuration files now have corresponding configuration
4304 directories: system.conf user.conf, logind.conf,
4305 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4306 resolved.conf, timesyncd.conf, journal-remote.conf, and
4307 journal-upload.conf. Note that distributions should use the
4308 configuration directories in /usr/lib/; the directories in
4309 /etc/ are reserved for the system administrator.
4311 * systemd-rfkill will no longer take the rfkill device name
4312 into account when storing rfkill state on disk, as the name
4313 might be dynamically assigned and not stable. Instead, the
4314 ID_PATH udev variable combined with the rfkill type (wlan,
4315 bluetooth, ...) is used.
4317 * A new service systemd-machine-id-commit.service has been
4318 added. When used on systems where /etc is read-only during
4319 boot, and /etc/machine-id is not initialized (but an empty
4320 file), this service will copy the temporary machine ID
4321 created as replacement into /etc after the system is fully
4322 booted up. This is useful for systems that are freshly
4323 installed with a non-initialized machine ID, but should get
4324 a fixed machine ID for subsequent boots.
4326 * networkd's .netdev files now provide a large set of
4327 configuration parameters for VXLAN devices. Similarly, the
4328 bridge port cost parameter is now configurable in .network
4329 files. There's also new support for configuring IP source
4330 routing. networkd .link files gained support for a new
4331 OriginalName= match that is useful to match against the
4332 original interface name the kernel assigned. .network files
4333 may include MTU= and MACAddress= fields for altering the MTU
4334 and MAC address while being connected to a specific network
4337 * The LUKS logic gained supported for configuring
4338 UUID-specific key files. There's also new support for naming
4339 LUKS device from the kernel command line, using the new
4340 luks.name= argument.
4342 * Timer units may now be transiently created via the bus API
4343 (this was previously already available for scope and service
4344 units). In addition it is now possible to create multiple
4345 transient units at the same time with a single bus call. The
4346 "systemd-run" tool has been updated to make use of this for
4347 running commands on a specified time, in at(1)-style.
4349 * tmpfiles gained support for "t" lines, for assigning
4350 extended attributes to files. Among other uses this may be
4351 used to assign SMACK labels to files.
4353 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4354 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4355 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4356 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4357 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4358 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4359 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4360 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4361 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4362 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4363 Michael Biebl, Michael Chapman, Michael Marineau, Michal
4364 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4365 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4366 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4367 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4368 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4369 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4370 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
4372 — Berlin, 2014-12-10
4376 * journalctl gained the new options -t/--identifier= to match
4377 on the syslog identifier (aka "tag"), as well as --utc to
4378 show log timestamps in the UTC timezone. journalctl now also
4379 accepts -n/--lines=all to disable line capping in a pager.
4381 * journalctl gained a new switch, --flush, that synchronously
4382 flushes logs from /run/log/journal to /var/log/journal if
4383 persistent storage is enabled. systemd-journal-flush.service
4384 now waits until the operation is complete.
4386 * Services can notify the manager before they start a reload
4387 (by sending RELOADING=1) or shutdown (by sending
4388 STOPPING=1). This allows the manager to track and show the
4389 internal state of daemons and closes a race condition when
4390 the process is still running but has closed its D-Bus
4393 * Services with Type=oneshot do not have to have any ExecStart
4396 * User units are now loaded also from
4397 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4398 /run/systemd/user directory that was already previously
4399 supported, but is under the control of the user.
4401 * Job timeouts (i.e. time-outs on the time a job that is
4402 queued stays in the run queue) can now optionally result in
4403 immediate reboot or power-off actions (JobTimeoutAction= and
4404 JobTimeoutRebootArgument=). This is useful on ".target"
4405 units, to limit the maximum time a target remains
4406 undispatched in the run queue, and to trigger an emergency
4407 operation in such a case. This is now used by default to
4408 turn off the system if boot-up (as defined by everything in
4409 basic.target) hangs and does not complete for at least
4410 15min. Also, if power-off or reboot hang for at least 30min
4411 an immediate power-off/reboot operation is triggered. This
4412 functionality is particularly useful to increase reliability
4413 on embedded devices, but also on laptops which might
4414 accidentally get powered on when carried in a backpack and
4415 whose boot stays stuck in a hard disk encryption passphrase
4418 * systemd-logind can be configured to also handle lid switch
4419 events even when the machine is docked or multiple displays
4420 are attached (HandleLidSwitchDocked= option).
4422 * A helper binary and a service have been added which can be
4423 used to resume from hibernation in the initramfs. A
4424 generator will parse the resume= option on the kernel
4425 command line to trigger resume.
4427 * A user console daemon systemd-consoled has been
4428 added. Currently, it is a preview, and will so far open a
4429 single terminal on each session of the user marked as
4430 Desktop=systemd-console.
4432 * Route metrics can be specified for DHCP routes added by
4435 * The SELinux context of socket-activated services can be set
4436 from the information provided by the networking stack
4437 (SELinuxContextFromNet= option).
4439 * Userspace firmware loading support has been removed and
4440 the minimum supported kernel version is thus bumped to 3.7.
4442 * Timeout for udev workers has been increased from 1 to 3
4443 minutes, but a warning will be printed after 1 minute to
4444 help diagnose kernel modules that take a long time to load.
4446 * Udev rules can now remove tags on devices with TAG-="foobar".
4448 * systemd's readahead implementation has been removed. In many
4449 circumstances it didn't give expected benefits even for
4450 rotational disk drives and was becoming less relevant in the
4451 age of SSDs. As none of the developers has been using
4452 rotating media anymore, and nobody stepped up to actively
4453 maintain this component of systemd it has now been removed.
4455 * Swap units can use Options= to specify discard options.
4456 Discard options specified for swaps in /etc/fstab are now
4459 * Docker containers are now detected as a separate type of
4462 * The Password Agent protocol gained support for queries where
4463 the user input is shown, useful e.g. for user names.
4464 systemd-ask-password gained a new --echo option to turn that
4467 * The default sysctl.d/ snippets will now set:
4469 net.core.default_qdisc = fq_codel
4471 This selects Fair Queuing Controlled Delay as the default
4472 queuing discipline for network interfaces. fq_codel helps
4473 fight the network bufferbloat problem. It is believed to be
4474 a good default with no tuning required for most workloads.
4475 Downstream distributions may override this choice. On 10Gbit
4476 servers that do not do forwarding, "fq" may perform better.
4477 Systems without a good clocksource should use "pfifo_fast".
4479 * If kdbus is enabled during build a new option BusPolicy= is
4480 available for service units, that allows locking all service
4481 processes into a stricter bus policy, in order to limit
4482 access to various bus services, or even hide most of them
4483 from the service's view entirely.
4485 * networkctl will now show the .network and .link file
4486 networkd has applied to a specific interface.
4488 * sd-login gained a new API call sd_session_get_desktop() to
4489 query which desktop environment has been selected for a
4492 * UNIX utmp support is now compile-time optional to support
4493 legacy-free systems.
4495 * systemctl gained two new commands "add-wants" and
4496 "add-requires" for pulling in units from specific targets
4499 * If the word "rescue" is specified on the kernel command line
4500 the system will now boot into rescue mode (aka
4501 rescue.target), which was previously available only by
4502 specifying "1" or "systemd.unit=rescue.target" on the kernel
4503 command line. This new kernel command line option nicely
4504 mirrors the already existing "emergency" kernel command line
4507 * New kernel command line options mount.usr=, mount.usrflags=,
4508 mount.usrfstype= have been added that match root=, rootflags=,
4509 rootfstype= but allow mounting a specific file system to
4512 * The $NOTIFY_SOCKET is now also passed to control processes of
4513 services, not only the main process.
4515 * This version reenables support for fsck's -l switch. This
4516 means at least version v2.25 of util-linux is required for
4517 operation, otherwise dead-locks on device nodes may
4518 occur. Again: you need to update util-linux to at least
4519 v2.25 when updating systemd to v217.
4521 * The "multi-seat-x" tool has been removed from systemd, as
4522 its functionality has been integrated into X servers 1.16,
4523 and the tool is hence redundant. It is recommended to update
4524 display managers invoking this tool to simply invoke X
4525 directly from now on, again.
4527 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4528 message flag has been added for all of systemd's PolicyKit
4529 authenticated method calls has been added. In particular
4530 this now allows optional interactive authorization via
4531 PolicyKit for many of PID1's privileged operations such as
4532 unit file enabling and disabling.
4534 * "udevadm hwdb --update" learnt a new switch "--usr" for
4535 placing the rebuilt hardware database in /usr instead of
4536 /etc. When used only hardware database entries stored in
4537 /usr will be used, and any user database entries in /etc are
4538 ignored. This functionality is useful for vendors to ship a
4539 pre-built database on systems where local configuration is
4540 unnecessary or unlikely.
4542 * Calendar time specifications in .timer units now also
4543 understand the strings "semi-annually", "quarterly" and
4544 "minutely" as shortcuts (in addition to the preexisting
4545 "anually", "hourly", ...).
4547 * systemd-tmpfiles will now correctly create files in /dev
4548 at boot which are marked for creation only at boot. It is
4549 recommended to always create static device nodes with 'c!'
4550 and 'b!', so that they are created only at boot and not
4551 overwritten at runtime.
4553 * When the watchdog logic is used for a service (WatchdogSec=)
4554 and the watchdog timeout is hit the service will now be
4555 terminated with SIGABRT (instead of just SIGTERM), in order
4556 to make sure a proper coredump and backtrace is
4557 generated. This ensures that hanging services will result in
4558 similar coredump/backtrace behaviour as services that hit a
4561 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4562 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4563 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4564 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4565 Herrmann, David Sommerseth, David Strauss, Emil Renner
4566 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4567 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4568 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4569 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4570 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4571 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4572 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4573 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4574 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4575 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4576 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4577 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4578 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4579 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4580 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4581 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
4584 — Berlin, 2014-10-28
4588 * timedated no longer reads NTP implementation unit names from
4589 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
4590 implementations should add a
4592 Conflicts=systemd-timesyncd.service
4594 to their unit files to take over and replace systemd's NTP
4595 default functionality.
4597 * systemd-sysusers gained a new line type "r" for configuring
4598 which UID/GID ranges to allocate system users/groups
4599 from. Lines of type "u" may now add an additional column
4600 that specifies the home directory for the system user to be
4601 created. Also, systemd-sysusers may now optionally read user
4602 information from STDIN instead of a file. This is useful for
4603 invoking it from RPM preinst scriptlets that need to create
4604 users before the first RPM file is installed since these
4605 files might need to be owned by them. A new
4606 %sysusers_create_inline RPM macro has been introduced to do
4607 just that. systemd-sysusers now updates the shadow files as
4608 well as the user/group databases, which should enhance
4609 compatibility with certain tools like grpck.
4611 * A number of bus APIs of PID 1 now optionally consult
4612 PolicyKit to permit access for otherwise unprivileged
4613 clients under certain conditions. Note that this currently
4614 doesn't support interactive authentication yet, but this is
4615 expected to be added eventually, too.
4617 * /etc/machine-info now has new fields for configuring the
4618 deployment environment of the machine, as well as the
4619 location of the machine. hostnamectl has been updated with
4620 new command to update these fields.
4622 * systemd-timesyncd has been updated to automatically acquire
4623 NTP server information from systemd-networkd, which might
4624 have been discovered via DHCP.
4626 * systemd-resolved now includes a caching DNS stub resolver
4627 and a complete LLMNR name resolution implementation. A new
4628 NSS module "nss-resolve" has been added which can be used
4629 instead of glibc's own "nss-dns" to resolve hostnames via
4630 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4631 be resolved via systemd-resolved D-Bus APIs. In contrast to
4632 the glibc internal resolver systemd-resolved is aware of
4633 multi-homed system, and keeps DNS server and caches separate
4634 and per-interface. Queries are sent simultaneously on all
4635 interfaces that have DNS servers configured, in order to
4636 properly handle VPNs and local LANs which might resolve
4637 separate sets of domain names. systemd-resolved may acquire
4638 DNS server information from systemd-networkd automatically,
4639 which in turn might have discovered them via DHCP. A tool
4640 "systemd-resolve-host" has been added that may be used to
4641 query the DNS logic in resolved. systemd-resolved implements
4642 IDNA and automatically uses IDNA or UTF-8 encoding depending
4643 on whether classic DNS or LLMNR is used as transport. In the
4644 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4645 implementation to systemd-resolved.
4647 * A new NSS module nss-mymachines has been added, that
4648 automatically resolves the names of all local registered
4649 containers to their respective IP addresses.
4651 * A new client tool "networkctl" for systemd-networkd has been
4652 added. It currently is entirely passive and will query
4653 networking configuration from udev, rtnetlink and networkd,
4654 and present it to the user in a very friendly
4655 way. Eventually, we hope to extend it to become a full
4656 control utility for networkd.
4658 * .socket units gained a new DeferAcceptSec= setting that
4659 controls the kernels' TCP_DEFER_ACCEPT sockopt for
4660 TCP. Similarly, support for controlling TCP keep-alive
4661 settings has been added (KeepAliveTimeSec=,
4662 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4663 turning off Nagle's algorithm on TCP has been added
4666 * logind learned a new session type "web", for use in projects
4667 like Cockpit which register web clients as PAM sessions.
4669 * timer units with at least one OnCalendar= setting will now
4670 be started only after time-sync.target has been
4671 reached. This way they will not elapse before the system
4672 clock has been corrected by a local NTP client or
4673 similar. This is particular useful on RTC-less embedded
4674 machines, that come up with an invalid system clock.
4676 * systemd-nspawn's --network-veth= switch should now result in
4677 stable MAC addresses for both the outer and the inner side
4680 * systemd-nspawn gained a new --volatile= switch for running
4681 container instances with /etc or /var unpopulated.
4683 * The kdbus client code has been updated to use the new Linux
4684 3.17 memfd subsystem instead of the old kdbus-specific one.
4686 * systemd-networkd's DHCP client and server now support
4687 FORCERENEW. There are also new configuration options to
4688 configure the vendor client identifier and broadcast mode
4691 * systemd will no longer inform the kernel about the current
4692 timezone, as this is necessarily incorrect and racy as the
4693 kernel has no understanding of DST and similar
4694 concepts. This hence means FAT timestamps will be always
4695 considered UTC, similar to what Android is already
4696 doing. Also, when the RTC is configured to the local time
4697 (rather than UTC) systemd will never synchronize back to it,
4698 as this might confuse Windows at a later boot.
4700 * systemd-analyze gained a new command "verify" for offline
4701 validation of unit files.
4703 * systemd-networkd gained support for a couple of additional
4704 settings for bonding networking setups. Also, the metric for
4705 statically configured routes may now be configured. For
4706 network interfaces where this is appropriate the peer IP
4707 address may now be configured.
4709 * systemd-networkd's DHCP client will no longer request
4710 broadcasting by default, as this tripped up some networks.
4711 For hardware where broadcast is required the feature should
4712 be switched back on using RequestBroadcast=yes.
4714 * systemd-networkd will now set up IPv4LL addresses (when
4715 enabled) even if DHCP is configured successfully.
4717 * udev will now default to respect network device names given
4718 by the kernel when the kernel indicates that these are
4719 predictable. This behavior can be tweaked by changing
4720 NamePolicy= in the relevant .link file.
4722 * A new library systemd-terminal has been added that
4723 implements full TTY stream parsing and rendering. This
4724 library is supposed to be used later on for implementing a
4725 full userspace VT subsystem, replacing the current kernel
4728 * A new tool systemd-journal-upload has been added to push
4729 journal data to a remote system running
4730 systemd-journal-remote.
4732 * journald will no longer forward all local data to another
4733 running syslog daemon. This change has been made because
4734 rsyslog (which appears to be the most commonly used syslog
4735 implementation these days) no longer makes use of this, and
4736 instead pulls the data out of the journal on its own. Since
4737 forwarding the messages to a non-existent syslog server is
4738 more expensive than we assumed we have now turned this
4739 off. If you run a syslog server that is not a recent rsyslog
4740 version, you have to turn this option on again
4741 (ForwardToSyslog= in journald.conf).
4743 * journald now optionally supports the LZ4 compressor for
4744 larger journal fields. This compressor should perform much
4745 better than XZ which was the previous default.
4747 * machinectl now shows the IP addresses of local containers,
4748 if it knows them, plus the interface name of the container.
4750 * A new tool "systemd-escape" has been added that makes it
4751 easy to escape strings to build unit names and similar.
4753 * sd_notify() messages may now include a new ERRNO= field
4754 which is parsed and collected by systemd and shown among the
4755 "systemctl status" output for a service.
4757 * A new component "systemd-firstboot" has been added that
4758 queries the most basic systemd information (timezone,
4759 hostname, root password) interactively on first
4760 boot. Alternatively it may also be used to provision these
4761 things offline on OS images installed into directories.
4763 * The default sysctl.d/ snippets will now set
4765 net.ipv4.conf.default.promote_secondaries=1
4767 This has the benefit of no flushing secondary IP addresses
4768 when primary addresses are removed.
4770 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4771 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4772 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4773 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4774 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4775 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4776 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4777 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4778 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4779 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4780 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4781 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4782 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4783 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4784 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4786 — Berlin, 2014-08-19
4790 * A new tool systemd-sysusers has been added. This tool
4791 creates system users and groups in /etc/passwd and
4792 /etc/group, based on static declarative system user/group
4793 definitions in /usr/lib/sysusers.d/. This is useful to
4794 enable factory resets and volatile systems that boot up with
4795 an empty /etc directory, and thus need system users and
4796 groups created during early boot. systemd now also ships
4797 with two default sysusers.d/ files for the most basic
4798 users and groups systemd and the core operating system
4801 * A new tmpfiles snippet has been added that rebuilds the
4802 essential files in /etc on boot, should they be missing.
4804 * A directive for ensuring automatic clean-up of
4805 /var/cache/man/ has been removed from the default
4806 configuration. This line should now be shipped by the man
4807 implementation. The necessary change has been made to the
4808 man-db implementation. Note that you need to update your man
4809 implementation to one that ships this line, otherwise no
4810 automatic clean-up of /var/cache/man will take place.
4812 * A new condition ConditionNeedsUpdate= has been added that
4813 may conditionalize services to only run when /etc or /var
4814 are "older" than the vendor operating system resources in
4815 /usr. This is useful for reconstructing or updating /etc
4816 after an offline update of /usr or a factory reset, on the
4817 next reboot. Services that want to run once after such an
4818 update or reset should use this condition and order
4819 themselves before the new systemd-update-done.service, which
4820 will mark the two directories as fully updated. A number of
4821 service files have been added making use of this, to rebuild
4822 the udev hardware database, the journald message catalog and
4823 dynamic loader cache (ldconfig). The systemd-sysusers tool
4824 described above also makes use of this now. With this in
4825 place it is now possible to start up a minimal operating
4826 system with /etc empty cleanly. For more information on the
4827 concepts involved see this recent blog story:
4829 http://0pointer.de/blog/projects/stateless.html
4831 * A new system group "input" has been introduced, and all
4832 input device nodes get this group assigned. This is useful
4833 for system-level software to get access to input devices. It
4834 complements what is already done for "audio" and "video".
4836 * systemd-networkd learnt minimal DHCPv4 server support in
4837 addition to the existing DHCPv4 client support. It also
4838 learnt DHCPv6 client and IPv6 Router Solicitation client
4839 support. The DHCPv4 client gained support for static routes
4840 passed in from the server. Note that the [DHCPv4] section
4841 known in older systemd-networkd versions has been renamed to
4842 [DHCP] and is now also used by the DHCPv6 client. Existing
4843 .network files using settings of this section should be
4844 updated, though compatibility is maintained. Optionally, the
4845 client hostname may now be sent to the DHCP server.
4847 * networkd gained support for vxlan virtual networks as well
4848 as tun/tap and dummy devices.
4850 * networkd gained support for automatic allocation of address
4851 ranges for interfaces from a system-wide pool of
4852 addresses. This is useful for dynamically managing a large
4853 number of interfaces with a single network configuration
4854 file. In particular this is useful to easily assign
4855 appropriate IP addresses to the veth links of a large number
4856 of nspawn instances.
4858 * RPM macros for processing sysusers, sysctl and binfmt
4859 drop-in snippets at package installation time have been
4862 * The /etc/os-release file should now be placed in
4863 /usr/lib/os-release. The old location is automatically
4864 created as symlink. /usr/lib is the more appropriate
4865 location of this file, since it shall actually describe the
4866 vendor operating system shipped in /usr, and not the
4867 configuration stored in /etc.
4869 * .mount units gained a new boolean SloppyOptions= setting
4870 that maps to mount(8)'s -s option which enables permissive
4871 parsing of unknown mount options.
4873 * tmpfiles learnt a new "L+" directive which creates a symlink
4874 but (unlike "L") deletes a pre-existing file first, should
4875 it already exist and not already be the correct
4876 symlink. Similarly, "b+", "c+" and "p+" directives have been
4877 added as well, which create block and character devices, as
4878 well as fifos in the filesystem, possibly removing any
4879 pre-existing files of different types.
4881 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4882 'argument' field (which so far specified the source to
4883 symlink/copy the files from) is now optional. If omitted the
4884 same file os copied from /usr/share/factory/ suffixed by the
4885 full destination path. This is useful for populating /etc
4886 with essential files, by copying them from vendor defaults
4887 shipped in /usr/share/factory/etc.
4889 * A new command "systemctl preset-all" has been added that
4890 applies the service preset settings to all installed unit
4891 files. A new switch --preset-mode= has been added that
4892 controls whether only enable or only disable operations
4895 * A new command "systemctl is-system-running" has been added
4896 that allows checking the overall state of the system, for
4897 example whether it is fully up and running.
4899 * When the system boots up with an empty /etc, the equivalent
4900 to "systemctl preset-all" is executed during early boot, to
4901 make sure all default services are enabled after a factory
4904 * systemd now contains a minimal preset file that enables the
4905 most basic services systemd ships by default.
4907 * Unit files' [Install] section gained a new DefaultInstance=
4908 field for defining the default instance to create if a
4909 template unit is enabled with no instance specified.
4911 * A new passive target cryptsetup-pre.target has been added
4912 that may be used by services that need to make they run and
4913 finish before the first LUKS cryptographic device is set up.
4915 * The /dev/loop-control and /dev/btrfs-control device nodes
4916 are now owned by the "disk" group by default, opening up
4917 access to this group.
4919 * systemd-coredump will now automatically generate a
4920 stack trace of all core dumps taking place on the system,
4921 based on elfutils' libdw library. This stack trace is logged
4924 * systemd-coredump may now optionally store coredumps directly
4925 on disk (in /var/lib/systemd/coredump, possibly compressed),
4926 instead of storing them unconditionally in the journal. This
4927 mode is the new default. A new configuration file
4928 /etc/systemd/coredump.conf has been added to configure this
4929 and other parameters of systemd-coredump.
4931 * coredumpctl gained a new "info" verb to show details about a
4932 specific coredump. A new switch "-1" has also been added
4933 that makes sure to only show information about the most
4934 recent entry instead of all entries. Also, as the tool is
4935 generally useful now the "systemd-" prefix of the binary
4936 name has been removed. Distributions that want to maintain
4937 compatibility with the old name should add a symlink from
4938 the old name to the new name.
4940 * journald's SplitMode= now defaults to "uid". This makes sure
4941 that unprivileged users can access their own coredumps with
4942 coredumpctl without restrictions.
4944 * New kernel command line options "systemd.wants=" (for
4945 pulling an additional unit during boot), "systemd.mask="
4946 (for masking a specific unit for the boot), and
4947 "systemd.debug-shell" (for enabling the debug shell on tty9)
4948 have been added. This is implemented in the new generator
4949 "systemd-debug-generator".
4951 * systemd-nspawn will now by default filter a couple of
4952 syscalls for containers, among them those required for
4953 kernel module loading, direct x86 IO port access, swap
4954 management, and kexec. Most importantly though
4955 open_by_handle_at() is now prohibited for containers,
4956 closing a hole similar to a recently discussed vulnerability
4957 in docker regarding access to files on file hierarchies the
4958 container should normally not have access to. Note that, for
4959 nspawn, we generally make no security claims anyway (and
4960 this is explicitly documented in the man page), so this is
4961 just a fix for one of the most obvious problems.
4963 * A new man page file-hierarchy(7) has been added that
4964 contains a minimized, modernized version of the file system
4965 layout systemd expects, similar in style to the FHS
4966 specification or hier(5). A new tool systemd-path(1) has
4967 been added to query many of these paths for the local
4970 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4971 longer done. Since the directory now has a per-user size
4972 limit, and is cleaned on logout this appears unnecessary,
4973 in particular since this now brings the lifecycle of this
4974 directory closer in line with how IPC objects are handled.
4976 * systemd.pc now exports a number of additional directories,
4977 including $libdir (which is useful to identify the library
4978 path for the primary architecture of the system), and a
4979 couple of drop-in directories.
4981 * udev's predictable network interface names now use the dev_port
4982 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4983 distinguish between ports of the same PCI function. dev_id should
4984 only be used for ports using the same HW address, hence the need
4987 * machined has been updated to export the OS version of a
4988 container (read from /etc/os-release and
4989 /usr/lib/os-release) on the bus. This is now shown in
4990 "machinectl status" for a machine.
4992 * A new service setting RestartForceExitStatus= has been
4993 added. If configured to a set of exit signals or process
4994 return values, the service will be restarted when the main
4995 daemon process exits with any of them, regardless of the
4998 * systemctl's -H switch for connecting to remote systemd
4999 machines has been extended so that it may be used to
5000 directly connect to a specific container on the
5001 host. "systemctl -H root@foobar:waldi" will now connect as
5002 user "root" to host "foobar", and then proceed directly to
5003 the container named "waldi". Note that currently you have to
5004 authenticate as user "root" for this to work, as entering
5005 containers is a privileged operation.
5007 Contributions from: Andreas Henriksson, Benjamin Steinwender,
5008 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
5009 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5010 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5011 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5012 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5013 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5014 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5015 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5016 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5017 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5018 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5020 — Berlin, 2014-07-03
5024 * As an experimental feature, udev now tries to lock the
5025 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5026 executes events for the disk or any of its partitions.
5027 Applications like partitioning programs can lock the
5028 disk device node (flock(LOCK_EX)) and claim temporary
5029 device ownership that way; udev will entirely skip all event
5030 handling for this disk and its partitions. If the disk
5031 was opened for writing, the close will trigger a partition
5032 table rescan in udev's "watch" facility, and if needed
5033 synthesize "change" events for the disk and all its partitions.
5034 This is now unconditionally enabled, and if it turns out to
5035 cause major problems, we might turn it on only for specific
5036 devices, or might need to disable it entirely. Device Mapper
5037 devices are excluded from this logic.
5039 * We temporarily dropped the "-l" switch for fsck invocations,
5040 since they collide with the flock() logic above. util-linux
5041 upstream has been changed already to avoid this conflict,
5042 and we will readd "-l" as soon as util-linux with this
5043 change has been released.
5045 * The dependency on libattr has been removed. Since a long
5046 time, the extended attribute calls have moved to glibc, and
5047 libattr is thus unnecessary.
5049 * Virtualization detection works without privileges now. This
5050 means the systemd-detect-virt binary no longer requires
5051 CAP_SYS_PTRACE file capabilities, and our daemons can run
5052 with fewer privileges.
5054 * systemd-networkd now runs under its own "systemd-network"
5055 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5056 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5057 loses the ability to write to files owned by root this way.
5059 * Similarly, systemd-resolved now runs under its own
5060 "systemd-resolve" user with no capabilities remaining.
5062 * Similarly, systemd-bus-proxyd now runs under its own
5063 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5065 * systemd-networkd gained support for setting up "veth"
5066 virtual Ethernet devices for container connectivity, as well
5067 as GRE and VTI tunnels.
5069 * systemd-networkd will no longer automatically attempt to
5070 manually load kernel modules necessary for certain tunnel
5071 transports. Instead, it is assumed the kernel loads them
5072 automatically when required. This only works correctly on
5073 very new kernels. On older kernels, please consider adding
5074 the kernel modules to /etc/modules-load.d/ as a work-around.
5076 * The resolv.conf file systemd-resolved generates has been
5077 moved to /run/systemd/resolve/. If you have a symlink from
5078 /etc/resolv.conf, it might be necessary to correct it.
5080 * Two new service settings, ProtectHome= and ProtectSystem=,
5081 have been added. When enabled, they will make the user data
5082 (such as /home) inaccessible or read-only and the system
5083 (such as /usr) read-only, for specific services. This allows
5084 very light-weight per-service sandboxing to avoid
5085 modifications of user data or system files from
5086 services. These two new switches have been enabled for all
5087 of systemd's long-running services, where appropriate.
5089 * Socket units gained new SocketUser= and SocketGroup=
5090 settings to set the owner user and group of AF_UNIX sockets
5091 and FIFOs in the file system.
5093 * Socket units gained a new RemoveOnStop= setting. If enabled,
5094 all FIFOS and sockets in the file system will be removed
5095 when the specific socket unit is stopped.
5097 * Socket units gained a new Symlinks= setting. It takes a list
5098 of symlinks to create to file system sockets or FIFOs
5099 created by the specific Unix sockets. This is useful to
5100 manage symlinks to socket nodes with the same life-cycle as
5103 * The /dev/log socket and /dev/initctl FIFO have been moved to
5104 /run, and have been replaced by symlinks. This allows
5105 connecting to these facilities even if PrivateDevices=yes is
5106 used for a service (which makes /dev/log itself unavailable,
5107 but /run is left). This also has the benefit of ensuring
5108 that /dev only contains device nodes, directories and
5109 symlinks, and nothing else.
5111 * sd-daemon gained two new calls sd_pid_notify() and
5112 sd_pid_notifyf(). They are similar to sd_notify() and
5113 sd_notifyf(), but allow overriding of the source PID of
5114 notification messages if permissions permit this. This is
5115 useful to send notify messages on behalf of a different
5116 process (for example, the parent process). The
5117 systemd-notify tool has been updated to make use of this
5118 when sending messages (so that notification messages now
5119 originate from the shell script invoking systemd-notify and
5120 not the systemd-notify process itself. This should minimize
5121 a race where systemd fails to associate notification
5122 messages to services when the originating process already
5125 * A new "on-abnormal" setting for Restart= has been added. If
5126 set, it will result in automatic restarts on all "abnormal"
5127 reasons for a process to exit, which includes unclean
5128 signals, core dumps, timeouts and watchdog timeouts, but
5129 does not include clean and unclean exit codes or clean
5130 signals. Restart=on-abnormal is an alternative for
5131 Restart=on-failure for services that shall be able to
5132 terminate and avoid restarts on certain errors, by
5133 indicating so with an unclean exit code. Restart=on-failure
5134 or Restart=on-abnormal is now the recommended setting for
5135 all long-running services.
5137 * If the InaccessibleDirectories= service setting points to a
5138 mount point (or if there are any submounts contained within
5139 it), it is now attempted to completely unmount it, to make
5140 the file systems truly unavailable for the respective
5143 * The ReadOnlyDirectories= service setting and
5144 systemd-nspawn's --read-only parameter are now recursively
5145 applied to all submounts, too.
5147 * Mount units may now be created transiently via the bus APIs.
5149 * The support for SysV and LSB init scripts has been removed
5150 from the systemd daemon itself. Instead, it is now
5151 implemented as a generator that creates native systemd units
5152 from these scripts when needed. This enables us to remove a
5153 substantial amount of legacy code from PID 1, following the
5154 fact that many distributions only ship a very small number
5155 of LSB/SysV init scripts nowadays.
5157 * Privileged Xen (dom0) domains are not considered
5158 virtualization anymore by the virtualization detection
5159 logic. After all, they generally have unrestricted access to
5160 the hardware and usually are used to manage the unprivileged
5163 * systemd-tmpfiles gained a new "C" line type, for copying
5164 files or entire directories.
5166 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
5167 lines. So far, they have been non-globbing versions of the
5168 latter, and have thus been redundant. In future, it is
5169 recommended to only use "z". "m" has hence been removed
5170 from the documentation, even though it stays supported.
5172 * A tmpfiles snippet to recreate the most basic structure in
5173 /var has been added. This is enough to create the /var/run →
5174 /run symlink and create a couple of structural
5175 directories. This allows systems to boot up with an empty or
5176 volatile /var. Of course, while with this change, the core OS
5177 now is capable with dealing with a volatile /var, not all
5178 user services are ready for it. However, we hope that sooner
5179 or later, many service daemons will be changed upstream so
5180 that they are able to automatically create their necessary
5181 directories in /var at boot, should they be missing. This is
5182 the first step to allow state-less systems that only require
5183 the vendor image for /usr to boot.
5185 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5186 empty tmpfs instance to a specific directory. This is
5187 particularly useful for making use of the automatic
5188 reconstruction of /var (see above), by passing --tmpfs=/var.
5190 * Access modes specified in tmpfiles snippets may now be
5191 prefixed with "~", which indicates that they shall be masked
5192 by whether the existing file or directory is currently
5193 writable, readable or executable at all. Also, if specified,
5194 the sgid/suid/sticky bits will be masked for all
5197 * A new passive target unit "network-pre.target" has been
5198 added which is useful for services that shall run before any
5199 network is configured, for example firewall scripts.
5201 * The "floppy" group that previously owned the /dev/fd*
5202 devices is no longer used. The "disk" group is now used
5203 instead. Distributions should probably deprecate usage of
5206 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5207 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5208 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5209 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5210 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5211 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5214 — Berlin, 2014-06-11
5218 * A new "systemd-timesyncd" daemon has been added for
5219 synchronizing the system clock across the network. It
5220 implements an SNTP client. In contrast to NTP
5221 implementations such as chrony or the NTP reference server,
5222 this only implements a client side, and does not bother with
5223 the full NTP complexity, focusing only on querying time from
5224 one remote server and synchronizing the local clock to
5225 it. Unless you intend to serve NTP to networked clients or
5226 want to connect to local hardware clocks, this simple NTP
5227 client should be more than appropriate for most
5228 installations. The daemon runs with minimal privileges, and
5229 has been hooked up with networkd to only operate when
5230 network connectivity is available. The daemon saves the
5231 current clock to disk every time a new NTP sync has been
5232 acquired, and uses this to possibly correct the system clock
5233 early at bootup, in order to accommodate for systems that
5234 lack an RTC such as the Raspberry Pi and embedded devices,
5235 and to make sure that time monotonically progresses on these
5236 systems, even if it is not always correct. To make use of
5237 this daemon, a new system user and group "systemd-timesync"
5238 needs to be created on installation of systemd.
5240 * The queue "seqnum" interface of libudev has been disabled, as
5241 it was generally incompatible with device namespacing as
5242 sequence numbers of devices go "missing" if the devices are
5243 part of a different namespace.
5245 * "systemctl list-timers" and "systemctl list-sockets" gained
5246 a --recursive switch for showing units of these types also
5247 for all local containers, similar in style to the already
5248 supported --recursive switch for "systemctl list-units".
5250 * A new RebootArgument= setting has been added for service
5251 units, which may be used to specify a kernel reboot argument
5252 to use when triggering reboots with StartLimitAction=.
5254 * A new FailureAction= setting has been added for service
5255 units which may be used to specify an operation to trigger
5256 when a service fails. This works similarly to
5257 StartLimitAction=, but unlike it, controls what is done
5258 immediately rather than only after several attempts to
5259 restart the service in question.
5261 * hostnamed got updated to also expose the kernel name,
5262 release, and version on the bus. This is useful for
5263 executing commands like hostnamectl with the -H switch.
5264 systemd-analyze makes use of this to properly display
5265 details when running non-locally.
5267 * The bootchart tool can now show cgroup information in the
5268 graphs it generates.
5270 * The CFS CPU quota cgroup attribute is now exposed for
5271 services. The new CPUQuota= switch has been added for this
5272 which takes a percentage value. Setting this will have the
5273 result that a service may never get more CPU time than the
5274 specified percentage, even if the machine is otherwise idle.
5276 * systemd-networkd learned IPIP and SIT tunnel support.
5278 * LSB init scripts exposing a dependency on $network will now
5279 get a dependency on network-online.target rather than simply
5280 network.target. This should bring LSB handling closer to
5281 what it was on SysV systems.
5283 * A new fsck.repair= kernel option has been added to control
5284 how fsck shall deal with unclean file systems at boot.
5286 * The (.ini) configuration file parser will now silently
5287 ignore sections whose name begins with "X-". This may be
5288 used to maintain application-specific extension sections in unit
5291 * machined gained a new API to query the IP addresses of
5292 registered containers. "machinectl status" has been updated
5293 to show these addresses in its output.
5295 * A new call sd_uid_get_display() has been added to the
5296 sd-login APIs for querying the "primary" session of a
5297 user. The "primary" session of the user is elected from the
5298 user's sessions and generally a graphical session is
5299 preferred over a text one.
5301 * A minimal systemd-resolved daemon has been added. It
5302 currently simply acts as a companion to systemd-networkd and
5303 manages resolv.conf based on per-interface DNS
5304 configuration, possibly supplied via DHCP. In the long run
5305 we hope to extend this into a local DNSSEC enabled DNS and
5308 * The systemd-networkd-wait-online tool is now enabled by
5309 default. It will delay network-online.target until a network
5310 connection has been configured. The tool primarily integrates
5311 with networkd, but will also make a best effort to make sense
5312 of network configuration performed in some other way.
5314 * Two new service options StartupCPUShares= and
5315 StartupBlockIOWeight= have been added that work similarly to
5316 CPUShares= and BlockIOWeight= however only apply during
5317 system startup. This is useful to prioritize certain services
5318 differently during bootup than during normal runtime.
5320 * hostnamed has been changed to prefer the statically
5321 configured hostname in /etc/hostname (unless set to
5322 'localhost' or empty) over any dynamic one supplied by
5323 dhcp. With this change, the rules for picking the hostname
5324 match more closely the rules of other configuration settings
5325 where the local administrator's configuration in /etc always
5326 overrides any other settings.
5328 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
5329 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5330 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5331 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5332 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5333 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5334 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5335 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5336 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
5337 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5338 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5339 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5340 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5341 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5342 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5343 Lindskog, WaLyong Cho, Will Woods, Zbigniew
5346 — Beijing, 2014-05-28
5350 * When restoring the screen brightness at boot, stay away from
5351 the darkest setting or from the lowest 5% of the available
5352 range, depending on which is the larger value of both. This
5353 should effectively protect the user from rebooting into a
5354 black screen, should the brightness have been set to minimum
5357 * sd-login gained a new sd_machine_get_class() call to
5358 determine the class ("vm" or "container") of a machine
5359 registered with machined.
5361 * sd-login gained new calls
5362 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5363 to query the identity of the peer of a local AF_UNIX
5364 connection. They operate similarly to their sd_pid_get_xyz()
5367 * PID 1 will now maintain a system-wide system state engine
5368 with the states "starting", "running", "degraded",
5369 "maintenance", "stopping". These states are bound to system
5370 startup, normal runtime, runtime with at least one failed
5371 service, rescue/emergency mode and system shutdown. This
5372 state is shown in the "systemctl status" output when no unit
5373 name is passed. It is useful to determine system state, in
5374 particularly when doing so for many systems or containers at
5377 * A new command "list-machines" has been added to "systemctl"
5378 that lists all local OS containers and shows their system
5379 state (see above), if systemd runs inside of them.
5381 * systemctl gained a new "-r" switch to recursively enumerate
5382 units on all local containers, when used with the
5383 "list-unit" command (which is the default one that is
5384 executed when no parameters are specified).
5386 * The GPT automatic partition discovery logic will now honour
5387 two GPT partition flags: one may be set on a partition to
5388 cause it to be mounted read-only, and the other may be set
5389 on a partition to ignore it during automatic discovery.
5391 * Two new GPT type UUIDs have been added for automatic root
5392 partition discovery, for 32-bit and 64-bit ARM. This is not
5393 particularly useful for discovering the root directory on
5394 these architectures during bare-metal boots (since UEFI is
5395 not common there), but still very useful to allow booting of
5396 ARM disk images in nspawn with the -i option.
5398 * MAC addresses of interfaces created with nspawn's
5399 --network-interface= switch will now be generated from the
5400 machine name, and thus be stable between multiple invocations
5403 * logind will now automatically remove all IPC objects owned
5404 by a user if she or he fully logs out. This makes sure that
5405 users who are logged out cannot continue to consume IPC
5406 resources. This covers SysV memory, semaphores and message
5407 queues as well as POSIX shared memory and message
5408 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5409 limits. With this functionality, that is corrected. This may
5410 be turned off by using the RemoveIPC= switch of logind.conf.
5412 * The systemd-machine-id-setup and tmpfiles tools gained a
5413 --root= switch to operate on a specific root directory,
5416 * journald can now forward logged messages to the TTYs of all
5417 logged in users ("wall"). This is the default for all
5418 emergency messages now.
5420 * A new tool systemd-journal-remote has been added to stream
5421 journal log messages across the network.
5423 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5424 controller trees are mounted into it. Note that the
5425 directories mounted beneath it are not read-only. This is a
5426 security measure and is particularly useful because glibc
5427 actually includes a search logic to pick any tmpfs it can
5428 find to implement shm_open() if /dev/shm is not available
5429 (which it might very well be in namespaced setups).
5431 * machinectl gained a new "poweroff" command to cleanly power
5432 down a local OS container.
5434 * The PrivateDevices= unit file setting will now also drop the
5435 CAP_MKNOD capability from the capability bound set, and
5436 imply DevicePolicy=closed.
5438 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5439 comprehensively on all long-running systemd services where
5440 this is appropriate.
5442 * systemd-udevd will now run in a disassociated mount
5443 namespace. To mount directories from udev rules, make sure to
5444 pull in mount units via SYSTEMD_WANTS properties.
5446 * The kdbus support gained support for uploading policy into
5447 the kernel. sd-bus gained support for creating "monitoring"
5448 connections that can eavesdrop into all bus communication
5449 for debugging purposes.
5451 * Timestamps may now be specified in seconds since the UNIX
5452 epoch Jan 1st, 1970 by specifying "@" followed by the value
5455 * Native tcpwrap support in systemd has been removed. tcpwrap
5456 is old code, not really maintained anymore and has serious
5457 shortcomings, and better options such as firewalls
5458 exist. For setups that require tcpwrap usage, please
5459 consider invoking your socket-activated service via tcpd,
5460 like on traditional inetd.
5462 * A new system.conf configuration option
5463 DefaultTimerAccuracySec= has been added that controls the
5464 default AccuracySec= setting of .timer units.
5466 * Timer units gained a new WakeSystem= switch. If enabled,
5467 timers configured this way will cause the system to resume
5468 from system suspend (if the system supports that, which most
5471 * Timer units gained a new Persistent= switch. If enabled,
5472 timers configured this way will save to disk when they have
5473 been last triggered. This information is then used on next
5474 reboot to possible execute overdue timer events, that
5475 could not take place because the system was powered off.
5476 This enables simple anacron-like behaviour for timer units.
5478 * systemctl's "list-timers" will now also list the time a
5479 timer unit was last triggered in addition to the next time
5480 it will be triggered.
5482 * systemd-networkd will now assign predictable IPv4LL
5483 addresses to its local interfaces.
5485 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5486 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5487 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5488 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5489 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5490 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5491 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5492 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5495 — Berlin, 2014-03-25
5499 * A new unit file setting RestrictAddressFamilies= has been
5500 added to restrict which socket address families unit
5501 processes gain access to. This takes address family names
5502 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5503 attack surface of services via exotic protocol stacks. This
5504 is built on seccomp system call filters.
5506 * Two new unit file settings RuntimeDirectory= and
5507 RuntimeDirectoryMode= have been added that may be used to
5508 manage a per-daemon runtime directories below /run. This is
5509 an alternative for setting up directory permissions with
5510 tmpfiles snippets, and has the advantage that the runtime
5511 directory's lifetime is bound to the daemon runtime and that
5512 the daemon starts up with an empty directory each time. This
5513 is particularly useful when writing services that drop
5514 privileges using the User= or Group= setting.
5516 * The DeviceAllow= unit setting now supports globbing for
5517 matching against device group names.
5519 * The systemd configuration file system.conf gained new
5520 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5521 DefaultMemoryAccounting= to globally turn on/off accounting
5522 for specific resources (cgroups) for all units. These
5523 settings may still be overridden individually in each unit
5526 * systemd-gpt-auto-generator is now able to discover /srv and
5527 root partitions in addition to /home and swap partitions. It
5528 also supports LUKS-encrypted partitions now. With this in
5529 place, automatic discovery of partitions to mount following
5530 the Discoverable Partitions Specification
5531 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
5532 is now a lot more complete. This allows booting without
5533 /etc/fstab and without root= on the kernel command line on
5534 systems prepared appropriately.
5536 * systemd-nspawn gained a new --image= switch which allows
5537 booting up disk images and Linux installations on any block
5538 device that follow the Discoverable Partitions Specification
5539 (see above). This means that installations made with
5540 appropriately updated installers may now be started and
5541 deployed using container managers, completely
5542 unmodified. (We hope that libvirt-lxc will add support for
5543 this feature soon, too.)
5545 * systemd-nspawn gained a new --network-macvlan= setting to
5546 set up a private macvlan interface for the
5547 container. Similarly, systemd-networkd gained a new
5548 Kind=macvlan setting in .netdev files.
5550 * systemd-networkd now supports configuring local addresses
5553 * A new tool systemd-network-wait-online has been added to
5554 synchronously wait for network connectivity using
5557 * The sd-bus.h bus API gained a new sd_bus_track object for
5558 tracking the life-cycle of bus peers. Note that sd-bus.h is
5559 still not a public API though (unless you specify
5560 --enable-kdbus on the configure command line, which however
5561 voids your warranty and you get no API stability guarantee).
5563 * The $XDG_RUNTIME_DIR runtime directories for each user are
5564 now individual tmpfs instances, which has the benefit of
5565 introducing separate pools for each user, with individual
5566 size limits, and thus making sure that unprivileged clients
5567 can no longer negatively impact the system or other users by
5568 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5569 RuntimeDirectorySize= has been introduced that allows
5570 controlling the default size limit for all users. It
5571 defaults to 10% of the available physical memory. This is no
5572 replacement for quotas on tmpfs though (which the kernel
5573 still does not support), as /dev/shm and /tmp are still
5574 shared resources used by both the system and unprivileged
5577 * logind will now automatically turn off automatic suspending
5578 on laptop lid close when more than one display is
5579 connected. This was previously expected to be implemented
5580 individually in desktop environments (such as GNOME),
5581 however has been added to logind now, in order to fix a
5582 boot-time race where a desktop environment might not have
5583 been started yet and thus not been able to take an inhibitor
5584 lock at the time where logind already suspends the system
5585 due to a closed lid.
5587 * logind will now wait at least 30s after each system
5588 suspend/resume cycle, and 3min after system boot before
5589 suspending the system due to a closed laptop lid. This
5590 should give USB docking stations and similar enough time to
5591 be probed and configured after system resume and boot in
5592 order to then act as suspend blocker.
5594 * systemd-run gained a new --property= setting which allows
5595 initialization of resource control properties (and others)
5596 for the created scope or service unit. Example: "systemd-run
5597 --property=BlockIOWeight=10 updatedb" may be used to run
5598 updatedb at a low block IO scheduling weight.
5600 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5601 now also work in --scope mode.
5603 * When systemd is compiled with kdbus support, basic support
5604 for enforced policies is now in place. (Note that enabling
5605 kdbus still voids your warranty and no API compatibility
5608 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5609 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5610 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5611 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5612 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5613 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5614 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5615 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5616 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5617 Zbigniew Jędrzejewski-Szmek
5619 — Berlin, 2014-03-12
5623 * systemd will now relabel /dev after loading the SMACK policy
5624 according to SMACK rules.
5626 * A new unit file option AppArmorProfile= has been added to
5627 set the AppArmor profile for the processes of a unit.
5629 * A new condition check ConditionArchitecture= has been added
5630 to conditionalize units based on the system architecture, as
5631 reported by uname()'s "machine" field.
5633 * systemd-networkd now supports matching on the system
5634 virtualization, architecture, kernel command line, host name
5637 * logind is now a lot more aggressive when suspending the
5638 machine due to a closed laptop lid. Instead of acting only
5639 on the lid close action, it will continuously watch the lid
5640 status and act on it. This is useful for laptops where the
5641 power button is on the outside of the chassis so that it can
5642 be reached without opening the lid (such as the Lenovo
5643 Yoga). On those machines, logind will now immediately
5644 re-suspend the machine if the power button has been
5645 accidentally pressed while the laptop was suspended and in a
5646 backpack or similar.
5648 * logind will now watch SW_DOCK switches and inhibit reaction
5649 to the lid switch if it is pressed. This means that logind
5650 will not suspend the machine anymore if the lid is closed
5651 and the system is docked, if the laptop supports SW_DOCK
5652 notifications via the input layer. Note that ACPI docking
5653 stations do not generate this currently. Also note that this
5654 logic is usually not fully sufficient and Desktop
5655 Environments should take a lid switch inhibitor lock when an
5656 external display is connected, as systemd will not watch
5659 * nspawn will now make use of the devices cgroup controller by
5660 default, and only permit creation of and access to the usual
5661 API device nodes like /dev/null or /dev/random, as well as
5662 access to (but not creation of) the pty devices.
5664 * We will now ship a default .network file for
5665 systemd-networkd that automatically configures DHCP for
5666 network interfaces created by nspawn's --network-veth or
5667 --network-bridge= switches.
5669 * systemd will now understand the usual M, K, G, T suffixes
5670 according to SI conventions (i.e. to the base 1000) when
5671 referring to throughput and hardware metrics. It will stay
5672 with IEC conventions (i.e. to the base 1024) for software
5673 metrics, according to what is customary according to
5674 Wikipedia. We explicitly document which base applies for
5675 each configuration option.
5677 * The DeviceAllow= setting in unit files now supports a syntax
5678 to whitelist an entire group of devices node majors at once,
5679 based on the /proc/devices listing. For example, with the
5680 string "char-pts", it is now possible to whitelist all
5681 current and future pseudo-TTYs at once.
5683 * sd-event learned a new "post" event source. Event sources of
5684 this type are triggered by the dispatching of any event
5685 source of a type that is not "post". This is useful for
5686 implementing clean-up and check event sources that are
5687 triggered by other work being done in the program.
5689 * systemd-networkd is no longer statically enabled, but uses
5690 the usual [Install] sections so that it can be
5691 enabled/disabled using systemctl. It still is enabled by
5694 * When creating a veth interface pair with systemd-nspawn, the
5695 host side will now be prefixed with "vb-" if
5696 --network-bridge= is used, and with "ve-" if --network-veth
5697 is used. This way, it is easy to distinguish these cases on
5698 the host, for example to apply different configuration to
5699 them with systemd-networkd.
5701 * The compatibility libraries for libsystemd-journal.so,
5702 libsystem-id128.so, libsystemd-login.so and
5703 libsystemd-daemon.so do not make use of IFUNC
5704 anymore. Instead, we now build libsystemd.so multiple times
5705 under these alternative names. This means that the footprint
5706 is drastically increased, but given that these are
5707 transitional compatibility libraries, this should not matter
5708 much. This change has been made necessary to support the ARM
5709 platform for these compatibility libraries, as the ARM
5710 toolchain is not really at the same level as the toolchain
5711 for other architectures like x86 and does not support
5712 IFUNC. Please make sure to use --enable-compat-libs only
5713 during a transitional period!
5715 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
5716 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5717 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5718 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5719 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5720 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5721 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5722 Zbigniew Jędrzejewski-Szmek
5724 — Berlin, 2014-02-24
5728 * A new component "systemd-networkd" has been added that can
5729 be used to configure local network interfaces statically or
5730 via DHCP. It is capable of bringing up bridges, VLANs, and
5731 bonding. Currently, no hook-ups for interactive network
5732 configuration are provided. Use this for your initrd,
5733 container, embedded, or server setup if you need a simple,
5734 yet powerful, network configuration solution. This
5735 configuration subsystem is quite nifty, as it allows wildcard
5736 hotplug matching in interfaces. For example, with a single
5737 configuration snippet, you can configure that all Ethernet
5738 interfaces showing up are automatically added to a bridge,
5739 or similar. It supports link-sensing and more.
5741 * A new tool "systemd-socket-proxyd" has been added which can
5742 act as a bidirectional proxy for TCP sockets. This is
5743 useful for adding socket activation support to services that
5744 do not actually support socket activation, including virtual
5745 machines and the like.
5747 * Add a new tool to save/restore rfkill state on
5750 * Save/restore state of keyboard backlights in addition to
5751 display backlights on shutdown/boot.
5753 * udev learned a new SECLABEL{} construct to label device
5754 nodes with a specific security label when they appear. For
5755 now, only SECLABEL{selinux} is supported, but the syntax is
5756 prepared for additional security frameworks.
5758 * udev gained a new scheme to configure link-level attributes
5759 from files in /etc/systemd/network/*.link. These files can
5760 match against MAC address, device path, driver name and type,
5761 and will apply attributes like the naming policy, link speed,
5762 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
5763 address assignment policy (randomized, ...).
5765 * The configuration of network interface naming rules for
5766 "permanent interface names" has changed: a new NamePolicy=
5767 setting in the [Link] section of .link files determines the
5768 priority of possible naming schemes (onboard, slot, MAC,
5769 path). The default value of this setting is determined by
5770 /usr/lib/net/links/99-default.link. Old
5771 80-net-name-slot.rules udev configuration file has been
5772 removed, so local configuration overriding this file should
5773 be adapted to override 99-default.link instead.
5775 * When the User= switch is used in a unit file, also
5776 initialize $SHELL= based on the user database entry.
5778 * systemd no longer depends on libdbus. All communication is
5779 now done with sd-bus, systemd's low-level bus library
5782 * kdbus support has been added to PID 1 itself. When kdbus is
5783 enabled, this causes PID 1 to set up the system bus and
5784 enable support for a new ".busname" unit type that
5785 encapsulates bus name activation on kdbus. It works a little
5786 bit like ".socket" units, except for bus names. A new
5787 generator has been added that converts classic dbus1 service
5788 activation files automatically into native systemd .busname
5791 * sd-bus: add a light-weight vtable implementation that allows
5792 defining objects on the bus with a simple static const
5793 vtable array of its methods, signals and properties.
5795 * systemd will not generate or install static dbus
5796 introspection data anymore to /usr/share/dbus-1/interfaces,
5797 as the precise format of these files is unclear, and
5798 nothing makes use of it.
5800 * A proxy daemon is now provided to proxy clients connecting
5801 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5802 compatibility with classic D-Bus.
5804 * A bus driver implementation has been added that supports the
5805 classic D-Bus bus driver calls on kdbus, also for
5806 compatibility purposes.
5808 * A new API "sd-event.h" has been added that implements a
5809 minimal event loop API built around epoll. It provides a
5810 couple of features that direct epoll usage is lacking:
5811 prioritization of events, scales to large numbers of timer
5812 events, per-event timer slack (accuracy), system-wide
5813 coalescing of timer events, exit handlers, watchdog
5814 supervision support using systemd's sd_notify() API, child
5817 * A new API "sd-rntl.h" has been added that provides an API
5818 around the route netlink interface of the kernel, similar in
5819 style to "sd-bus.h".
5821 * A new API "sd-dhcp-client.h" has been added that provides a
5822 small DHCPv4 client-side implementation. This is used by
5825 * There is a new kernel command line option
5826 "systemd.restore_state=0|1". When set to "0", none of the
5827 systemd tools will restore saved runtime state to hardware
5828 devices. More specifically, the rfkill and backlight states
5831 * The FsckPassNo= compatibility option in mount/service units
5832 has been removed. The fstab generator will now add the
5833 necessary dependencies automatically, and does not require
5834 PID1's support for that anymore.
5836 * journalctl gained a new switch, --list-boots, that lists
5837 recent boots with their times and boot IDs.
5839 * The various tools like systemctl, loginctl, timedatectl,
5840 busctl, systemd-run, ... have gained a new switch "-M" to
5841 connect to a specific, local OS container (as direct
5842 connection, without requiring SSH). This works on any
5843 container that is registered with machined, such as those
5844 created by libvirt-lxc or nspawn.
5846 * systemd-run and systemd-analyze also gained support for "-H"
5847 to connect to remote hosts via SSH. This is particularly
5848 useful for systemd-run because it enables queuing of jobs
5849 onto remote systems.
5851 * machinectl gained a new command "login" to open a getty
5852 login in any local container. This works with any container
5853 that is registered with machined (such as those created by
5854 libvirt-lxc or nspawn), and which runs systemd inside.
5856 * machinectl gained a new "reboot" command that may be used to
5857 trigger a reboot on a specific container that is registered
5858 with machined. This works on any container that runs an init
5859 system of some kind.
5861 * systemctl gained a new "list-timers" command to print a nice
5862 listing of installed timer units with the times they elapse
5865 * Alternative reboot() parameters may now be specified on the
5866 "systemctl reboot" command line and are passed to the
5867 reboot() system call.
5869 * systemctl gained a new --job-mode= switch to configure the
5870 mode to queue a job with. This is a more generic version of
5871 --fail, --irreversible, and --ignore-dependencies, which are
5872 still available but not advertised anymore.
5874 * /etc/systemd/system.conf gained new settings to configure
5875 various default timeouts of units, as well as the default
5876 start limit interval and burst. These may still be overridden
5879 * PID1 will now export on the bus profile data of the security
5880 policy upload process (such as the SELinux policy upload to
5883 * journald: when forwarding logs to the console, include
5884 timestamps (following the setting in
5885 /sys/module/printk/parameters/time).
5887 * OnCalendar= in timer units now understands the special
5888 strings "yearly" and "annually". (Both are equivalent)
5890 * The accuracy of timer units is now configurable with the new
5891 AccuracySec= setting. It defaults to 1min.
5893 * A new dependency type JoinsNamespaceOf= has been added that
5894 allows running two services within the same /tmp and network
5895 namespace, if PrivateNetwork= or PrivateTmp= are used.
5897 * A new command "cat" has been added to systemctl. It outputs
5898 the original unit file of a unit, and concatenates the
5899 contents of additional "drop-in" unit file snippets, so that
5900 the full configuration is shown.
5902 * systemctl now supports globbing on the various "list-xyz"
5903 commands, like "list-units" or "list-sockets", as well as on
5904 those commands which take multiple unit names.
5906 * journalctl's --unit= switch gained support for globbing.
5908 * All systemd daemons now make use of the watchdog logic so
5909 that systemd automatically notices when they hang.
5911 * If the $container_ttys environment variable is set,
5912 getty-generator will automatically spawn a getty for each
5913 listed tty. This is useful for container managers to request
5914 login gettys to be spawned on as many ttys as needed.
5916 * %h, %s, %U specifier support is not available anymore when
5917 used in unit files for PID 1. This is because NSS calls are
5918 not safe from PID 1. They stay available for --user
5919 instances of systemd, and as special case for the root user.
5921 * loginctl gained a new "--no-legend" switch to turn off output
5924 * The "sd-login.h" API gained three new calls:
5925 sd_session_is_remote(), sd_session_get_remote_user(),
5926 sd_session_get_remote_host() to query information about
5929 * The udev hardware database now also carries vendor/product
5930 information of SDIO devices.
5932 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5933 determine whether watchdog notifications are requested by
5936 * Socket-activated per-connection services now include a
5937 short description of the connection parameters in the
5940 * tmpfiles gained a new "--boot" option. When this is not used,
5941 only lines where the command character is not suffixed with
5942 "!" are executed. When this option is specified, those
5943 options are executed too. This partitions tmpfiles
5944 directives into those that can be safely executed at any
5945 time, and those which should be run only at boot (for
5946 example, a line that creates /run/nologin).
5948 * A new API "sd-resolve.h" has been added which provides a simple
5949 asynchronous wrapper around glibc NSS host name resolution
5950 calls, such as getaddrinfo(). In contrast to glibc's
5951 getaddrinfo_a(), it does not use signals. In contrast to most
5952 other asynchronous name resolution libraries, this one does
5953 not reimplement DNS, but reuses NSS, so that alternate
5954 host name resolution systems continue to work, such as mDNS,
5955 LDAP, etc. This API is based on libasyncns, but it has been
5956 cleaned up for inclusion in systemd.
5958 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5959 "sd-daemon.h" are no longer found in individual libraries
5960 libsystemd-journal.so, libsystemd-login.so,
5961 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
5962 merged them into a single library, libsystemd.so, which
5963 provides all symbols. The reason for this is cyclic
5964 dependencies, as these libraries tend to use each other's
5965 symbols. So far, we have managed to workaround that by linking
5966 a copy of a good part of our code into each of these
5967 libraries again and again, which, however, makes certain
5968 things hard to do, like sharing static variables. Also, it
5969 substantially increases footprint. With this change, there
5970 is only one library for the basic APIs systemd
5971 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5972 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5973 library as well, however are subject to the --enable-kdbus
5974 switch (see below). Note that "sd-dhcp-client.h" is not part
5975 of this library (this is because it only consumes, never
5976 provides, services of/to other APIs). To make the transition
5977 easy from the separate libraries to the unified one, we
5978 provide the --enable-compat-libs compile-time switch which
5979 will generate stub libraries that are compatible with the
5980 old ones but redirect all calls to the new one.
5982 * All of the kdbus logic and the new APIs "sd-bus.h",
5983 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
5984 and "sd-utf8.h" are compile-time optional via the
5985 "--enable-kdbus" switch, and they are not compiled in by
5986 default. To make use of kdbus, you have to explicitly enable
5987 the switch. Note however, that neither the kernel nor the
5988 userspace API for all of this is considered stable yet. We
5989 want to maintain the freedom to still change the APIs for
5990 now. By specifying this build-time switch, you acknowledge
5991 that you are aware of the instability of the current
5994 * Also, note that while kdbus is pretty much complete,
5995 it lacks one thing: proper policy support. This means you
5996 can build a fully working system with all features; however,
5997 it will be highly insecure. Policy support will be added in
5998 one of the next releases, at the same time that we will
5999 declare the APIs stable.
6001 * When the kernel command line argument "kdbus" is specified,
6002 systemd will automatically load the kdbus.ko kernel module. At
6003 this stage of development, it is only useful for testing kdbus
6004 and should not be used in production. Note: if "--enable-kdbus"
6005 is specified, and the kdbus.ko kernel module is available, and
6006 "kdbus" is added to the kernel command line, the entire system
6007 runs with kdbus instead of dbus-daemon, with the above mentioned
6008 problem of missing the system policy enforcement. Also a future
6009 version of kdbus.ko or a newer systemd will not be compatible with
6010 each other, and will unlikely be able to boot the machine if only
6011 one of them is updated.
6013 * systemctl gained a new "import-environment" command which
6014 uploads the caller's environment (or parts thereof) into the
6015 service manager so that it is inherited by services started
6016 by the manager. This is useful to upload variables like
6017 $DISPLAY into the user service manager.
6019 * A new PrivateDevices= switch has been added to service units
6020 which allows running a service with a namespaced /dev
6021 directory that does not contain any device nodes for
6022 physical devices. More specifically, it only includes devices
6023 such as /dev/null, /dev/urandom, and /dev/zero which are API
6026 * logind has been extended to support behaviour like VT
6027 switching on seats that do not support a VT. This makes
6028 multi-session available on seats that are not the first seat
6029 (seat0), and on systems where kernel support for VTs has
6030 been disabled at compile-time.
6032 * If a process holds a delay lock for system sleep or shutdown
6033 and fails to release it in time, we will now log its
6034 identity. This makes it easier to identify processes that
6035 cause slow suspends or power-offs.
6037 * When parsing /etc/crypttab, support for a new key-slot=
6038 option as supported by Debian is added. It allows indicating
6039 which LUKS slot to use on disk, speeding up key loading.
6041 * The sd_journal_sendv() API call has been checked and
6042 officially declared to be async-signal-safe so that it may
6043 be invoked from signal handlers for logging purposes.
6045 * Boot-time status output is now enabled automatically after a
6046 short timeout if boot does not progress, in order to give
6047 the user an indication what she or he is waiting for.
6049 * The boot-time output has been improved to show how much time
6050 remains until jobs expire.
6052 * The KillMode= switch in service units gained a new possible
6053 value "mixed". If set, and the unit is shut down, then the
6054 initial SIGTERM signal is sent only to the main daemon
6055 process, while the following SIGKILL signal is sent to
6056 all remaining processes of the service.
6058 * When a scope unit is registered, a new property "Controller"
6059 may be set. If set to a valid bus name, systemd will send a
6060 RequestStop() signal to this name when it would like to shut
6061 down the scope. This may be used to hook manager logic into
6062 the shutdown logic of scope units. Also, scope units may now
6063 be put in a special "abandoned" state, in which case the
6064 manager process which created them takes no further
6065 responsibilities for it.
6067 * When reading unit files, systemd will now verify
6068 the access mode of these files, and warn about certain
6069 suspicious combinations. This has been added to make it
6070 easier to track down packaging bugs where unit files are
6071 marked executable or world-writable.
6073 * systemd-nspawn gained a new "--setenv=" switch to set
6074 container-wide environment variables. The similar option in
6075 systemd-activate was renamed from "--environment=" to
6076 "--setenv=" for consistency.
6078 * systemd-nspawn has been updated to create a new kdbus domain
6079 for each container that is invoked, thus allowing each
6080 container to have its own set of system and user buses,
6081 independent of the host.
6083 * systemd-nspawn gained a new --drop-capability= switch to run
6084 the container with less capabilities than the default. Both
6085 --drop-capability= and --capability= now take the special
6086 string "all" for dropping or keeping all capabilities.
6088 * systemd-nspawn gained new switches for executing containers
6089 with specific SELinux labels set.
6091 * systemd-nspawn gained a new --quiet switch to not generate
6092 any additional output but the container's own console
6095 * systemd-nspawn gained a new --share-system switch to run a
6096 container without PID namespacing enabled.
6098 * systemd-nspawn gained a new --register= switch to control
6099 whether the container is registered with systemd-machined or
6100 not. This is useful for containers that do not run full
6101 OS images, but only specific apps.
6103 * systemd-nspawn gained a new --keep-unit which may be used
6104 when invoked as the only program from a service unit, and
6105 results in registration of the unit service itself in
6106 systemd-machined, instead of a newly opened scope unit.
6108 * systemd-nspawn gained a new --network-interface= switch for
6109 moving arbitrary interfaces to the container. The new
6110 --network-veth switch creates a virtual Ethernet connection
6111 between host and container. The new --network-bridge=
6112 switch then allows assigning the host side of this virtual
6113 Ethernet connection to a bridge device.
6115 * systemd-nspawn gained a new --personality= switch for
6116 setting the kernel personality for the container. This is
6117 useful when running a 32-bit container on a 64-bit host. A
6118 similar option Personality= is now also available for service
6121 * logind will now also track a "Desktop" identifier for each
6122 session which encodes the desktop environment of it. This is
6123 useful for desktop environments that want to identify
6124 multiple running sessions of itself easily.
6126 * A new SELinuxContext= setting for service units has been
6127 added that allows setting a specific SELinux execution
6128 context for a service.
6130 * Most systemd client tools will now honour $SYSTEMD_LESS for
6131 settings of the "less" pager. By default, these tools will
6132 override $LESS to allow certain operations to work, such as
6133 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6134 influence this logic.
6136 * systemd's "seccomp" hook-up has been changed to make use of
6137 the libseccomp library instead of using its own
6138 implementation. This has benefits for portability among
6141 * For usage together with SystemCallFilter=, a new
6142 SystemCallErrorNumber= setting has been introduced that
6143 allows configuration of a system error number to be returned
6144 on filtered system calls, instead of immediately killing the
6145 process. Also, SystemCallArchitectures= has been added to
6146 limit access to system calls of a particular architecture
6147 (in order to turn off support for unused secondary
6148 architectures). There is also a global
6149 SystemCallArchitectures= setting in system.conf now to turn
6150 off support for non-native system calls system-wide.
6152 * systemd requires a kernel with a working name_to_handle_at(),
6153 please see the kernel config requirements in the README file.
6155 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6156 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6157 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6158 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6159 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6160 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6161 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6162 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6163 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6164 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6165 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6166 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6167 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6168 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6169 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6170 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6171 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6172 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6173 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6174 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6175 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6176 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6177 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6178 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6180 — Berlin, 2014-02-20
6184 * logind has gained support for facilitating privileged input
6185 and drm device access for unprivileged clients. This work is
6186 useful to allow Wayland display servers (and similar
6187 programs, such as kmscon) to run under the user's ID and
6188 access input and drm devices which are normally
6189 protected. When this is used (and the kernel is new enough)
6190 logind will "mute" IO on the file descriptors passed to
6191 Wayland as long as it is in the background and "unmute" it
6192 if it returns into the foreground. This allows secure
6193 session switching without allowing background sessions to
6194 eavesdrop on input and display data. This also introduces
6195 session switching support if VT support is turned off in the
6196 kernel, and on seats that are not seat0.
6198 * A new kernel command line option luks.options= is understood
6199 now which allows specifying LUKS options for usage for LUKS
6200 encrypted partitions specified with luks.uuid=.
6202 * tmpfiles.d(5) snippets may now use specifier expansion in
6203 path names. More specifically %m, %b, %H, %v, are now
6204 replaced by the local machine id, boot id, hostname, and
6205 kernel version number.
6207 * A new tmpfiles.d(5) command "m" has been introduced which
6208 may be used to change the owner/group/access mode of a file
6209 or directory if it exists, but do nothing if it does not.
6211 * This release removes high-level support for the
6212 MemorySoftLimit= cgroup setting. The underlying kernel
6213 cgroup attribute memory.soft_limit= is currently badly
6214 designed and likely to be removed from the kernel API in its
6215 current form, hence we should not expose it for now.
6217 * The memory.use_hierarchy cgroup attribute is now enabled for
6218 all cgroups systemd creates in the memory cgroup
6219 hierarchy. This option is likely to be come the built-in
6220 default in the kernel anyway, and the non-hierarchical mode
6221 never made much sense in the intrinsically hierarchical
6224 * A new field _SYSTEMD_SLICE= is logged along with all journal
6225 messages containing the slice a message was generated
6226 from. This is useful to allow easy per-customer filtering of
6227 logs among other things.
6229 * systemd-journald will no longer adjust the group of journal
6230 files it creates to the "systemd-journal" group. Instead we
6231 rely on the journal directory to be owned by the
6232 "systemd-journal" group, and its setgid bit set, so that the
6233 kernel file system layer will automatically enforce that
6234 journal files inherit this group assignment. The reason for
6235 this change is that we cannot allow NSS look-ups from
6236 journald which would be necessary to resolve
6237 "systemd-journal" to a numeric GID, because this might
6238 create deadlocks if NSS involves synchronous queries to
6239 other daemons (such as nscd, or sssd) which in turn are
6240 logging clients of journald and might block on it, which
6241 would then dead lock. A tmpfiles.d(5) snippet included in
6242 systemd will make sure the setgid bit and group are
6243 properly set on the journal directory if it exists on every
6244 boot. However, we recommend adjusting it manually after
6245 upgrades too (or from RPM scriptlets), so that the change is
6246 not delayed until next reboot.
6248 * Backlight and random seed files in /var/lib/ have moved into
6249 the /var/lib/systemd/ directory, in order to centralize all
6250 systemd generated files in one directory.
6252 * Boot time performance measurements (as displayed by
6253 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6254 performance information if that's available to determine how
6255 much time BIOS and boot loader initialization required. With
6256 a sufficiently new BIOS you hence no longer need to boot
6257 with Gummiboot to get access to such information.
6259 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6260 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6261 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6262 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6263 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6264 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6265 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6267 — Berlin, 2013-10-02
6271 * The Restart= option for services now understands a new
6272 on-watchdog setting, which will restart the service
6273 automatically if the service stops sending out watchdog keep
6274 alive messages (as configured with WatchdogSec=).
6276 * The getty generator (which is responsible for bringing up a
6277 getty on configured serial consoles) will no longer only
6278 start a getty on the primary kernel console but on all
6279 others, too. This makes the order in which console= is
6280 specified on the kernel command line less important.
6282 * libsystemd-logind gained a new sd_session_get_vt() call to
6283 retrieve the VT number of a session.
6285 * If the option "tries=0" is set for an entry of /etc/crypttab
6286 its passphrase is queried indefinitely instead of any
6287 maximum number of tries.
6289 * If a service with a configure PID file terminates its PID
6290 file will now be removed automatically if it still exists
6291 afterwards. This should put an end to stale PID files.
6293 * systemd-run will now also take relative binary path names
6294 for execution and no longer insists on absolute paths.
6296 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6297 paths that are optionally prefixed with "-" to indicate that
6298 it should not be considered a failure if they do not exist.
6300 * journalctl -o (and similar commands) now understands a new
6301 output mode "short-precise", it is similar to "short" but
6302 shows timestamps with usec accuracy.
6304 * The option "discard" (as known from Debian) is now
6305 synonymous to "allow-discards" in /etc/crypttab. In fact,
6306 "discard" is preferred now (since it is easier to remember
6309 * Some licensing clean-ups were made, so that more code is now
6310 LGPL-2.1 licensed than before.
6312 * A minimal tool to save/restore the display backlight
6313 brightness across reboots has been added. It will store the
6314 backlight setting as late as possible at shutdown, and
6315 restore it as early as possible during reboot.
6317 * A logic to automatically discover and enable home and swap
6318 partitions on GPT disks has been added. With this in place
6319 /etc/fstab becomes optional for many setups as systemd can
6320 discover certain partitions located on the root disk
6321 automatically. Home partitions are recognized under their
6322 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6323 partitions are recognized under their GPT type ID
6324 0657fd6da4ab43c484e50933c84b4f4f.
6326 * systemd will no longer pass any environment from the kernel
6327 or initrd to system services. If you want to set an
6328 environment for all services, do so via the kernel command
6329 line systemd.setenv= assignment.
6331 * The systemd-sysctl tool no longer natively reads the file
6332 /etc/sysctl.conf. If desired, the file should be symlinked
6333 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6334 legacy support by a symlink rather than built-in code, it
6335 also makes the otherwise hidden order of application of the
6336 different files visible. (Note that this partly reverts to a
6337 pre-198 application order of sysctl knobs!)
6339 * The "systemctl set-log-level" and "systemctl dump" commands
6340 have been moved to systemd-analyze.
6342 * systemd-run learned the new --remain-after-exit switch,
6343 which causes the scope unit not to be cleaned up
6344 automatically after the process terminated.
6346 * tmpfiles learned a new --exclude-prefix= switch to exclude
6347 certain paths from operation.
6349 * journald will now automatically flush all messages to disk
6350 as soon as a message at the log level CRIT, ALERT or EMERG
6353 Contributions from: Andrew Cook, Brandon Philips, Christian
6354 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6355 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6356 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6357 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6358 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6359 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6360 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6361 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6362 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6363 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6364 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6365 William Giokas, Zbigniew Jędrzejewski-Szmek
6367 — Berlin, 2013-09-13
6371 * The documentation has been updated to cover the various new
6372 concepts introduced with 205.
6374 * Unit files now understand the new %v specifier which
6375 resolves to the kernel version string as returned by "uname
6378 * systemctl now supports filtering the unit list output by
6379 load state, active state and sub state, using the new
6382 * "systemctl status" will now show the results of the
6383 condition checks (like ConditionPathExists= and similar) of
6384 the last start attempts of the unit. They are also logged to
6387 * "journalctl -b" may now be used to look for boot output of a
6388 specific boot. Try "journalctl -b -1" for the previous boot,
6389 but the syntax is substantially more powerful.
6391 * "journalctl --show-cursor" has been added which prints the
6392 cursor string the last shown log line. This may then be used
6393 with the new "journalctl --after-cursor=" switch to continue
6394 browsing logs from that point on.
6396 * "journalctl --force" may now be used to force regeneration
6399 * Creation of "dead" device nodes has been moved from udev
6400 into kmod and tmpfiles. Previously, udev would read the kmod
6401 databases to pre-generate dead device nodes based on meta
6402 information contained in kernel modules, so that these would
6403 be auto-loaded on access rather then at boot. As this
6404 does not really have much to do with the exposing actual
6405 kernel devices to userspace this has always been slightly
6406 alien in the udev codebase. Following the new scheme kmod
6407 will now generate a runtime snippet for tmpfiles from the
6408 module meta information and it now is tmpfiles' job to the
6409 create the nodes. This also allows overriding access and
6410 other parameters for the nodes using the usual tmpfiles
6411 facilities. As side effect this allows us to remove the
6412 CAP_SYS_MKNOD capability bit from udevd entirely.
6414 * logind's device ACLs may now be applied to these "dead"
6415 devices nodes too, thus finally allowing managed access to
6416 devices such as /dev/snd/sequencer without loading the
6417 backing module right-away.
6419 * A new RPM macro has been added that may be used to apply
6420 tmpfiles configuration during package installation.
6422 * systemd-detect-virt and ConditionVirtualization= now can
6423 detect User-Mode-Linux machines (UML).
6425 * journald will now implicitly log the effective capabilities
6426 set of processes in the message metadata.
6428 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6430 * The initrd interface has been simplified (more specifically,
6431 support for passing performance data via environment
6432 variables and fsck results via files in /run has been
6433 removed). These features were non-essential, and are
6434 nowadays available in a much nicer way by having systemd in
6435 the initrd serialize its state and have the hosts systemd
6436 deserialize it again.
6438 * The udev "keymap" data files and tools to apply keyboard
6439 specific mappings of scan to key codes, and force-release
6440 scan code lists have been entirely replaced by a udev
6441 "keyboard" builtin and a hwdb data file.
6443 * systemd will now honour the kernel's "quiet" command line
6444 argument also during late shutdown, resulting in a
6445 completely silent shutdown when used.
6447 * There's now an option to control the SO_REUSEPORT socket
6448 option in .socket units.
6450 * Instance units will now automatically get a per-template
6451 subslice of system.slice unless something else is explicitly
6452 configured. For example, instances of sshd@.service will now
6453 implicitly be placed in system-sshd.slice rather than
6454 system.slice as before.
6456 * Test coverage support may now be enabled at build time.
6458 Contributions from: Dave Reisner, Frederic Crozat, Harald
6459 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6460 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6461 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6462 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6463 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6464 Giokas, Zbigniew Jędrzejewski-Szmek
6466 — Berlin, 2013-07-23
6470 * Two new unit types have been introduced:
6472 Scope units are very similar to service units, however, are
6473 created out of pre-existing processes — instead of PID 1
6474 forking off the processes. By using scope units it is
6475 possible for system services and applications to group their
6476 own child processes (worker processes) in a powerful way
6477 which then maybe used to organize them, or kill them
6478 together, or apply resource limits on them.
6480 Slice units may be used to partition system resources in an
6481 hierarchical fashion and then assign other units to them. By
6482 default there are now three slices: system.slice (for all
6483 system services), user.slice (for all user sessions),
6484 machine.slice (for VMs and containers).
6486 Slices and scopes have been introduced primarily in
6487 context of the work to move cgroup handling to a
6488 single-writer scheme, where only PID 1
6489 creates/removes/manages cgroups.
6491 * There's a new concept of "transient" units. In contrast to
6492 normal units these units are created via an API at runtime,
6493 not from configuration from disk. More specifically this
6494 means it is now possible to run arbitrary programs as
6495 independent services, with all execution parameters passed
6496 in via bus APIs rather than read from disk. Transient units
6497 make systemd substantially more dynamic then it ever was,
6498 and useful as a general batch manager.
6500 * logind has been updated to make use of scope and slice units
6501 for managing user sessions. As a user logs in he will get
6502 his own private slice unit, to which all sessions are added
6503 as scope units. We also added support for automatically
6504 adding an instance of user@.service for the user into the
6505 slice. Effectively logind will no longer create cgroup
6506 hierarchies on its own now, it will defer entirely to PID 1
6507 for this by means of scope, service and slice units. Since
6508 user sessions this way become entities managed by PID 1
6509 the output of "systemctl" is now a lot more comprehensive.
6511 * A new mini-daemon "systemd-machined" has been added which
6512 may be used by virtualization managers to register local
6513 VMs/containers. nspawn has been updated accordingly, and
6514 libvirt will be updated shortly. machined will collect a bit
6515 of meta information about the VMs/containers, and assign
6516 them their own scope unit (see above). The collected
6517 meta-data is then made available via the "machinectl" tool,
6518 and exposed in "ps" and similar tools. machined/machinectl
6519 is compile-time optional.
6521 * As discussed earlier, the low-level cgroup configuration
6522 options ControlGroup=, ControlGroupModify=,
6523 ControlGroupPersistent=, ControlGroupAttribute= have been
6524 removed. Please use high-level attribute settings instead as
6525 well as slice units.
6527 * A new bus call SetUnitProperties() has been added to alter
6528 various runtime parameters of a unit. This is primarily
6529 useful to alter cgroup parameters dynamically in a nice way,
6530 but will be extended later on to make more properties
6531 modifiable at runtime. systemctl gained a new set-properties
6532 command that wraps this call.
6534 * A new tool "systemd-run" has been added which can be used to
6535 run arbitrary command lines as transient services or scopes,
6536 while configuring a number of settings via the command
6537 line. This tool is currently very basic, however already
6538 very useful. We plan to extend this tool to even allow
6539 queuing of execution jobs with time triggers from the
6540 command line, similar in fashion to "at".
6542 * nspawn will now inform the user explicitly that kernels with
6543 audit enabled break containers, and suggest the user to turn
6546 * Support for detecting the IMA and AppArmor security
6547 frameworks with ConditionSecurity= has been added.
6549 * journalctl gained a new "-k" switch for showing only kernel
6550 messages, mimicking dmesg output; in addition to "--user"
6551 and "--system" switches for showing only user's own logs
6554 * systemd-delta can now show information about drop-in
6555 snippets extending unit files.
6557 * libsystemd-bus has been substantially updated but is still
6558 not available as public API.
6560 * systemd will now look for the "debug" argument on the kernel
6561 command line and enable debug logging, similar to what
6562 "systemd.log_level=debug" already did before.
6564 * "systemctl set-default", "systemctl get-default" has been
6565 added to configure the default.target symlink, which
6566 controls what to boot into by default.
6568 * "systemctl set-log-level" has been added as a convenient
6569 way to raise and lower systemd logging threshold.
6571 * "systemd-analyze plot" will now show the time the various
6572 generators needed for execution, as well as information
6573 about the unit file loading.
6575 * libsystemd-journal gained a new sd_journal_open_files() call
6576 for opening specific journal files. journactl also gained a
6577 new switch to expose this new functionality. Previously we
6578 only supported opening all files from a directory, or all
6579 files from the system, as opening individual files only is
6580 racy due to journal file rotation.
6582 * systemd gained the new DefaultEnvironment= setting in
6583 /etc/systemd/system.conf to set environment variables for
6586 * If a privileged process logs a journal message with the
6587 OBJECT_PID= field set, then journald will automatically
6588 augment this with additional OBJECT_UID=, OBJECT_GID=,
6589 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6590 system services want to log events about specific client
6591 processes. journactl/systemctl has been updated to make use
6592 of this information if all log messages regarding a specific
6595 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6596 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6597 Reisner, David Coppa, David King, David Strauss, Eelco
6598 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6599 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6600 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6601 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6602 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6603 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6604 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6605 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6606 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6607 Łukasz Stelmach, 장동준
6611 * The Python bindings gained some minimal support for the APIs
6612 exposed by libsystemd-logind.
6614 * ConditionSecurity= gained support for detecting SMACK. Since
6615 this condition already supports SELinux and AppArmor we only
6616 miss IMA for this. Patches welcome!
6618 Contributions from: Karol Lewandowski, Lennart Poettering,
6619 Zbigniew Jędrzejewski-Szmek
6623 * systemd-nspawn will now create /etc/resolv.conf if
6624 necessary, before bind-mounting the host's file onto it.
6626 * systemd-nspawn will now store meta information about a
6627 container on the container's cgroup as extended attribute
6628 fields, including the root directory.
6630 * The cgroup hierarchy has been reworked in many ways. All
6631 objects any of the components systemd creates in the cgroup
6632 tree are now suffixed. More specifically, user sessions are
6633 now placed in cgroups suffixed with ".session", users in
6634 cgroups suffixed with ".user", and nspawn containers in
6635 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6636 names are now escaped in a simple scheme to avoid collision
6637 of userspace object names with kernel filenames. This work
6638 is preparation for making these objects relocatable in the
6639 cgroup tree, in order to allow easy resource partitioning of
6640 these objects without causing naming conflicts.
6642 * systemctl list-dependencies gained the new switches
6643 --plain, --reverse, --after and --before.
6645 * systemd-inhibit now shows the process name of processes that
6646 have taken an inhibitor lock.
6648 * nss-myhostname will now also resolve "localhost"
6649 implicitly. This makes /etc/hosts an optional file and
6650 nicely handles that on IPv6 ::1 maps to both "localhost" and
6653 * libsystemd-logind.so gained a new call
6654 sd_get_machine_names() to enumerate running containers and
6655 VMs (currently only supported by very new libvirt and
6656 nspawn). sd_login_monitor can now be used to watch
6657 VMs/containers coming and going.
6659 * .include is not allowed recursively anymore, and only in
6660 unit files. Usually it is better to use drop-in snippets in
6661 .d/*.conf anyway, as introduced with systemd 198.
6663 * systemd-analyze gained a new "critical-chain" command that
6664 determines the slowest chain of units run during system
6665 boot-up. It is very useful for tracking down where
6666 optimizing boot time is the most beneficial.
6668 * systemd will no longer allow manipulating service paths in
6669 the name=systemd:/system cgroup tree using ControlGroup= in
6670 units. (But is still fine with it in all other dirs.)
6672 * There's a new systemd-nspawn@.service service file that may
6673 be used to easily run nspawn containers as system
6674 services. With the container's root directory in
6675 /var/lib/container/foobar it is now sufficient to run
6676 "systemctl start systemd-nspawn@foobar.service" to boot it.
6678 * systemd-cgls gained a new parameter "--machine" to list only
6679 the processes within a certain container.
6681 * ConditionSecurity= now can check for "apparmor". We still
6682 are lacking checks for SMACK and IMA for this condition
6683 check though. Patches welcome!
6685 * A new configuration file /etc/systemd/sleep.conf has been
6686 added that may be used to configure which kernel operation
6687 systemd is supposed to execute when "suspend", "hibernate"
6688 or "hybrid-sleep" is requested. This makes the new kernel
6689 "freeze" state accessible to the user.
6691 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6692 the passed argument if applicable.
6694 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6695 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6696 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6697 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6698 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6699 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6700 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6705 * The output of 'systemctl list-jobs' got some polishing. The
6706 '--type=' argument may now be passed more than once. A new
6707 command 'systemctl list-sockets' has been added which shows
6708 a list of kernel sockets systemd is listening on with the
6709 socket units they belong to, plus the units these socket
6712 * The experimental libsystemd-bus library got substantial
6713 updates to work in conjunction with the (also experimental)
6714 kdbus kernel project. It works well enough to exchange
6715 messages with some sophistication. Note that kdbus is not
6716 ready yet, and the library is mostly an elaborate test case
6717 for now, and not installable.
6719 * systemd gained a new unit 'systemd-static-nodes.service'
6720 that generates static device nodes earlier during boot, and
6721 can run in conjunction with udev.
6723 * libsystemd-login gained a new call sd_pid_get_user_unit()
6724 to retrieve the user systemd unit a process is running
6725 in. This is useful for systems where systemd is used as
6728 * systemd-nspawn now places all containers in the new /machine
6729 top-level cgroup directory in the name=systemd
6730 hierarchy. libvirt will soon do the same, so that we get a
6731 uniform separation of /system, /user and /machine for system
6732 services, user processes and containers/virtual
6733 machines. This new cgroup hierarchy is also useful to stick
6734 stable names to specific container instances, which can be
6735 recognized later this way (this name may be controlled
6736 via systemd-nspawn's new -M switch). libsystemd-login also
6737 gained a new call sd_pid_get_machine_name() to retrieve the
6738 name of the container/VM a specific process belongs to.
6740 * bootchart can now store its data in the journal.
6742 * libsystemd-journal gained a new call
6743 sd_journal_add_conjunction() for AND expressions to the
6744 matching logic. This can be used to express more complex
6745 logical expressions.
6747 * journactl can now take multiple --unit= and --user-unit=
6750 * The cryptsetup logic now understands the "luks.key=" kernel
6751 command line switch for specifying a file to read the
6752 decryption key from. Also, if a configured key file is not
6753 found the tool will now automatically fall back to prompting
6756 * Python systemd.journal module was updated to wrap recently
6757 added functions from libsystemd-journal. The interface was
6758 changed to bring the low level interface in s.j._Reader
6759 closer to the C API, and the high level interface in
6760 s.j.Reader was updated to wrap and convert all data about
6763 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6764 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6765 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6766 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6767 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6768 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6772 * journalctl --update-catalog now understands a new --root=
6773 option to operate on catalogs found in a different root
6776 * During shutdown after systemd has terminated all running
6777 services a final killing loop kills all remaining left-over
6778 processes. We will now print the name of these processes
6779 when we send SIGKILL to them, since this usually indicates a
6782 * If /etc/crypttab refers to password files stored on
6783 configured mount points automatic dependencies will now be
6784 generated to ensure the specific mount is established first
6785 before the key file is attempted to be read.
6787 * 'systemctl status' will now show information about the
6788 network sockets a socket unit is listening on.
6790 * 'systemctl status' will also shown information about any
6791 drop-in configuration file for units. (Drop-In configuration
6792 files in this context are files such as
6793 /etc/systemd/systemd/foobar.service.d/*.conf)
6795 * systemd-cgtop now optionally shows summed up CPU times of
6796 cgroups. Press '%' while running cgtop to switch between
6797 percentage and absolute mode. This is useful to determine
6798 which cgroups use up the most CPU time over the entire
6799 runtime of the system. systemd-cgtop has also been updated
6800 to be 'pipeable' for processing with further shell tools.
6802 * 'hostnamectl set-hostname' will now allow setting of FQDN
6805 * The formatting and parsing of time span values has been
6806 changed. The parser now understands fractional expressions
6807 such as "5.5h". The formatter will now output fractional
6808 expressions for all time spans under 1min, i.e. "5.123456s"
6809 rather than "5s 123ms 456us". For time spans under 1s
6810 millisecond values are shown, for those under 1ms
6811 microsecond values are shown. This should greatly improve
6812 all time-related output of systemd.
6814 * libsystemd-login and libsystemd-journal gained new
6815 functions for querying the poll() events mask and poll()
6816 timeout value for integration into arbitrary event
6819 * localectl gained the ability to list available X11 keymaps
6820 (models, layouts, variants, options).
6822 * 'systemd-analyze dot' gained the ability to filter for
6823 specific units via shell-style globs, to create smaller,
6824 more useful graphs. I.e. it is now possible to create simple
6825 graphs of all the dependencies between only target units, or
6826 of all units that Avahi has dependencies with.
6828 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6829 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6830 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6831 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6832 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6833 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6834 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6838 * The boot-time readahead implementation for rotating media
6839 will now read the read-ahead data in multiple passes which
6840 consist of all read requests made in equidistant time
6841 intervals. This means instead of strictly reading read-ahead
6842 data in its physical order on disk we now try to find a
6843 middle ground between physical and access time order.
6845 * /etc/os-release files gained a new BUILD_ID= field for usage
6846 on operating systems that provide continuous builds of OS
6849 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6850 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6851 William Douglas, Zbigniew Jędrzejewski-Szmek
6855 * systemd-python gained an API exposing libsystemd-daemon.
6857 * The SMACK setup logic gained support for uploading CIPSO
6860 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6861 ReadOnlyDirectories= and InaccessibleDirectories= has
6862 changed. The private /tmp and /var/tmp directories are now
6863 shared by all processes of a service (which means
6864 ExecStartPre= may now leave data in /tmp that ExecStart= of
6865 the same service can still access). When a service is
6866 stopped its temporary directories are immediately deleted
6867 (normal clean-up with tmpfiles is still done in addition to
6870 * By default, systemd will now set a couple of sysctl
6871 variables in the kernel: the safe sysrq options are turned
6872 on, IP route verification is turned on, and source routing
6873 disabled. The recently added hardlink and softlink
6874 protection of the kernel is turned on. These settings should
6875 be reasonably safe, and good defaults for all new systems.
6877 * The predictable network naming logic may now be turned off
6878 with a new kernel command line switch: net.ifnames=0.
6880 * A new libsystemd-bus module has been added that implements a
6881 pretty complete D-Bus client library. For details see:
6883 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
6885 * journald will now explicitly flush the journal files to disk
6886 at the latest 5min after each write. The file will then also
6887 be marked offline until the next write. This should increase
6888 reliability in case of a crash. The synchronization delay
6889 can be configured via SyncIntervalSec= in journald.conf.
6891 * There's a new remote-fs-setup.target unit that can be used
6892 to pull in specific services when at least one remote file
6893 system is to be mounted.
6895 * There are new targets timers.target and paths.target as
6896 canonical targets to pull user timer and path units in
6897 from. This complements sockets.target with a similar
6898 purpose for socket units.
6900 * libudev gained a new call udev_device_set_attribute_value()
6901 to set sysfs attributes of a device.
6903 * The udev daemon now sets the default number of worker
6904 processes executed in parallel based on the number of available
6905 CPUs instead of the amount of available RAM. This is supposed
6906 to provide a more reliable default and limit a too aggressive
6907 parallelism for setups with 1000s of devices connected.
6909 Contributions from: Auke Kok, Colin Walters, Cristian
6910 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6911 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6912 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6913 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6914 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6915 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6916 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6917 Zbigniew Jędrzejewski-Szmek
6921 * Configuration of unit files may now be extended via drop-in
6922 files without having to edit/override the unit files
6923 themselves. More specifically, if the administrator wants to
6924 change one value for a service file foobar.service he can
6925 now do so by dropping in a configuration snippet into
6926 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
6927 will load all these snippets and apply them on top of the
6928 main unit configuration file, possibly extending or
6929 overriding its settings. Using these drop-in snippets is
6930 generally nicer than the two earlier options for changing
6931 unit files locally: copying the files from
6932 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6933 them there; or creating a new file in /etc/systemd/system/
6934 that incorporates the original one via ".include". Drop-in
6935 snippets into these .d/ directories can be placed in any
6936 directory systemd looks for units in, and the usual
6937 overriding semantics between /usr/lib, /etc and /run apply
6940 * Most unit file settings which take lists of items can now be
6941 reset by assigning the empty string to them. For example,
6942 normally, settings such as Environment=FOO=BAR append a new
6943 environment variable assignment to the environment block,
6944 each time they are used. By assigning Environment= the empty
6945 string the environment block can be reset to empty. This is
6946 particularly useful with the .d/*.conf drop-in snippets
6947 mentioned above, since this adds the ability to reset list
6948 settings from vendor unit files via these drop-ins.
6950 * systemctl gained a new "list-dependencies" command for
6951 listing the dependencies of a unit recursively.
6953 * Inhibitors are now honored and listed by "systemctl
6954 suspend", "systemctl poweroff" (and similar) too, not only
6955 GNOME. These commands will also list active sessions by
6958 * Resource limits (as exposed by the various control group
6959 controllers) can now be controlled dynamically at runtime
6960 for all units. More specifically, you can now use a command
6961 like "systemctl set-cgroup-attr foobar.service cpu.shares
6962 2000" to alter the CPU shares a specific service gets. These
6963 settings are stored persistently on disk, and thus allow the
6964 administrator to easily adjust the resource usage of
6965 services with a few simple commands. This dynamic resource
6966 management logic is also available to other programs via the
6967 bus. Almost any kernel cgroup attribute and controller is
6970 * systemd-vconsole-setup will now copy all font settings to
6971 all allocated VTs, where it previously applied them only to
6974 * libsystemd-login gained the new sd_session_get_tty() API
6977 * This release drops support for a few legacy or
6978 distribution-specific LSB facility names when parsing init
6979 scripts: $x-display-manager, $mail-transfer-agent,
6980 $mail-transport-agent, $mail-transfer-agent, $smtp,
6981 $null. Also, the mail-transfer-agent.target unit backing
6982 this has been removed. Distributions which want to retain
6983 compatibility with this should carry the burden for
6984 supporting this themselves and patch support for these back
6985 in, if they really need to. Also, the facilities $syslog and
6986 $local_fs are now ignored, since systemd does not support
6987 early-boot LSB init scripts anymore, and these facilities
6988 are implied anyway for normal services. syslog.target has
6991 * There are new bus calls on PID1's Manager object for
6992 cancelling jobs, and removing snapshot units. Previously,
6993 both calls were only available on the Job and Snapshot
6996 * systemd-journal-gatewayd gained SSL support.
6998 * The various "environment" files, such as /etc/locale.conf
6999 now support continuation lines with a backslash ("\") as
7000 last character in the line, similarly in style (but different)
7001 to how this is supported in shells.
7003 * For normal user processes the _SYSTEMD_USER_UNIT= field is
7004 now implicitly appended to every log entry logged. systemctl
7005 has been updated to filter by this field when operating on a
7006 user systemd instance.
7008 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
7009 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7010 the container. This makes it easier to boot unmodified
7011 Fedora systems in a container, which however still requires
7012 audit=0 to be passed on the kernel command line. Auditing in
7013 kernel and userspace is unfortunately still too broken in
7014 context of containers, hence we recommend compiling it out
7015 of the kernel or using audit=0. Hopefully this will be fixed
7016 one day for good in the kernel.
7018 * nspawn gained the new --bind= and --bind-ro= parameters to
7019 bind mount specific directories from the host into the
7022 * nspawn will now mount its own devpts file system instance
7023 into the container, in order not to leak pty devices from
7024 the host into the container.
7026 * systemd will now read the firmware boot time performance
7027 information from the EFI variables, if the used boot loader
7028 supports this, and takes it into account for boot performance
7029 analysis via "systemd-analyze". This is currently supported
7030 only in conjunction with Gummiboot, but could be supported
7031 by other boot loaders too. For details see:
7033 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
7035 * A new generator has been added that automatically mounts the
7036 EFI System Partition (ESP) to /boot, if that directory
7037 exists, is empty, and no other file system has been
7038 configured to be mounted there.
7040 * logind will now send out PrepareForSleep(false) out
7041 unconditionally, after coming back from suspend. This may be
7042 used by applications as asynchronous notification for
7043 system resume events.
7045 * "systemctl unlock-sessions" has been added, that allows
7046 unlocking the screens of all user sessions at once, similar
7047 to how "systemctl lock-sessions" already locked all users
7048 sessions. This is backed by a new D-Bus call UnlockSessions().
7050 * "loginctl seat-status" will now show the master device of a
7051 seat. (i.e. the device of a seat that needs to be around for
7052 the seat to be considered available, usually the graphics
7055 * tmpfiles gained a new "X" line type, that allows
7056 configuration of files and directories (with wildcards) that
7057 shall be excluded from automatic cleanup ("aging").
7059 * udev default rules set the device node permissions now only
7060 at "add" events, and do not change them any longer with a
7061 later "change" event.
7063 * The log messages for lid events and power/sleep keypresses
7064 now carry a message ID.
7066 * We now have a substantially larger unit test suite, but this
7067 continues to be work in progress.
7069 * udevadm hwdb gained a new --root= parameter to change the
7070 root directory to operate relative to.
7072 * logind will now issue a background sync() request to the kernel
7073 early at shutdown, so that dirty buffers are flushed to disk early
7074 instead of at the last moment, in order to optimize shutdown
7077 * A new bootctl tool has been added that is an interface for
7078 certain boot loader operations. This is currently a preview
7079 and is likely to be extended into a small mechanism daemon
7080 like timedated, localed, hostnamed, and can be used by
7081 graphical UIs to enumerate available boot options, and
7082 request boot into firmware operations.
7084 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7085 the rest of the package. It also has been updated to work
7086 correctly in initrds.
7088 * Policykit previously has been runtime optional, and is now
7089 also compile time optional via a configure switch.
7091 * systemd-analyze has been reimplemented in C. Also "systemctl
7092 dot" has moved into systemd-analyze.
7094 * "systemctl status" with no further parameters will now print
7095 the status of all active or failed units.
7097 * Operations such as "systemctl start" can now be executed
7098 with a new mode "--irreversible" which may be used to queue
7099 operations that cannot accidentally be reversed by a later
7100 job queuing. This is by default used to make shutdown
7101 requests more robust.
7103 * The Python API of systemd now gained a new module for
7104 reading journal files.
7106 * A new tool kernel-install has been added that can install
7107 kernel images according to the Boot Loader Specification:
7109 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7111 * Boot time console output has been improved to provide
7112 animated boot time output for hanging jobs.
7114 * A new tool systemd-activate has been added which can be used
7115 to test socket activation with, directly from the command
7116 line. This should make it much easier to test and debug
7117 socket activation in daemons.
7119 * journalctl gained a new "--reverse" (or -r) option to show
7120 journal output in reverse order (i.e. newest line first).
7122 * journalctl gained a new "--pager-end" (or -e) option to jump
7123 to immediately jump to the end of the journal in the
7124 pager. This is only supported in conjunction with "less".
7126 * journalctl gained a new "--user-unit=" option, that works
7127 similarly to "--unit=" but filters for user units rather than
7130 * A number of unit files to ease adoption of systemd in
7131 initrds has been added. This moves some minimal logic from
7132 the various initrd implementations into systemd proper.
7134 * The journal files are now owned by a new group
7135 "systemd-journal", which exists specifically to allow access
7136 to the journal, and nothing else. Previously, we used the
7137 "adm" group for that, which however possibly covers more
7138 than just journal/log file access. This new group is now
7139 already used by systemd-journal-gatewayd to ensure this
7140 daemon gets access to the journal files and as little else
7141 as possible. Note that "make install" will also set FS ACLs
7142 up for /var/log/journal to give "adm" and "wheel" read
7143 access to it, in addition to "systemd-journal" which owns
7144 the journal files. We recommend that packaging scripts also
7145 add read access to "adm" + "wheel" to /var/log/journal, and
7146 all existing/future journal files. To normal users and
7147 administrators little changes, however packagers need to
7148 ensure to create the "systemd-journal" system group at
7149 package installation time.
7151 * The systemd-journal-gatewayd now runs as unprivileged user
7152 systemd-journal-gateway:systemd-journal-gateway. Packaging
7153 scripts need to create these system user/group at
7156 * timedated now exposes a new boolean property CanNTP that
7157 indicates whether a local NTP service is available or not.
7159 * systemd-detect-virt will now also detect xen PVs
7161 * The pstore file system is now mounted by default, if it is
7164 * In addition to the SELinux and IMA policies we will now also
7165 load SMACK policies at early boot.
7167 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7168 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7169 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7170 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7171 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7172 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7173 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7174 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7175 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7176 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7177 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7178 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7179 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7180 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7184 * Timer units now support calendar time events in addition to
7185 monotonic time events. That means you can now trigger a unit
7186 based on a calendar time specification such as "Thu,Fri
7187 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7188 or fifth day of any month of the year 2013, given that it is
7189 a thursday or friday. This brings timer event support
7190 considerably closer to cron's capabilities. For details on
7191 the supported calendar time specification language see
7194 * udev now supports a number of different naming policies for
7195 network interfaces for predictable names, and a combination
7196 of these policies is now the default. Please see this wiki
7197 document for details:
7199 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
7201 * Auke Kok's bootchart implementation has been added to the
7202 systemd tree. It is an optional component that can graph the
7203 boot in quite some detail. It is one of the best bootchart
7204 implementations around and minimal in its code and
7207 * nss-myhostname has been integrated into the systemd source
7208 tree. nss-myhostname guarantees that the local hostname
7209 always stays resolvable via NSS. It has been a weak
7210 requirement of systemd-hostnamed since a long time, and
7211 since its code is actually trivial we decided to just
7212 include it in systemd's source tree. It can be turned off
7213 with a configure switch.
7215 * The read-ahead logic is now capable of properly detecting
7216 whether a btrfs file system is on SSD or rotating media, in
7217 order to optimize the read-ahead scheme. Previously, it was
7218 only capable of detecting this on traditional file systems
7221 * In udev, additional device properties are now read from the
7222 IAB in addition to the OUI database. Also, Bluetooth company
7223 identities are attached to the devices as well.
7225 * In service files %U may be used as specifier that is
7226 replaced by the configured user name of the service.
7228 * nspawn may now be invoked without a controlling TTY. This
7229 makes it suitable for invocation as its own service. This
7230 may be used to set up a simple containerized server system
7231 using only core OS tools.
7233 * systemd and nspawn can now accept socket file descriptors
7234 when they are started for socket activation. This enables
7235 implementation of socket activated nspawn
7236 containers. i.e. think about autospawning an entire OS image
7237 when the first SSH or HTTP connection is received. We expect
7238 that similar functionality will also be added to libvirt-lxc
7241 * journalctl will now suppress ANSI color codes when
7242 presenting log data.
7244 * systemctl will no longer show control group information for
7245 a unit if the control group is empty anyway.
7247 * logind can now automatically suspend/hibernate/shutdown the
7250 * /etc/machine-info and hostnamed now also expose the chassis
7251 type of the system. This can be used to determine whether
7252 the local system is a laptop, desktop, handset or
7253 tablet. This information may either be configured by the
7254 user/vendor or is automatically determined from ACPI and DMI
7255 information if possible.
7257 * A number of PolicyKit actions are now bound together with
7258 "imply" rules. This should simplify creating UIs because
7259 many actions will now authenticate similar ones as well.
7261 * Unit files learnt a new condition ConditionACPower= which
7262 may be used to conditionalize a unit depending on whether an
7263 AC power source is connected or not, of whether the system
7264 is running on battery power.
7266 * systemctl gained a new "is-failed" verb that may be used in
7267 shell scripts and suchlike to check whether a specific unit
7268 is in the "failed" state.
7270 * The EnvironmentFile= setting in unit files now supports file
7271 globbing, and can hence be used to easily read a number of
7272 environment files at once.
7274 * systemd will no longer detect and recognize specific
7275 distributions. All distribution-specific #ifdeffery has been
7276 removed, systemd is now fully generic and
7277 distribution-agnostic. Effectively, not too much is lost as
7278 a lot of the code is still accessible via explicit configure
7279 switches. However, support for some distribution specific
7280 legacy configuration file formats has been dropped. We
7281 recommend distributions to simply adopt the configuration
7282 files everybody else uses now and convert the old
7283 configuration from packaging scripts. Most distributions
7284 already did that. If that's not possible or desirable,
7285 distributions are welcome to forward port the specific
7286 pieces of code locally from the git history.
7288 * When logging a message about a unit systemd will now always
7289 log the unit name in the message meta data.
7291 * localectl will now also discover system locale data that is
7292 not stored in locale archives, but directly unpacked.
7294 * logind will no longer unconditionally use framebuffer
7295 devices as seat masters, i.e. as devices that are required
7296 to be existing before a seat is considered preset. Instead,
7297 it will now look for all devices that are tagged as
7298 "seat-master" in udev. By default, framebuffer devices will
7299 be marked as such, but depending on local systems, other
7300 devices might be marked as well. This may be used to
7301 integrate graphics cards using closed source drivers (such
7302 as NVidia ones) more nicely into logind. Note however, that
7303 we recommend using the open source NVidia drivers instead,
7304 and no udev rules for the closed-source drivers will be
7305 shipped from us upstream.
7307 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7308 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7309 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7310 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7311 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7312 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7313 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7314 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7315 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7316 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7317 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7318 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7323 * udev gained support for loading additional device properties
7324 from an indexed database that is keyed by vendor/product IDs
7325 and similar device identifiers. For the beginning this
7326 "hwdb" is populated with data from the well-known PCI and
7327 USB database, but also includes PNP, ACPI and OID data. In
7328 the longer run this indexed database shall grow into
7329 becoming the one central database for non-essential
7330 userspace device metadata. Previously, data from the PCI/USB
7331 database was only attached to select devices, since the
7332 lookup was a relatively expensive operation due to O(n) time
7333 complexity (with n being the number of entries in the
7334 database). Since this is now O(1), we decided to add in this
7335 data for all devices where this is available, by
7336 default. Note that the indexed database needs to be rebuilt
7337 when new data files are installed. To achieve this you need
7338 to update your packaging scripts to invoke "udevadm hwdb
7339 --update" after installation of hwdb data files. For
7340 RPM-based distributions we introduced the new
7341 %udev_hwdb_update macro for this purpose.
7343 * The Journal gained support for the "Message Catalog", an
7344 indexed database to link up additional information with
7345 journal entries. For further details please check:
7347 https://www.freedesktop.org/wiki/Software/systemd/catalog
7349 The indexed message catalog database also needs to be
7350 rebuilt after installation of message catalog files. Use
7351 "journalctl --update-catalog" for this. For RPM-based
7352 distributions we introduced the %journal_catalog_update
7353 macro for this purpose.
7355 * The Python Journal bindings gained support for the standard
7356 Python logging framework.
7358 * The Journal API gained new functions for checking whether
7359 the underlying file system of a journal file is capable of
7360 properly reporting file change notifications, or whether
7361 applications that want to reflect journal changes "live"
7362 need to recheck journal files continuously in appropriate
7365 * It is now possible to set the "age" field for tmpfiles
7366 entries to 0, indicating that files matching this entry
7367 shall always be removed when the directories are cleaned up.
7369 * coredumpctl gained a new "gdb" verb which invokes gdb
7370 right-away on the selected coredump.
7372 * There's now support for "hybrid sleep" on kernels that
7373 support this, in addition to "suspend" and "hibernate". Use
7374 "systemctl hybrid-sleep" to make use of this.
7376 * logind's HandleSuspendKey= setting (and related settings)
7377 now gained support for a new "lock" setting to simply
7378 request the screen lock on all local sessions, instead of
7379 actually executing a suspend or hibernation.
7381 * systemd will now mount the EFI variables file system by
7384 * Socket units now gained support for configuration of the
7385 SMACK security label.
7387 * timedatectl will now output the time of the last and next
7388 daylight saving change.
7390 * We dropped support for various legacy and distro-specific
7391 concepts, such as insserv, early-boot SysV services
7392 (i.e. those for non-standard runlevels such as 'b' or 'S')
7393 or ArchLinux /etc/rc.conf support. We recommend the
7394 distributions who still need support this to either continue
7395 to maintain the necessary patches downstream, or find a
7396 different solution. (Talk to us if you have questions!)
7398 * Various systemd components will now bypass PolicyKit checks
7399 for root and otherwise handle properly if PolicyKit is not
7400 found to be around. This should fix most issues for
7401 PolicyKit-less systems. Quite frankly this should have been
7402 this way since day one. It is absolutely our intention to
7403 make systemd work fine on PolicyKit-less systems, and we
7404 consider it a bug if something does not work as it should if
7405 PolicyKit is not around.
7407 * For embedded systems it is now possible to build udev and
7408 systemd without blkid and/or kmod support.
7410 * "systemctl switch-root" is now capable of switching root
7411 more than once. I.e. in addition to transitions from the
7412 initrd to the host OS it is now possible to transition to
7413 further OS images from the host. This is useful to implement
7414 offline updating tools.
7416 * Various other additions have been made to the RPM macros
7417 shipped with systemd. Use %udev_rules_update() after
7418 installing new udev rules files. %_udevhwdbdir,
7419 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7420 %_sysctldir are now available which resolve to the right
7421 directories for packages to place various data files in.
7423 * journalctl gained the new --full switch (in addition to
7424 --all, to disable ellipsation for long messages.
7426 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7427 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7428 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7429 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7430 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7431 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7432 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7433 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7434 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7438 * journalctl gained new --since= and --until= switches to
7439 filter by time. It also now supports nice filtering for
7440 units via --unit=/-u.
7442 * Type=oneshot services may use ExecReload= and do the
7445 * The journal daemon now supports time-based rotation and
7446 vacuuming, in addition to the usual disk-space based
7449 * The journal will now index the available field values for
7450 each field name. This enables clients to show pretty drop
7451 downs of available match values when filtering. The bash
7452 completion of journalctl has been updated
7453 accordingly. journalctl gained a new switch -F to list all
7454 values a certain field takes in the journal database.
7456 * More service events are now written as structured messages
7457 to the journal, and made recognizable via message IDs.
7459 * The timedated, localed and hostnamed mini-services which
7460 previously only provided support for changing time, locale
7461 and hostname settings from graphical DEs such as GNOME now
7462 also have a minimal (but very useful) text-based client
7463 utility each. This is probably the nicest way to changing
7464 these settings from the command line now, especially since
7465 it lists available options and is fully integrated with bash
7468 * There's now a new tool "systemd-coredumpctl" to list and
7469 extract coredumps from the journal.
7471 * We now install a README each in /var/log/ and
7472 /etc/rc.d/init.d explaining where the system logs and init
7473 scripts went. This hopefully should help folks who go to
7474 that dirs and look into the otherwise now empty void and
7475 scratch their heads.
7477 * When user-services are invoked (by systemd --user) the
7478 $MANAGERPID env var is set to the PID of systemd.
7480 * SIGRTMIN+24 when sent to a --user instance will now result
7481 in immediate termination of systemd.
7483 * gatewayd received numerous feature additions such as a
7484 "follow" mode, for live syncing and filtering.
7486 * browse.html now allows filtering and showing detailed
7487 information on specific entries. Keyboard navigation and
7488 mouse screen support has been added.
7490 * gatewayd/journalctl now supports HTML5/JSON
7491 Server-Sent-Events as output.
7493 * The SysV init script compatibility logic will now
7494 heuristically determine whether a script supports the
7495 "reload" verb, and only then make this available as
7498 * "systemctl status --follow" has been removed, use "journalctl
7501 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7502 have been removed since they are hardly useful to be
7505 * And I'd like to take the opportunity to specifically mention
7506 Zbigniew for his great contributions. Zbigniew, you rock!
7508 Contributions from: Andrew Eikum, Christian Hesse, Colin
7509 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
7510 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7511 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7512 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7513 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7514 Jędrzejewski-Szmek, Сковорода Никита Андреевич
7518 * If /etc/vconsole.conf is non-existent or empty we will no
7519 longer load any console font or key map at boot by
7520 default. Instead the kernel defaults will be left
7521 intact. This is definitely the right thing to do, as no
7522 configuration should mean no configuration, and hard-coding
7523 font names that are different on all archs is probably a bad
7524 idea. Also, the kernel default key map and font should be
7525 good enough for most cases anyway, and mostly identical to
7526 the userspace fonts/key maps we previously overloaded them
7527 with. If distributions want to continue to default to a
7528 non-kernel font or key map they should ship a default
7529 /etc/vconsole.conf with the appropriate contents.
7531 Contributions from: Colin Walters, Daniel J Walsh, Dave
7532 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7533 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7537 * journalctl gained a new --cursor= switch to show entries
7538 starting from the specified location in the journal.
7540 * We now enforce a size limit on journal entry fields exported
7541 with "-o json" in journalctl. Fields larger than 4K will be
7542 assigned null. This can be turned off with --all.
7544 * An (optional) journal gateway daemon is now available as
7545 "systemd-journal-gatewayd.service". This service provides
7546 access to the journal via HTTP and JSON. This functionality
7547 will be used to implement live log synchronization in both
7548 pull and push modes, but has various other users too, such
7549 as easy log access for debugging of embedded devices. Right
7550 now it is already useful to retrieve the journal via HTTP:
7552 # systemctl start systemd-journal-gatewayd.service
7553 # wget http://localhost:19531/entries
7555 This will download the journal contents in a
7556 /var/log/messages compatible format. The same as JSON:
7558 # curl -H"Accept: application/json" http://localhost:19531/entries
7560 This service is also accessible via a web browser where a
7561 single static HTML5 app is served that uses the JSON logic
7562 to enable the user to do some basic browsing of the
7563 journal. This will be extended later on. Here's an example
7564 screenshot of this app in its current state:
7566 http://0pointer.de/public/journal-gatewayd
7568 Contributions from: Kay Sievers, Lennart Poettering, Robert
7569 Milasan, Tom Gundersen
7573 * The bash completion logic is now available for journalctl
7576 * We do not mount the "cpuset" controller anymore together with
7577 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7578 started if no parameters are assigned to it. "cpuset" hence
7579 broke code that assumed it could create "cpu" groups and
7582 * journalctl -f will now subscribe to terminal size changes,
7583 and line break accordingly.
7585 Contributions from: Dave Reisner, Kay Sievers, Lennart
7586 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
7590 * nspawn will now create a symlink /etc/localtime in the
7591 container environment, copying the host's timezone
7592 setting. Previously this has been done via a bind mount, but
7593 since symlinks cannot be bind mounted this has now been
7594 changed to create/update the appropriate symlink.
7596 * journalctl -n's line number argument is now optional, and
7597 will default to 10 if omitted.
7599 * journald will now log the maximum size the journal files may
7600 take up on disk. This is particularly useful if the default
7601 built-in logic of determining this parameter from the file
7602 system size is used. Use "systemctl status
7603 systemd-journald.service" to see this information.
7605 * The multi-seat X wrapper tool has been stripped down. As X
7606 is now capable of enumerating graphics devices via udev in a
7607 seat-aware way the wrapper is not strictly necessary
7608 anymore. A stripped down temporary stop-gap is still shipped
7609 until the upstream display managers have been updated to
7610 fully support the new X logic. Expect this wrapper to be
7611 removed entirely in one of the next releases.
7613 * HandleSleepKey= in logind.conf has been split up into
7614 HandleSuspendKey= and HandleHibernateKey=. The old setting
7615 is not available anymore. X11 and the kernel are
7616 distinguishing between these keys and we should too. This
7617 also means the inhibition lock for these keys has been split
7620 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7621 Poettering, Lukas Nykryn, Václav Pavlín
7625 * Whenever a unit changes state we will now log this to the
7626 journal and show along the unit's own log output in
7629 * ConditionPathIsMountPoint= can now properly detect bind
7630 mount points too. (Previously, a bind mount of one file
7631 system to another place in the same file system could not be
7632 detected as mount, since they shared struct stat's st_dev
7635 * We will now mount the cgroup controllers cpu, cpuacct,
7636 cpuset and the controllers net_cls, net_prio together by
7639 * nspawn containers will now have a virtualized boot
7640 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7641 over with a randomized ID at container initialization). This
7642 has the effect of making "journalctl -b" do the right thing
7645 * The JSON output journal serialization has been updated not
7646 to generate "endless" list objects anymore, but rather one
7647 JSON object per line. This is more in line how most JSON
7648 parsers expect JSON objects. The new output mode
7649 "json-pretty" has been added to provide similar output, but
7650 neatly aligned for readability by humans.
7652 * We dropped all explicit sync() invocations in the shutdown
7653 code. The kernel does this implicitly anyway in the kernel
7654 reboot() syscall. halt(8)'s -n option is now a compatibility
7657 * We now support virtualized reboot() in containers, as
7658 supported by newer kernels. We will fall back to exit() if
7659 CAP_SYS_REBOOT is not available to the container. Also,
7660 nspawn makes use of this now and will actually reboot the
7661 container if the containerized OS asks for that.
7663 * journalctl will only show local log output by default
7664 now. Use --merge (-m) to show remote log output, too.
7666 * libsystemd-journal gained the new sd_journal_get_usage()
7667 call to determine the current disk usage of all journal
7668 files. This is exposed in the new "journalctl --disk-usage"
7671 * journald gained a new configuration setting SplitMode= in
7672 journald.conf which may be used to control how user journals
7673 are split off. See journald.conf(5) for details.
7675 * A new condition type ConditionFileNotEmpty= has been added.
7677 * tmpfiles' "w" lines now support file globbing, to write
7678 multiple files at once.
7680 * We added Python bindings for the journal submission
7681 APIs. More Python APIs for a number of selected APIs will
7682 likely follow. Note that we intend to add native bindings
7683 only for the Python language, as we consider it common
7684 enough to deserve bindings shipped within systemd. There are
7685 various projects outside of systemd that provide bindings
7686 for languages such as PHP or Lua.
7688 * Many conditions will now resolve specifiers such as %i. In
7689 addition, PathChanged= and related directives of .path units
7690 now support specifiers as well.
7692 * There's now a new RPM macro definition for the system preset
7695 * journald will now warn if it ca not forward a message to the
7696 syslog daemon because its socket is full.
7698 * timedated will no longer write or process /etc/timezone,
7699 except on Debian. As we do not support late mounted /usr
7700 anymore /etc/localtime always being a symlink is now safe,
7701 and hence the information in /etc/timezone is not necessary
7704 * logind will now always reserve one VT for a text getty (VT6
7705 by default). Previously if more than 6 X sessions where
7706 started they took up all the VTs with auto-spawned gettys,
7707 so that no text gettys were available anymore.
7709 * udev will now automatically inform the btrfs kernel logic
7710 about btrfs RAID components showing up. This should make
7711 simple hotplug based btrfs RAID assembly work.
7713 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7714 (but not for its children which will stay at the kernel
7715 default). This should allow setups with a lot more listening
7718 * systemd will now always pass the configured timezone to the
7719 kernel at boot. timedated will do the same when the timezone
7722 * logind's inhibition logic has been updated. By default,
7723 logind will now handle the lid switch, the power and sleep
7724 keys all the time, even in graphical sessions. If DEs want
7725 to handle these events on their own they should take the new
7726 handle-power-key, handle-sleep-key and handle-lid-switch
7727 inhibitors during their runtime. A simple way to achieve
7728 that is to invoke the DE wrapped in an invocation of:
7730 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7732 * Access to unit operations is now checked via SELinux taking
7733 the unit file label and client process label into account.
7735 * systemd will now notify the administrator in the journal
7736 when he over-mounts a non-empty directory.
7738 * There are new specifiers that are resolved in unit files,
7739 for the host name (%H), the machine ID (%m) and the boot ID
7742 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
7743 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7744 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7745 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7746 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7747 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7748 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7752 * Support for reading structured kernel messages from
7753 /dev/kmsg has now been added and is enabled by default.
7755 * Support for reading kernel messages from /proc/kmsg has now
7756 been removed. If you want kernel messages in the journal
7757 make sure to run a recent kernel (>= 3.5) that supports
7758 reading structured messages from /dev/kmsg (see
7759 above). /proc/kmsg is now exclusive property of classic
7760 syslog daemons again.
7762 * The libudev API gained the new
7763 udev_device_new_from_device_id() call.
7765 * The logic for file system namespace (ReadOnlyDirectory=,
7766 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7767 require pivot_root() anymore. This means fewer temporary
7768 directories are created below /tmp for this feature.
7770 * nspawn containers will now see and receive all submounts
7771 made on the host OS below the root file system of the
7774 * Forward Secure Sealing is now supported for Journal files,
7775 which provide cryptographical sealing of journal files so
7776 that attackers cannot alter log history anymore without this
7777 being detectable. Lennart will soon post a blog story about
7778 this explaining it in more detail.
7780 * There are two new service settings RestartPreventExitStatus=
7781 and SuccessExitStatus= which allow configuration of exit
7782 status (exit code or signal) which will be excepted from the
7783 restart logic, resp. consider successful.
7785 * journalctl gained the new --verify switch that can be used
7786 to check the integrity of the structure of journal files and
7787 (if Forward Secure Sealing is enabled) the contents of
7790 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7791 and similar symlinks pre-created. This makes running shells
7792 as container init process a lot more fun.
7794 * The fstab support can now handle PARTUUID= and PARTLABEL=
7797 * A new ConditionHost= condition has been added to match
7798 against the hostname (with globs) and machine ID. This is
7799 useful for clusters where a single OS image is used to
7800 provision a large number of hosts which shall run slightly
7801 different sets of services.
7803 * Services which hit the restart limit will now be placed in a
7806 Contributions from: Bertram Poettering, Dave Reisner, Huang
7807 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7808 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7812 * When running in --user mode systemd will now become a
7813 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7814 tree a lot more organized.
7816 * A new PartOf= unit dependency type has been introduced that
7817 may be used to group services in a natural way.
7819 * "systemctl enable" may now be used to enable instances of
7822 * journalctl now prints error log levels in red, and
7823 warning/notice log levels in bright white. It also supports
7824 filtering by log level now.
7826 * cgtop gained a new -n switch (similar to top), to configure
7827 the maximum number of iterations to run for. It also gained
7828 -b, to run in batch mode (accepting no input).
7830 * The suffix ".service" may now be omitted on most systemctl
7831 command lines involving service unit names.
7833 * There's a new bus call in logind to lock all sessions, as
7834 well as a loginctl verb for it "lock-sessions".
7836 * libsystemd-logind.so gained a new call sd_journal_perror()
7837 that works similar to libc perror() but logs to the journal
7838 and encodes structured information about the error number.
7840 * /etc/crypttab entries now understand the new keyfile-size=
7843 * shutdown(8) now can send a (configurable) wall message when
7844 a shutdown is cancelled.
7846 * The mount propagation mode for the root file system will now
7847 default to "shared", which is useful to make containers work
7848 nicely out-of-the-box so that they receive new mounts from
7849 the host. This can be undone locally by running "mount
7850 --make-rprivate /" if needed.
7852 * The prefdm.service file has been removed. Distributions
7853 should maintain this unit downstream if they intend to keep
7854 it around. However, we recommend writing normal unit files
7855 for display managers instead.
7857 * Since systemd is a crucial part of the OS we will now
7858 default to a number of compiler switches that improve
7859 security (hardening) such as read-only relocations, stack
7860 protection, and suchlike.
7862 * The TimeoutSec= setting for services is now split into
7863 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7864 of individual time outs for the start and the stop phase of
7867 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7868 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7869 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7870 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7871 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7872 Gundersen, Zbigniew Jędrzejewski-Szmek
7876 * The journal and id128 C APIs are now fully documented as man
7879 * Extra safety checks have been added when transitioning from
7880 the initial RAM disk to the main system to avoid accidental
7883 * /etc/crypttab entries now understand the new keyfile-offset=
7886 * systemctl -t can now be used to filter by unit load state.
7888 * The journal C API gained the new sd_journal_wait() call to
7889 make writing synchronous journal clients easier.
7891 * journalctl gained the new -D switch to show journals from a
7894 * journalctl now displays a special marker between log
7895 messages of two different boots.
7897 * The journal is now explicitly flushed to /var via a service
7898 systemd-journal-flush.service, rather than implicitly simply
7899 by seeing /var/log/journal to be writable.
7901 * journalctl (and the journal C APIs) can now match for much
7902 more complex expressions, with alternatives and
7905 * When transitioning from the initial RAM disk to the main
7906 system we will now kill all processes in a killing spree to
7907 ensure no processes stay around by accident.
7909 * Three new specifiers may be used in unit files: %u, %h, %s
7910 resolve to the user name, user home directory resp. user
7911 shell. This is useful for running systemd user instances.
7913 * We now automatically rotate journal files if their data
7914 object hash table gets a fill level > 75%. We also size the
7915 hash table based on the configured maximum file size. This
7916 together should lower hash collisions drastically and thus
7917 speed things up a bit.
7919 * journalctl gained the new "--header" switch to introspect
7920 header data of journal files.
7922 * A new setting SystemCallFilters= has been added to services
7923 which may be used to apply blacklists or whitelists to
7924 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7926 * nspawn gained a new --link-journal= switch (and quicker: -j)
7927 to link the container journal with the host. This makes it
7928 very easy to centralize log viewing on the host for all
7929 guests while still keeping the journal files separated.
7931 * Many bugfixes and optimizations
7933 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7934 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7935 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7940 * Several tools now understand kernel command line arguments,
7941 which are only read when run in an initial RAM disk. They
7942 usually follow closely their normal counterparts, but are
7945 * There's a new tool to analyze the readahead files that are
7946 automatically generated at boot. Use:
7948 /usr/lib/systemd/systemd-readahead analyze /.readahead
7950 * We now provide an early debug shell on tty9 if this enabled. Use:
7952 systemctl enable debug-shell.service
7954 * All plymouth related units have been moved into the Plymouth
7955 package. Please make sure to upgrade your Plymouth version
7958 * systemd-tmpfiles now supports getting passed the basename of
7959 a configuration file only, in which case it will look for it
7960 in all appropriate directories automatically.
7962 * udevadm info now takes a /dev or /sys path as argument, and
7963 does the right thing. Example:
7965 udevadm info /dev/sda
7966 udevadm info /sys/class/block/sda
7968 * systemctl now prints a warning if a unit is stopped but a
7969 unit that might trigger it continues to run. Example: a
7970 service is stopped but the socket that activates it is left
7973 * "systemctl status" will now mention if the log output was
7974 shortened due to rotation since a service has been started.
7976 * The journal API now exposes functions to determine the
7977 "cutoff" times due to rotation.
7979 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7980 immediately flushing of runtime logs to /var if possible,
7981 resp. for triggering immediate rotation of the journal
7984 * It is now considered an error if a service is attempted to
7985 be stopped that is not loaded.
7987 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7989 * systemd-analyze now supports Python 3
7991 * tmpfiles now supports cleaning up directories via aging
7992 where the first level dirs are always kept around but
7993 directories beneath it automatically aged. This is enabled
7994 by prefixing the age field with '~'.
7996 * Seat objects now expose CanGraphical, CanTTY properties
7997 which is required to deal with very fast bootups where the
7998 display manager might be running before the graphics drivers
7999 completed initialization.
8001 * Seat objects now expose a State property.
8003 * We now include RPM macros for service enabling/disabling
8004 based on the preset logic. We recommend RPM based
8005 distributions to make use of these macros if possible. This
8006 makes it simpler to reuse RPM spec files across
8009 * We now make sure that the collected systemd unit name is
8010 always valid when services log to the journal via
8013 * There's a new man page kernel-command-line(7) detailing all
8014 command line options we understand.
8016 * The fstab generator may now be disabled at boot by passing
8017 fstab=0 on the kernel command line.
8019 * A new kernel command line option modules-load= is now understood
8020 to load a specific kernel module statically, early at boot.
8022 * Unit names specified on the systemctl command line are now
8023 automatically escaped as needed. Also, if file system or
8024 device paths are specified they are automatically turned
8025 into the appropriate mount or device unit names. Example:
8027 systemctl status /home
8028 systemctl status /dev/sda
8030 * The SysVConsole= configuration option has been removed from
8031 system.conf parsing.
8033 * The SysV search path is no longer exported on the D-Bus
8036 * The Names= option has been removed from unit file parsing.
8038 * There's a new man page bootup(7) detailing the boot process.
8040 * Every unit and every generator we ship with systemd now
8041 comes with full documentation. The self-explanatory boot is
8044 * A couple of services gained "systemd-" prefixes in their
8045 name if they wrap systemd code, rather than only external
8046 code. Among them fsck@.service which is now
8047 systemd-fsck@.service.
8049 * The HaveWatchdog property has been removed from the D-Bus
8052 * systemd.confirm_spawn= on the kernel command line should now
8055 * There's a new man page crypttab(5) which details all options
8056 we actually understand.
8058 * systemd-nspawn gained a new --capability= switch to pass
8059 additional capabilities to the container.
8061 * timedated will now read known NTP implementation unit names
8062 from /usr/lib/systemd/ntp-units.d/*.list,
8063 systemd-timedated-ntp.target has been removed.
8065 * journalctl gained a new switch "-b" that lists log data of
8066 the current boot only.
8068 * The notify socket is in the abstract namespace again, in
8069 order to support daemons which chroot() at start-up.
8071 * There is a new Storage= configuration option for journald
8072 which allows configuration of where log data should go. This
8073 also provides a way to disable journal logging entirely, so
8074 that data collected is only forwarded to the console, the
8075 kernel log buffer or another syslog implementation.
8077 * Many bugfixes and optimizations
8079 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8080 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8081 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8082 Shawn Landden, Tom Gundersen
8086 * "systemctl help <unit>" now shows the man page if one is
8089 * Several new man pages have been added.
8091 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8092 MaxLevelConsole= can now be specified in
8093 journald.conf. These options allow reducing the amount of
8094 data stored on disk or forwarded by the log level.
8096 * TimerSlackNSec= can now be specified in system.conf for
8097 PID1. This allows system-wide power savings.
8099 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8100 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8105 * logind is now capable of (optionally) handling power and
8106 sleep keys as well as the lid switch.
8108 * journalctl now understands the syntax "journalctl
8109 /usr/bin/avahi-daemon" to get all log output of a specific
8112 * CapabilityBoundingSet= in system.conf now also influences
8113 the capability bound set of usermode helpers of the kernel.
8115 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8116 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8117 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8118 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8122 * Note that we skipped 139 releases here in order to set the
8123 new version to something that is greater than both udev's
8124 and systemd's most recent version number.
8126 * udev: all udev sources are merged into the systemd source tree now.
8127 All future udev development will happen in the systemd tree. It
8128 is still fully supported to use the udev daemon and tools without
8129 systemd running, like in initramfs or other init systems. Building
8130 udev though, will require the *build* of the systemd tree, but
8131 udev can be properly *run* without systemd.
8133 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8134 should be used to create dead device nodes as workarounds for broken
8137 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8138 no longer supported. udev_monitor_new_from_netlink() needs to be
8139 used to subscribe to events.
8141 * udev: when udevd is started by systemd, processes which are left
8142 behind by forking them off of udev rules, are unconditionally cleaned
8143 up and killed now after the event handling has finished. Services or
8144 daemons must be started as systemd services. Services can be
8145 pulled-in by udev to get started, but they can no longer be directly
8146 forked by udev rules.
8148 * udev: the daemon binary is called systemd-udevd now and installed
8149 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8150 to adapt to that, create symlink, or rename the binary after building
8153 * libudev no longer provides these symbols:
8154 udev_monitor_from_socket()
8155 udev_queue_get_failed_list_entry()
8156 udev_get_{dev,sys,run}_path()
8157 The versions number was bumped and symbol versioning introduced.
8159 * systemd-loginctl and systemd-journalctl have been renamed
8160 to loginctl and journalctl to match systemctl.
8162 * The config files: /etc/systemd/systemd-logind.conf and
8163 /etc/systemd/systemd-journald.conf have been renamed to
8164 logind.conf and journald.conf. Package updates should rename
8165 the files to the new names on upgrade.
8167 * For almost all files the license is now LGPL2.1+, changed
8168 from the previous GPL2.0+. Exceptions are some minor stuff
8169 of udev (which will be changed to LGPL2.1 eventually, too),
8170 and the MIT licensed sd-daemon.[ch] library that is suitable
8171 to be used as drop-in files.
8173 * systemd and logind now handle system sleep states, in
8174 particular suspending and hibernating.
8176 * logind now implements a sleep/shutdown/idle inhibiting logic
8177 suitable for a variety of uses. Soonishly Lennart will blog
8178 about this in more detail.
8180 * var-run.mount and var-lock.mount are no longer provided
8181 (which previously bind mounted these directories to their new
8182 places). Distributions which have not converted these
8183 directories to symlinks should consider stealing these files
8184 from git history and add them downstream.
8186 * We introduced the Documentation= field for units and added
8187 this to all our shipped units. This is useful to make it
8188 easier to explore the boot and the purpose of the various
8191 * All smaller setup units (such as
8192 systemd-vconsole-setup.service) now detect properly if they
8193 are run in a container and are skipped when
8194 appropriate. This guarantees an entirely noise-free boot in
8195 Linux container environments such as systemd-nspawn.
8197 * A framework for implementing offline system updates is now
8198 integrated, for details see:
8199 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
8201 * A new service type Type=idle is available now which helps us
8202 avoiding ugly interleaving of getty output and boot status
8205 * There's now a system-wide CapabilityBoundingSet= option to
8206 globally reduce the set of capabilities for the
8207 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8208 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8209 even CAP_NET_ADMIN system-wide for secure systems.
8211 * There are now system-wide DefaultLimitXXX= options to
8212 globally change the defaults of the various resource limits
8213 for all units started by PID 1.
8215 * Harald Hoyer's systemd test suite has been integrated into
8216 systemd which allows easy testing of systemd builds in qemu
8217 and nspawn. (This is really awesome! Ask us for details!)
8219 * The fstab parser is now implemented as generator, not inside
8222 * systemctl will now warn you if .mount units generated from
8223 /etc/fstab are out of date due to changes in fstab that
8224 have not been read by systemd yet.
8226 * systemd is now suitable for usage in initrds. Dracut has
8227 already been updated to make use of this. With this in place
8228 initrds get a slight bit faster but primarily are much
8229 easier to introspect and debug since "systemctl status" in
8230 the host system can be used to introspect initrd services,
8231 and the journal from the initrd is kept around too.
8233 * systemd-delta has been added, a tool to explore differences
8234 between user/admin configuration and vendor defaults.
8236 * PrivateTmp= now affects both /tmp and /var/tmp.
8238 * Boot time status messages are now much prettier and feature
8239 proper english language. Booting up systemd has never been
8242 * Read-ahead pack files now include the inode number of all
8243 files to pre-cache. When the inode changes the pre-caching
8244 is not attempted. This should be nicer to deal with updated
8245 packages which might result in changes of read-ahead
8248 * We now temporaritly lower the kernel's read_ahead_kb variable
8249 when collecting read-ahead data to ensure the kernel's
8250 built-in read-ahead does not add noise to our measurements
8251 of necessary blocks to pre-cache.
8253 * There's now RequiresMountsFor= to add automatic dependencies
8254 for all mounts necessary for a specific file system path.
8256 * MountAuto= and SwapAuto= have been removed from
8257 system.conf. Mounting file systems at boot has to take place
8260 * nspawn now learned a new switch --uuid= to set the machine
8261 ID on the command line.
8263 * nspawn now learned the -b switch to automatically search
8266 * vt102 is now the default TERM for serial TTYs, upgraded from
8269 * systemd-logind now works on VT-less systems.
8271 * The build tree has been reorganized. The individual
8272 components now have directories of their own.
8274 * A new condition type ConditionPathIsReadWrite= is now available.
8276 * nspawn learned the new -C switch to create cgroups for the
8277 container in other hierarchies.
8279 * We now have support for hardware watchdogs, configurable in
8282 * The scheduled shutdown logic now has a public API.
8284 * We now mount /tmp as tmpfs by default, but this can be
8285 masked and /etc/fstab can override it.
8287 * Since udisks does not make use of /media anymore we are not
8288 mounting a tmpfs on it anymore.
8290 * journalctl gained a new --local switch to only interleave
8291 locally generated journal files.
8293 * We can now load the IMA policy at boot automatically.
8295 * The GTK tools have been split off into a systemd-ui.
8297 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8298 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8299 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8300 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8301 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8302 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8303 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8304 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8305 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8310 * This is mostly a bugfix release
8312 * Support optional initialization of the machine ID from the
8313 KVM or container configured UUID.
8315 * Support immediate reboots with "systemctl reboot -ff"
8317 * Show /etc/os-release data in systemd-analyze output
8319 * Many bugfixes for the journal, including endianness fixes and
8320 ensuring that disk space enforcement works
8322 * sd-login.h is C++ compatible again
8324 * Extend the /etc/os-release format on request of the Debian
8327 * We now refuse non-UTF8 strings used in various configuration
8328 and unit files. This is done to ensure we do not pass invalid
8329 data over D-Bus or expose it elsewhere.
8331 * Register Mimo USB Screens as suitable for automatic seat
8334 * Read SELinux client context from journal clients in a race
8337 * Reorder configuration file lookup order. /etc now always
8338 overrides /run in order to allow the administrator to always
8339 and unconditionally override vendor-supplied or
8340 automatically generated data.
8342 * The various user visible bits of the journal now have man
8343 pages. We still lack man pages for the journal API calls
8346 * We now ship all man pages in HTML format again in the
8349 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8350 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8351 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8356 * This is mostly a bugfix release
8358 * systems lacking /etc/os-release are no longer supported.
8360 * Various functionality updates to libsystemd-login.so
8362 * Track class of PAM logins to distinguish greeters from
8365 Contributions from: Kay Sievers, Lennart Poettering, Michael
8370 * This is an important bugfix release for v41.
8372 * Building man pages is now optional which should be useful
8373 for those building systemd from git but unwilling to install
8376 * Watchdog support for supervising services is now usable. In
8377 a future release support for hardware watchdogs
8378 (i.e. /dev/watchdog) will be added building on this.
8380 * Service start rate limiting is now configurable and can be
8381 turned off per service. When a start rate limit is hit a
8382 reboot can automatically be triggered.
8384 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8386 Contributions from: Benjamin Franzke, Bill Nottingham,
8387 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8388 Schmidt, Michał Górny, Piotr Drąg
8392 * The systemd binary is installed /usr/lib/systemd/systemd now;
8393 An existing /sbin/init symlink needs to be adapted with the
8396 * The code that loads kernel modules has been ported to invoke
8397 libkmod directly, instead of modprobe. This means we do not
8398 support systems with module-init-tools anymore.
8400 * Watchdog support is now already useful, but still not
8403 * A new kernel command line option systemd.setenv= is
8404 understood to set system wide environment variables
8405 dynamically at boot.
8407 * We now limit the set of capabilities of systemd-journald.
8409 * We now set SIGPIPE to ignore by default, since it only is
8410 useful in shell pipelines, and has little use in general
8411 code. This can be disabled with IgnoreSIPIPE=no in unit
8414 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8415 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8420 * This is mostly a bugfix release
8422 * We now expose the reason why a service failed in the
8423 "Result" D-Bus property.
8425 * Rudimentary service watchdog support (will be completed over
8426 the next few releases.)
8428 * When systemd forks off in order execute some service we will
8429 now immediately changes its argv[0] to reflect which process
8430 it will execute. This is useful to minimize the time window
8431 with a generic argv[0], which makes bootcharts more useful
8433 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8434 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8435 Mike Kazantsev, Ray Strode
8439 * This is mostly a test release, but incorporates many
8442 * New systemd-cgtop tool to show control groups by their
8445 * Linking against libacl for ACLs is optional again. If
8446 disabled, support tracking device access for active logins
8447 goes becomes unavailable, and so does access to the user
8448 journals by the respective users.
8450 * If a group "adm" exists, journal files are automatically
8451 owned by them, thus allow members of this group full access
8452 to the system journal as well as all user journals.
8454 * The journal now stores the SELinux context of the logging
8455 client for all entries.
8457 * Add C++ inclusion guards to all public headers
8459 * New output mode "cat" in the journal to print only text
8460 messages, without any meta data like date or time.
8462 * Include tiny X server wrapper as a temporary stop-gap to
8463 teach XOrg udev display enumeration. This is used by display
8464 managers such as gdm, and will go away as soon as XOrg
8465 learned native udev hotplugging for display devices.
8467 * Add new systemd-cat tool for executing arbitrary programs
8468 with STDERR/STDOUT connected to the journal. Can also act as
8469 BSD logger replacement, and does so by default.
8471 * Optionally store all locally generated coredumps in the
8472 journal along with meta data.
8474 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8475 writing short strings to files (for usage for /sys), and for
8476 creating symlinks, character and block device nodes.
8478 * New unit file option ControlGroupPersistent= to make cgroups
8479 persistent, following the mechanisms outlined in
8480 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
8482 * Support multiple local RTCs in a sane way
8484 * No longer monopolize IO when replaying readahead data on
8485 rotating disks, since we might starve non-file-system IO to
8486 death, since fanotify() will not see accesses done by blkid,
8489 * Do not show kernel threads in systemd-cgls anymore, unless
8490 requested with new -k switch.
8492 Contributions from: Dan Horák, Kay Sievers, Lennart
8493 Poettering, Michal Schmidt
8497 * This is mostly a test release, but incorporates many
8500 * The git repository moved to:
8501 git://anongit.freedesktop.org/systemd/systemd
8502 ssh://git.freedesktop.org/git/systemd/systemd
8504 * First release with the journal
8505 http://0pointer.de/blog/projects/the-journal.html
8507 * The journal replaces both systemd-kmsg-syslogd and
8508 systemd-stdout-bridge.
8510 * New sd_pid_get_unit() API call in libsystemd-logind
8512 * Many systemadm clean-ups
8514 * Introduce remote-fs-pre.target which is ordered before all
8515 remote mounts and may be used to start services before all
8518 * Added Mageia support
8520 * Add bash completion for systemd-loginctl
8522 * Actively monitor PID file creation for daemons which exit in
8523 the parent process before having finished writing the PID
8524 file in the daemon process. Daemons which do this need to be
8525 fixed (i.e. PID file creation must have finished before the
8526 parent exits), but we now react a bit more gracefully to them.
8528 * Add colourful boot output, mimicking the well-known output
8529 of existing distributions.
8531 * New option PassCredentials= for socket units, for
8532 compatibility with a recent kernel ABI breakage.
8534 * /etc/rc.local is now hooked in via a generator binary, and
8535 thus will no longer act as synchronization point during
8538 * systemctl list-unit-files now supports --root=.
8540 * systemd-tmpfiles now understands two new commands: z, Z for
8541 relabelling files according to the SELinux database. This is
8542 useful to apply SELinux labels to specific files in /sys,
8545 * Output of SysV services is now forwarded to both the console
8546 and the journal by default, not only just the console.
8548 * New man pages for all APIs from libsystemd-login.
8550 * The build tree got reorganized and the build system is a
8551 lot more modular allowing embedded setups to specifically
8552 select the components of systemd they are interested in.
8554 * Support for Linux systems lacking the kernel VT subsystem is
8557 * configure's --with-rootdir= got renamed to
8558 --with-rootprefix= to follow the naming used by udev and
8561 * Unless specified otherwise we will now install to /usr instead
8562 of /usr/local by default.
8564 * Processes with '@' in argv[0][0] are now excluded from the
8565 final shut-down killing spree, following the logic explained
8567 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
8569 * All processes remaining in a service cgroup when we enter
8570 the START or START_PRE states are now killed with
8571 SIGKILL. That means it is no longer possible to spawn
8572 background processes from ExecStart= lines (which was never
8573 supported anyway, and bad style).
8575 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8576 reloading of units together.
8578 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
8579 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8580 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8581 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8582 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek