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 * The Boot Loader Specification has been added to the source tree.
307 https://github.com/systemd/systemd/blob/master/doc/BOOT_LOADER_SPECIFICATION.md
309 While moving it into our source tree we have updated it and further
310 changes are now accepted through the usual github PR workflow.
312 * pam_elogind will now look for PAM userdata fields elogind.memory_max,
313 elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
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 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
337 when the system is booted in UEFI "secure mode".
339 * A new unit "system-update-pre.target" is added, which defines an
340 optional synchronization point for offline system updates, as
341 implemented by the pre-existing "system-update.target" unit. It
342 allows ordering services before the service that executes the actual
343 update process in a generic way.
345 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
346 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
347 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
348 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
349 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
350 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
351 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
352 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
353 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
354 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
355 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
356 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
357 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
358 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
359 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
360 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
361 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
362 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
363 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
364 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
365 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
366 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
367 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
368 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
369 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
370 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
371 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
372 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
373 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
379 * The MemoryAccounting= unit property now defaults to on. After
380 discussions with the upstream control group maintainers we learnt
381 that the negative impact of cgroup memory accounting on current
382 kernels is finally relatively minimal, so that it should be safe to
383 enable this by default without affecting system performance. Besides
384 memory accounting only task accounting is turned on by default, all
385 other forms of resource accounting (CPU, IO, IP) remain off for now,
386 because it's not clear yet that their impact is small enough to move
387 from opt-in to opt-out. We recommend downstreams to leave memory
388 accounting on by default if kernel 4.14 or higher is primarily
389 used. On very resource constrained systems or when support for old
390 kernels is a necessity, -Dmemory-accounting-default=false can be used
391 to revert this change.
393 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
394 %udev_rules_update) and the journal catalog (%journal_catalog_update)
395 from the upgrade scriptlets of individual packages now do nothing.
396 Transfiletriggers have been added which will perform those updates
397 once at the end of the transaction.
399 Similar transfiletriggers have been added to execute any sysctl.d
400 and binfmt.d rules. Thus, it should be unnecessary to provide any
401 scriptlets to execute this configuration from package installation
404 * elogind-sysusers gained a mode where the configuration to execute is
405 specified on the command line, but this configuration is not executed
406 directly, but instead it is merged with the configuration on disk,
407 and the result is executed. This is useful for package installation
408 scripts which want to create the user before installing any files on
409 disk (in case some of those files are owned by that user), while
410 still allowing local admin overrides.
412 This functionality is exposed to rpm scriptlets through a new
413 %sysusers_create_package macro. Old %sysusers_create and
414 %sysusers_create_inline macros are deprecated.
416 A transfiletrigger for sysusers.d configuration is now installed,
417 which means that it should be unnecessary to call elogind-sysusers from
418 package installation scripts, unless the package installs any files
419 owned by those newly-created users, in which case
420 %sysusers_create_package should be used.
422 * Analogous change has been done for elogind-tmpfiles: it gained a mode
423 where the command-line configuration is merged with the configuration
424 on disk. This is exposed as the new %tmpfiles_create_package macro,
425 and %tmpfiles_create is deprecated. A transfiletrigger is installed
426 for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
427 from package installation scripts.
429 * sysusers.d configuration for a user may now also specify the group
430 number, in addition to the user number ("u username 123:456"), or
431 without the user number ("u username -:456").
433 * Configution items for elogind-sysusers can now be specified as
434 positional arguments when the new --inline switch is used.
436 * The login shell of users created through sysusers.d may now be
437 specified (previously, it was always /bin/sh for root and
438 /sbin/nologin for other users).
440 * elogind-analyze gained a new --global switch to look at global user
441 configuration. It also gained a unit-paths verb to list the unit load
442 paths that are compiled into elogind (which can be used with
443 --elogind, --user, or --global).
445 * udevadm trigger gained a new --settle/-w option to wait for any
446 triggered events to finish (but just those, and not any other events
447 which are triggered meanwhile).
449 * The action that elogind takes when the lid is closed and the
450 machine is connected to external power can now be configured using
451 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
452 was determined by HandleLidSwitch=, and, for backwards compatibility,
453 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
455 * journalctl will periodically call sd_journal_process() to make it
456 resilient against inotify queue overruns when journal files are
457 rotated very quickly.
459 * Two new functions in libelogind — sd_bus_get_n_queued_read and
460 sd_bus_get_n_queued_write — may be used to check the number of
461 pending bus messages.
463 * elogind gained a new
464 org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
465 which can be used to migrate foreign processes to scope and service
466 units. The primary user for this new API is elogind itself: the
467 elogind --user instance uses this call of the elogind --system
468 instance to migrate processes if it itself gets the request to
469 migrate processes and the kernel refuses this due to access
470 restrictions. Thanks to this "elogind-run --scope --user …" works
471 again in pure cgroups v2 environments when invoked from the user
474 * A new TemporaryFileSystem= setting can be used to mask out part of
475 the real file system tree with tmpfs mounts. This may be combined
476 with BindPaths= and BindReadOnlyPaths= to hide files or directories
477 not relevant to the unit, while still allowing some paths lower in
478 the tree to be accessed.
480 ProtectHome=tmpfs may now be used to hide user home and runtime
481 directories from units, in a way that is mostly equivalent to
482 "TemporaryFileSystem=/home /run/user /root".
484 * Non-service units are now started with KeyringMode=shared by default.
485 This means that mount and swapon and other mount tools have access
486 to keys in the main keyring.
488 * /sys/fs/bpf is now mounted automatically.
490 * QNX virtualization is now detected by elogind-detect-virt and may
491 be used in ConditionVirtualization=.
493 * IPAccounting= may now be enabled also for slice units.
495 * A new -Dsplit-bin= build configuration switch may be used to specify
496 whether bin and sbin directories are merged, or if they should be
497 included separately in $PATH and various listings of executable
498 directories. The build configuration scripts will try to autodetect
499 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
500 system, but distributions are encouraged to configure this
503 * A new -Dok-color= build configuration switch may be used to change
504 the colour of "OK" status messages.
506 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
507 PrivateNetwork=yes was buggy in previous versions of elogind. This
508 means that after the upgrade and daemon-reexec, any such units must
511 * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
512 will not exclude read-only files owned by root from cleanup.
514 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
515 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
516 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
517 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
518 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
519 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
520 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
521 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
522 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
523 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
524 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
525 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
526 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
527 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
528 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
529 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
535 * Some keyboards come with a zoom see-saw or rocker which until now got
536 mapped to the Linux "zoomin/out" keys in hwdb. However, these
537 keycodes are not recognized by any major desktop. They now produce
538 Up/Down key events so that they can be used for scrolling.
540 * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
541 slightly: previously, if an argument was specified for lines of this
542 type (i.e. the right-most column was set) this string was appended to
543 existing files each time elogind-tmpfiles was run. This behaviour was
544 different from what the documentation said, and not particularly
545 useful, as repeated elogind-tmpfiles invocations would not be
546 idempotent and grow such files without bounds. With this release
547 behaviour has been altered slightly, to match what the documentation
548 says: lines of this type only have an effect if the indicated files
549 don't exist yet, and only then the argument string is written to the
552 * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
553 elogind-tmpfiles behaviour: previously, read-only files owned by root
554 were always excluded from the file "aging" algorithm (i.e. the
555 automatic clean-up of directories like /tmp based on
556 atime/mtime/ctime). We intend to drop this restriction, and age files
557 by default even when owned by root and read-only. This behaviour was
558 inherited from older tools, but there have been requests to remove
559 it, and it's not obvious why this restriction was made in the first
560 place. Please speak up now, if you are aware of software that reqires
561 this behaviour, otherwise we'll remove the restriction in v238.
563 * A new environment variable $SYSTEMD_OFFLINE is now understood by
564 systemctl. It takes a boolean argument. If on, systemctl assumes it
565 operates on an "offline" OS tree, and will not attempt to talk to the
566 service manager. Previously, this mode was implicitly enabled if a
567 chroot() environment was detected, and this new environment variable
568 now provides explicit control.
570 * .path and .socket units may now be created transiently, too.
571 Previously only service, mount, automount and timer units were
572 supported as transient units. The elogind-run tool has been updated
573 to expose this new functionality, you may hence use it now to bind
574 arbitrary commands to path or socket activation on-the-fly from the
575 command line. Moreover, almost all properties are now exposed for the
576 unit types that already supported transient operation.
578 * The elogind-mount command gained support for a new --owner= parameter
579 which takes a user name, which is then resolved and included in uid=
580 and gid= mount options string of the file system to mount.
582 * A new unit condition ConditionControlGroupController= has been added
583 that checks whether a specific cgroup controller is available.
585 * Unit files, udev's .link files, and elogind-networkd's .netdev and
586 .network files all gained support for a new condition
587 ConditionKernelVersion= for checking against specific kernel
590 * In elogind-networkd, the [IPVLAN] section in .netdev files gained
591 support for configuring device flags in the Flags= setting. In the
592 same files, the [Tunnel] section gained support for configuring
593 AllowLocalRemote=. The [Route] section in .network files gained
594 support for configuring InitialCongestionWindow=,
595 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
596 understands RapidCommit=.
598 * elogind-networkd's DHCPv6 support gained support for Prefix
601 * sd-bus gained support for a new "watch-bind" feature. When this
602 feature is enabled, an sd_bus connection may be set up to connect to
603 an AF_UNIX socket in the file system as soon as it is created. This
604 functionality is useful for writing early-boot services that
605 automatically connect to the system bus as soon as it is started,
606 without ugly time-based polling. elogind-networkd and
607 elogind-resolved have been updated to make use of this
608 functionality. busctl exposes this functionality in a new
609 --watch-bind= command line switch.
611 * sd-bus will now optionally synthesize a local "Connected" signal as
612 soon as a D-Bus connection is set up fully. This message mirrors the
613 already existing "Disconnected" signal which is synthesized when the
614 connection is terminated. This signal is generally useful but
615 particularly handy in combination with the "watch-bind" feature
616 described above. Synthesizing of this message has to be requested
617 explicitly through the new API call sd_bus_set_connected_signal(). In
618 addition a new call sd_bus_is_ready() has been added that checks
619 whether a connection is fully set up (i.e. between the "Connected" and
620 "Disconnected" signals).
622 * sd-bus gained two new calls sd_bus_request_name_async() and
623 sd_bus_release_name_async() for asynchronously registering bus
624 names. Similar, there is now sd_bus_add_match_async() for installing
625 a signal match asynchronously. All of elogind's own services have
626 been updated to make use of these calls. Doing these operations
627 asynchronously has two benefits: it reduces the risk of deadlocks in
628 case of cyclic dependencies between bus services, and it speeds up
629 service initialization since synchronization points for bus
630 round-trips are removed.
632 * sd-bus gained two new calls sd_bus_match_signal() and
633 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
634 and sd_bus_add_match_async() but instead of taking a D-Bus match
635 string take match fields as normal function parameters.
637 * sd-bus gained two new calls sd_bus_set_sender() and
638 sd_bus_message_set_sender() for setting the sender name of outgoing
639 messages (either for all outgoing messages or for just one specific
640 one). These calls are only useful in direct connections as on
641 brokered connections the broker fills in the sender anyway,
642 overwriting whatever the client filled in.
644 * sd-event gained a new pseudo-handle that may be specified on all API
645 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
646 used this refers to the default event loop object of the calling
647 thread. Note however that this does not implicitly allocate one —
648 which has to be done prior by using sd_event_default(). Similarly
649 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
650 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
651 to the default bus of the specified type of the calling thread. Here
652 too this does not implicitly allocate bus connection objects, this
653 has to be done prior with sd_bus_default() and friends.
655 * sd-event gained a new call pair
656 sd_event_source_{get|set}_io_fd_own(). This may be used to request
657 automatic closure of the file descriptor an IO event source watches
658 when the event source is destroyed.
660 * elogind-networkd gained support for natively configuring WireGuard
663 * In previous versions elogind synthesized user records both for the
664 "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
665 internally. In order to simplify distribution-wide renames of the
666 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
667 new transitional flag file has been added: if
668 /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
669 user and group record within the elogind codebase is disabled.
671 * elogind-notify gained a new --uid= option for selecting the source
672 user/UID to use for notification messages sent to the service
675 * journalctl gained a new --grep= option to list only entries in which
676 the message matches a certain pattern. By default matching is case
677 insensitive if the pattern is lowercase, and case sensitive
678 otherwise. Option --case-sensitive=yes|no can be used to override
679 this an specify case sensitivity or case insensitivity.
681 * There's now a "elogind-analyze service-watchdogs" command for printing
682 the current state of the service runtime watchdog, and optionally
683 enabling or disabling the per-service watchdogs system-wide if given a
684 boolean argument (i.e. the concept you configure in WatchdogSec=), for
685 debugging purposes. There's also a kernel command line option
686 elogind.service_watchdogs= for controlling the same.
688 * Two new "log-level" and "log-target" options for elogind-analyze were
689 added that merge the now deprecated get-log-level, set-log-level and
690 get-log-target, set-log-target pairs. The deprecated options are still
691 understood for backwards compatibility. The two new options print the
692 current value when no arguments are given, and set them when a
693 level/target is given as an argument.
695 * sysusers.d's "u" lines now optionally accept both a UID and a GID
696 specification, separated by a ":" character, in order to create users
697 where UID and GID do not match.
699 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
700 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
701 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
702 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
703 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
704 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
705 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
706 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
707 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
708 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
709 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
710 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
711 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
712 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
713 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
714 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
715 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
716 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
717 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
718 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
725 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
726 in v235 has been extended to also set the dummy.ko module option
727 numdummies=0, preventing the kernel from automatically creating
728 dummy0. All dummy interfaces must now be explicitly created.
730 * Unknown '%' specifiers in configuration files are now rejected. This
731 applies to units and tmpfiles.d configuration. Any percent characters
732 that are followed by a letter or digit that are not supposed to be
733 interpreted as the beginning of a specifier should be escaped by
734 doubling ("%%"). (So "size=5%" is still accepted, as well as
735 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
736 valid specifiers today.)
738 * systemd-resolved now maintains a new dynamic
739 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
740 recommended to make /etc/resolv.conf a symlink to it. This file
741 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
742 includes dynamically acquired search domains, achieving more correct
743 DNS resolution by software that bypasses local DNS APIs such as NSS.
745 * The "uaccess" udev tag has been dropped from /dev/kvm and
746 /dev/dri/renderD*. These devices now have the 0666 permissions by
747 default (but this may be changed at build-time). /dev/dri/renderD*
748 will now be owned by the "render" group along with /dev/kfd.
750 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
751 systemd-journal-gatewayd.service and
752 systemd-journal-upload.service. This means "nss-systemd" must be
753 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
754 services are resolved properly.
756 * In /etc/fstab two new mount options are now understood:
757 x-systemd.makefs and x-systemd.growfs. The former has the effect that
758 the configured file system is formatted before it is mounted, the
759 latter that the file system is resized to the full block device size
760 after it is mounted (i.e. if the file system is smaller than the
761 partition it resides on, it's grown). This is similar to the fsck
762 logic in /etc/fstab, and pulls in systemd-makefs@.service and
763 systemd-growfs@.service as necessary, similar to
764 systemd-fsck@.service. Resizing is currently only supported on ext4
767 * In systemd-networkd, the IPv6 RA logic now optionally may announce
768 DNS server and domain information.
770 * Support for the LUKS2 on-disk format for encrypted partitions has
771 been added. This requires libcryptsetup2 during compilation and
774 * The systemd --user instance will now signal "readiness" when its
775 basic.target unit has been reached, instead of when the run queue ran
776 empty for the first time.
778 * Tmpfiles.d with user configuration are now also supported.
779 systemd-tmpfiles gained a new --user switch, and snippets placed in
780 ~/.config/user-tmpfiles.d/ and corresponding directories will be
781 executed by systemd-tmpfiles --user running in the new
782 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
783 running in the user session.
785 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
786 %S resolves to the top-level state directory (/var/lib for the system
787 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
788 top-level cache directory (/var/cache for the system instance,
789 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
790 logs directory (/var/log for the system instance,
791 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
792 existing %t specifier, that resolves to the top-level runtime
793 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
796 * journalctl learnt a new parameter --output-fields= for limiting the
797 set of journal fields to output in verbose and JSON output modes.
799 * systemd-timesyncd's configuration file gained a new option
800 RootDistanceMaxSec= for setting the maximum root distance of servers
801 it'll use, as well as the new options PollIntervalMinSec= and
802 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
804 * bootctl gained a new command "list" for listing all available boot
805 menu items on systems that follow the boot loader specification.
807 * systemctl gained a new --dry-run switch that shows what would be done
808 instead of doing it, and is currently supported by the shutdown and
811 * ConditionSecurity= can now detect the TOMOYO security module.
813 * Unit file [Install] sections are now also respected in unit drop-in
814 files. This is intended to be used by drop-ins under /usr/lib/.
816 * systemd-firstboot may now also set the initial keyboard mapping.
818 * Udev "changed" events for devices which are exposed as systemd
819 .device units are now propagated to units specified in
820 ReloadPropagatedFrom= as reload requests.
822 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
823 unit template name (i.e. a name in the form of 'foobar@.service',
824 without the instance component between the '@' and - the '.'), then
825 the escaped sysfs path of the device is automatically used as the
828 * SystemCallFilter= in unit files has been extended so that an "errno"
829 can be specified individually for each system call. Example:
830 SystemCallFilter=~uname:EILSEQ.
832 * The cgroup delegation logic has been substantially updated. Delegate=
833 now optionally takes a list of controllers (instead of a boolean, as
834 before), which lists the controllers to delegate at least.
836 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
838 * A new LogLevelMax= setting configures the maximum log level any
839 process of the service may log at (i.e. anything with a lesser
840 priority than what is specified is automatically dropped). A new
841 LogExtraFields= setting allows configuration of additional journal
842 fields to attach to all log records generated by any of the unit's
845 * New StandardInputData= and StandardInputText= settings along with the
846 new option StandardInput=data may be used to configure textual or
847 binary data that shall be passed to the executed service process via
848 standard input, encoded in-line in the unit file.
850 * StandardInput=, StandardOutput= and StandardError= may now be used to
851 connect stdin/stdout/stderr of executed processes directly with a
852 file or AF_UNIX socket in the file system, using the new "file:" option.
854 * A new unit file option CollectMode= has been added, that allows
855 tweaking the garbage collection logic for units. It may be used to
856 tell systemd to garbage collect units that have failed automatically
857 (normally it only GCs units that exited successfully). systemd-run
858 and systemd-mount expose this new functionality with a new -G option.
860 * "machinectl bind" may now be used to bind mount non-directories
861 (i.e. regularfiles, devices, fifos, sockets).
863 * systemd-analyze gained a new verb "calendar" for validating and
864 testing calendar time specifications to use for OnCalendar= in timer
865 units. Besides validating the expression it will calculate the next
866 time the specified expression would elapse.
868 * In addition to the pre-existing FailureAction= unit file setting
869 there's now SuccessAction=, for configuring a shutdown action to
870 execute when a unit completes successfully. This is useful in
871 particular inside containers that shall terminate after some workload
872 has been completed. Also, both options are now supported for all unit
873 types, not just services.
875 * networkds's IP rule support gained two new options
876 IncomingInterface= and OutgoingInterface= for configuring the incoming
877 and outgoing interfaces of configured rules. systemd-networkd also
878 gained support for "vxcan" network devices.
880 * networkd gained a new setting RequiredForOnline=, taking a
881 boolean. If set, systemd-wait-online will take it into consideration
882 when determining that the system is up, otherwise it will ignore the
883 interface for this purpose.
885 * The sd_notify() protocol gained support for a new operation: with
886 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
887 store again, ahead of POLLHUP or POLLERR when they are removed
890 * A new document doc/UIDS-GIDS.md has been added to the source tree,
891 that documents the UID/GID range and assignment assumptions and
892 requirements of systemd.
894 * The watchdog device PID 1 will ping may now be configured through the
895 WatchdogDevice= configuration file setting, or by setting the
896 systemd.watchdog_service= kernel commandline option.
898 * systemd-resolved's gained support for registering DNS-SD services on
899 the local network using MulticastDNS. Services may either be
900 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
901 the same dir below /run, /usr/lib), or through its D-Bus API.
903 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
904 extend the effective start, runtime, and stop time. The service must
905 continue to send EXTEND_TIMEOUT_USEC within the period specified to
906 prevent the service manager from making the service as timedout.
908 * elogind-resolved's DNSSEC support gained support for RFC 8080
909 (Ed25519 keys and signatures).
911 * The elogind-resolve command line tool gained a new set of options
912 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
913 --set-nta= and --revert to configure per-interface DNS configuration
914 dynamically during runtime. It's useful for pushing DNS information
915 into elogind-resolved from DNS hook scripts that various interface
916 managing software supports (such as pppd).
918 * elogind-nspawn gained a new --network-namespace-path= command line
919 option, which may be used to make a container join an existing
920 network namespace, by specifying a path to a "netns" file.
922 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
923 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
924 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
925 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
926 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
927 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
928 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
929 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
930 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
931 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
932 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
933 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
934 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
935 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
936 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
937 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
938 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
939 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
940 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
941 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
942 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
943 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
944 Jędrzejewski-Szmek, Zeal Jagannatha
950 * INCOMPATIBILITY: systemd-logind.service and other long-running
951 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
952 communication with the outside. This generally improves security of
953 the system, and is in almost all cases a safe and good choice, as
954 these services do not and should not provide any network-facing
955 functionality. However, systemd-logind uses the glibc NSS API to
956 query the user database. This creates problems on systems where NSS
957 is set up to directly consult network services for user database
958 lookups. In particular, this creates incompatibilities with the
959 "nss-nis" module, which attempts to directly contact the NIS/YP
960 network servers it is configured for, and will now consistently
961 fail. In such cases, it is possible to turn off IP sandboxing for
962 systemd-logind.service (set IPAddressDeny= in its [Service] section
963 to the empty string, via a .d/ unit file drop-in). Downstream
964 distributions might want to update their nss-nis packaging to include
965 such a drop-in snippet, accordingly, to hide this incompatibility
966 from the user. Another option is to make use of glibc's nscd service
967 to proxy such network requests through a privilege-separated, minimal
968 local caching daemon, or to switch to more modern technologies such
969 sssd, whose NSS hook-ups generally do not involve direct network
970 access. In general, we think it's definitely time to question the
971 implementation choices of nss-nis, i.e. whether it's a good idea
972 today to embed a network-facing loadable module into all local
973 processes that need to query the user database, including the most
974 trivial and benign ones, such as "ls". For more details about
975 IPAddressDeny= see below.
977 * A new modprobe.d drop-in is now shipped by default that sets the
978 bonding module option max_bonds=0. This overrides the kernel default,
979 to avoid conflicts and ambiguity as to whether or not bond0 should be
980 managed by systemd-networkd or not. This resolves multiple issues
981 with bond0 properties not being applied, when bond0 is configured
982 with systemd-networkd. Distributors may choose to not package this,
983 however in that case users will be prevented from correctly managing
984 bond0 interface using systemd-networkd.
986 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
987 which print the logging level and target of the system manager. They
988 complement the existing "set-log-level" and "set-log-target" verbs
989 used to change those values.
991 * journald.conf gained a new boolean setting ReadKMsg= which defaults
992 to on. If turned off kernel log messages will not be read by
993 systemd-journald or included in the logs. It also gained a new
994 setting LineMax= for configuring the maximum line length in
995 STDOUT/STDERR log streams. The new default for this value is 48K, up
996 from the previous hardcoded 2048.
998 * A new unit setting RuntimeDirectoryPreserve= has been added, which
999 allows more detailed control of what to do with a runtime directory
1000 configured with RuntimeDirectory= (i.e. a directory below /run or
1001 $XDG_RUNTIME_DIR) after a unit is stopped.
1003 * The RuntimeDirectory= setting for units gained support for creating
1004 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
1005 one top-level directory.
1007 * Units gained new options StateDirectory=, CacheDirectory=,
1008 LogsDirectory= and ConfigurationDirectory= which are closely related
1009 to RuntimeDirectory= but manage per-service directories below
1010 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
1011 possible to write unit files which when activated automatically gain
1012 properly owned service specific directories in these locations, thus
1013 making unit files self-contained and increasing compatibility with
1014 stateless systems and factory reset where /etc or /var are
1015 unpopulated at boot. Matching these new settings there's also
1016 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1017 ConfigurationDirectoryMode= for configuring the access mode of these
1018 directories. These settings are particularly useful in combination
1019 with DynamicUser=yes as they provide secure, properly-owned,
1020 writable, and stateful locations for storage, excluded from the
1021 sandbox that such services live in otherwise.
1023 * Automake support has been removed from this release. systemd is now
1026 * systemd-journald will now aggressively cache client metadata during
1027 runtime, speeding up log write performance under pressure. This comes
1028 at a small price though: as much of the metadata is read
1029 asynchronously from /proc/ (and isn't implicitly attached to log
1030 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1031 metadata stored alongside a log entry might be slightly
1032 out-of-date. Previously it could only be slightly newer than the log
1033 message. The time window is small however, and given that the kernel
1034 is unlikely to be improved anytime soon in this regard, this appears
1037 * nss-myhostname/systemd-resolved will now by default synthesize an
1038 A/AAAA resource record for the "_gateway" hostname, pointing to the
1039 current default IP gateway. Previously it did that for the "gateway"
1040 name, hampering adoption, as some distributions wanted to leave that
1041 host name open for local use. The old behaviour may still be
1042 requested at build time.
1044 * systemd-networkd's [Address] section in .network files gained a new
1045 Scope= setting for configuring the IP address scope. The [Network]
1046 section gained a new boolean setting ConfigureWithoutCarrier= that
1047 tells systemd-networkd to ignore link sensing when configuring the
1048 device. The [DHCP] section gained a new Anonymize= boolean option for
1049 turning on a number of options suggested in RFC 7844. A new
1050 [RoutingPolicyRule] section has been added for configuring the IP
1051 routing policy. The [Route] section has gained support for a new
1052 Type= setting which permits configuring
1053 blackhole/unreachable/prohibit routes.
1055 * The [VRF] section in .netdev files gained a new Table= setting for
1056 configuring the routing table to use. The [Tunnel] section gained a
1057 new Independent= boolean field for configuring tunnels independent of
1058 an underlying network interface. The [Bridge] section gained a new
1059 GroupForwardMask= option for configuration of propagation of link
1060 local frames between bridge ports.
1062 * The WakeOnLan= setting in .link files gained support for a number of
1063 new modes. A new TCP6SegmentationOffload= setting has been added for
1064 configuring TCP/IPv6 hardware segmentation offload.
1066 * The IPv6 RA sender implementation may now optionally send out RDNSS
1067 and RDNSSL records to supply DNS configuration to peers.
1069 * systemd-nspawn gained support for a new --system-call-filter= command
1070 line option for adding and removing entries in the default system
1071 call filter it applies. Moreover systemd-nspawn has been changed to
1072 implement a system call whitelist instead of a blacklist.
1074 * systemd-run gained support for a new --pipe command line option. If
1075 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1076 are directly passed on to the activated transient service
1077 executable. This allows invoking arbitrary processes as systemd
1078 services (for example to take benefit of dependency management,
1079 accounting management, resource management or log management that is
1080 done automatically for services) — while still allowing them to be
1081 integrated in a classic UNIX shell pipeline.
1083 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1084 using ReloadPropagationTo= is configured, a reload is now propagated
1085 to configured units. (Previously this was only done on explicitly
1086 requested reloads, using "systemctl reload" or an equivalent
1089 * For each service unit a restart counter is now kept: it is increased
1090 each time the service is restarted due to Restart=, and may be
1091 queried using "systemctl show -p NRestarts …".
1093 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1094 @signal and @timer have been added, for usage with SystemCallFilter=
1095 in unit files and the new --system-call-filter= command line option
1096 of systemd-nspawn (see above).
1098 * ExecStart= lines in unit files gained two new modifiers: when a
1099 command line is prefixed with "!" the command will be executed as
1100 configured, except for the credentials applied by
1101 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1102 "+", but does still apply namespacing options unlike "+". There's
1103 also "!!" now, which is mostly identical, but becomes a NOP on
1104 systems that support ambient capabilities. This is useful to write
1105 unit files that work with ambient capabilities where possible but
1106 automatically fall back to traditional privilege dropping mechanisms
1107 on systems where this is not supported.
1109 * ListenNetlink= settings in socket units now support RDMA netlink
1112 * A new unit file setting LockPersonality= has been added which permits
1113 locking down the chosen execution domain ("personality") of a service
1116 * A new special target "getty-pre.target" has been added, which is
1117 ordered before all text logins, and may be used to order services
1118 before textual logins acquire access to the console.
1120 * systemd will now attempt to load the virtio-rng.ko kernel module very
1121 early on if a VM environment supporting this is detected. This should
1122 improve entropy during early boot in virtualized environments.
1124 * A _netdev option is now supported in /etc/crypttab that operates in a
1125 similar way as the same option in /etc/fstab: it permits configuring
1126 encrypted devices that need to be ordered after the network is up.
1127 Following this logic, two new special targets
1128 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
1129 added that are to cryptsetup.target what remote-fs.target and
1130 remote-fs-pre.target are to local-fs.target.
1132 * Service units gained a new UnsetEnvironment= setting which permits
1133 unsetting specific environment variables for services that are
1134 normally passed to it (for example in order to mask out locale
1135 settings for specific services that can't deal with it).
1137 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1138 traffic accounting (packet count as well as byte count) is done for
1139 the service, and shown as part of "systemctl status" or "systemd-run
1142 * Service units acquired two new options IPAddressAllow= and
1143 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1144 for configuring a simple IP access control list for all sockets of
1145 the unit. These options are available also on .slice and .socket
1146 units, permitting flexible access list configuration for individual
1147 services as well as groups of services (as defined by a slice unit),
1148 including system-wide. Note that IP ACLs configured this way are
1149 enforced on every single IPv4 and IPv6 socket created by any process
1150 of the service unit, and apply to ingress as well as egress traffic.
1152 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
1153 structured log message is generated each time the unit is stopped,
1154 containing information about the consumed resources of this
1157 * A new setting KeyringMode= has been added to unit files, which may be
1158 used to control how the kernel keyring is set up for executed
1161 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1162 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1163 behaviour (that is: these commands return immediately after the
1164 operation was enqueued instead of waiting for the operation to
1165 complete). Previously, "systemctl poweroff" and "systemctl reboot"
1166 were asynchronous on systems using systemd-logind (i.e. almost
1167 always, and like they were on sysvinit), and the other three commands
1168 were unconditionally synchronous. With this release this is cleaned
1169 up, and callers will see the same asynchronous behaviour on all
1170 systems for all five operations.
1172 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1175 * .timer units now accept calendar specifications in other timezones
1176 than UTC or the local timezone.
1178 * The tmpfiles snippet var.conf has been changed to create
1179 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1180 the "utmp" group already, and it appears to be generally understood
1181 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1182 databases. Previously this was implemented correctly for all these
1183 databases excepts btmp, which has been opened up like this now
1184 too. Note that while the other databases are world-readable
1185 (i.e. 0644), btmp is not and remains more restrictive.
1187 * The systemd-resolve tool gained a new --reset-server-features
1188 switch. When invoked like this systemd-resolved will forget
1189 everything it learnt about the features supported by the configured
1190 upstream DNS servers, and restarts the feature probing logic on the
1191 next resolver look-up for them at the highest feature level
1194 * The status dump systemd-resolved sends to the logs upon receiving
1195 SIGUSR1 now also includes information about all DNS servers it is
1196 configured to use, and the features levels it probed for them.
1198 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1199 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
1200 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1201 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1202 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1203 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1204 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1205 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1206 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1207 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1208 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1209 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1210 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1211 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1212 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1213 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1214 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1215 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1216 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1217 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1219 — Berlin, 2017-10-06
1223 * Meson is now supported as build system in addition to Automake. It is
1224 our plan to remove Automake in one of our next releases, so that
1225 Meson becomes our exclusive build system. Hence, please start using
1226 the Meson build system in your downstream packaging. There's plenty
1227 of documentation around how to use Meson, the extremely brief
1230 ./autogen.sh && ./configure && make && sudo make install
1234 meson build && ninja -C build && sudo ninja -C build install
1236 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1237 which permits configuring a timeout on the time a job is
1238 running. This is particularly useful for setting timeouts on jobs for
1241 * Unit files gained two new options ConditionUser= and ConditionGroup=
1242 for conditionalizing units based on the identity of the user/group
1243 running a systemd user instance.
1245 * systemd-networkd now understands a new FlowLabel= setting in the
1246 [VXLAN] section of .network files, as well as a Priority= in
1247 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1248 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1249 gained support for configuration of GENEVE links, and IPv6 address
1250 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1252 * .link files now understand a new Port= setting.
1254 * systemd-networkd's DHCP support gained support for DHCP option 119
1255 (domain search list).
1257 * systemd-networkd gained support for serving IPv6 address ranges using
1258 the Router Advertisement protocol. The new .network configuration
1259 section [IPv6Prefix] may be used to configure the ranges to
1260 serve. This is implemented based on a new, minimal, native server
1261 implementation of RA.
1263 * journalctl's --output= switch gained support for a new parameter
1264 "short-iso-precise" for a mode where timestamps are shown as precise
1267 * systemd-udevd's "net_id" builtin may now generate stable network
1268 interface names from IBM PowerVM VIO devices as well as ACPI platform
1271 * MulticastDNS support in systemd-resolved may now be explicitly
1272 enabled/disabled using the new MulticastDNS= configuration file
1275 * systemd-resolved may now optionally use libidn2 instead of the libidn
1276 for processing internationalized domain names. Support for libidn2
1277 should be considered experimental and should not be enabled by
1280 * "machinectl pull-tar" and related call may now do verification of
1281 downloaded images using SUSE-style .sha256 checksum files in addition
1282 to the already existing support for validating using Ubuntu-style
1285 * sd-bus gained support for a new sd_bus_message_appendv() call which
1286 is va_list equivalent of sd_bus_message_append().
1288 * sd-boot gained support for validating images using SHIM/MOK.
1290 * The SMACK code learnt support for "onlycap".
1292 * systemd-mount --umount is now much smarter in figuring out how to
1293 properly unmount a device given its mount or device path.
1295 * The code to call libnss_dns as a fallback from libnss_resolve when
1296 the communication with systemd-resolved fails was removed. This
1297 fallback was redundant and interfered with the [!UNAVAIL=return]
1298 suffix. See nss-resolve(8) for the recommended configuration.
1300 * systemd-logind may now be restarted without losing state. It stores
1301 the file descriptors for devices it manages in the system manager
1302 using the FDSTORE= mechanism. Please note that further changes in
1303 other components may be required to make use of this (for example
1304 Xorg has code to listen for stops of systemd-logind and terminate
1305 itself when logind is stopped or restarted, in order to avoid using
1306 stale file descriptors for graphical devices, which is now
1307 counterproductive and must be reverted in order for restarts of
1308 systemd-logind to be safe. See
1309 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1311 * All kernel install plugins are called with the environment variable
1312 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1313 /etc/machine-id. If the file is missing or empty, the variable is
1314 empty and BOOT_DIR_ABS is the path of a temporary directory which is
1315 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
1316 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1318 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
1319 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1320 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1321 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1322 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
1323 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1324 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1325 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1326 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
1327 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1328 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1329 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1330 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1331 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1332 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1333 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1334 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1335 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1336 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1337 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1338 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1339 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1340 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1341 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1342 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
1343 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1344 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
1345 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1346 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1349 — Berlin, 2017-07-12
1353 * The "hybrid" control group mode has been modified to improve
1354 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1355 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1356 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1357 cgroups-v1 hierarchy), the only externally visible change being that
1358 the cgroups-v2 hierarchy is also mounted, to
1359 /sys/fs/cgroup/unified. This should provide a large degree of
1360 compatibility with "legacy" cgroups-v1, while taking benefit of the
1361 better management capabilities of cgroups-v2.
1363 * The default control group setup mode may be selected both a boot-time
1364 via a set of kernel command line parameters (specifically:
1365 systemd.unified_cgroup_hierarchy= and
1366 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1367 default selected on the configure command line
1368 (--with-default-hierarchy=). The upstream default is "hybrid"
1369 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1370 this will change in a future systemd version to be "unified" (pure
1371 cgroups-v2 mode). The third option for the compile time option is
1372 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1373 distributions to default to "hybrid" mode for release distributions,
1374 starting with v233. We recommend "unified" for development
1375 distributions (specifically: distributions such as Fedora's rawhide)
1376 as that's where things are headed in the long run. Use "legacy" for
1377 greatest stability and compatibility only.
1379 * Note one current limitation of "unified" and "hybrid" control group
1380 setup modes: the kernel currently does not permit the systemd --user
1381 instance (i.e. unprivileged code) to migrate processes between two
1382 disconnected cgroup subtrees, even if both are managed and owned by
1383 the user. This effectively means "systemd-run --user --scope" doesn't
1384 work when invoked from outside of any "systemd --user" service or
1385 scope. Specifically, it is not supported from session scopes. We are
1386 working on fixing this in a future systemd version. (See #3388 for
1387 further details about this.)
1389 * DBus policy files are now installed into /usr rather than /etc. Make
1390 sure your system has dbus >= 1.9.18 running before upgrading to this
1391 version, or override the install path with --with-dbuspolicydir= .
1393 * All python scripts shipped with systemd (specifically: the various
1394 tests written in Python) now require Python 3.
1396 * systemd unit tests can now run standalone (without the source or
1397 build directories), and can be installed into /usr/lib/systemd/tests/
1398 with 'make install-tests'.
1400 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1401 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1404 * Support for the %c, %r, %R specifiers in unit files has been
1405 removed. Specifiers are not supposed to be dependent on configuration
1406 in the unit file itself (so that they resolve the same regardless
1407 where used in the unit files), but these specifiers were influenced
1408 by the Slice= option.
1410 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1411 all cases. If distributions want to use a different shell for this
1412 purpose (for example Fedora's /sbin/sushell) they need to specify
1413 this explicitly at configure time using --with-debug-shell=.
1415 * The confirmation spawn prompt has been reworked to offer the
1418 (c)ontinue, proceed without asking anymore
1419 (D)ump, show the state of the unit
1420 (f)ail, don't execute the command and pretend it failed
1422 (i)nfo, show a short summary of the unit
1423 (j)obs, show jobs that are in progress
1424 (s)kip, don't execute the command and pretend it succeeded
1425 (y)es, execute the command
1427 The 'n' choice for the confirmation spawn prompt has been removed,
1428 because its meaning was confusing.
1430 The prompt may now also be redirected to an alternative console by
1431 specifying the console as parameter to systemd.confirm_spawn=.
1433 * Services of Type=notify require a READY=1 notification to be sent
1434 during startup. If no such message is sent, the service now fails,
1435 even if the main process exited with a successful exit code.
1437 * Services that fail to start up correctly now always have their
1438 ExecStopPost= commands executed. Previously, they'd enter "failed"
1439 state directly, without executing these commands.
1441 * The option MulticastDNS= of network configuration files has acquired
1442 an actual implementation. With MulticastDNS=yes a host can resolve
1443 names of remote hosts and reply to mDNS A and AAAA requests.
1445 * When units are about to be started an additional check is now done to
1446 ensure that all dependencies of type BindsTo= (when used in
1447 combination with After=) have been started.
1449 * systemd-analyze gained a new verb "syscall-filter" which shows which
1450 system call groups are defined for the SystemCallFilter= unit file
1451 setting, and which system calls they contain.
1453 * A new system call filter group "@filesystem" has been added,
1454 consisting of various file system related system calls. Group
1455 "@reboot" has been added, covering reboot, kexec and shutdown related
1456 calls. Finally, group "@swap" has been added covering swap
1457 configuration related calls.
1459 * A new unit file option RestrictNamespaces= has been added that may be
1460 used to restrict access to the various process namespace types the
1461 Linux kernel provides. Specifically, it may be used to take away the
1462 right for a service unit to create additional file system, network,
1463 user, and other namespaces. This sandboxing option is particularly
1464 relevant due to the high amount of recently discovered namespacing
1465 related vulnerabilities in the kernel.
1467 * systemd-udev's .link files gained support for a new AutoNegotiation=
1468 setting for configuring Ethernet auto-negotiation.
1470 * systemd-networkd's .network files gained support for a new
1471 ListenPort= setting in the [DHCP] section to explicitly configure the
1472 UDP client port the DHCP client shall listen on.
1474 * .network files gained a new Unmanaged= boolean setting for explicitly
1475 excluding one or more interfaces from management by systemd-networkd.
1477 * The systemd-networkd ProxyARP= option has been renamed to
1478 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
1479 renamed to ReduceARPProxy=. The old names continue to be available
1482 * systemd-networkd gained support for configuring IPv6 Proxy NDP
1483 addresses via the new IPv6ProxyNDPAddress= .network file setting.
1485 * systemd-networkd's bonding device support gained support for two new
1486 configuration options ActiveSlave= and PrimarySlave=.
1488 * The various options in the [Match] section of .network files gained
1489 support for negative matching.
1491 * New systemd-specific mount options are now understood in /etc/fstab:
1493 x-systemd.mount-timeout= may be used to configure the maximum
1494 permitted runtime of the mount command.
1496 x-systemd.device-bound may be set to bind a mount point to its
1497 backing device unit, in order to automatically remove a mount point
1498 if its backing device is unplugged. This option may also be
1499 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
1500 on the block device, which is now automatically set for all CDROM
1501 drives, so that mounted CDs are automatically unmounted when they are
1502 removed from the drive.
1504 x-systemd.after= and x-systemd.before= may be used to explicitly
1505 order a mount after or before another unit or mount point.
1507 * Enqueued start jobs for device units are now automatically garbage
1508 collected if there are no jobs waiting for them anymore.
1510 * systemctl list-jobs gained two new switches: with --after, for every
1511 queued job the jobs it's waiting for are shown; with --before the
1512 jobs which it's blocking are shown.
1514 * systemd-nspawn gained support for ephemeral boots from disk images
1515 (or in other words: --ephemeral and --image= may now be
1516 combined). Moreover, ephemeral boots are now supported for normal
1517 directories, even if the backing file system is not btrfs. Of course,
1518 if the file system does not support file system snapshots or
1519 reflinks, the initial copy operation will be relatively expensive, but
1520 this should still be suitable for many use cases.
1522 * Calendar time specifications in .timer units now support
1523 specifications relative to the end of a month by using "~" instead of
1524 "-" as separator between month and day. For example, "*-02~03" means
1525 "the third last day in February". In addition a new syntax for
1526 repeated events has been added using the "/" character. For example,
1527 "9..17/2:00" means "every two hours from 9am to 5pm".
1529 * systemd-socket-proxyd gained a new parameter --connections-max= for
1530 configuring the maximum number of concurrent connections.
1532 * sd-id128 gained a new API for generating unique IDs for the host in a
1533 way that does not leak the machine ID. Specifically,
1534 sd_id128_get_machine_app_specific() derives an ID based on the
1535 machine ID a in well-defined, non-reversible, stable way. This is
1536 useful whenever an identifier for the host is needed but where the
1537 identifier shall not be useful to identify the system beyond the
1538 scope of the application itself. (Internally this uses HMAC-SHA256 as
1539 keyed hash function using the machine ID as input.)
1541 * NotifyAccess= gained a new supported value "exec". When set
1542 notifications are accepted from all processes systemd itself invoked,
1543 including all control processes.
1545 * .nspawn files gained support for defining overlay mounts using the
1546 Overlay= and OverlayReadOnly= options. Previously this functionality
1547 was only available on the systemd-nspawn command line.
1549 * systemd-nspawn's --bind= and --overlay= options gained support for
1550 bind/overlay mounts whose source lies within the container tree by
1551 prefixing the source path with "+".
1553 * systemd-nspawn's --bind= and --overlay= options gained support for
1554 automatically allocating a temporary source directory in /var/tmp
1555 that is removed when the container dies. Specifically, if the source
1556 directory is specified as empty string this mechanism is selected. An
1557 example usage is --overlay=+/var::/var, which creates an overlay
1558 mount based on the original /var contained in the image, overlayed
1559 with a temporary directory in the host's /var/tmp. This way changes
1560 to /var are automatically flushed when the container shuts down.
1562 * systemd-nspawn --image= option does now permit raw file system block
1563 devices (in addition to images containing partition tables, as
1566 * The disk image dissection logic in systemd-nspawn gained support for
1567 automatically setting up LUKS encrypted as well as Verity protected
1568 partitions. When a container is booted from an encrypted image the
1569 passphrase is queried at start-up time. When a container with Verity
1570 data is started, the root hash is search in a ".roothash" file
1571 accompanying the disk image (alternatively, pass the root hash via
1572 the new --root-hash= command line option).
1574 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
1575 be used to dissect disk images the same way as systemd-nspawn does
1576 it, following the Bootable Partition Specification. It may even be
1577 used to mount disk images with complex partition setups (including
1578 LUKS and Verity partitions) to a local host directory, in order to
1579 inspect them. This tool is not considered public API (yet), and is
1580 thus not installed into /usr/bin. Please do not rely on its
1581 existence, since it might go away or be changed in later systemd
1584 * A new generator "systemd-verity-generator" has been added, similar in
1585 style to "systemd-cryptsetup-generator", permitting automatic setup of
1586 Verity root partitions when systemd boots up. In order to make use of
1587 this your partition setup should follow the Discoverable Partitions
1588 Specification, and the GPT partition ID of the root file system
1589 partition should be identical to the upper 128bit of the Verity root
1590 hash. The GPT partition ID of the Verity partition protecting it
1591 should be the lower 128bit of the Verity root hash. If the partition
1592 image follows this model it is sufficient to specify a single
1593 "roothash=" kernel command line argument to both configure which root
1594 image and verity partition to use as well as the root hash for
1595 it. Note that systemd-nspawn's Verity support follows the same
1596 semantics, meaning that disk images with proper Verity data in place
1597 may be booted in containers with systemd-nspawn as well as on
1598 physical systems via the verity generator. Also note that the "mkosi"
1599 tool available at https://github.com/systemd/mkosi has been updated
1600 to generate Verity protected disk images following this scheme. In
1601 fact, it has been updated to generate disk images that optionally
1602 implement a complete UEFI SecureBoot trust chain, involving a signed
1603 kernel and initrd image that incorporates such a root hash as well as
1604 a Verity-enabled root partition.
1606 * The hardware database (hwdb) udev supports has been updated to carry
1607 accelerometer quirks.
1609 * All system services are now run with a fresh kernel keyring set up
1610 for them. The invocation ID is stored by default in it, thus
1611 providing a safe, non-overridable way to determine the invocation
1614 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
1615 options for bind mounting arbitrary paths in a service-specific
1616 way. When these options are used, arbitrary host or service files and
1617 directories may be mounted to arbitrary locations in the service's
1620 * Documentation has been added that lists all of systemd's low-level
1621 environment variables:
1623 https://github.com/systemd/systemd/blob/master/doc/ENVIRONMENT.md
1625 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
1626 whether a specific socket file descriptor matches a specified socket
1629 * systemd-firstboot has been updated to check for the
1630 systemd.firstboot= kernel command line option. It accepts a boolean
1631 and when set to false the first boot questions are skipped.
1633 * systemd-fstab-generator has been updated to check for the
1634 systemd.volatile= kernel command line option, which either takes an
1635 optional boolean parameter or the special value "state". If used the
1636 system may be booted in a "volatile" boot mode. Specifically,
1637 "systemd.volatile" is used, the root directory will be mounted as
1638 tmpfs, and only /usr is mounted from the actual root file system. If
1639 "systemd.volatile=state" is used, the root directory will be mounted
1640 as usual, but /var is mounted as tmpfs. This concept provides similar
1641 functionality as systemd-nspawn's --volatile= option, but provides it
1642 on physical boots. Use this option for implementing stateless
1643 systems, or testing systems with all state and/or configuration reset
1644 to the defaults. (Note though that many distributions are not
1645 prepared to boot up without a populated /etc or /var, though.)
1647 * systemd-gpt-auto-generator gained support for LUKS encrypted root
1648 partitions. Previously it only supported LUKS encrypted partitions
1649 for all other uses, except for the root partition itself.
1651 * Socket units gained support for listening on AF_VSOCK sockets for
1652 communication in virtualized QEMU environments.
1654 * The "configure" script gained a new option --with-fallback-hostname=
1655 for specifying the fallback hostname to use if none is configured in
1656 /etc/hostname. For example, by specifying
1657 --with-fallback-hostname=fedora it is possible to default to a
1658 hostname of "fedora" on pristine installations.
1660 * systemd-cgls gained support for a new --unit= switch for listing only
1661 the control groups of a specific unit. Similar --user-unit= has been
1662 added for listing only the control groups of a specific user unit.
1664 * systemd-mount gained a new --umount switch for unmounting a mount or
1665 automount point (and all mount/automount points below it).
1667 * systemd will now refuse full configuration reloads (via systemctl
1668 daemon-reload and related calls) unless at least 16MiB of free space
1669 are available in /run. This is a safety precaution in order to ensure
1670 that generators can safely operate after the reload completed.
1672 * A new unit file option RootImage= has been added, which has a similar
1673 effect as RootDirectory= but mounts the service's root directory from
1674 a disk image instead of plain directory. This logic reuses the same
1675 image dissection and mount logic that systemd-nspawn already uses,
1676 and hence supports any disk images systemd-nspawn supports, including
1677 those following the Discoverable Partition Specification, as well as
1678 Verity enabled images. This option enables systemd to run system
1679 services directly off disk images acting as resource bundles,
1680 possibly even including full integrity data.
1682 * A new MountAPIVFS= unit file option has been added, taking a boolean
1683 argument. If enabled /proc, /sys and /dev (collectively called the
1684 "API VFS") will be mounted for the service. This is only relevant if
1685 RootDirectory= or RootImage= is used for the service, as these mounts
1686 are of course in place in the host mount namespace anyway.
1688 * systemd-nspawn gained support for a new --pivot-root= switch. If
1689 specified the root directory within the container image is pivoted to
1690 the specified mount point, while the original root disk is moved to a
1691 different place. This option enables booting of ostree images
1692 directly with systemd-nspawn.
1694 * The systemd build scripts will no longer complain if the NTP server
1695 addresses are not changed from the defaults. Google now supports
1696 these NTP servers officially. We still recommend downstreams to
1697 properly register an NTP pool with the NTP pool project though.
1699 * coredumpctl gained a new "--reverse" option for printing the list
1700 of coredumps in reverse order.
1702 * coredumpctl will now show additional information about truncated and
1703 inaccessible coredumps, as well as coredumps that are still being
1704 processed. It also gained a new --quiet switch for suppressing
1705 additional informational message in its output.
1707 * coredumpctl gained support for only showing coredumps newer and/or
1708 older than specific timestamps, using the new --since= and --until=
1709 options, reminiscent of journalctl's options by the same name.
1711 * The systemd-coredump logic has been improved so that it may be reused
1712 to collect backtraces in non-compiled languages, for example in
1713 scripting languages such as Python.
1715 * machinectl will now show the UID shift of local containers, if user
1716 namespacing is enabled for them.
1718 * systemd will now optionally run "environment generator" binaries at
1719 configuration load time. They may be used to add environment
1720 variables to the environment block passed to services invoked. One
1721 user environment generator is shipped by default that sets up
1722 environment variables based on files dropped into /etc/environment.d
1723 and ~/.config/environment.d/.
1725 * systemd-resolved now includes the new, recently published 2017 DNSSEC
1728 * hostnamed has been updated to report a new chassis type of
1729 "convertible" to cover "foldable" laptops that can both act as a
1730 tablet and as a laptop, such as various Lenovo Yoga devices.
1732 Contributions from: Adrián López, Alexander Galanin, Alexander
1733 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
1734 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
1735 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
1736 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
1737 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
1738 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
1739 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
1740 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
1741 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
1742 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
1743 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
1744 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
1745 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
1746 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
1747 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
1748 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
1749 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
1750 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
1751 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
1752 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
1753 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
1754 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
1755 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
1756 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
1757 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
1758 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
1759 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
1762 — Berlin, 2017-03-01
1766 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
1767 RestrictAddressFamilies= enabled. These sandboxing options should
1768 generally be compatible with the various external udev call-out
1769 binaries we are aware of, however there may be exceptions, in
1770 particular when exotic languages for these call-outs are used. In
1771 this case, consider turning off these settings locally.
1773 * The new RemoveIPC= option can be used to remove IPC objects owned by
1774 the user or group of a service when that service exits.
1776 * The new ProtectKernelModules= option can be used to disable explicit
1777 load and unload operations of kernel modules by a service. In
1778 addition access to /usr/lib/modules is removed if this option is set.
1780 * ProtectSystem= option gained a new value "strict", which causes the
1781 whole file system tree with the exception of /dev, /proc, and /sys,
1782 to be remounted read-only for a service.
1784 * The new ProtectKernelTunables= option can be used to disable
1785 modification of configuration files in /sys and /proc by a service.
1786 Various directories and files are remounted read-only, so access is
1787 restricted even if the file permissions would allow it.
1789 * The new ProtectControlGroups= option can be used to disable write
1790 access by a service to /sys/fs/cgroup.
1792 * Various systemd services have been hardened with
1793 ProtectKernelTunables=yes, ProtectControlGroups=yes,
1794 RestrictAddressFamilies=.
1796 * Support for dynamically creating users for the lifetime of a service
1797 has been added. If DynamicUser=yes is specified, user and group IDs
1798 will be allocated from the range 61184..65519 for the lifetime of the
1799 service. They can be resolved using the new nss-systemd.so NSS
1800 module. The module must be enabled in /etc/nsswitch.conf. Services
1801 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
1802 any resources allocated by the service will be cleaned up when the
1803 service exits. They also have ProtectHome=read-only and
1804 ProtectSystem=strict enabled, so they are not able to make any
1805 permanent modifications to the system.
1807 * The nss-systemd module also always resolves root and nobody, making
1808 it possible to have no /etc/passwd or /etc/group files in minimal
1809 container or chroot environments.
1811 * Services may be started with their own user namespace using the new
1812 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
1813 under which the service is running are mapped. All other users are
1816 * Support for the cgroup namespace has been added to systemd-nspawn. If
1817 supported by kernel, the container system started by systemd-nspawn
1818 will have its own view of the cgroup hierarchy. This new behaviour
1819 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
1821 * The new MemorySwapMax= option can be used to limit the maximum swap
1822 usage under the unified cgroup hierarchy.
1824 * Support for the CPU controller in the unified cgroup hierarchy has
1825 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
1826 options. This controller requires out-of-tree patches for the kernel
1827 and the support is provisional.
1829 * Mount and automount units may now be created transiently
1830 (i.e. dynamically at runtime via the bus API, instead of requiring
1831 unit files in the file system).
1833 * systemd-mount is a new tool which may mount file systems – much like
1834 mount(8), optionally pulling in additional dependencies through
1835 transient .mount and .automount units. For example, this tool
1836 automatically runs fsck on a backing block device before mounting,
1837 and allows the automount logic to be used dynamically from the
1838 command line for establishing mount points. This tool is particularly
1839 useful when dealing with removable media, as it will ensure fsck is
1840 run – if necessary – before the first access and that the file system
1841 is quickly unmounted after each access by utilizing the automount
1842 logic. This maximizes the chance that the file system on the
1843 removable media stays in a clean state, and if it isn't in a clean
1844 state is fixed automatically.
1846 * LazyUnmount=yes option for mount units has been added to expose the
1847 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
1850 * /efi will be used as the mount point of the EFI boot partition, if
1851 the directory is present, and the mount point was not configured
1852 through other means (e.g. fstab). If /efi directory does not exist,
1853 /boot will be used as before. This makes it easier to automatically
1854 mount the EFI partition on systems where /boot is used for something
1857 * When operating on GPT disk images for containers, systemd-nspawn will
1858 now mount the ESP to /boot or /efi according to the same rules as PID
1859 1 running on a host. This allows tools like "bootctl" to operate
1860 correctly within such containers, in order to make container images
1861 bootable on physical systems.
1863 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
1865 * Two new user session targets have been added to support running
1866 graphical sessions under the systemd --user instance:
1867 graphical-session.target and graphical-session-pre.target. See
1868 systemd.special(7) for a description of how those targets should be
1871 * The vconsole initialization code has been significantly reworked to
1872 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
1873 support unicode keymaps. Font and keymap configuration will now be
1874 copied to all allocated virtual consoles.
1876 * FreeBSD's bhyve virtualization is now detected.
1878 * Information recorded in the journal for core dumps now includes the
1879 contents of /proc/mountinfo and the command line of the process at
1880 the top of the process hierarchy (which is usually the init process
1883 * systemd-journal-gatewayd learned the --directory= option to serve
1884 files from the specified location.
1886 * journalctl --root=… can be used to peruse the journal in the
1887 /var/log/ directories inside of a container tree. This is similar to
1888 the existing --machine= option, but does not require the container to
1891 * The hardware database has been extended to support
1892 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
1895 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
1896 specify the click rate for mice which include a horizontal wheel with
1897 a click rate that is different than the one for the vertical wheel.
1899 * systemd-run gained a new --wait option that makes service execution
1900 synchronous. (Specifically, the command will not return until the
1901 specified service binary exited.)
1903 * systemctl gained a new --wait option that causes the start command to
1904 wait until the units being started have terminated again.
1906 * A new journal output mode "short-full" has been added which displays
1907 timestamps with abbreviated English day names and adds a timezone
1908 suffix. Those timestamps include more information than the default
1909 "short" output mode, and can be passed directly to journalctl's
1910 --since= and --until= options.
1912 * /etc/resolv.conf will be bind-mounted into containers started by
1913 systemd-nspawn, if possible, so any changes to resolv.conf contents
1914 are automatically propagated to the container.
1916 * The number of instances for socket-activated services originating
1917 from a single IP address can be limited with
1918 MaxConnectionsPerSource=, extending the existing setting of
1921 * systemd-networkd gained support for vcan ("Virtual CAN") interface
1924 * .netdev and .network configuration can now be extended through
1927 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
1928 Segmentation Offload, Generic Receive Offload, Large Receive Offload
1929 can be enabled and disabled using the new UDPSegmentationOffload=,
1930 TCPSegmentationOffload=, GenericSegmentationOffload=,
1931 GenericReceiveOffload=, LargeReceiveOffload= options in the
1932 [Link] section of .link files.
1934 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
1935 Port VLAN ID can be configured for bridge devices using the new STP=,
1936 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
1937 section of .netdev files.
1939 * The route table to which routes received over DHCP or RA should be
1940 added can be configured with the new RouteTable= option in the [DHCP]
1941 and [IPv6AcceptRA] sections of .network files.
1943 * The Address Resolution Protocol can be disabled on links managed by
1944 systemd-networkd using the ARP=no setting in the [Link] section of
1947 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
1948 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
1949 encode information about the result and exit codes of the current
1950 service runtime cycle.
1952 * systemd-sysctl will now configure kernel parameters in the order
1953 they occur in the configuration files. This matches what sysctl
1954 has been traditionally doing.
1956 * kernel-install "plugins" that are executed to perform various
1957 tasks after a new kernel is added and before an old one is removed
1958 can now return a special value to terminate the procedure and
1959 prevent any later plugins from running.
1961 * Journald's SplitMode=login setting has been deprecated. It has been
1962 removed from documentation, and its use is discouraged. In a future
1963 release it will be completely removed, and made equivalent to current
1964 default of SplitMode=uid.
1966 * Storage=both option setting in /etc/systemd/coredump.conf has been
1967 removed. With fast LZ4 compression storing the core dump twice is not
1970 * The --share-system systemd-nspawn option has been replaced with an
1971 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
1972 this functionality is discouraged. In addition the variables
1973 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
1974 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
1975 individual namespaces.
1977 * "machinectl list" now shows the IP address of running containers in
1978 the output, as well as OS release information.
1980 * "loginctl list" now shows the TTY of each session in the output.
1982 * sd-bus gained new API calls sd_bus_track_set_recursive(),
1983 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
1984 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
1985 tracking objects in a "recursive" mode, where a single client can be
1986 counted multiple times, if it takes multiple references.
1988 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
1989 sd_bus_get_exit_on_disconnect(). They may be used to make a
1990 process using sd-bus automatically exit if the bus connection is
1993 * Bus clients of the service manager may now "pin" loaded units into
1994 memory, by taking an explicit reference on them. This is useful to
1995 ensure the client can retrieve runtime data about the service even
1996 after the service completed execution. Taking such a reference is
1997 available only for privileged clients and should be helpful to watch
1998 running services in a race-free manner, and in particular collect
1999 information about exit statuses and results.
2001 * The nss-resolve module has been changed to strictly return UNAVAIL
2002 when communication via D-Bus with resolved failed, and NOTFOUND when
2003 a lookup completed but was negative. This means it is now possible to
2004 neatly configure fallbacks using nsswitch.conf result checking
2005 expressions. Taking benefit of this, the new recommended
2006 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
2008 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
2010 * A new setting CtrlAltDelBurstAction= has been added to
2011 /etc/systemd/system.conf which may be used to configure the precise
2012 behaviour if the user on the console presses Ctrl-Alt-Del more often
2013 than 7 times in 2s. Previously this would unconditionally result in
2014 an expedited, immediate reboot. With this new setting the precise
2015 operation may be configured in more detail, and also turned off
2018 * In .netdev files two new settings RemoteChecksumTx= and
2019 RemoteChecksumRx= are now understood that permit configuring the
2020 remote checksumming logic for VXLAN networks.
2022 * The service manager learnt a new "invocation ID" concept for invoked
2023 services. Each runtime cycle of a service will get a new invocation
2024 ID (a 128bit random UUID) assigned that identifies the current
2025 run of the service uniquely and globally. A new invocation ID
2026 is generated each time a service starts up. The journal will store
2027 the invocation ID of a service along with any logged messages, thus
2028 making the invocation ID useful for matching the online runtime of a
2029 service with the offline log data it generated in a safe way without
2030 relying on synchronized timestamps. In many ways this new service
2031 invocation ID concept is similar to the kernel's boot ID concept that
2032 uniquely and globally identifies the runtime of each boot. The
2033 invocation ID of a service is passed to the service itself via an
2034 environment variable ($INVOCATION_ID). A new bus call
2035 GetUnitByInvocationID() has been added that is similar to GetUnit()
2036 but instead of retrieving the bus path for a unit by its name
2037 retrieves it by its invocation ID. The returned path is valid only as
2038 long as the passed invocation ID is current.
2040 * systemd-resolved gained a new "DNSStubListener" setting in
2041 resolved.conf. It either takes a boolean value or the special values
2042 "udp" and "tcp", and configures whether to enable the stub DNS
2043 listener on 127.0.0.53:53.
2045 * IP addresses configured via networkd may now carry additional
2046 configuration settings supported by the kernel. New options include:
2047 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2048 PrefixRoute=, AutoJoin=.
2050 * The PAM configuration fragment file for "user@.service" shipped with
2051 systemd (i.e. the --user instance of systemd) has been stripped to
2052 the minimum necessary to make the system boot. Previously, it
2053 contained Fedora-specific stanzas that did not apply to other
2054 distributions. It is expected that downstream distributions add
2055 additional configuration lines, matching their needs to this file,
2056 using it only as rough template of what systemd itself needs. Note
2057 that this reduced fragment does not even include an invocation of
2058 pam_limits which most distributions probably want to add, even though
2059 systemd itself does not need it. (There's also the new build time
2060 option --with-pamconfdir=no to disable installation of the PAM
2063 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2064 capability is now also dropped from its set (in addition to
2065 CAP_SYS_MKNOD as before).
2067 * In service unit files it is now possible to connect a specific named
2068 file descriptor with stdin/stdout/stdout of an executed service. The
2069 name may be specified in matching .socket units using the
2070 FileDescriptorName= setting.
2072 * A number of journal settings may now be configured on the kernel
2073 command line. Specifically, the following options are now understood:
2074 systemd.journald.max_level_console=,
2075 systemd.journald.max_level_store=,
2076 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2077 systemd.journald.max_level_wall=.
2079 * "systemctl is-enabled --full" will now show by which symlinks a unit
2080 file is enabled in the unit dependency tree.
2082 * Support for VeraCrypt encrypted partitions has been added to the
2083 "cryptsetup" logic and /etc/crypttab.
2085 * systemd-detect-virt gained support for a new --private-users switch
2086 that checks whether the invoking processes are running inside a user
2087 namespace. Similar, a new special value "private-users" for the
2088 existing ConditionVirtualization= setting has been added, permitting
2089 skipping of specific units in user namespace environments.
2091 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2092 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2093 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2094 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2095 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2096 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2097 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2098 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2099 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2100 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2101 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2102 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2103 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2104 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2105 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2106 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2107 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2108 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2109 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2110 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2111 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2112 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2113 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2114 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2115 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2116 Jędrzejewski-Szmek, Zeal Jagannatha
2118 — Santa Fe, 2016-11-03
2122 * In service units the various ExecXYZ= settings have been extended
2123 with an additional special character as first argument of the
2124 assigned value: if the character '+' is used the specified command
2125 line it will be run with full privileges, regardless of User=,
2126 Group=, CapabilityBoundingSet= and similar options. The effect is
2127 similar to the existing PermissionsStartOnly= option, but allows
2128 configuration of this concept for each executed command line
2131 * Services may now alter the service watchdog timeout at runtime by
2132 sending a WATCHDOG_USEC= message via sd_notify().
2134 * MemoryLimit= and related unit settings now optionally take percentage
2135 specifications. The percentage is taken relative to the amount of
2136 physical memory in the system (or in case of containers, the assigned
2137 amount of memory). This allows scaling service resources neatly with
2138 the amount of RAM available on the system. Similarly, systemd-logind's
2139 RuntimeDirectorySize= option now also optionally takes percentage
2142 * In similar fashion TasksMax= takes percentage values now, too. The
2143 value is taken relative to the configured maximum number of processes
2144 on the system. The per-service task maximum has been changed to 15%
2145 using this functionality. (Effectively this is an increase of 512 →
2146 4915 for service units, given the kernel's default pid_max setting.)
2148 * Calendar time specifications in .timer units now understand a ".."
2149 syntax for time ranges. Example: "4..7:10" may now be used for
2150 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2153 * The InaccessableDirectories=, ReadOnlyDirectories= and
2154 ReadWriteDirectories= unit file settings have been renamed to
2155 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2156 applied to all kinds of file nodes, and not just directories, with
2157 the exception of symlinks. Specifically these settings may now be
2158 used on block and character device nodes, UNIX sockets and FIFOS as
2159 well as regular files. The old names of these settings remain
2160 available for compatibility.
2162 * systemd will now log about all service processes it kills forcibly
2163 (using SIGKILL) because they remained after the clean shutdown phase
2164 of the service completed. This should help identifying services that
2165 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2166 systemd-logind's configuration a similar log message is generated for
2167 processes killed at the end of each session due to this setting.
2169 * systemd will now set the $JOURNAL_STREAM environment variable for all
2170 services whose stdout/stderr are connected to the Journal (which
2171 effectively means by default: all services). The variable contains
2172 the device and inode number of the file descriptor used for
2173 stdout/stderr. This may be used by invoked programs to detect whether
2174 their stdout/stderr is connected to the Journal, in which case they
2175 can switch over to direct Journal communication, thus being able to
2176 pass extended, structured metadata along with their log messages. As
2177 one example, this is now used by glib's logging primitives.
2179 * When using systemd's default tmp.mount unit for /tmp, the mount point
2180 will now be established with the "nosuid" and "nodev" options. This
2181 avoids privilege escalation attacks that put traps and exploits into
2182 /tmp. However, this might cause problems if you e. g. put container
2183 images or overlays into /tmp; if you need this, override tmp.mount's
2184 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2187 * systemd now supports the "memory" cgroup controller also on
2190 * The systemd-cgtop tool now optionally takes a control group path as
2191 command line argument. If specified, the control group list shown is
2192 limited to subgroups of that group.
2194 * The SystemCallFilter= unit file setting gained support for
2195 pre-defined, named system call filter sets. For example
2196 SystemCallFilter=@clock is now an effective way to make all clock
2197 changing-related system calls unavailable to a service. A number of
2198 similar pre-defined groups are defined. Writing system call filters
2199 for system services is simplified substantially with this new
2200 concept. Accordingly, all of systemd's own, long-running services now
2201 enable system call filtering based on this, by default.
2203 * A new service setting MemoryDenyWriteExecute= has been added, taking
2204 a boolean value. If turned on, a service may no longer create memory
2205 mappings that are writable and executable at the same time. This
2206 enhances security for services where this is enabled as it becomes
2207 harder to dynamically write and then execute memory in exploited
2208 service processes. This option has been enabled for all of systemd's
2209 own long-running services.
2211 * A new RestrictRealtime= service setting has been added, taking a
2212 boolean argument. If set the service's processes may no longer
2213 acquire realtime scheduling. This improves security as realtime
2214 scheduling may otherwise be used to easily freeze the system.
2216 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2217 value. This may be used for requesting that the system manager inside
2218 of the container reports start-up completion to nspawn which then
2219 propagates this notification further to the service manager
2220 supervising nspawn itself. A related option NotifyReady= in .nspawn
2221 files has been added too. This functionality allows ordering of the
2222 start-up of multiple containers using the usual systemd ordering
2225 * machinectl gained a new command "stop" that is an alias for
2228 * systemd-resolved gained support for contacting DNS servers on
2229 link-local IPv6 addresses.
2231 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2232 its caches. A method call for requesting the same operation has been
2233 added to the bus API too, and is made available via "systemd-resolve
2236 * systemd-resolve gained a new --status switch. If passed a brief
2237 summary of the used DNS configuration with per-interface information
2240 * resolved.conf gained a new Cache= boolean option, defaulting to
2241 on. If turned off local DNS caching is disabled. This comes with a
2242 performance penalty in particular when DNSSEC is enabled. Note that
2243 resolved disables its internal caching implicitly anyway, when the
2244 configured DNS server is on a host-local IP address such as ::1 or
2245 127.0.0.1, thus automatically avoiding double local caching.
2247 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2248 for DNS requests. This improves compatibility with local programs
2249 that do not use the libc NSS or systemd-resolved's bus APIs for name
2250 resolution. This minimal DNS service is only available to local
2251 programs and does not implement the full DNS protocol, but enough to
2252 cover local DNS clients. A new, static resolv.conf file, listing just
2253 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2254 now recommended to make /etc/resolv.conf a symlink to this file in
2255 order to route all DNS lookups to systemd-resolved, regardless if
2256 done via NSS, the bus API or raw DNS packets. Note that this local
2257 DNS service is not as fully featured as the libc NSS or
2258 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2259 used to deliver link-local address information (as this implies
2260 sending a local interface index along), LLMNR/mDNS support via this
2261 interface is severely restricted. It is thus strongly recommended for
2262 all applications to use the libc NSS API or native systemd-resolved
2265 * systemd-networkd's bridge support learned a new setting
2266 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2267 in .network files has been added for configuring VLAN bridging in
2268 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2270 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2271 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2272 now be acquired without relying on DHCPv6. Two new options
2273 UseDomains= and UseDNS= have been added to configure this behaviour.
2275 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2276 renamed IPv6AcceptRA=, without altering its behaviour. The old
2277 setting name remains available for compatibility reasons.
2279 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2280 Key=, InputKey= and OutputKey=.
2282 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2283 interface configuration.
2285 * "systemctl edit" may now be used to create new unit files by
2286 specifying the --force switch.
2288 * sd-event gained a new function sd_event_get_iteration() for
2289 requesting the current iteration counter of the event loop. It starts
2290 at zero and is increased by one with each event loop iteration.
2292 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2293 file. It can be used in lieu of %systemd_requires in packages which
2294 don't use any systemd functionality and are intended to be installed
2295 in minimal containers without systemd present. This macro provides
2296 ordering dependencies to ensure that if the package is installed in
2297 the same rpm transaction as systemd, systemd will be installed before
2298 the scriptlets for the package are executed, allowing unit presets
2301 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2302 been added to simplify packaging of generators.
2304 * The os-release file gained VERSION_CODENAME field for the
2305 distribution nickname (e.g. VERSION_CODENAME=woody).
2307 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2308 can be set to disable parsing of metadata and the creation
2309 of persistent symlinks for that device.
2311 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2312 to make them available to logged-in users has been reverted.
2314 * Much of the common code of the various systemd components is now
2315 built into an internal shared library libsystemd-shared-231.so
2316 (incorporating the systemd version number in the name, to be updated
2317 with future releases) that the components link to. This should
2318 decrease systemd footprint both in memory during runtime and on
2319 disk. Note that the shared library is not for public use, and is
2320 neither API nor ABI stable, but is likely to change with every new
2321 released update. Packagers need to make sure that binaries
2322 linking to libsystemd-shared.so are updated in step with the
2325 * Configuration for "mkosi" is now part of the systemd
2326 repository. mkosi is a tool to easily build legacy-free OS images,
2327 and is available on github: https://github.com/systemd/mkosi. If
2328 "mkosi" is invoked in the build tree a new raw OS image is generated
2329 incorporating the systemd sources currently being worked on and a
2330 clean, fresh distribution installation. The generated OS image may be
2331 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
2332 UEFI PC. This functionality is particularly useful to easily test
2333 local changes made to systemd in a pristine, defined environment. See
2334 doc/HACKING for details.
2336 * configure learned the --with-support-url= option to specify the
2337 distribution's bugtracker.
2339 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2340 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2341 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2342 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2343 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2344 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2345 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2346 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2347 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2348 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2349 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2350 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2351 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2352 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2353 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2354 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
2355 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2356 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
2357 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
2359 — Berlin, 2016-07-25
2363 * DNSSEC is now turned on by default in systemd-resolved (in
2364 "allow-downgrade" mode), but may be turned off during compile time by
2365 passing "--with-default-dnssec=no" to "configure" (and of course,
2366 during runtime with DNSSEC= in resolved.conf). We recommend
2367 downstreams to leave this on at least during development cycles and
2368 report any issues with the DNSSEC logic upstream. We are very
2369 interested in collecting feedback about the DNSSEC validator and its
2370 limitations in the wild. Note however, that DNSSEC support is
2371 probably nothing downstreams should turn on in stable distros just
2372 yet, as it might create incompatibilities with a few DNS servers and
2373 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2374 automatically whenever we detect such incompatible setups, but there
2375 might be systems we do not cover yet. Hence: please help us testing
2376 the DNSSEC code, leave this on where you can, report back, but then
2377 again don't consider turning this on in your stable, LTS or
2378 production release just yet. (Note that you have to enable
2379 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2380 and its DNSSEC mode for host name resolution from local
2383 * systemd-resolve conveniently resolves DANE records with the --tlsa
2384 option and OPENPGPKEY records with the --openpgp option. It also
2385 supports dumping raw DNS record data via the new --raw= switch.
2387 * systemd-logind will now by default terminate user processes that are
2388 part of the user session scope unit (session-XX.scope) when the user
2389 logs out. This behavior is controlled by the KillUserProcesses=
2390 setting in logind.conf, and the previous default of "no" is now
2391 changed to "yes". This means that user sessions will be properly
2392 cleaned up after, but additional steps are necessary to allow
2393 intentionally long-running processes to survive logout.
2395 While the user is logged in at least once, user@.service is running,
2396 and any service that should survive the end of any individual login
2397 session can be started at a user service or scope using systemd-run.
2398 systemd-run(1) man page has been extended with an example which shows
2399 how to run screen in a scope unit underneath user@.service. The same
2400 command works for tmux.
2402 After the user logs out of all sessions, user@.service will be
2403 terminated too, by default, unless the user has "lingering" enabled.
2404 To effectively allow users to run long-term tasks even if they are
2405 logged out, lingering must be enabled for them. See loginctl(1) for
2406 details. The default polkit policy was modified to allow users to
2407 set lingering for themselves without authentication.
2409 Previous defaults can be restored at compile time by the
2410 --without-kill-user-processes option to "configure".
2412 * systemd-logind gained new configuration settings SessionsMax= and
2413 InhibitorsMax=, both with a default of 8192. It will not register new
2414 user sessions or inhibitors above this limit.
2416 * systemd-logind will now reload configuration on SIGHUP.
2418 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
2419 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
2420 enable. Also, support for the "io" cgroup controller in the unified
2421 hierarchy has been added, so that the "memory", "pids" and "io" are
2422 now the controllers that are supported on the unified hierarchy.
2424 WARNING: it is not possible to use previous systemd versions with
2425 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2426 is necessary to also update systemd in the initramfs if using the
2427 unified hierarchy. An updated SELinux policy is also required.
2429 * LLDP support has been extended, and both passive (receive-only) and
2430 active (sender) modes are supported. Passive mode ("routers-only") is
2431 enabled by default in systemd-networkd. Active LLDP mode is enabled
2432 by default for containers on the internal network. The "networkctl
2433 lldp" command may be used to list information gathered. "networkctl
2434 status" will also show basic LLDP information on connected peers now.
2436 * The IAID and DUID unique identifier sent in DHCP requests may now be
2437 configured for the system and each .network file managed by
2438 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2440 * systemd-networkd gained support for configuring proxy ARP support for
2441 each interface, via the ProxyArp= setting in .network files. It also
2442 gained support for configuring the multicast querier feature of
2443 bridge devices, via the new MulticastQuerier= setting in .netdev
2444 files. Similarly, snooping on the IGMP traffic can be controlled
2445 via the new setting MulticastSnooping=.
2447 A new setting PreferredLifetime= has been added for addresses
2448 configured in .network file to configure the lifetime intended for an
2451 The systemd-networkd DHCP server gained the option EmitRouter=, which
2452 defaults to yes, to configure whether the DHCP Option 3 (Router)
2455 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
2456 systemd-socket-activate and installed into /usr/bin. It is now fully
2459 * systemd-journald now uses separate threads to flush changes to disk
2460 when closing journal files, thus reducing impact of slow disk I/O on
2461 logging performance.
2463 * The sd-journal API gained two new calls
2464 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2465 can be used to open journal files using file descriptors instead of
2466 file or directory paths. sd_journal_open_container() has been
2467 deprecated, sd_journal_open_directory_fd() should be used instead
2468 with the flag SD_JOURNAL_OS_ROOT.
2470 * journalctl learned a new output mode "-o short-unix" that outputs log
2471 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2472 UTC). It also gained support for a new --no-hostname setting to
2473 suppress the hostname column in the family of "short" output modes.
2475 * systemd-ask-password now optionally skips printing of the password to
2476 stdout with --no-output which can be useful in scripts.
2478 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
2479 (devices tagged with ID_MAKER_TOOL) are now tagged with
2480 "uaccess" and are available to logged in users.
2482 * The DeviceAllow= unit setting now supports specifiers (with "%").
2484 * "systemctl show" gained a new --value switch, which allows print a
2485 only the contents of a specific unit property, without also printing
2486 the property's name. Similar support was added to "show*" verbs
2487 of loginctl and machinectl that output "key=value" lists.
2489 * A new unit type "generated" was added for files dynamically generated
2490 by generator tools. Similarly, a new unit type "transient" is used
2491 for unit files created using the runtime API. "systemctl enable" will
2492 refuse to operate on such files.
2494 * A new command "systemctl revert" has been added that may be used to
2495 revert to the vendor version of a unit file, in case local changes
2496 have been made by adding drop-ins or overriding the unit file.
2498 * "machinectl clean" gained a new verb to automatically remove all or
2499 just hidden container images.
2501 * systemd-tmpfiles gained support for a new line type "e" for emptying
2502 directories, if they exist, without creating them if they don't.
2504 * systemd-nspawn gained support for automatically patching the UID/GIDs
2505 of the owners and the ACLs of all files and directories in a
2506 container tree to match the UID/GID user namespacing range selected
2507 for the container invocation. This mode is enabled via the new
2508 --private-users-chown switch. It also gained support for
2509 automatically choosing a free, previously unused UID/GID range when
2510 starting a container, via the new --private-users=pick setting (which
2511 implies --private-users-chown). Together, these options for the first
2512 time make user namespacing for nspawn containers fully automatic and
2513 thus deployable. The systemd-nspawn@.service template unit file has
2514 been changed to use this functionality by default.
2516 * systemd-nspawn gained a new --network-zone= switch, that allows
2517 creating ad-hoc virtual Ethernet links between multiple containers,
2518 that only exist as long as at least one container referencing them is
2519 running. This allows easy connecting of multiple containers with a
2520 common link that implements an Ethernet broadcast domain. Each of
2521 these network "zones" may be named relatively freely by the user, and
2522 may be referenced by any number of containers, but each container may
2523 only reference one of these "zones". On the lower level, this is
2524 implemented by an automatically managed bridge network interface for
2525 each zone, that is created when the first container referencing its
2526 zone is created and removed when the last one referencing its zone
2529 * The default start timeout may now be configured on the kernel command
2530 line via systemd.default_timeout_start_sec=. It was already
2531 configurable via the DefaultTimeoutStartSec= option in
2532 /etc/systemd/system.conf.
2534 * Socket units gained a new TriggerLimitIntervalSec= and
2535 TriggerLimitBurst= setting to configure a limit on the activation
2536 rate of the socket unit.
2538 * The LimitNICE= setting now optionally takes normal UNIX nice values
2539 in addition to the raw integer limit value. If the specified
2540 parameter is prefixed with "+" or "-" and is in the range -20..19 the
2541 value is understood as UNIX nice value. If not prefixed like this it
2542 is understood as raw RLIMIT_NICE limit.
2544 * Note that the effect of the PrivateDevices= unit file setting changed
2545 slightly with this release: the per-device /dev file system will be
2546 mounted read-only from this version on, and will have "noexec"
2547 set. This (minor) change of behavior might cause some (exceptional)
2548 legacy software to break, when PrivateDevices=yes is set for its
2549 service. Please leave PrivateDevices= off if you run into problems
2552 * systemd-bootchart has been split out to a separate repository:
2553 https://github.com/systemd/systemd-bootchart
2555 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
2556 merged into the kernel in its current form.
2558 * The compatibility libraries libsystemd-daemon.so,
2559 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
2560 which have been deprecated since systemd-209 have been removed along
2561 with the corresponding pkg-config files. All symbols provided by
2562 those libraries are provided by libsystemd.so.
2564 * The Capabilities= unit file setting has been removed (it is ignored
2565 for backwards compatibility). AmbientCapabilities= and
2566 CapabilityBoundingSet= should be used instead.
2568 * A new special target has been added, initrd-root-device.target,
2569 which creates a synchronization point for dependencies of the root
2570 device in early userspace. Initramfs builders must ensure that this
2571 target is now included in early userspace.
2573 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
2574 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
2575 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
2576 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
2577 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
2578 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
2579 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
2580 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
2581 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
2582 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
2583 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
2584 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
2585 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
2586 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
2587 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
2588 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
2589 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
2590 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
2591 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
2592 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
2593 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
2594 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
2595 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
2596 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
2597 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
2600 — Fairfax, 2016-05-21
2604 * The systemd-resolved DNS resolver service has gained a substantial
2605 set of new features, most prominently it may now act as a DNSSEC
2606 validating stub resolver. DNSSEC mode is currently turned off by
2607 default, but is expected to be turned on by default in one of the
2608 next releases. For now, we invite everybody to test the DNSSEC logic
2609 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
2610 service also gained a full set of D-Bus interfaces, including calls
2611 to configure DNS and DNSSEC settings per link (for use by external
2612 network management software). systemd-resolved and systemd-networkd
2613 now distinguish between "search" and "routing" domains. The former
2614 are used to qualify single-label names, the latter are used purely
2615 for routing lookups within certain domains to specific links.
2616 resolved now also synthesizes RRs for all entries from /etc/hosts.
2618 * The systemd-resolve tool (which is a client utility for
2619 systemd-resolved) has been improved considerably and is now fully
2620 supported and documented. Hence it has moved from /usr/lib/systemd to
2623 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
2626 * The coredump collection logic has been reworked: when a coredump is
2627 collected it is now written to disk, compressed and processed
2628 (including stacktrace extraction) from a new instantiated service
2629 systemd-coredump@.service, instead of directly from the
2630 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
2631 processing large coredumps can take up a substantial amount of
2632 resources and time, and this previously happened entirely outside of
2633 systemd's service supervision. With the new logic the core_pattern
2634 hook only does minimal metadata collection before passing off control
2635 to the new instantiated service, which is configured with a time
2636 limit, a nice level and other settings to minimize negative impact on
2637 the rest of the system. Also note that the new logic will honour the
2638 RLIMIT_CORE setting of the crashed process, which now allows users
2639 and processes to turn off coredumping for their processes by setting
2642 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
2643 and all forked processes by default. Previously, PID 1 would leave
2644 the setting at "0" for all processes, as set by the kernel. Note that
2645 the resource limit traditionally has no effect on the generated
2646 coredumps on the system if the /proc/sys/kernel/core_pattern hook
2647 logic is used. Since the limit is now honoured (see above) its
2648 default has been changed so that the coredumping logic is enabled by
2649 default for all processes, while allowing specific opt-out.
2651 * When the stacktrace is extracted from processes of system users, this
2652 is now done as "systemd-coredump" user, in order to sandbox this
2653 potentially security sensitive parsing operation. (Note that when
2654 processing coredumps of normal users this is done under the user ID
2655 of process that crashed, as before.) Packagers should take notice
2656 that it is now necessary to create the "systemd-coredump" system user
2657 and group at package installation time.
2659 * The systemd-activate socket activation testing tool gained support
2660 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
2661 and --seqpacket switches. It also has been extended to support both
2662 new-style and inetd-style file descriptor passing. Use the new
2663 --inetd switch to request inetd-style file descriptor passing.
2665 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
2666 variable, which takes a boolean value. If set to false, ANSI color
2667 output is disabled in the tools even when run on a terminal that
2670 * The VXLAN support in networkd now supports two new settings
2671 DestinationPort= and PortRange=.
2673 * A new systemd.machine_id= kernel command line switch has been added,
2674 that may be used to set the machine ID in /etc/machine-id if it is
2675 not initialized yet. This command line option has no effect if the
2676 file is already initialized.
2678 * systemd-nspawn gained a new --as-pid2 switch that invokes any
2679 specified command line as PID 2 rather than PID 1 in the
2680 container. In this mode PID 1 is a minimal stub init process that
2681 implements the special POSIX and Linux semantics of PID 1 regarding
2682 signal and child process management. Note that this stub init process
2683 is implemented in nspawn itself and requires no support from the
2684 container image. This new logic is useful to support running
2685 arbitrary commands in the container, as normal processes are
2686 generally not prepared to run as PID 1.
2688 * systemd-nspawn gained a new --chdir= switch for setting the current
2689 working directory for the process started in the container.
2691 * "journalctl /dev/sda" will now output all kernel log messages for
2692 specified device from the current boot, in addition to all devices
2693 that are parents of it. This should make log output about devices
2694 pretty useful, as long as kernel drivers attach enough metadata to
2695 the log messages. (The usual SATA drivers do.)
2697 * The sd-journal API gained two new calls
2698 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
2699 that report whether log data from /run or /var has been found.
2701 * journalctl gained a new switch "--fields" that prints all journal
2702 record field names currently in use in the journal. This is backed
2703 by two new sd-journal API calls sd_journal_enumerate_fields() and
2704 sd_journal_restart_fields().
2706 * Most configurable timeouts in systemd now expect an argument of
2707 "infinity" to turn them off, instead of "0" as before. The semantics
2708 from now on is that a timeout of "0" means "now", and "infinity"
2709 means "never". To maintain backwards compatibility, "0" continues to
2710 turn off previously existing timeout settings.
2712 * "systemctl reload-or-try-restart" has been renamed to "systemctl
2713 try-reload-or-restart" to clarify what it actually does: the "try"
2714 logic applies to both reloading and restarting, not just restarting.
2715 The old name continues to be accepted for compatibility.
2717 * On boot-up, when PID 1 detects that the system clock is behind the
2718 release date of the systemd version in use, the clock is now set
2719 to the latter. Previously, this was already done in timesyncd, in order
2720 to avoid running with clocks set to the various clock epochs such as
2721 1902, 1938 or 1970. With this change the logic is now done in PID 1
2722 in addition to timesyncd during early boot-up, so that it is enforced
2723 before the first process is spawned by systemd. Note that the logic
2724 in timesyncd remains, as it is more comprehensive and ensures
2725 clock monotonicity by maintaining a persistent timestamp file in
2726 /var. Since /var is generally not available in earliest boot or the
2727 initrd, this part of the logic remains in timesyncd, and is not done
2730 * Support for tweaking details in net_cls.class_id through the
2731 NetClass= configuration directive has been removed, as the kernel
2732 people have decided to deprecate that controller in cgroup v2.
2733 Userspace tools such as nftables are moving over to setting rules
2734 that are specific to the full cgroup path of a task, which obsoletes
2735 these controllers anyway. The NetClass= directive is kept around for
2736 legacy compatibility reasons. For a more in-depth description of the
2737 kernel change, please refer to the respective upstream commit:
2739 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
2741 * A new service setting RuntimeMaxSec= has been added that may be used
2742 to specify a maximum runtime for a service. If the timeout is hit, the
2743 service is terminated and put into a failure state.
2745 * A new service setting AmbientCapabilities= has been added. It allows
2746 configuration of additional Linux process capabilities that are
2747 passed to the activated processes. This is only available on very
2750 * The process resource limit settings in service units may now be used
2751 to configure hard and soft limits individually.
2753 * The various libsystemd APIs such as sd-bus or sd-event now publicly
2754 expose support for gcc's __attribute__((cleanup())) C extension.
2755 Specifically, for many object destructor functions alternative
2756 versions have been added that have names suffixed with "p" and take a
2757 pointer to a pointer to the object to destroy, instead of just a
2758 pointer to the object itself. This is useful because these destructor
2759 functions may be used directly as parameters to the cleanup
2760 construct. Internally, systemd has been a heavy user of this GCC
2761 extension for a long time, and with this change similar support is
2762 now available to consumers of the library outside of systemd. Note
2763 that by using this extension in your sources compatibility with old
2764 and strictly ANSI compatible C compilers is lost. However, all gcc or
2765 LLVM versions of recent years support this extension.
2767 * Timer units gained support for a new setting RandomizedDelaySec= that
2768 allows configuring some additional randomized delay to the configured
2769 time. This is useful to spread out timer events to avoid load peaks in
2770 clusters or larger setups.
2772 * Calendar time specifications now support sub-second accuracy.
2774 * Socket units now support listening on SCTP and UDP-lite protocol
2777 * The sd-event API now comes with a full set of man pages.
2779 * Older versions of systemd contained experimental support for
2780 compressing journal files and coredumps with the LZ4 compressor that
2781 was not compatible with the lz4 binary (due to API limitations of the
2782 lz4 library). This support has been removed; only support for files
2783 compatible with the lz4 binary remains. This LZ4 logic is now
2784 officially supported and no longer considered experimental.
2786 * The dkr image import logic has been removed again from importd. dkr's
2787 micro-services focus doesn't fit into the machine image focus of
2788 importd, and quickly got out of date with the upstream dkr API.
2790 * Creation of the /run/lock/lockdev/ directory was dropped from
2791 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
2792 been available for many years. If you still need this, you need to
2793 create your own tmpfiles.d config file with:
2795 d /run/lock/lockdev 0775 root lock -
2797 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
2798 and RebootArgument= have been moved from the [Service] section of
2799 unit files to [Unit], and they are now supported on all unit types,
2800 not just service units. Of course, elogind will continue to
2801 understand these settings also at the old location, in order to
2802 maintain compatibility.
2804 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
2805 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
2806 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
2807 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
2808 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
2809 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
2810 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
2811 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
2812 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
2813 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
2814 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
2815 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
2816 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
2817 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
2818 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
2819 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
2820 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
2821 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
2822 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2824 — Berlin, 2016-02-11
2828 * A number of properties previously only settable in unit
2829 files are now also available as properties to set when
2830 creating transient units programmatically via the bus, as it
2831 is exposed with systemd-run's --property=
2832 setting. Specifically, these are: SyslogIdentifier=,
2833 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
2834 EnvironmentFile=, ReadWriteDirectories=,
2835 ReadOnlyDirectories=, InaccessibleDirectories=,
2836 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
2838 * When creating transient services via the bus API it is now
2839 possible to pass in a set of file descriptors to use as
2840 STDIN/STDOUT/STDERR for the invoked process.
2842 * Slice units may now be created transiently via the bus APIs,
2843 similar to the way service and scope units may already be
2844 created transiently.
2846 * Wherever systemd expects a calendar timestamp specification
2847 (like in journalctl's --since= and --until= switches) UTC
2848 timestamps are now supported. Timestamps suffixed with "UTC"
2849 are now considered to be in Universal Time Coordinated
2850 instead of the local timezone. Also, timestamps may now
2851 optionally be specified with sub-second accuracy. Both of
2852 these additions also apply to recurring calendar event
2853 specification, such as OnCalendar= in timer units.
2855 * journalctl gained a new "--sync" switch that asks the
2856 journal daemon to write all so far unwritten log messages to
2857 disk and sync the files, before returning.
2859 * systemd-tmpfiles learned two new line types "q" and "Q" that
2860 operate like "v", but also set up a basic btrfs quota
2861 hierarchy when used on a btrfs file system with quota
2864 * tmpfiles' "v", "q" and "Q" will now create a plain directory
2865 instead of a subvolume (even on a btrfs file system) if the
2866 root directory is a plain directory, and not a
2867 subvolume. This should simplify things with certain chroot()
2868 environments which are not aware of the concept of btrfs
2871 * systemd-detect-virt gained a new --chroot switch to detect
2872 whether execution takes place in a chroot() environment.
2874 * CPUAffinity= now takes CPU index ranges in addition to
2877 * The various memory-related resource limit settings (such as
2878 LimitAS=) now understand the usual K, M, G, ... suffixes to
2879 the base of 1024 (IEC). Similar, the time-related resource
2880 limit settings understand the usual min, h, day, ...
2883 * There's a new system.conf setting DefaultTasksMax= to
2884 control the default TasksMax= setting for services and
2885 scopes running on the system. (TasksMax= is the primary
2886 setting that exposes the "pids" cgroup controller on systemd
2887 and was introduced in the previous systemd release.) The
2888 setting now defaults to 512, which means services that are
2889 not explicitly configured otherwise will only be able to
2890 create 512 processes or threads at maximum, from this
2891 version on. Note that this means that thread- or
2892 process-heavy services might need to be reconfigured to set
2893 TasksMax= to a higher value. It is sufficient to set
2894 TasksMax= in these specific unit files to a higher value, or
2895 even "infinity". Similar, there's now a logind.conf setting
2896 UserTasksMax= that defaults to 4096 and limits the total
2897 number of processes or tasks each user may own
2898 concurrently. nspawn containers also have the TasksMax=
2899 value set by default now, to 8192. Note that all of this
2900 only has an effect if the "pids" cgroup controller is
2901 enabled in the kernel. The general benefit of these changes
2902 should be a more robust and safer system, that provides a
2903 certain amount of per-service fork() bomb protection.
2905 * systemd-nspawn gained the new --network-veth-extra= switch
2906 to define additional and arbitrarily-named virtual Ethernet
2907 links between the host and the container.
2909 * A new service execution setting PassEnvironment= has been
2910 added that allows importing select environment variables
2911 from PID1's environment block into the environment block of
2914 * Timer units gained support for a new RemainAfterElapse=
2915 setting which takes a boolean argument. It defaults to on,
2916 exposing behaviour unchanged to previous releases. If set to
2917 off, timer units are unloaded after they elapsed if they
2918 cannot elapse again. This is particularly useful for
2919 transient timer units, which shall not stay around longer
2920 than until they first elapse.
2922 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
2923 default now (the kernel default is 16). This is beneficial
2924 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
2925 allows substantially larger numbers of queued
2926 datagrams. This should increase the capability of systemd to
2927 parallelize boot-up, as logging and sd_notify() are unlikely
2928 to stall execution anymore. If you need to change the value
2929 from the new defaults, use the usual sysctl.d/ snippets.
2931 * The compression framing format used by the journal or
2932 coredump processing has changed to be in line with what the
2933 official LZ4 tools generate. LZ4 compression support in
2934 systemd was considered unsupported previously, as the format
2935 was not compatible with the normal tools. With this release
2936 this has changed now, and it is hence safe for downstream
2937 distributions to turn it on. While not compressing as well
2938 as the XZ, LZ4 is substantially faster, which makes
2939 it a good default choice for the compression logic in the
2940 journal and in coredump handling.
2942 * Any reference to /etc/mtab has been dropped from
2943 systemd. The file has been obsolete since a while, but
2944 systemd refused to work on systems where it was incorrectly
2945 set up (it should be a symlink or non-existent). Please make
2946 sure to update to util-linux 2.27.1 or newer in conjunction
2947 with this systemd release, which also drops any reference to
2948 /etc/mtab. If you maintain a distribution make sure that no
2949 software you package still references it, as this is a
2950 likely source of bugs. There's also a glibc bug pending,
2951 asking for removal of any reference to this obsolete file:
2953 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
2955 Note that only util-linux versions built with
2956 --enable-libmount-force-mountinfo are supported.
2958 * Support for the ".snapshot" unit type has been removed. This
2959 feature turned out to be little useful and little used, and
2960 has now been removed from the core and from systemctl.
2962 * The dependency types RequiresOverridable= and
2963 RequisiteOverridable= have been removed from systemd. They
2964 have been used only very sparingly to our knowledge and
2965 other options that provide a similar effect (such as
2966 systemctl --mode=ignore-dependencies) are much more useful
2967 and commonly used. Moreover, they were only half-way
2968 implemented as the option to control behaviour regarding
2969 these dependencies was never added to systemctl. By removing
2970 these dependency types the execution engine becomes a bit
2971 simpler. Unit files that use these dependencies should be
2972 changed to use the non-Overridable dependency types
2973 instead. In fact, when parsing unit files with these
2974 options, that's what systemd will automatically convert them
2975 too, but it will also warn, asking users to fix the unit
2976 files accordingly. Removal of these dependency types should
2977 only affect a negligible number of unit files in the wild.
2979 * Behaviour of networkd's IPForward= option changed
2980 (again). It will no longer maintain a per-interface setting,
2981 but propagate one way from interfaces where this is enabled
2982 to the global kernel setting. The global setting will be
2983 enabled when requested by a network that is set up, but
2984 never be disabled again. This change was made to make sure
2985 IPv4 and IPv6 behaviour regarding packet forwarding is
2986 similar (as the Linux IPv6 stack does not support
2987 per-interface control of this setting) and to minimize
2990 * In unit files the behaviour of %u, %U, %h, %s has
2991 changed. These specifiers will now unconditionally resolve
2992 to the various user database fields of the user that the
2993 systemd instance is running as, instead of the user
2994 configured in the specific unit via User=. Note that this
2995 effectively doesn't change much, as resolving of these
2996 specifiers was already turned off in the --system instance
2997 of systemd, as we cannot do NSS lookups from PID 1. In the
2998 --user instance of systemd these specifiers where correctly
2999 resolved, but hardly made any sense, since the user instance
3000 lacks privileges to do user switches anyway, and User= is
3001 hence useless. Moreover, even in the --user instance of
3002 systemd behaviour was awkward as it would only take settings
3003 from User= assignment placed before the specifier into
3004 account. In order to unify and simplify the logic around
3005 this the specifiers will now always resolve to the
3006 credentials of the user invoking the manager (which in case
3007 of PID 1 is the root user).
3009 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
3010 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
3011 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
3012 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3013 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3014 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3015 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3016 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3017 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3018 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3019 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3020 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3021 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3022 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3025 — Berlin, 2015-11-18
3029 * systemd now depends on util-linux v2.27. More specifically,
3030 the newly added mount monitor feature in libmount now
3031 replaces systemd's former own implementation.
3033 * libmount mandates /etc/mtab not to be regular file, and
3034 systemd now enforces this condition at early boot.
3035 /etc/mtab has been deprecated and warned about for a very
3036 long time, so systems running systemd should already have
3037 stopped having this file around as anything else than a
3038 symlink to /proc/self/mounts.
3040 * Support for the "pids" cgroup controller has been added. It
3041 allows accounting the number of tasks in a cgroup and
3042 enforcing limits on it. This adds two new setting
3043 TasksAccounting= and TasksMax= to each unit, as well as a
3044 global option DefaultTasksAccounting=.
3046 * Support for the "net_cls" cgroup controller has been added.
3047 It allows assigning a net class ID to each task in the
3048 cgroup, which can then be used in firewall rules and traffic
3049 shaping configurations. Note that the kernel netfilter net
3050 class code does not currently work reliably for ingress
3051 packets on unestablished sockets.
3053 This adds a new config directive called NetClass= to CGroup
3054 enabled units. Allowed values are positive numbers for fixed
3055 assignments and "auto" for picking a free value
3058 * 'systemctl is-system-running' now returns 'offline' if the
3059 system is not booted with systemd. This command can now be
3060 used as a substitute for 'systemd-notify --booted'.
3062 * Watchdog timeouts have been increased to 3 minutes for all
3063 in-tree service files. Apparently, disk IO issues are more
3064 frequent than we hoped, and user reported >1 minute waiting
3067 * 'machine-id-commit' functionality has been merged into
3068 'machine-id-setup --commit'. The separate binary has been
3071 * The WorkingDirectory= directive in unit files may now be set
3072 to the special value '~'. In this case, the working
3073 directory is set to the home directory of the user
3074 configured in User=.
3076 * "machinectl shell" will now open the shell in the home
3077 directory of the selected user by default.
3079 * The CrashChVT= configuration file setting is renamed to
3080 CrashChangeVT=, following our usual logic of not
3081 abbreviating unnecessarily. The old directive is still
3082 supported for compat reasons. Also, this directive now takes
3083 an integer value between 1 and 63, or a boolean value. The
3084 formerly supported '-1' value for disabling stays around for
3087 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
3088 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
3089 RootDirectory= properties can now be set for transient
3092 * The systemd-analyze tool gained a new "set-log-target" verb
3093 to change the logging target the system manager logs to
3094 dynamically during runtime. This is similar to how
3095 "systemd-analyze set-log-level" already changes the log
3098 * In nspawn /sys is now mounted as tmpfs, with only a selected
3099 set of subdirectories mounted in from the real sysfs. This
3100 enhances security slightly, and is useful for ensuring user
3101 namespaces work correctly.
3103 * Support for USB FunctionFS activation has been added. This
3104 allows implementation of USB gadget services that are
3105 activated as soon as they are requested, so that they don't
3106 have to run continuously, similar to classic socket
3109 * The "systemctl exit" command now optionally takes an
3110 additional parameter that sets the exit code to return from
3111 the systemd manager when exiting. This is only relevant when
3112 running the systemd user instance, or when running the
3113 system instance in a container.
3115 * sd-bus gained the new API calls sd_bus_path_encode_many()
3116 and sd_bus_path_decode_many() that allow easy encoding and
3117 decoding of multiple identifier strings inside a D-Bus
3118 object path. Another new call sd_bus_default_flush_close()
3119 has been added to flush and close per-thread default
3122 * systemd-cgtop gained support for a -M/--machine= switch to
3123 show the control groups within a certain container only.
3125 * "systemctl kill" gained support for an optional --fail
3126 switch. If specified the requested operation will fail of no
3127 processes have been killed, because the unit had no
3128 processes attached, or similar.
3130 * A new systemd.crash_reboot=1 kernel command line option has
3131 been added that triggers a reboot after crashing. This can
3132 also be set through CrashReboot= in systemd.conf.
3134 * The RuntimeDirectory= setting now understands unit
3135 specifiers like %i or %f.
3137 * A new (still internal) library API sd-ipv4acd has been added,
3138 that implements address conflict detection for IPv4. It's
3139 based on code from sd-ipv4ll, and will be useful for
3140 detecting DHCP address conflicts.
3142 * File descriptors passed during socket activation may now be
3143 named. A new API sd_listen_fds_with_names() is added to
3144 access the names. The default names may be overridden,
3145 either in the .socket file using the FileDescriptorName=
3146 parameter, or by passing FDNAME= when storing the file
3147 descriptors using sd_notify().
3149 * systemd-networkd gained support for:
3151 - Setting the IPv6 Router Advertisement settings via
3152 IPv6AcceptRouterAdvertisements= in .network files.
3154 - Configuring the HelloTimeSec=, MaxAgeSec= and
3155 ForwardDelaySec= bridge parameters in .netdev files.
3157 - Configuring PreferredSource= for static routes in
3160 * The "ask-password" framework used to query for LUKS harddisk
3161 passwords or SSL passwords during boot gained support for
3162 caching passwords in the kernel keyring, if it is
3163 available. This makes sure that the user only has to type in
3164 a passphrase once if there are multiple objects to unlock
3165 with the same one. Previously, such password caching was
3166 available only when Plymouth was used; this moves the
3167 caching logic into the systemd codebase itself. The
3168 "systemd-ask-password" utility gained a new --keyname=
3169 switch to control which kernel keyring key to use for
3170 caching a password in. This functionality is also useful for
3171 enabling display managers such as gdm to automatically
3172 unlock the user's GNOME keyring if its passphrase, the
3173 user's password and the harddisk password are the same, if
3174 gdm-autologin is used.
3176 * When downloading tar or raw images using "machinectl
3177 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3178 file is now also downloaded, if it is available and stored
3179 next to the image file.
3181 * Units of type ".socket" gained a new boolean setting
3182 Writable= which is only useful in conjunction with
3183 ListenSpecial=. If true, enables opening the specified
3184 special file in O_RDWR mode rather than O_RDONLY mode.
3186 * systemd-rfkill has been reworked to become a singleton
3187 service that is activated through /dev/rfkill on each rfkill
3188 state change and saves the settings to disk. This way,
3189 systemd-rfkill is now compatible with devices that exist
3190 only intermittendly, and even restores state if the previous
3191 system shutdown was abrupt rather than clean.
3193 * The journal daemon gained support for vacuuming old journal
3194 files controlled by the number of files that shall remain,
3195 in addition to the already existing control by size and by
3196 date. This is useful as journal interleaving performance
3197 degrades with too many separate journal files, and allows
3198 putting an effective limit on them. The new setting defaults
3199 to 100, but this may be changed by setting SystemMaxFiles=
3200 and RuntimeMaxFiles= in journald.conf. Also, the
3201 "journalctl" tool gained the new --vacuum-files= switch to
3202 manually vacuum journal files to leave only the specified
3203 number of files in place.
3205 * udev will now create /dev/disk/by-path links for ATA devices
3206 on kernels where that is supported.
3208 * Galician, Serbian, Turkish and Korean translations were added.
3210 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3211 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3212 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3213 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3214 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3215 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3216 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3217 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3218 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3219 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3220 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3221 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3222 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3223 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3224 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3225 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3226 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3227 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3229 — Berlin, 2015-10-07
3233 * The DHCP implementation of systemd-networkd gained a set of
3236 - The DHCP server now supports emitting DNS and NTP
3237 information. It may be enabled and configured via
3238 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3239 and NTP information is enabled, but no servers are
3240 configured, the corresponding uplink information (if there
3241 is any) is propagated.
3243 - Server and client now support transmission and reception
3244 of timezone information. It can be configured via the
3245 newly introduced network options UseTimezone=,
3246 EmitTimezone=, and Timezone=. Transmission of timezone
3247 information is enabled between host and containers by
3248 default now: the container will change its local timezone
3249 to what the host has set.
3251 - Lease timeouts can now be configured via
3252 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3254 - The DHCP server improved on the stability of
3255 leases. Clients are more likely to get the same lease
3256 information back, even if the server loses state.
3258 - The DHCP server supports two new configuration options to
3259 control the lease address pool metrics, PoolOffset= and
3262 * The encapsulation limit of tunnels in systemd-networkd may
3263 now be configured via 'EncapsulationLimit='. It allows
3264 modifying the maximum additional levels of encapsulation
3265 that are permitted to be prepended to a packet.
3267 * systemd now supports the concept of user buses replacing
3268 session buses, if used with dbus-1.10 (and enabled via dbus
3269 --enable-user-session). It previously only supported this on
3270 kdbus-enabled systems, and this release expands this to
3271 'dbus-daemon' systems.
3273 * systemd-networkd now supports predictable interface names
3276 * systemd now optionally supports the new Linux kernel
3277 "unified" control group hierarchy. If enabled via the kernel
3278 command-line option 'systemd.unified_cgroup_hierarchy=1',
3279 systemd will try to mount the unified cgroup hierarchy
3280 directly on /sys/fs/cgroup. If not enabled, or not
3281 available, systemd will fall back to the legacy cgroup
3282 hierarchy setup, as before. Host system and containers can
3283 mix and match legacy and unified hierarchies as they
3284 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
3285 environment variable to individually select the hierarchy to
3286 use for executed containers. By default, nspawn will use the
3287 unified hierarchy for the containers if the host uses the
3288 unified hierarchy, and the legacy hierarchy otherwise.
3289 Please note that at this point the unified hierarchy is an
3290 experimental kernel feature and is likely to change in one
3291 of the next kernel releases. Therefore, it should not be
3292 enabled by default in downstream distributions yet. The
3293 minimum required kernel version for the unified hierarchy to
3294 work is 4.2. Note that when the unified hierarchy is used
3295 for the first time delegated access to controllers is
3296 safe. Because of this systemd-nspawn containers will get
3297 access to controllers now, as will systemd user
3298 sessions. This means containers and user sessions may now
3299 manage their own resources, partitioning up what the system
3302 * A new special scope unit "init.scope" has been introduced
3303 that encapsulates PID 1 of the system. It may be used to
3304 determine resource usage and enforce resource limits on PID
3305 1 itself. PID 1 hence moved out of the root of the control
3308 * The cgtop tool gained support for filtering out kernel
3309 threads when counting tasks in a control group. Also, the
3310 count of processes is now recursively summed up by
3311 default. Two options -k and --recursive= have been added to
3312 revert to old behaviour. The tool has also been updated to
3313 work correctly in containers now.
3315 * systemd-nspawn's --bind= and --bind-ro= options have been
3316 extended to allow creation of non-recursive bind mounts.
3318 * libsystemd gained two new calls sd_pid_get_cgroup() and
3319 sd_peer_get_cgroup() which return the control group path of
3320 a process or peer of a connected AF_UNIX socket. This
3321 function call is particularly useful when implementing
3322 delegated subtrees support in the control group hierarchy.
3324 * The "sd-event" event loop API of libsystemd now supports
3325 correct dequeuing of real-time signals, without losing
3328 * When systemd requests a PolicyKit decision when managing
3329 units it will now add additional fields to the request,
3330 including unit name and desired operation. This enables more
3331 powerful PolicyKit policies, that make decisions depending
3332 on these parameters.
3334 * nspawn learnt support for .nspawn settings files, that may
3335 accompany the image files or directories of containers, and
3336 may contain additional settings for the container. This is
3337 an alternative to configuring container parameters via the
3338 nspawn command line.
3340 Contributions from: Cristian Rodríguez, Daniel Mack, David
3341 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3342 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3343 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3344 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3345 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3346 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
3347 Andersen, Tom Gundersen, Torstein Husebø
3349 — Berlin, 2015-09-08
3353 * machinectl gained a new verb 'shell' which opens a fresh
3354 shell on the target container or the host. It is similar to
3355 the existing 'login' command of machinectl, but spawns the
3356 shell directly without prompting for username or
3357 password. The pseudo machine '.host' now refers to the local
3358 host and is used by default. Hence, 'machinectl shell' can
3359 be used as replacement for 'su -' which spawns a session as
3360 a fresh systemd unit in a way that is fully isolated from
3361 the originating session.
3363 * systemd-networkd learned to cope with private-zone DHCP
3364 options and allows other programs to query the values.
3366 * SELinux access control when enabling/disabling units is no
3367 longer enforced with this release. The previous
3368 implementation was incorrect, and a new corrected
3369 implementation is not yet available. As unit file operations
3370 are still protected via PolicyKit and D-Bus policy this is
3371 not a security problem. Yet, distributions which care about
3372 optimal SELinux support should probably not stabilize on
3375 * sd-bus gained support for matches of type "arg0has=", that
3376 test for membership of strings in string arrays sent in bus
3379 * systemd-resolved now dumps the contents of its DNS and LLMNR
3380 caches to the logs on reception of the SIGUSR1 signal. This
3381 is useful to debug DNS behaviour.
3383 * The coredumpctl tool gained a new --directory= option to
3384 operate on journal files in a specific directory.
3386 * "systemctl reboot" and related commands gained a new
3387 "--message=" option which may be used to set a free-text
3388 wall message when shutting down or rebooting the
3389 system. This message is also logged, which is useful for
3390 figuring out the reason for a reboot or shutdown a
3393 * The "systemd-resolve-host" tool's -i switch now takes
3394 network interface numbers as alternative to interface names.
3396 * A new unit file setting for services has been introduced:
3397 UtmpMode= allows configuration of how precisely systemd
3398 handles utmp and wtmp entries for the service if this is
3399 enabled. This allows writing services that appear similar to
3400 user sessions in the output of the "w", "who", "last" and
3403 * systemd-resolved will now locally synthesize DNS resource
3404 records for the "localhost" and "gateway" domains as well as
3405 the local hostname. This should ensure that clients querying
3406 RRs via resolved will get similar results as those going via
3407 NSS, if nss-myhostname is enabled.
3409 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3410 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3411 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3412 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3413 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3414 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3415 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3416 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3417 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3418 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3419 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3420 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3422 — Berlin, 2015-08-27
3426 * The systemd-efi-boot-generator functionality was merged into
3427 systemd-gpt-auto-generator.
3429 * systemd-networkd now supports Group Policy for vxlan
3430 devices. It can be enabled via the new boolean configuration
3431 option called 'GroupPolicyExtension='.
3433 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3434 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3435 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3437 — Berlin, 2015-07-31
3441 * The python-systemd code has been removed from the systemd repository.
3442 A new repository has been created which accommodates the code from
3443 now on, and we kindly ask distributions to create a separate package
3444 for this: https://github.com/systemd/python-systemd
3446 * The systemd daemon will now reload its main configuration
3447 (/etc/systemd/system.conf) on daemon-reload.
3449 * sd-dhcp now exposes vendor specific extensions via
3450 sd_dhcp_lease_get_vendor_specific().
3452 * systemd-networkd gained a number of new configuration options.
3454 - A new boolean configuration option for TAP devices called
3455 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
3456 device, thus allowing to send and receive GSO packets.
3458 - A new tunnel configuration option called 'CopyDSCP='.
3459 If enabled, the DSCP field of ip6 tunnels is copied into the
3460 decapsulated packet.
3462 - A set of boolean bridge configuration options were added.
3463 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3464 and 'UnicastFlood=' are now parsed by networkd and applied to the
3465 respective bridge link device via the respective IFLA_BRPORT_*
3468 - A new string configuration option to override the hostname sent
3469 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3470 is true, networkd will use the configured hostname instead of the
3471 system hostname when sending DHCP requests.
3473 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3474 networkd will configure the IPv6 flow-label of the tunnel device
3475 according to RFC2460.
3477 - The 'macvtap' virtual network devices are now supported, similar to
3478 the already supported 'macvlan' devices.
3480 * systemd-resolved now implements RFC5452 to improve resilience against
3481 cache poisoning. Additionally, source port randomization is enabled
3482 by default to further protect against DNS spoofing attacks.
3484 * nss-mymachines now supports translating UIDs and GIDs of running
3485 containers with user-namespaces enabled. If a container 'foo'
3486 translates a host uid 'UID' to the container uid 'TUID', then
3487 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
3488 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
3489 mapped as 'vg-foo-TGID'.
3491 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
3492 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
3493 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
3494 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
3495 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
3496 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
3497 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
3498 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
3499 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
3500 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
3502 — Berlin, 2015-07-29
3506 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
3507 There are no known issues with current sysfs, and udev does not need
3508 or should be used to work around such bugs.
3510 * udev does no longer enable USB HID power management. Several reports
3511 indicate, that some devices cannot handle that setting.
3513 * The udev accelerometer helper was removed. The functionality
3514 is now fully included in iio-sensor-proxy. But this means,
3515 older iio-sensor-proxy versions will no longer provide
3516 accelerometer/orientation data with this systemd version.
3517 Please upgrade iio-sensor-proxy to version 1.0.
3519 * networkd gained a new configuration option IPv6PrivacyExtensions=
3520 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
3521 for Stateless Address") on selected networks.
3523 * For the sake of fewer build-time dependencies and less code in the
3524 main repository, the python bindings are about to be removed in the
3525 next release. A new repository has been created which accommodates
3526 the code from now on, and we kindly ask distributions to create a
3527 separate package for this. The removal will take place in v223.
3529 https://github.com/systemd/python-systemd
3531 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
3532 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
3533 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
3534 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
3535 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
3536 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
3537 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
3538 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
3539 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3540 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
3542 — Berlin, 2015-07-07
3546 * The sd-bus.h and sd-event.h APIs have now been declared
3547 stable and have been added to the official interface of
3548 libsystemd.so. sd-bus implements an alternative D-Bus client
3549 library, that is relatively easy to use, very efficient and
3550 supports both classic D-Bus as well as kdbus as transport
3551 backend. sd-event is a generic event loop abstraction that
3552 is built around Linux epoll, but adds features such as event
3553 prioritization or efficient timer handling. Both APIs are good
3554 choices for C programs looking for a bus and/or event loop
3555 implementation that is minimal and does not have to be
3556 portable to other kernels.
3558 * kdbus support is no longer compile-time optional. It is now
3559 always built-in. However, it can still be disabled at
3560 runtime using the kdbus=0 kernel command line setting, and
3561 that setting may be changed to default to off, by specifying
3562 --disable-kdbus at build-time. Note though that the kernel
3563 command line setting has no effect if the kdbus.ko kernel
3564 module is not installed, in which case kdbus is (obviously)
3565 also disabled. We encourage all downstream distributions to
3566 begin testing kdbus by adding it to the kernel images in the
3567 development distributions, and leaving kdbus support in
3570 * The minimal required util-linux version has been bumped to
3573 * Support for chkconfig (--enable-chkconfig) was removed in
3574 favor of calling an abstraction tool
3575 /lib/systemd/systemd-sysv-install. This needs to be
3576 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
3577 in README for details.
3579 * If there's a systemd unit and a SysV init script for the
3580 same service name, and the user executes "systemctl enable"
3581 for it (or a related call), then this will now enable both
3582 (or execute the related operation on both), not just the
3585 * The libudev API documentation has been converted from gtkdoc
3588 * gudev has been removed from the systemd tree, it is now an
3591 * The systemd-cgtop tool learnt a new --raw switch to generate
3592 "raw" (machine parsable) output.
3594 * networkd's IPForwarding= .network file setting learnt the
3595 new setting "kernel", which ensures that networkd does not
3596 change the IP forwarding sysctl from the default kernel
3599 * The systemd-logind bus API now exposes a new boolean
3600 property "Docked" that reports whether logind considers the
3601 system "docked", i.e. connected to a docking station or not.
3603 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
3604 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
3605 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
3606 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
3607 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
3608 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
3609 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
3610 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
3611 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
3612 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
3613 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
3614 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
3615 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
3616 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
3617 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
3618 Fink, Zbigniew Jędrzejewski-Szmek
3620 — Berlin, 2015-06-19
3624 * The gudev library has been extracted into a separate repository
3625 available at: https://git.gnome.org/browse/libgudev/
3626 It is now managed as part of the Gnome project. Distributions
3627 are recommended to pass --disable-gudev to systemd and use
3628 gudev from the Gnome project instead. gudev is still included
3629 in systemd, for now. It will be removed soon, though. Please
3630 also see the announcement-thread on systemd-devel:
3631 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
3633 * systemd now exposes a CPUUsageNSec= property for each
3634 service unit on the bus, that contains the overall consumed
3635 CPU time of a service (the sum of what each process of the
3636 service consumed). This value is only available if
3637 CPUAccounting= is turned on for a service, and is then shown
3638 in the "systemctl status" output.
3640 * Support for configuring alternative mappings of the old SysV
3641 runlevels to systemd targets has been removed. They are now
3642 hardcoded in a way that runlevels 2, 3, 4 all map to
3643 multi-user.target and 5 to graphical.target (which
3644 previously was already the default behaviour).
3646 * The auto-mounter logic gained support for mount point
3647 expiry, using a new TimeoutIdleSec= setting in .automount
3648 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
3650 * The EFI System Partition (ESP) as mounted to /boot by
3651 systemd-efi-boot-generator will now be unmounted
3652 automatically after 2 minutes of not being used. This should
3653 minimize the risk of ESP corruptions.
3655 * New /etc/fstab options x-systemd.requires= and
3656 x-systemd.requires-mounts-for= are now supported to express
3657 additional dependencies for mounts. This is useful for
3658 journalling file systems that support external journal
3659 devices or overlay file systems that require underlying file
3660 systems to be mounted.
3662 * systemd does not support direct live-upgrades (via systemctl
3663 daemon-reexec) from versions older than v44 anymore. As no
3664 distribution we are aware of shipped such old versions in a
3665 stable release this should not be problematic.
3667 * When systemd forks off a new per-connection service instance
3668 it will now set the $REMOTE_ADDR environment variable to the
3669 remote IP address, and $REMOTE_PORT environment variable to
3670 the remote IP port. This behaviour is similar to the
3671 corresponding environment variables defined by CGI.
3673 * systemd-networkd gained support for uplink failure
3674 detection. The BindCarrier= option allows binding interface
3675 configuration dynamically to the link sense of other
3676 interfaces. This is useful to achieve behaviour like in
3679 * systemd-networkd gained support for configuring the DHCP
3680 client identifier to use when requesting leases.
3682 * systemd-networkd now has a per-network UseNTP= option to
3683 configure whether NTP server information acquired via DHCP
3684 is passed on to services like systemd-timesyncd.
3686 * systemd-networkd gained support for vti6 tunnels.
3688 * Note that systemd-networkd manages the sysctl variable
3689 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
3690 it is configured for since v219. The variable controls IP
3691 forwarding, and is a per-interface alternative to the global
3692 /proc/sys/net/ipv[46]/ip_forward. This setting is
3693 configurable in the IPForward= option, which defaults to
3694 "no". This means if networkd is used for an interface it is
3695 no longer sufficient to set the global sysctl option to turn
3696 on IP forwarding! Instead, the .network file option
3697 IPForward= needs to be turned on! Note that the
3698 implementation of this behaviour was broken in v219 and has
3701 * Many bonding and vxlan options are now configurable in
3704 * systemd-nspawn gained a new --property= setting to set unit
3705 properties for the container scope. This is useful for
3706 setting resource parameters (e.g. "CPUShares=500") on
3707 containers started from the command line.
3709 * systemd-nspawn gained a new --private-users= switch to make
3710 use of user namespacing available on recent Linux kernels.
3712 * systemd-nspawn may now be called as part of a shell pipeline
3713 in which case the pipes used for stdin and stdout are passed
3714 directly to the process invoked in the container, without
3715 indirection via a pseudo tty.
3717 * systemd-nspawn gained a new switch to control the UNIX
3718 signal to use when killing the init process of the container
3721 * systemd-nspawn gained a new --overlay= switch for mounting
3722 overlay file systems into the container using the new kernel
3725 * When a container image is imported via systemd-importd and
3726 the host file system is not btrfs, a loopback block device
3727 file is created in /var/lib/machines.raw with a btrfs file
3728 system inside. It is then mounted to /var/lib/machines to
3729 enable btrfs features for container management. The loopback
3730 file and btrfs file system is grown as needed when container
3731 images are imported via systemd-importd.
3733 * systemd-machined/systemd-importd gained support for btrfs
3734 quota, to enforce container disk space limits on disk. This
3735 is exposed in "machinectl set-limit".
3737 * systemd-importd now can import containers from local .tar,
3738 .raw and .qcow2 images, and export them to .tar and .raw. It
3739 can also import dkr v2 images now from the network (on top
3742 * systemd-importd gained support for verifying downloaded
3743 images with gpg2 (previously only gpg1 was supported).
3745 * systemd-machined, systemd-logind, systemd: most bus calls
3746 are now accessible to unprivileged processes via
3747 PolicyKit. Also, systemd-logind will now allow users to kill
3748 their own sessions without further privileges or
3751 * systemd-shutdownd has been removed. This service was
3752 previously responsible for implementing scheduled shutdowns
3753 as exposed in /usr/bin/shutdown's time parameter. This
3754 functionality has now been moved into systemd-logind and is
3755 accessible via a bus interface.
3757 * "systemctl reboot" gained a new switch --firmware-setup that
3758 can be used to reboot into the EFI firmware setup, if that
3759 is available. systemd-logind now exposes an API on the bus
3760 to trigger such reboots, in case graphical desktop UIs want
3761 to cover this functionality.
3763 * "systemctl enable", "systemctl disable" and "systemctl mask"
3764 now support a new "--now" switch. If specified the units
3765 that are enabled will also be started, and the ones
3766 disabled/masked also stopped.
3768 * The Gummiboot EFI boot loader tool has been merged into
3769 systemd, and renamed to "systemd-boot". The bootctl tool has been
3770 updated to support systemd-boot.
3772 * An EFI kernel stub has been added that may be used to create
3773 kernel EFI binaries that contain not only the actual kernel,
3774 but also an initrd, boot splash, command line and OS release
3775 information. This combined binary can then be signed as a
3776 single image, so that the firmware can verify it all in one
3777 step. systemd-boot has special support for EFI binaries created
3778 like this and can extract OS release information from them
3779 and show them in the boot menu. This functionality is useful
3780 to implement cryptographically verified boot schemes.
3782 * Optional support has been added to systemd-fsck to pass
3783 fsck's progress report to an AF_UNIX socket in the file
3786 * udev will no longer create device symlinks for all block
3787 devices by default. A blacklist for excluding special block
3788 devices from this logic has been turned into a whitelist
3789 that requires picking block devices explicitly that require
3792 * A new (currently still internal) API sd-device.h has been
3793 added to libsystemd. This modernized API is supposed to
3794 replace libudev eventually. In fact, already much of libudev
3795 is now just a wrapper around sd-device.h.
3797 * A new hwdb database for storing metadata about pointing
3798 stick devices has been added.
3800 * systemd-tmpfiles gained support for setting file attributes
3801 similar to the "chattr" tool with new 'h' and 'H' lines.
3803 * systemd-journald will no longer unconditionally set the
3804 btrfs NOCOW flag on new journal files. This is instead done
3805 with tmpfiles snippet using the new 'h' line type. This
3806 allows easy disabling of this logic, by masking the
3807 journal-nocow.conf tmpfiles file.
3809 * systemd-journald will now translate audit message types to
3810 human readable identifiers when writing them to the
3811 journal. This should improve readability of audit messages.
3813 * The LUKS logic gained support for the offset= and skip=
3814 options in /etc/crypttab, as previously implemented by
3817 * /usr/lib/os-release gained a new optional field VARIANT= for
3818 distributions that support multiple variants (such as a
3819 desktop edition, a server edition, ...)
3821 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
3822 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
3823 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
3824 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
3825 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
3826 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
3827 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
3828 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
3829 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
3830 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
3831 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
3832 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
3833 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
3834 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
3835 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
3836 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
3837 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
3838 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
3839 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
3840 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
3841 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
3842 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
3843 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
3844 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
3845 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
3846 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
3847 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
3849 — Berlin, 2015-05-22
3853 * Introduce a new API "sd-hwdb.h" for querying the hardware
3854 metadata database. With this minimal interface one can query
3855 and enumerate the udev hwdb, decoupled from the old libudev
3856 library. libudev's interface for this is now only a wrapper
3857 around sd-hwdb. A new tool systemd-hwdb has been added to
3858 interface with and update the database.
3860 * When any of systemd's tools copies files (for example due to
3861 tmpfiles' C lines) a btrfs reflink will attempted first,
3862 before bytewise copying is done.
3864 * systemd-nspawn gained a new --ephemeral switch. When
3865 specified a btrfs snapshot is taken of the container's root
3866 directory, and immediately removed when the container
3867 terminates again. Thus, a container can be started whose
3868 changes never alter the container's root directory, and are
3869 lost on container termination. This switch can also be used
3870 for starting a container off the root file system of the
3871 host without affecting the host OS. This switch is only
3872 available on btrfs file systems.
3874 * systemd-nspawn gained a new --template= switch. It takes the
3875 path to a container tree to use as template for the tree
3876 specified via --directory=, should that directory be
3877 missing. This allows instantiating containers dynamically,
3878 on first run. This switch is only available on btrfs file
3881 * When a .mount unit refers to a mount point on which multiple
3882 mounts are stacked, and the .mount unit is stopped all of
3883 the stacked mount points will now be unmounted until no
3884 mount point remains.
3886 * systemd now has an explicit notion of supported and
3887 unsupported unit types. Jobs enqueued for unsupported unit
3888 types will now fail with an "unsupported" error code. More
3889 specifically .swap, .automount and .device units are not
3890 supported in containers, .busname units are not supported on
3891 non-kdbus systems. .swap and .automount are also not
3892 supported if their respective kernel compile time options
3895 * machinectl gained support for two new "copy-from" and
3896 "copy-to" commands for copying files from a running
3897 container to the host or vice versa.
3899 * machinectl gained support for a new "bind" command to bind
3900 mount host directories into local containers. This is
3901 currently only supported for nspawn containers.
3903 * networkd gained support for configuring bridge forwarding
3904 database entries (fdb) from .network files.
3906 * A new tiny daemon "systemd-importd" has been added that can
3907 download container images in tar, raw, qcow2 or dkr formats,
3908 and make them available locally in /var/lib/machines, so
3909 that they can run as nspawn containers. The daemon can GPG
3910 verify the downloads (not supported for dkr, since it has no
3911 provisions for verifying downloads). It will transparently
3912 decompress bz2, xz, gzip compressed downloads if necessary,
3913 and restore sparse files on disk. The daemon uses privilege
3914 separation to ensure the actual download logic runs with
3915 fewer privileges than the daemon itself. machinectl has
3916 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
3917 make the functionality of importd available to the
3918 user. With this in place the Fedora and Ubuntu "Cloud"
3919 images can be downloaded and booted as containers unmodified
3920 (the Fedora images lack the appropriate GPG signature files
3921 currently, so they cannot be verified, but this will change
3922 soon, hopefully). Note that downloading images is currently
3923 only fully supported on btrfs.
3925 * machinectl is now able to list container images found in
3926 /var/lib/machines, along with some metadata about sizes of
3927 disk and similar. If the directory is located on btrfs and
3928 quota is enabled, this includes quota display. A new command
3929 "image-status" has been added that shows additional
3930 information about images.
3932 * machinectl is now able to clone container images
3933 efficiently, if the underlying file system (btrfs) supports
3934 it, with the new "machinectl clone" command. It also
3935 gained commands for renaming and removing images, as well as
3936 marking them read-only or read-write (supported also on
3937 legacy file systems).
3939 * networkd gained support for collecting LLDP network
3940 announcements, from hardware that supports this. This is
3941 shown in networkctl output.
3943 * systemd-run gained support for a new -t (--pty) switch for
3944 invoking a binary on a pty whose input and output is
3945 connected to the invoking terminal. This allows executing
3946 processes as system services while interactively
3947 communicating with them via the terminal. Most interestingly
3948 this is supported across container boundaries. Invoking
3949 "systemd-run -t /bin/bash" is an alternative to running a
3950 full login session, the difference being that the former
3951 will not register a session, nor go through the PAM session
3954 * tmpfiles gained support for a new "v" line type for creating
3955 btrfs subvolumes. If the underlying file system is a legacy
3956 file system, this automatically degrades to creating a
3957 normal directory. Among others /var/lib/machines is now
3958 created like this at boot, should it be missing.
3960 * The directory /var/lib/containers/ has been deprecated and
3961 been replaced by /var/lib/machines. The term "machines" has
3962 been used in the systemd context as generic term for both
3963 VMs and containers, and hence appears more appropriate for
3964 this, as the directory can also contain raw images bootable
3967 * systemd-nspawn when invoked with -M but without --directory=
3968 or --image= is now capable of searching for the container
3969 root directory, subvolume or disk image automatically, in
3970 /var/lib/machines. systemd-nspawn@.service has been updated
3971 to make use of this, thus allowing it to be used for raw
3974 * A new machines.target unit has been introduced that is
3975 supposed to group all containers/VMs invoked as services on
3976 the system. systemd-nspawn@.service has been updated to
3977 integrate with that.
3979 * machinectl gained a new "start" command, for invoking a
3980 container as a service. "machinectl start foo" is mostly
3981 equivalent to "systemctl start systemd-nspawn@foo.service",
3982 but handles escaping in a nicer way.
3984 * systemd-nspawn will now mount most of the cgroupfs tree
3985 read-only into each container, with the exception of the
3986 container's own subtree in the name=systemd hierarchy.
3988 * journald now sets the special FS_NOCOW file flag for its
3989 journal files. This should improve performance on btrfs, by
3990 avoiding heavy fragmentation when journald's write-pattern
3991 is used on COW file systems. It degrades btrfs' data
3992 integrity guarantees for the files to the same levels as for
3993 ext3/ext4 however. This should be OK though as journald does
3994 its own data integrity checks and all its objects are
3995 checksummed on disk. Also, journald should handle btrfs disk
3996 full events a lot more gracefully now, by processing SIGBUS
3997 errors, and not relying on fallocate() anymore.
3999 * When journald detects that journal files it is writing to
4000 have been deleted it will immediately start new journal
4003 * systemd now provides a way to store file descriptors
4004 per-service in PID 1. This is useful for daemons to ensure
4005 that fds they require are not lost during a daemon
4006 restart. The fds are passed to the daemon on the next
4007 invocation in the same way socket activation fds are
4008 passed. This is now used by journald to ensure that the
4009 various sockets connected to all the system's stdout/stderr
4010 are not lost when journald is restarted. File descriptors
4011 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
4012 an extension to sd_notify(). Note that a limit is enforced
4013 on the number of fds a service can store in PID 1, and it
4014 defaults to 0, so that no fds may be stored, unless this is
4015 explicitly turned on.
4017 * The default TERM variable to use for units connected to a
4018 terminal, when no other value is explicitly is set is now
4019 vt220 rather than vt102. This should be fairly safe still,
4020 but allows PgUp/PgDn work.
4022 * The /etc/crypttab option header= as known from Debian is now
4025 * "loginctl user-status" and "loginctl session-status" will
4026 now show the last 10 lines of log messages of the
4027 user/session following the status output. Similar,
4028 "machinectl status" will show the last 10 log lines
4029 associated with a virtual machine or container
4030 service. (Note that this is usually not the log messages
4031 done in the VM/container itself, but simply what the
4032 container manager logs. For nspawn this includes all console
4035 * "loginctl session-status" without further argument will now
4036 show the status of the session of the caller. Similar,
4037 "lock-session", "unlock-session", "activate",
4038 "enable-linger", "disable-linger" may now be called without
4039 session/user parameter in which case they apply to the
4040 caller's session/user.
4042 * An X11 session scriptlet is now shipped that uploads
4043 $DISPLAY and $XAUTHORITY into the environment of the systemd
4044 --user daemon if a session begins. This should improve
4045 compatibility with X11 enabled applications run as systemd
4048 * Generators are now subject to masking via /etc and /run, the
4049 same way as unit files.
4051 * networkd .network files gained support for configuring
4052 per-link IPv4/IPv6 packet forwarding as well as IPv4
4053 masquerading. This is by default turned on for veth links to
4054 containers, as registered by systemd-nspawn. This means that
4055 nspawn containers run with --network-veth will now get
4056 automatic routed access to the host's networks without any
4057 further configuration or setup, as long as networkd runs on
4060 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4061 or UDP posts of a container on the host. With this in place
4062 it is possible to run containers with private veth links
4063 (--network-veth), and have their functionality exposed on
4064 the host as if their services were running directly on the
4067 * systemd-nspawn's --network-veth switch now gained a short
4068 version "-n", since with the changes above it is now truly
4069 useful out-of-the-box. The systemd-nspawn@.service has been
4070 updated to make use of it too by default.
4072 * systemd-nspawn will now maintain a per-image R/W lock, to
4073 ensure that the same image is not started more than once
4074 writable. (It's OK to run an image multiple times
4075 simultaneously in read-only mode.)
4077 * systemd-nspawn's --image= option is now capable of
4078 dissecting and booting MBR and GPT disk images that contain
4079 only a single active Linux partition. Previously it
4080 supported only GPT disk images with proper GPT type
4081 IDs. This allows running cloud images from major
4082 distributions directly with systemd-nspawn, without
4085 * In addition to collecting mouse dpi data in the udev
4086 hardware database, there's now support for collecting angle
4087 information for mouse scroll wheels. The database is
4088 supposed to guarantee similar scrolling behavior on mice
4089 that it knows about. There's also support for collecting
4090 information about Touchpad types.
4092 * udev's input_id built-in will now also collect touch screen
4093 dimension data and attach it to probed devices.
4095 * /etc/os-release gained support for a Distribution Privacy
4098 * networkd gained support for creating "ipvlan", "gretap",
4099 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4101 * systemd-tmpfiles gained support for "a" lines for setting
4104 * systemd-nspawn will now mount /tmp in the container to
4105 tmpfs, automatically.
4107 * systemd now exposes the memory.usage_in_bytes cgroup
4108 attribute and shows it for each service in the "systemctl
4109 status" output, if available.
4111 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4112 immediate reboot is triggered. This useful if shutdown is
4113 hung and is unable to complete, to expedite the
4114 operation. Note that this kind of reboot will still unmount
4115 all file systems, and hence should not result in fsck being
4118 * A .device unit for an optical block device will now be
4119 considered active only when a medium is in the drive. Also,
4120 mount units are now bound to their backing devices thus
4121 triggering automatic unmounting when devices become
4122 unavailable. With this in place systemd will now
4123 automatically unmount left-over mounts when a CD-ROM is
4124 ejected or an USB stick is yanked from the system.
4126 * networkd-wait-online now has support for waiting for
4127 specific interfaces only (with globbing), and for giving up
4128 after a configurable timeout.
4130 * networkd now exits when idle. It will be automatically
4131 restarted as soon as interfaces show up, are removed or
4132 change state. networkd will stay around as long as there is
4133 at least one DHCP state machine or similar around, that keep
4136 * networkd may now configure IPv6 link-local addressing in
4137 addition to IPv4 link-local addressing.
4139 * The IPv6 "token" for use in SLAAC may now be configured for
4140 each .network interface in networkd.
4142 * Routes configured with networkd may now be assigned a scope
4145 * networkd's [Match] sections now support globbing and lists
4146 of multiple space-separated matches per item.
4148 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
4149 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4150 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4151 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4152 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4153 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4154 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4155 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4156 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4157 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4158 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4159 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4160 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4161 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4162 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
4163 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4164 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4165 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4166 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4167 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4168 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4169 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
4170 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4171 Hoffmann, Zbigniew Jędrzejewski-Szmek
4173 — Berlin, 2015-02-16
4177 * When querying unit file enablement status (for example via
4178 "systemctl is-enabled"), a new state "indirect" is now known
4179 which indicates that a unit might not be enabled itself, but
4180 another unit listed in its Also= setting might be.
4182 * Similar to the various existing ConditionXYZ= settings for
4183 units, there are now matching AssertXYZ= settings. While
4184 failing conditions cause a unit to be skipped, but its job
4185 to succeed, failing assertions declared like this will cause
4186 a unit start operation and its job to fail.
4188 * hostnamed now knows a new chassis type "embedded".
4190 * systemctl gained a new "edit" command. When used on a unit
4191 file, this allows extending unit files with .d/ drop-in
4192 configuration snippets or editing the full file (after
4193 copying it from /usr/lib to /etc). This will invoke the
4194 user's editor (as configured with $EDITOR), and reload the
4195 modified configuration after editing.
4197 * "systemctl status" now shows the suggested enablement state
4198 for a unit, as declared in the (usually vendor-supplied)
4199 system preset files.
4201 * nss-myhostname will now resolve the single-label host name
4202 "gateway" to the locally configured default IP routing
4203 gateways, ordered by their metrics. This assigns a stable
4204 name to the used gateways, regardless which ones are
4205 currently configured. Note that the name will only be
4206 resolved after all other name sources (if nss-myhostname is
4207 configured properly) and should hence not negatively impact
4208 systems that use the single-label host name "gateway" in
4211 * systemd-inhibit now allows filtering by mode when listing
4214 * Scope and service units gained a new "Delegate" boolean
4215 property, which, when set, allows processes running inside the
4216 unit to further partition resources. This is primarily
4217 useful for systemd user instances as well as container
4220 * journald will now pick up audit messages directly from
4221 the kernel, and log them like any other log message. The
4222 audit fields are split up and fully indexed. This means that
4223 journalctl in many ways is now a (nicer!) alternative to
4224 ausearch, the traditional audit client. Note that this
4225 implements only a minimal audit client. If you want the
4226 special audit modes like reboot-on-log-overflow, please use
4227 the traditional auditd instead, which can be used in
4228 parallel to journald.
4230 * The ConditionSecurity= unit file option now understands the
4231 special string "audit" to check whether auditing is
4234 * journalctl gained two new commands --vacuum-size= and
4235 --vacuum-time= to delete old journal files until the
4236 remaining ones take up no more than the specified size on disk,
4237 or are not older than the specified time.
4239 * A new, native PPPoE library has been added to sd-network,
4240 systemd's library of light-weight networking protocols. This
4241 library will be used in a future version of networkd to
4242 enable PPPoE communication without an external pppd daemon.
4244 * The busctl tool now understands a new "capture" verb that
4245 works similar to "monitor", but writes a packet capture
4246 trace to STDOUT that can be redirected to a file which is
4247 compatible with libcap's capture file format. This can then
4248 be loaded in Wireshark and similar tools to inspect bus
4251 * The busctl tool now understands a new "tree" verb that shows
4252 the object trees of a specific service on the bus, or of all
4255 * The busctl tool now understands a new "introspect" verb that
4256 shows all interfaces and members of objects on the bus,
4257 including their signature and values. This is particularly
4258 useful to get more information about bus objects shown by
4259 the new "busctl tree" command.
4261 * The busctl tool now understands new verbs "call",
4262 "set-property" and "get-property" for invoking bus method
4263 calls, setting and getting bus object properties in a
4266 * busctl gained a new --augment-creds= argument that controls
4267 whether the tool shall augment credential information it
4268 gets from the bus with data from /proc, in a possibly
4271 * nspawn's --link-journal= switch gained two new values
4272 "try-guest" and "try-host" that work like "guest" and
4273 "host", but do not fail if the host has no persistent
4274 journalling enabled. -j is now equivalent to
4275 --link-journal=try-guest.
4277 * macvlan network devices created by nspawn will now have
4278 stable MAC addresses.
4280 * A new SmackProcessLabel= unit setting has been added, which
4281 controls the SMACK security label processes forked off by
4282 the respective unit shall use.
4284 * If compiled with --enable-xkbcommon, systemd-localed will
4285 verify x11 keymap settings by compiling the given keymap. It
4286 will spew out warnings if the compilation fails. This
4287 requires libxkbcommon to be installed.
4289 * When a coredump is collected, a larger number of metadata
4290 fields is now collected and included in the journal records
4291 created for it. More specifically, control group membership,
4292 environment variables, memory maps, working directory,
4293 chroot directory, /proc/$PID/status, and a list of open file
4294 descriptors is now stored in the log entry.
4296 * The udev hwdb now contains DPI information for mice. For
4299 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4301 * All systemd programs that read standalone configuration
4302 files in /etc now also support a corresponding series of
4303 .conf.d configuration directories in /etc/, /run/,
4304 /usr/local/lib/, /usr/lib/, and (if configured with
4305 --enable-split-usr) /lib/. In particular, the following
4306 configuration files now have corresponding configuration
4307 directories: system.conf user.conf, logind.conf,
4308 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4309 resolved.conf, timesyncd.conf, journal-remote.conf, and
4310 journal-upload.conf. Note that distributions should use the
4311 configuration directories in /usr/lib/; the directories in
4312 /etc/ are reserved for the system administrator.
4314 * systemd-rfkill will no longer take the rfkill device name
4315 into account when storing rfkill state on disk, as the name
4316 might be dynamically assigned and not stable. Instead, the
4317 ID_PATH udev variable combined with the rfkill type (wlan,
4318 bluetooth, ...) is used.
4320 * A new service systemd-machine-id-commit.service has been
4321 added. When used on systems where /etc is read-only during
4322 boot, and /etc/machine-id is not initialized (but an empty
4323 file), this service will copy the temporary machine ID
4324 created as replacement into /etc after the system is fully
4325 booted up. This is useful for systems that are freshly
4326 installed with a non-initialized machine ID, but should get
4327 a fixed machine ID for subsequent boots.
4329 * networkd's .netdev files now provide a large set of
4330 configuration parameters for VXLAN devices. Similarly, the
4331 bridge port cost parameter is now configurable in .network
4332 files. There's also new support for configuring IP source
4333 routing. networkd .link files gained support for a new
4334 OriginalName= match that is useful to match against the
4335 original interface name the kernel assigned. .network files
4336 may include MTU= and MACAddress= fields for altering the MTU
4337 and MAC address while being connected to a specific network
4340 * The LUKS logic gained supported for configuring
4341 UUID-specific key files. There's also new support for naming
4342 LUKS device from the kernel command line, using the new
4343 luks.name= argument.
4345 * Timer units may now be transiently created via the bus API
4346 (this was previously already available for scope and service
4347 units). In addition it is now possible to create multiple
4348 transient units at the same time with a single bus call. The
4349 "systemd-run" tool has been updated to make use of this for
4350 running commands on a specified time, in at(1)-style.
4352 * tmpfiles gained support for "t" lines, for assigning
4353 extended attributes to files. Among other uses this may be
4354 used to assign SMACK labels to files.
4356 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4357 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4358 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4359 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4360 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4361 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4362 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4363 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4364 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4365 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4366 Michael Biebl, Michael Chapman, Michael Marineau, Michal
4367 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4368 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4369 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4370 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4371 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4372 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4373 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
4375 — Berlin, 2014-12-10
4379 * journalctl gained the new options -t/--identifier= to match
4380 on the syslog identifier (aka "tag"), as well as --utc to
4381 show log timestamps in the UTC timezone. journalctl now also
4382 accepts -n/--lines=all to disable line capping in a pager.
4384 * journalctl gained a new switch, --flush, that synchronously
4385 flushes logs from /run/log/journal to /var/log/journal if
4386 persistent storage is enabled. systemd-journal-flush.service
4387 now waits until the operation is complete.
4389 * Services can notify the manager before they start a reload
4390 (by sending RELOADING=1) or shutdown (by sending
4391 STOPPING=1). This allows the manager to track and show the
4392 internal state of daemons and closes a race condition when
4393 the process is still running but has closed its D-Bus
4396 * Services with Type=oneshot do not have to have any ExecStart
4399 * User units are now loaded also from
4400 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4401 /run/systemd/user directory that was already previously
4402 supported, but is under the control of the user.
4404 * Job timeouts (i.e. time-outs on the time a job that is
4405 queued stays in the run queue) can now optionally result in
4406 immediate reboot or power-off actions (JobTimeoutAction= and
4407 JobTimeoutRebootArgument=). This is useful on ".target"
4408 units, to limit the maximum time a target remains
4409 undispatched in the run queue, and to trigger an emergency
4410 operation in such a case. This is now used by default to
4411 turn off the system if boot-up (as defined by everything in
4412 basic.target) hangs and does not complete for at least
4413 15min. Also, if power-off or reboot hang for at least 30min
4414 an immediate power-off/reboot operation is triggered. This
4415 functionality is particularly useful to increase reliability
4416 on embedded devices, but also on laptops which might
4417 accidentally get powered on when carried in a backpack and
4418 whose boot stays stuck in a hard disk encryption passphrase
4421 * systemd-logind can be configured to also handle lid switch
4422 events even when the machine is docked or multiple displays
4423 are attached (HandleLidSwitchDocked= option).
4425 * A helper binary and a service have been added which can be
4426 used to resume from hibernation in the initramfs. A
4427 generator will parse the resume= option on the kernel
4428 command line to trigger resume.
4430 * A user console daemon systemd-consoled has been
4431 added. Currently, it is a preview, and will so far open a
4432 single terminal on each session of the user marked as
4433 Desktop=systemd-console.
4435 * Route metrics can be specified for DHCP routes added by
4438 * The SELinux context of socket-activated services can be set
4439 from the information provided by the networking stack
4440 (SELinuxContextFromNet= option).
4442 * Userspace firmware loading support has been removed and
4443 the minimum supported kernel version is thus bumped to 3.7.
4445 * Timeout for udev workers has been increased from 1 to 3
4446 minutes, but a warning will be printed after 1 minute to
4447 help diagnose kernel modules that take a long time to load.
4449 * Udev rules can now remove tags on devices with TAG-="foobar".
4451 * systemd's readahead implementation has been removed. In many
4452 circumstances it didn't give expected benefits even for
4453 rotational disk drives and was becoming less relevant in the
4454 age of SSDs. As none of the developers has been using
4455 rotating media anymore, and nobody stepped up to actively
4456 maintain this component of systemd it has now been removed.
4458 * Swap units can use Options= to specify discard options.
4459 Discard options specified for swaps in /etc/fstab are now
4462 * Docker containers are now detected as a separate type of
4465 * The Password Agent protocol gained support for queries where
4466 the user input is shown, useful e.g. for user names.
4467 systemd-ask-password gained a new --echo option to turn that
4470 * The default sysctl.d/ snippets will now set:
4472 net.core.default_qdisc = fq_codel
4474 This selects Fair Queuing Controlled Delay as the default
4475 queuing discipline for network interfaces. fq_codel helps
4476 fight the network bufferbloat problem. It is believed to be
4477 a good default with no tuning required for most workloads.
4478 Downstream distributions may override this choice. On 10Gbit
4479 servers that do not do forwarding, "fq" may perform better.
4480 Systems without a good clocksource should use "pfifo_fast".
4482 * If kdbus is enabled during build a new option BusPolicy= is
4483 available for service units, that allows locking all service
4484 processes into a stricter bus policy, in order to limit
4485 access to various bus services, or even hide most of them
4486 from the service's view entirely.
4488 * networkctl will now show the .network and .link file
4489 networkd has applied to a specific interface.
4491 * sd-login gained a new API call sd_session_get_desktop() to
4492 query which desktop environment has been selected for a
4495 * UNIX utmp support is now compile-time optional to support
4496 legacy-free systems.
4498 * systemctl gained two new commands "add-wants" and
4499 "add-requires" for pulling in units from specific targets
4502 * If the word "rescue" is specified on the kernel command line
4503 the system will now boot into rescue mode (aka
4504 rescue.target), which was previously available only by
4505 specifying "1" or "systemd.unit=rescue.target" on the kernel
4506 command line. This new kernel command line option nicely
4507 mirrors the already existing "emergency" kernel command line
4510 * New kernel command line options mount.usr=, mount.usrflags=,
4511 mount.usrfstype= have been added that match root=, rootflags=,
4512 rootfstype= but allow mounting a specific file system to
4515 * The $NOTIFY_SOCKET is now also passed to control processes of
4516 services, not only the main process.
4518 * This version reenables support for fsck's -l switch. This
4519 means at least version v2.25 of util-linux is required for
4520 operation, otherwise dead-locks on device nodes may
4521 occur. Again: you need to update util-linux to at least
4522 v2.25 when updating systemd to v217.
4524 * The "multi-seat-x" tool has been removed from systemd, as
4525 its functionality has been integrated into X servers 1.16,
4526 and the tool is hence redundant. It is recommended to update
4527 display managers invoking this tool to simply invoke X
4528 directly from now on, again.
4530 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
4531 message flag has been added for all of systemd's PolicyKit
4532 authenticated method calls has been added. In particular
4533 this now allows optional interactive authorization via
4534 PolicyKit for many of PID1's privileged operations such as
4535 unit file enabling and disabling.
4537 * "udevadm hwdb --update" learnt a new switch "--usr" for
4538 placing the rebuilt hardware database in /usr instead of
4539 /etc. When used only hardware database entries stored in
4540 /usr will be used, and any user database entries in /etc are
4541 ignored. This functionality is useful for vendors to ship a
4542 pre-built database on systems where local configuration is
4543 unnecessary or unlikely.
4545 * Calendar time specifications in .timer units now also
4546 understand the strings "semi-annually", "quarterly" and
4547 "minutely" as shortcuts (in addition to the preexisting
4548 "anually", "hourly", ...).
4550 * systemd-tmpfiles will now correctly create files in /dev
4551 at boot which are marked for creation only at boot. It is
4552 recommended to always create static device nodes with 'c!'
4553 and 'b!', so that they are created only at boot and not
4554 overwritten at runtime.
4556 * When the watchdog logic is used for a service (WatchdogSec=)
4557 and the watchdog timeout is hit the service will now be
4558 terminated with SIGABRT (instead of just SIGTERM), in order
4559 to make sure a proper coredump and backtrace is
4560 generated. This ensures that hanging services will result in
4561 similar coredump/backtrace behaviour as services that hit a
4564 Contributions from: Andreas Henriksson, Andrei Borzenkov,
4565 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
4566 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
4567 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
4568 Herrmann, David Sommerseth, David Strauss, Emil Renner
4569 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
4570 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
4571 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
4572 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
4573 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
4574 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
4575 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
4576 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
4577 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
4578 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
4579 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
4580 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
4581 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
4582 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
4583 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
4584 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
4587 — Berlin, 2014-10-28
4591 * timedated no longer reads NTP implementation unit names from
4592 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
4593 implementations should add a
4595 Conflicts=systemd-timesyncd.service
4597 to their unit files to take over and replace systemd's NTP
4598 default functionality.
4600 * systemd-sysusers gained a new line type "r" for configuring
4601 which UID/GID ranges to allocate system users/groups
4602 from. Lines of type "u" may now add an additional column
4603 that specifies the home directory for the system user to be
4604 created. Also, systemd-sysusers may now optionally read user
4605 information from STDIN instead of a file. This is useful for
4606 invoking it from RPM preinst scriptlets that need to create
4607 users before the first RPM file is installed since these
4608 files might need to be owned by them. A new
4609 %sysusers_create_inline RPM macro has been introduced to do
4610 just that. systemd-sysusers now updates the shadow files as
4611 well as the user/group databases, which should enhance
4612 compatibility with certain tools like grpck.
4614 * A number of bus APIs of PID 1 now optionally consult
4615 PolicyKit to permit access for otherwise unprivileged
4616 clients under certain conditions. Note that this currently
4617 doesn't support interactive authentication yet, but this is
4618 expected to be added eventually, too.
4620 * /etc/machine-info now has new fields for configuring the
4621 deployment environment of the machine, as well as the
4622 location of the machine. hostnamectl has been updated with
4623 new command to update these fields.
4625 * systemd-timesyncd has been updated to automatically acquire
4626 NTP server information from systemd-networkd, which might
4627 have been discovered via DHCP.
4629 * systemd-resolved now includes a caching DNS stub resolver
4630 and a complete LLMNR name resolution implementation. A new
4631 NSS module "nss-resolve" has been added which can be used
4632 instead of glibc's own "nss-dns" to resolve hostnames via
4633 systemd-resolved. Hostnames, addresses and arbitrary RRs may
4634 be resolved via systemd-resolved D-Bus APIs. In contrast to
4635 the glibc internal resolver systemd-resolved is aware of
4636 multi-homed system, and keeps DNS server and caches separate
4637 and per-interface. Queries are sent simultaneously on all
4638 interfaces that have DNS servers configured, in order to
4639 properly handle VPNs and local LANs which might resolve
4640 separate sets of domain names. systemd-resolved may acquire
4641 DNS server information from systemd-networkd automatically,
4642 which in turn might have discovered them via DHCP. A tool
4643 "systemd-resolve-host" has been added that may be used to
4644 query the DNS logic in resolved. systemd-resolved implements
4645 IDNA and automatically uses IDNA or UTF-8 encoding depending
4646 on whether classic DNS or LLMNR is used as transport. In the
4647 next releases we intend to add a DNSSEC and mDNS/DNS-SD
4648 implementation to systemd-resolved.
4650 * A new NSS module nss-mymachines has been added, that
4651 automatically resolves the names of all local registered
4652 containers to their respective IP addresses.
4654 * A new client tool "networkctl" for systemd-networkd has been
4655 added. It currently is entirely passive and will query
4656 networking configuration from udev, rtnetlink and networkd,
4657 and present it to the user in a very friendly
4658 way. Eventually, we hope to extend it to become a full
4659 control utility for networkd.
4661 * .socket units gained a new DeferAcceptSec= setting that
4662 controls the kernels' TCP_DEFER_ACCEPT sockopt for
4663 TCP. Similarly, support for controlling TCP keep-alive
4664 settings has been added (KeepAliveTimeSec=,
4665 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
4666 turning off Nagle's algorithm on TCP has been added
4669 * logind learned a new session type "web", for use in projects
4670 like Cockpit which register web clients as PAM sessions.
4672 * timer units with at least one OnCalendar= setting will now
4673 be started only after time-sync.target has been
4674 reached. This way they will not elapse before the system
4675 clock has been corrected by a local NTP client or
4676 similar. This is particular useful on RTC-less embedded
4677 machines, that come up with an invalid system clock.
4679 * systemd-nspawn's --network-veth= switch should now result in
4680 stable MAC addresses for both the outer and the inner side
4683 * systemd-nspawn gained a new --volatile= switch for running
4684 container instances with /etc or /var unpopulated.
4686 * The kdbus client code has been updated to use the new Linux
4687 3.17 memfd subsystem instead of the old kdbus-specific one.
4689 * systemd-networkd's DHCP client and server now support
4690 FORCERENEW. There are also new configuration options to
4691 configure the vendor client identifier and broadcast mode
4694 * systemd will no longer inform the kernel about the current
4695 timezone, as this is necessarily incorrect and racy as the
4696 kernel has no understanding of DST and similar
4697 concepts. This hence means FAT timestamps will be always
4698 considered UTC, similar to what Android is already
4699 doing. Also, when the RTC is configured to the local time
4700 (rather than UTC) systemd will never synchronize back to it,
4701 as this might confuse Windows at a later boot.
4703 * systemd-analyze gained a new command "verify" for offline
4704 validation of unit files.
4706 * systemd-networkd gained support for a couple of additional
4707 settings for bonding networking setups. Also, the metric for
4708 statically configured routes may now be configured. For
4709 network interfaces where this is appropriate the peer IP
4710 address may now be configured.
4712 * systemd-networkd's DHCP client will no longer request
4713 broadcasting by default, as this tripped up some networks.
4714 For hardware where broadcast is required the feature should
4715 be switched back on using RequestBroadcast=yes.
4717 * systemd-networkd will now set up IPv4LL addresses (when
4718 enabled) even if DHCP is configured successfully.
4720 * udev will now default to respect network device names given
4721 by the kernel when the kernel indicates that these are
4722 predictable. This behavior can be tweaked by changing
4723 NamePolicy= in the relevant .link file.
4725 * A new library systemd-terminal has been added that
4726 implements full TTY stream parsing and rendering. This
4727 library is supposed to be used later on for implementing a
4728 full userspace VT subsystem, replacing the current kernel
4731 * A new tool systemd-journal-upload has been added to push
4732 journal data to a remote system running
4733 systemd-journal-remote.
4735 * journald will no longer forward all local data to another
4736 running syslog daemon. This change has been made because
4737 rsyslog (which appears to be the most commonly used syslog
4738 implementation these days) no longer makes use of this, and
4739 instead pulls the data out of the journal on its own. Since
4740 forwarding the messages to a non-existent syslog server is
4741 more expensive than we assumed we have now turned this
4742 off. If you run a syslog server that is not a recent rsyslog
4743 version, you have to turn this option on again
4744 (ForwardToSyslog= in journald.conf).
4746 * journald now optionally supports the LZ4 compressor for
4747 larger journal fields. This compressor should perform much
4748 better than XZ which was the previous default.
4750 * machinectl now shows the IP addresses of local containers,
4751 if it knows them, plus the interface name of the container.
4753 * A new tool "systemd-escape" has been added that makes it
4754 easy to escape strings to build unit names and similar.
4756 * sd_notify() messages may now include a new ERRNO= field
4757 which is parsed and collected by systemd and shown among the
4758 "systemctl status" output for a service.
4760 * A new component "systemd-firstboot" has been added that
4761 queries the most basic systemd information (timezone,
4762 hostname, root password) interactively on first
4763 boot. Alternatively it may also be used to provision these
4764 things offline on OS images installed into directories.
4766 * The default sysctl.d/ snippets will now set
4768 net.ipv4.conf.default.promote_secondaries=1
4770 This has the benefit of no flushing secondary IP addresses
4771 when primary addresses are removed.
4773 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
4774 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
4775 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
4776 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
4777 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
4778 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
4779 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
4780 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
4781 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
4782 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
4783 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
4784 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
4785 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
4786 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
4787 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
4789 — Berlin, 2014-08-19
4793 * A new tool systemd-sysusers has been added. This tool
4794 creates system users and groups in /etc/passwd and
4795 /etc/group, based on static declarative system user/group
4796 definitions in /usr/lib/sysusers.d/. This is useful to
4797 enable factory resets and volatile systems that boot up with
4798 an empty /etc directory, and thus need system users and
4799 groups created during early boot. systemd now also ships
4800 with two default sysusers.d/ files for the most basic
4801 users and groups systemd and the core operating system
4804 * A new tmpfiles snippet has been added that rebuilds the
4805 essential files in /etc on boot, should they be missing.
4807 * A directive for ensuring automatic clean-up of
4808 /var/cache/man/ has been removed from the default
4809 configuration. This line should now be shipped by the man
4810 implementation. The necessary change has been made to the
4811 man-db implementation. Note that you need to update your man
4812 implementation to one that ships this line, otherwise no
4813 automatic clean-up of /var/cache/man will take place.
4815 * A new condition ConditionNeedsUpdate= has been added that
4816 may conditionalize services to only run when /etc or /var
4817 are "older" than the vendor operating system resources in
4818 /usr. This is useful for reconstructing or updating /etc
4819 after an offline update of /usr or a factory reset, on the
4820 next reboot. Services that want to run once after such an
4821 update or reset should use this condition and order
4822 themselves before the new systemd-update-done.service, which
4823 will mark the two directories as fully updated. A number of
4824 service files have been added making use of this, to rebuild
4825 the udev hardware database, the journald message catalog and
4826 dynamic loader cache (ldconfig). The systemd-sysusers tool
4827 described above also makes use of this now. With this in
4828 place it is now possible to start up a minimal operating
4829 system with /etc empty cleanly. For more information on the
4830 concepts involved see this recent blog story:
4832 http://0pointer.de/blog/projects/stateless.html
4834 * A new system group "input" has been introduced, and all
4835 input device nodes get this group assigned. This is useful
4836 for system-level software to get access to input devices. It
4837 complements what is already done for "audio" and "video".
4839 * systemd-networkd learnt minimal DHCPv4 server support in
4840 addition to the existing DHCPv4 client support. It also
4841 learnt DHCPv6 client and IPv6 Router Solicitation client
4842 support. The DHCPv4 client gained support for static routes
4843 passed in from the server. Note that the [DHCPv4] section
4844 known in older systemd-networkd versions has been renamed to
4845 [DHCP] and is now also used by the DHCPv6 client. Existing
4846 .network files using settings of this section should be
4847 updated, though compatibility is maintained. Optionally, the
4848 client hostname may now be sent to the DHCP server.
4850 * networkd gained support for vxlan virtual networks as well
4851 as tun/tap and dummy devices.
4853 * networkd gained support for automatic allocation of address
4854 ranges for interfaces from a system-wide pool of
4855 addresses. This is useful for dynamically managing a large
4856 number of interfaces with a single network configuration
4857 file. In particular this is useful to easily assign
4858 appropriate IP addresses to the veth links of a large number
4859 of nspawn instances.
4861 * RPM macros for processing sysusers, sysctl and binfmt
4862 drop-in snippets at package installation time have been
4865 * The /etc/os-release file should now be placed in
4866 /usr/lib/os-release. The old location is automatically
4867 created as symlink. /usr/lib is the more appropriate
4868 location of this file, since it shall actually describe the
4869 vendor operating system shipped in /usr, and not the
4870 configuration stored in /etc.
4872 * .mount units gained a new boolean SloppyOptions= setting
4873 that maps to mount(8)'s -s option which enables permissive
4874 parsing of unknown mount options.
4876 * tmpfiles learnt a new "L+" directive which creates a symlink
4877 but (unlike "L") deletes a pre-existing file first, should
4878 it already exist and not already be the correct
4879 symlink. Similarly, "b+", "c+" and "p+" directives have been
4880 added as well, which create block and character devices, as
4881 well as fifos in the filesystem, possibly removing any
4882 pre-existing files of different types.
4884 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
4885 'argument' field (which so far specified the source to
4886 symlink/copy the files from) is now optional. If omitted the
4887 same file os copied from /usr/share/factory/ suffixed by the
4888 full destination path. This is useful for populating /etc
4889 with essential files, by copying them from vendor defaults
4890 shipped in /usr/share/factory/etc.
4892 * A new command "systemctl preset-all" has been added that
4893 applies the service preset settings to all installed unit
4894 files. A new switch --preset-mode= has been added that
4895 controls whether only enable or only disable operations
4898 * A new command "systemctl is-system-running" has been added
4899 that allows checking the overall state of the system, for
4900 example whether it is fully up and running.
4902 * When the system boots up with an empty /etc, the equivalent
4903 to "systemctl preset-all" is executed during early boot, to
4904 make sure all default services are enabled after a factory
4907 * systemd now contains a minimal preset file that enables the
4908 most basic services systemd ships by default.
4910 * Unit files' [Install] section gained a new DefaultInstance=
4911 field for defining the default instance to create if a
4912 template unit is enabled with no instance specified.
4914 * A new passive target cryptsetup-pre.target has been added
4915 that may be used by services that need to make they run and
4916 finish before the first LUKS cryptographic device is set up.
4918 * The /dev/loop-control and /dev/btrfs-control device nodes
4919 are now owned by the "disk" group by default, opening up
4920 access to this group.
4922 * systemd-coredump will now automatically generate a
4923 stack trace of all core dumps taking place on the system,
4924 based on elfutils' libdw library. This stack trace is logged
4927 * systemd-coredump may now optionally store coredumps directly
4928 on disk (in /var/lib/systemd/coredump, possibly compressed),
4929 instead of storing them unconditionally in the journal. This
4930 mode is the new default. A new configuration file
4931 /etc/systemd/coredump.conf has been added to configure this
4932 and other parameters of systemd-coredump.
4934 * coredumpctl gained a new "info" verb to show details about a
4935 specific coredump. A new switch "-1" has also been added
4936 that makes sure to only show information about the most
4937 recent entry instead of all entries. Also, as the tool is
4938 generally useful now the "systemd-" prefix of the binary
4939 name has been removed. Distributions that want to maintain
4940 compatibility with the old name should add a symlink from
4941 the old name to the new name.
4943 * journald's SplitMode= now defaults to "uid". This makes sure
4944 that unprivileged users can access their own coredumps with
4945 coredumpctl without restrictions.
4947 * New kernel command line options "systemd.wants=" (for
4948 pulling an additional unit during boot), "systemd.mask="
4949 (for masking a specific unit for the boot), and
4950 "systemd.debug-shell" (for enabling the debug shell on tty9)
4951 have been added. This is implemented in the new generator
4952 "systemd-debug-generator".
4954 * systemd-nspawn will now by default filter a couple of
4955 syscalls for containers, among them those required for
4956 kernel module loading, direct x86 IO port access, swap
4957 management, and kexec. Most importantly though
4958 open_by_handle_at() is now prohibited for containers,
4959 closing a hole similar to a recently discussed vulnerability
4960 in docker regarding access to files on file hierarchies the
4961 container should normally not have access to. Note that, for
4962 nspawn, we generally make no security claims anyway (and
4963 this is explicitly documented in the man page), so this is
4964 just a fix for one of the most obvious problems.
4966 * A new man page file-hierarchy(7) has been added that
4967 contains a minimized, modernized version of the file system
4968 layout systemd expects, similar in style to the FHS
4969 specification or hier(5). A new tool systemd-path(1) has
4970 been added to query many of these paths for the local
4973 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
4974 longer done. Since the directory now has a per-user size
4975 limit, and is cleaned on logout this appears unnecessary,
4976 in particular since this now brings the lifecycle of this
4977 directory closer in line with how IPC objects are handled.
4979 * systemd.pc now exports a number of additional directories,
4980 including $libdir (which is useful to identify the library
4981 path for the primary architecture of the system), and a
4982 couple of drop-in directories.
4984 * udev's predictable network interface names now use the dev_port
4985 sysfs attribute, introduced in linux 3.15 instead of dev_id to
4986 distinguish between ports of the same PCI function. dev_id should
4987 only be used for ports using the same HW address, hence the need
4990 * machined has been updated to export the OS version of a
4991 container (read from /etc/os-release and
4992 /usr/lib/os-release) on the bus. This is now shown in
4993 "machinectl status" for a machine.
4995 * A new service setting RestartForceExitStatus= has been
4996 added. If configured to a set of exit signals or process
4997 return values, the service will be restarted when the main
4998 daemon process exits with any of them, regardless of the
5001 * systemctl's -H switch for connecting to remote systemd
5002 machines has been extended so that it may be used to
5003 directly connect to a specific container on the
5004 host. "systemctl -H root@foobar:waldi" will now connect as
5005 user "root" to host "foobar", and then proceed directly to
5006 the container named "waldi". Note that currently you have to
5007 authenticate as user "root" for this to work, as entering
5008 containers is a privileged operation.
5010 Contributions from: Andreas Henriksson, Benjamin Steinwender,
5011 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
5012 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5013 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5014 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5015 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5016 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5017 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5018 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5019 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5020 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5021 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5023 — Berlin, 2014-07-03
5027 * As an experimental feature, udev now tries to lock the
5028 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5029 executes events for the disk or any of its partitions.
5030 Applications like partitioning programs can lock the
5031 disk device node (flock(LOCK_EX)) and claim temporary
5032 device ownership that way; udev will entirely skip all event
5033 handling for this disk and its partitions. If the disk
5034 was opened for writing, the close will trigger a partition
5035 table rescan in udev's "watch" facility, and if needed
5036 synthesize "change" events for the disk and all its partitions.
5037 This is now unconditionally enabled, and if it turns out to
5038 cause major problems, we might turn it on only for specific
5039 devices, or might need to disable it entirely. Device Mapper
5040 devices are excluded from this logic.
5042 * We temporarily dropped the "-l" switch for fsck invocations,
5043 since they collide with the flock() logic above. util-linux
5044 upstream has been changed already to avoid this conflict,
5045 and we will readd "-l" as soon as util-linux with this
5046 change has been released.
5048 * The dependency on libattr has been removed. Since a long
5049 time, the extended attribute calls have moved to glibc, and
5050 libattr is thus unnecessary.
5052 * Virtualization detection works without privileges now. This
5053 means the systemd-detect-virt binary no longer requires
5054 CAP_SYS_PTRACE file capabilities, and our daemons can run
5055 with fewer privileges.
5057 * systemd-networkd now runs under its own "systemd-network"
5058 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5059 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5060 loses the ability to write to files owned by root this way.
5062 * Similarly, systemd-resolved now runs under its own
5063 "systemd-resolve" user with no capabilities remaining.
5065 * Similarly, systemd-bus-proxyd now runs under its own
5066 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5068 * systemd-networkd gained support for setting up "veth"
5069 virtual Ethernet devices for container connectivity, as well
5070 as GRE and VTI tunnels.
5072 * systemd-networkd will no longer automatically attempt to
5073 manually load kernel modules necessary for certain tunnel
5074 transports. Instead, it is assumed the kernel loads them
5075 automatically when required. This only works correctly on
5076 very new kernels. On older kernels, please consider adding
5077 the kernel modules to /etc/modules-load.d/ as a work-around.
5079 * The resolv.conf file systemd-resolved generates has been
5080 moved to /run/systemd/resolve/. If you have a symlink from
5081 /etc/resolv.conf, it might be necessary to correct it.
5083 * Two new service settings, ProtectHome= and ProtectSystem=,
5084 have been added. When enabled, they will make the user data
5085 (such as /home) inaccessible or read-only and the system
5086 (such as /usr) read-only, for specific services. This allows
5087 very light-weight per-service sandboxing to avoid
5088 modifications of user data or system files from
5089 services. These two new switches have been enabled for all
5090 of systemd's long-running services, where appropriate.
5092 * Socket units gained new SocketUser= and SocketGroup=
5093 settings to set the owner user and group of AF_UNIX sockets
5094 and FIFOs in the file system.
5096 * Socket units gained a new RemoveOnStop= setting. If enabled,
5097 all FIFOS and sockets in the file system will be removed
5098 when the specific socket unit is stopped.
5100 * Socket units gained a new Symlinks= setting. It takes a list
5101 of symlinks to create to file system sockets or FIFOs
5102 created by the specific Unix sockets. This is useful to
5103 manage symlinks to socket nodes with the same life-cycle as
5106 * The /dev/log socket and /dev/initctl FIFO have been moved to
5107 /run, and have been replaced by symlinks. This allows
5108 connecting to these facilities even if PrivateDevices=yes is
5109 used for a service (which makes /dev/log itself unavailable,
5110 but /run is left). This also has the benefit of ensuring
5111 that /dev only contains device nodes, directories and
5112 symlinks, and nothing else.
5114 * sd-daemon gained two new calls sd_pid_notify() and
5115 sd_pid_notifyf(). They are similar to sd_notify() and
5116 sd_notifyf(), but allow overriding of the source PID of
5117 notification messages if permissions permit this. This is
5118 useful to send notify messages on behalf of a different
5119 process (for example, the parent process). The
5120 systemd-notify tool has been updated to make use of this
5121 when sending messages (so that notification messages now
5122 originate from the shell script invoking systemd-notify and
5123 not the systemd-notify process itself. This should minimize
5124 a race where systemd fails to associate notification
5125 messages to services when the originating process already
5128 * A new "on-abnormal" setting for Restart= has been added. If
5129 set, it will result in automatic restarts on all "abnormal"
5130 reasons for a process to exit, which includes unclean
5131 signals, core dumps, timeouts and watchdog timeouts, but
5132 does not include clean and unclean exit codes or clean
5133 signals. Restart=on-abnormal is an alternative for
5134 Restart=on-failure for services that shall be able to
5135 terminate and avoid restarts on certain errors, by
5136 indicating so with an unclean exit code. Restart=on-failure
5137 or Restart=on-abnormal is now the recommended setting for
5138 all long-running services.
5140 * If the InaccessibleDirectories= service setting points to a
5141 mount point (or if there are any submounts contained within
5142 it), it is now attempted to completely unmount it, to make
5143 the file systems truly unavailable for the respective
5146 * The ReadOnlyDirectories= service setting and
5147 systemd-nspawn's --read-only parameter are now recursively
5148 applied to all submounts, too.
5150 * Mount units may now be created transiently via the bus APIs.
5152 * The support for SysV and LSB init scripts has been removed
5153 from the systemd daemon itself. Instead, it is now
5154 implemented as a generator that creates native systemd units
5155 from these scripts when needed. This enables us to remove a
5156 substantial amount of legacy code from PID 1, following the
5157 fact that many distributions only ship a very small number
5158 of LSB/SysV init scripts nowadays.
5160 * Privileged Xen (dom0) domains are not considered
5161 virtualization anymore by the virtualization detection
5162 logic. After all, they generally have unrestricted access to
5163 the hardware and usually are used to manage the unprivileged
5166 * systemd-tmpfiles gained a new "C" line type, for copying
5167 files or entire directories.
5169 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
5170 lines. So far, they have been non-globbing versions of the
5171 latter, and have thus been redundant. In future, it is
5172 recommended to only use "z". "m" has hence been removed
5173 from the documentation, even though it stays supported.
5175 * A tmpfiles snippet to recreate the most basic structure in
5176 /var has been added. This is enough to create the /var/run →
5177 /run symlink and create a couple of structural
5178 directories. This allows systems to boot up with an empty or
5179 volatile /var. Of course, while with this change, the core OS
5180 now is capable with dealing with a volatile /var, not all
5181 user services are ready for it. However, we hope that sooner
5182 or later, many service daemons will be changed upstream so
5183 that they are able to automatically create their necessary
5184 directories in /var at boot, should they be missing. This is
5185 the first step to allow state-less systems that only require
5186 the vendor image for /usr to boot.
5188 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5189 empty tmpfs instance to a specific directory. This is
5190 particularly useful for making use of the automatic
5191 reconstruction of /var (see above), by passing --tmpfs=/var.
5193 * Access modes specified in tmpfiles snippets may now be
5194 prefixed with "~", which indicates that they shall be masked
5195 by whether the existing file or directory is currently
5196 writable, readable or executable at all. Also, if specified,
5197 the sgid/suid/sticky bits will be masked for all
5200 * A new passive target unit "network-pre.target" has been
5201 added which is useful for services that shall run before any
5202 network is configured, for example firewall scripts.
5204 * The "floppy" group that previously owned the /dev/fd*
5205 devices is no longer used. The "disk" group is now used
5206 instead. Distributions should probably deprecate usage of
5209 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5210 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5211 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5212 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5213 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5214 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5217 — Berlin, 2014-06-11
5221 * A new "systemd-timesyncd" daemon has been added for
5222 synchronizing the system clock across the network. It
5223 implements an SNTP client. In contrast to NTP
5224 implementations such as chrony or the NTP reference server,
5225 this only implements a client side, and does not bother with
5226 the full NTP complexity, focusing only on querying time from
5227 one remote server and synchronizing the local clock to
5228 it. Unless you intend to serve NTP to networked clients or
5229 want to connect to local hardware clocks, this simple NTP
5230 client should be more than appropriate for most
5231 installations. The daemon runs with minimal privileges, and
5232 has been hooked up with networkd to only operate when
5233 network connectivity is available. The daemon saves the
5234 current clock to disk every time a new NTP sync has been
5235 acquired, and uses this to possibly correct the system clock
5236 early at bootup, in order to accommodate for systems that
5237 lack an RTC such as the Raspberry Pi and embedded devices,
5238 and to make sure that time monotonically progresses on these
5239 systems, even if it is not always correct. To make use of
5240 this daemon, a new system user and group "systemd-timesync"
5241 needs to be created on installation of systemd.
5243 * The queue "seqnum" interface of libudev has been disabled, as
5244 it was generally incompatible with device namespacing as
5245 sequence numbers of devices go "missing" if the devices are
5246 part of a different namespace.
5248 * "systemctl list-timers" and "systemctl list-sockets" gained
5249 a --recursive switch for showing units of these types also
5250 for all local containers, similar in style to the already
5251 supported --recursive switch for "systemctl list-units".
5253 * A new RebootArgument= setting has been added for service
5254 units, which may be used to specify a kernel reboot argument
5255 to use when triggering reboots with StartLimitAction=.
5257 * A new FailureAction= setting has been added for service
5258 units which may be used to specify an operation to trigger
5259 when a service fails. This works similarly to
5260 StartLimitAction=, but unlike it, controls what is done
5261 immediately rather than only after several attempts to
5262 restart the service in question.
5264 * hostnamed got updated to also expose the kernel name,
5265 release, and version on the bus. This is useful for
5266 executing commands like hostnamectl with the -H switch.
5267 systemd-analyze makes use of this to properly display
5268 details when running non-locally.
5270 * The bootchart tool can now show cgroup information in the
5271 graphs it generates.
5273 * The CFS CPU quota cgroup attribute is now exposed for
5274 services. The new CPUQuota= switch has been added for this
5275 which takes a percentage value. Setting this will have the
5276 result that a service may never get more CPU time than the
5277 specified percentage, even if the machine is otherwise idle.
5279 * systemd-networkd learned IPIP and SIT tunnel support.
5281 * LSB init scripts exposing a dependency on $network will now
5282 get a dependency on network-online.target rather than simply
5283 network.target. This should bring LSB handling closer to
5284 what it was on SysV systems.
5286 * A new fsck.repair= kernel option has been added to control
5287 how fsck shall deal with unclean file systems at boot.
5289 * The (.ini) configuration file parser will now silently
5290 ignore sections whose name begins with "X-". This may be
5291 used to maintain application-specific extension sections in unit
5294 * machined gained a new API to query the IP addresses of
5295 registered containers. "machinectl status" has been updated
5296 to show these addresses in its output.
5298 * A new call sd_uid_get_display() has been added to the
5299 sd-login APIs for querying the "primary" session of a
5300 user. The "primary" session of the user is elected from the
5301 user's sessions and generally a graphical session is
5302 preferred over a text one.
5304 * A minimal systemd-resolved daemon has been added. It
5305 currently simply acts as a companion to systemd-networkd and
5306 manages resolv.conf based on per-interface DNS
5307 configuration, possibly supplied via DHCP. In the long run
5308 we hope to extend this into a local DNSSEC enabled DNS and
5311 * The systemd-networkd-wait-online tool is now enabled by
5312 default. It will delay network-online.target until a network
5313 connection has been configured. The tool primarily integrates
5314 with networkd, but will also make a best effort to make sense
5315 of network configuration performed in some other way.
5317 * Two new service options StartupCPUShares= and
5318 StartupBlockIOWeight= have been added that work similarly to
5319 CPUShares= and BlockIOWeight= however only apply during
5320 system startup. This is useful to prioritize certain services
5321 differently during bootup than during normal runtime.
5323 * hostnamed has been changed to prefer the statically
5324 configured hostname in /etc/hostname (unless set to
5325 'localhost' or empty) over any dynamic one supplied by
5326 dhcp. With this change, the rules for picking the hostname
5327 match more closely the rules of other configuration settings
5328 where the local administrator's configuration in /etc always
5329 overrides any other settings.
5331 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
5332 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5333 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5334 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5335 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5336 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5337 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5338 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5339 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
5340 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5341 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5342 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5343 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5344 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5345 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5346 Lindskog, WaLyong Cho, Will Woods, Zbigniew
5349 — Beijing, 2014-05-28
5353 * When restoring the screen brightness at boot, stay away from
5354 the darkest setting or from the lowest 5% of the available
5355 range, depending on which is the larger value of both. This
5356 should effectively protect the user from rebooting into a
5357 black screen, should the brightness have been set to minimum
5360 * sd-login gained a new sd_machine_get_class() call to
5361 determine the class ("vm" or "container") of a machine
5362 registered with machined.
5364 * sd-login gained new calls
5365 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5366 to query the identity of the peer of a local AF_UNIX
5367 connection. They operate similarly to their sd_pid_get_xyz()
5370 * PID 1 will now maintain a system-wide system state engine
5371 with the states "starting", "running", "degraded",
5372 "maintenance", "stopping". These states are bound to system
5373 startup, normal runtime, runtime with at least one failed
5374 service, rescue/emergency mode and system shutdown. This
5375 state is shown in the "systemctl status" output when no unit
5376 name is passed. It is useful to determine system state, in
5377 particularly when doing so for many systems or containers at
5380 * A new command "list-machines" has been added to "systemctl"
5381 that lists all local OS containers and shows their system
5382 state (see above), if systemd runs inside of them.
5384 * systemctl gained a new "-r" switch to recursively enumerate
5385 units on all local containers, when used with the
5386 "list-unit" command (which is the default one that is
5387 executed when no parameters are specified).
5389 * The GPT automatic partition discovery logic will now honour
5390 two GPT partition flags: one may be set on a partition to
5391 cause it to be mounted read-only, and the other may be set
5392 on a partition to ignore it during automatic discovery.
5394 * Two new GPT type UUIDs have been added for automatic root
5395 partition discovery, for 32-bit and 64-bit ARM. This is not
5396 particularly useful for discovering the root directory on
5397 these architectures during bare-metal boots (since UEFI is
5398 not common there), but still very useful to allow booting of
5399 ARM disk images in nspawn with the -i option.
5401 * MAC addresses of interfaces created with nspawn's
5402 --network-interface= switch will now be generated from the
5403 machine name, and thus be stable between multiple invocations
5406 * logind will now automatically remove all IPC objects owned
5407 by a user if she or he fully logs out. This makes sure that
5408 users who are logged out cannot continue to consume IPC
5409 resources. This covers SysV memory, semaphores and message
5410 queues as well as POSIX shared memory and message
5411 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5412 limits. With this functionality, that is corrected. This may
5413 be turned off by using the RemoveIPC= switch of logind.conf.
5415 * The systemd-machine-id-setup and tmpfiles tools gained a
5416 --root= switch to operate on a specific root directory,
5419 * journald can now forward logged messages to the TTYs of all
5420 logged in users ("wall"). This is the default for all
5421 emergency messages now.
5423 * A new tool systemd-journal-remote has been added to stream
5424 journal log messages across the network.
5426 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5427 controller trees are mounted into it. Note that the
5428 directories mounted beneath it are not read-only. This is a
5429 security measure and is particularly useful because glibc
5430 actually includes a search logic to pick any tmpfs it can
5431 find to implement shm_open() if /dev/shm is not available
5432 (which it might very well be in namespaced setups).
5434 * machinectl gained a new "poweroff" command to cleanly power
5435 down a local OS container.
5437 * The PrivateDevices= unit file setting will now also drop the
5438 CAP_MKNOD capability from the capability bound set, and
5439 imply DevicePolicy=closed.
5441 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5442 comprehensively on all long-running systemd services where
5443 this is appropriate.
5445 * systemd-udevd will now run in a disassociated mount
5446 namespace. To mount directories from udev rules, make sure to
5447 pull in mount units via SYSTEMD_WANTS properties.
5449 * The kdbus support gained support for uploading policy into
5450 the kernel. sd-bus gained support for creating "monitoring"
5451 connections that can eavesdrop into all bus communication
5452 for debugging purposes.
5454 * Timestamps may now be specified in seconds since the UNIX
5455 epoch Jan 1st, 1970 by specifying "@" followed by the value
5458 * Native tcpwrap support in systemd has been removed. tcpwrap
5459 is old code, not really maintained anymore and has serious
5460 shortcomings, and better options such as firewalls
5461 exist. For setups that require tcpwrap usage, please
5462 consider invoking your socket-activated service via tcpd,
5463 like on traditional inetd.
5465 * A new system.conf configuration option
5466 DefaultTimerAccuracySec= has been added that controls the
5467 default AccuracySec= setting of .timer units.
5469 * Timer units gained a new WakeSystem= switch. If enabled,
5470 timers configured this way will cause the system to resume
5471 from system suspend (if the system supports that, which most
5474 * Timer units gained a new Persistent= switch. If enabled,
5475 timers configured this way will save to disk when they have
5476 been last triggered. This information is then used on next
5477 reboot to possible execute overdue timer events, that
5478 could not take place because the system was powered off.
5479 This enables simple anacron-like behaviour for timer units.
5481 * systemctl's "list-timers" will now also list the time a
5482 timer unit was last triggered in addition to the next time
5483 it will be triggered.
5485 * systemd-networkd will now assign predictable IPv4LL
5486 addresses to its local interfaces.
5488 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
5489 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
5490 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
5491 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
5492 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
5493 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
5494 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
5495 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
5498 — Berlin, 2014-03-25
5502 * A new unit file setting RestrictAddressFamilies= has been
5503 added to restrict which socket address families unit
5504 processes gain access to. This takes address family names
5505 like "AF_INET" or "AF_UNIX", and is useful to minimize the
5506 attack surface of services via exotic protocol stacks. This
5507 is built on seccomp system call filters.
5509 * Two new unit file settings RuntimeDirectory= and
5510 RuntimeDirectoryMode= have been added that may be used to
5511 manage a per-daemon runtime directories below /run. This is
5512 an alternative for setting up directory permissions with
5513 tmpfiles snippets, and has the advantage that the runtime
5514 directory's lifetime is bound to the daemon runtime and that
5515 the daemon starts up with an empty directory each time. This
5516 is particularly useful when writing services that drop
5517 privileges using the User= or Group= setting.
5519 * The DeviceAllow= unit setting now supports globbing for
5520 matching against device group names.
5522 * The systemd configuration file system.conf gained new
5523 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
5524 DefaultMemoryAccounting= to globally turn on/off accounting
5525 for specific resources (cgroups) for all units. These
5526 settings may still be overridden individually in each unit
5529 * systemd-gpt-auto-generator is now able to discover /srv and
5530 root partitions in addition to /home and swap partitions. It
5531 also supports LUKS-encrypted partitions now. With this in
5532 place, automatic discovery of partitions to mount following
5533 the Discoverable Partitions Specification
5534 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
5535 is now a lot more complete. This allows booting without
5536 /etc/fstab and without root= on the kernel command line on
5537 systems prepared appropriately.
5539 * systemd-nspawn gained a new --image= switch which allows
5540 booting up disk images and Linux installations on any block
5541 device that follow the Discoverable Partitions Specification
5542 (see above). This means that installations made with
5543 appropriately updated installers may now be started and
5544 deployed using container managers, completely
5545 unmodified. (We hope that libvirt-lxc will add support for
5546 this feature soon, too.)
5548 * systemd-nspawn gained a new --network-macvlan= setting to
5549 set up a private macvlan interface for the
5550 container. Similarly, systemd-networkd gained a new
5551 Kind=macvlan setting in .netdev files.
5553 * systemd-networkd now supports configuring local addresses
5556 * A new tool systemd-network-wait-online has been added to
5557 synchronously wait for network connectivity using
5560 * The sd-bus.h bus API gained a new sd_bus_track object for
5561 tracking the life-cycle of bus peers. Note that sd-bus.h is
5562 still not a public API though (unless you specify
5563 --enable-kdbus on the configure command line, which however
5564 voids your warranty and you get no API stability guarantee).
5566 * The $XDG_RUNTIME_DIR runtime directories for each user are
5567 now individual tmpfs instances, which has the benefit of
5568 introducing separate pools for each user, with individual
5569 size limits, and thus making sure that unprivileged clients
5570 can no longer negatively impact the system or other users by
5571 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
5572 RuntimeDirectorySize= has been introduced that allows
5573 controlling the default size limit for all users. It
5574 defaults to 10% of the available physical memory. This is no
5575 replacement for quotas on tmpfs though (which the kernel
5576 still does not support), as /dev/shm and /tmp are still
5577 shared resources used by both the system and unprivileged
5580 * logind will now automatically turn off automatic suspending
5581 on laptop lid close when more than one display is
5582 connected. This was previously expected to be implemented
5583 individually in desktop environments (such as GNOME),
5584 however has been added to logind now, in order to fix a
5585 boot-time race where a desktop environment might not have
5586 been started yet and thus not been able to take an inhibitor
5587 lock at the time where logind already suspends the system
5588 due to a closed lid.
5590 * logind will now wait at least 30s after each system
5591 suspend/resume cycle, and 3min after system boot before
5592 suspending the system due to a closed laptop lid. This
5593 should give USB docking stations and similar enough time to
5594 be probed and configured after system resume and boot in
5595 order to then act as suspend blocker.
5597 * systemd-run gained a new --property= setting which allows
5598 initialization of resource control properties (and others)
5599 for the created scope or service unit. Example: "systemd-run
5600 --property=BlockIOWeight=10 updatedb" may be used to run
5601 updatedb at a low block IO scheduling weight.
5603 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
5604 now also work in --scope mode.
5606 * When systemd is compiled with kdbus support, basic support
5607 for enforced policies is now in place. (Note that enabling
5608 kdbus still voids your warranty and no API compatibility
5611 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
5612 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5613 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
5614 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
5615 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
5616 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
5617 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
5618 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
5619 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
5620 Zbigniew Jędrzejewski-Szmek
5622 — Berlin, 2014-03-12
5626 * systemd will now relabel /dev after loading the SMACK policy
5627 according to SMACK rules.
5629 * A new unit file option AppArmorProfile= has been added to
5630 set the AppArmor profile for the processes of a unit.
5632 * A new condition check ConditionArchitecture= has been added
5633 to conditionalize units based on the system architecture, as
5634 reported by uname()'s "machine" field.
5636 * systemd-networkd now supports matching on the system
5637 virtualization, architecture, kernel command line, host name
5640 * logind is now a lot more aggressive when suspending the
5641 machine due to a closed laptop lid. Instead of acting only
5642 on the lid close action, it will continuously watch the lid
5643 status and act on it. This is useful for laptops where the
5644 power button is on the outside of the chassis so that it can
5645 be reached without opening the lid (such as the Lenovo
5646 Yoga). On those machines, logind will now immediately
5647 re-suspend the machine if the power button has been
5648 accidentally pressed while the laptop was suspended and in a
5649 backpack or similar.
5651 * logind will now watch SW_DOCK switches and inhibit reaction
5652 to the lid switch if it is pressed. This means that logind
5653 will not suspend the machine anymore if the lid is closed
5654 and the system is docked, if the laptop supports SW_DOCK
5655 notifications via the input layer. Note that ACPI docking
5656 stations do not generate this currently. Also note that this
5657 logic is usually not fully sufficient and Desktop
5658 Environments should take a lid switch inhibitor lock when an
5659 external display is connected, as systemd will not watch
5662 * nspawn will now make use of the devices cgroup controller by
5663 default, and only permit creation of and access to the usual
5664 API device nodes like /dev/null or /dev/random, as well as
5665 access to (but not creation of) the pty devices.
5667 * We will now ship a default .network file for
5668 systemd-networkd that automatically configures DHCP for
5669 network interfaces created by nspawn's --network-veth or
5670 --network-bridge= switches.
5672 * systemd will now understand the usual M, K, G, T suffixes
5673 according to SI conventions (i.e. to the base 1000) when
5674 referring to throughput and hardware metrics. It will stay
5675 with IEC conventions (i.e. to the base 1024) for software
5676 metrics, according to what is customary according to
5677 Wikipedia. We explicitly document which base applies for
5678 each configuration option.
5680 * The DeviceAllow= setting in unit files now supports a syntax
5681 to whitelist an entire group of devices node majors at once,
5682 based on the /proc/devices listing. For example, with the
5683 string "char-pts", it is now possible to whitelist all
5684 current and future pseudo-TTYs at once.
5686 * sd-event learned a new "post" event source. Event sources of
5687 this type are triggered by the dispatching of any event
5688 source of a type that is not "post". This is useful for
5689 implementing clean-up and check event sources that are
5690 triggered by other work being done in the program.
5692 * systemd-networkd is no longer statically enabled, but uses
5693 the usual [Install] sections so that it can be
5694 enabled/disabled using systemctl. It still is enabled by
5697 * When creating a veth interface pair with systemd-nspawn, the
5698 host side will now be prefixed with "vb-" if
5699 --network-bridge= is used, and with "ve-" if --network-veth
5700 is used. This way, it is easy to distinguish these cases on
5701 the host, for example to apply different configuration to
5702 them with systemd-networkd.
5704 * The compatibility libraries for libsystemd-journal.so,
5705 libsystem-id128.so, libsystemd-login.so and
5706 libsystemd-daemon.so do not make use of IFUNC
5707 anymore. Instead, we now build libsystemd.so multiple times
5708 under these alternative names. This means that the footprint
5709 is drastically increased, but given that these are
5710 transitional compatibility libraries, this should not matter
5711 much. This change has been made necessary to support the ARM
5712 platform for these compatibility libraries, as the ARM
5713 toolchain is not really at the same level as the toolchain
5714 for other architectures like x86 and does not support
5715 IFUNC. Please make sure to use --enable-compat-libs only
5716 during a transitional period!
5718 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
5719 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
5720 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
5721 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
5722 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
5723 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
5724 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
5725 Zbigniew Jędrzejewski-Szmek
5727 — Berlin, 2014-02-24
5731 * A new component "systemd-networkd" has been added that can
5732 be used to configure local network interfaces statically or
5733 via DHCP. It is capable of bringing up bridges, VLANs, and
5734 bonding. Currently, no hook-ups for interactive network
5735 configuration are provided. Use this for your initrd,
5736 container, embedded, or server setup if you need a simple,
5737 yet powerful, network configuration solution. This
5738 configuration subsystem is quite nifty, as it allows wildcard
5739 hotplug matching in interfaces. For example, with a single
5740 configuration snippet, you can configure that all Ethernet
5741 interfaces showing up are automatically added to a bridge,
5742 or similar. It supports link-sensing and more.
5744 * A new tool "systemd-socket-proxyd" has been added which can
5745 act as a bidirectional proxy for TCP sockets. This is
5746 useful for adding socket activation support to services that
5747 do not actually support socket activation, including virtual
5748 machines and the like.
5750 * Add a new tool to save/restore rfkill state on
5753 * Save/restore state of keyboard backlights in addition to
5754 display backlights on shutdown/boot.
5756 * udev learned a new SECLABEL{} construct to label device
5757 nodes with a specific security label when they appear. For
5758 now, only SECLABEL{selinux} is supported, but the syntax is
5759 prepared for additional security frameworks.
5761 * udev gained a new scheme to configure link-level attributes
5762 from files in /etc/systemd/network/*.link. These files can
5763 match against MAC address, device path, driver name and type,
5764 and will apply attributes like the naming policy, link speed,
5765 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
5766 address assignment policy (randomized, ...).
5768 * The configuration of network interface naming rules for
5769 "permanent interface names" has changed: a new NamePolicy=
5770 setting in the [Link] section of .link files determines the
5771 priority of possible naming schemes (onboard, slot, MAC,
5772 path). The default value of this setting is determined by
5773 /usr/lib/net/links/99-default.link. Old
5774 80-net-name-slot.rules udev configuration file has been
5775 removed, so local configuration overriding this file should
5776 be adapted to override 99-default.link instead.
5778 * When the User= switch is used in a unit file, also
5779 initialize $SHELL= based on the user database entry.
5781 * systemd no longer depends on libdbus. All communication is
5782 now done with sd-bus, systemd's low-level bus library
5785 * kdbus support has been added to PID 1 itself. When kdbus is
5786 enabled, this causes PID 1 to set up the system bus and
5787 enable support for a new ".busname" unit type that
5788 encapsulates bus name activation on kdbus. It works a little
5789 bit like ".socket" units, except for bus names. A new
5790 generator has been added that converts classic dbus1 service
5791 activation files automatically into native systemd .busname
5794 * sd-bus: add a light-weight vtable implementation that allows
5795 defining objects on the bus with a simple static const
5796 vtable array of its methods, signals and properties.
5798 * systemd will not generate or install static dbus
5799 introspection data anymore to /usr/share/dbus-1/interfaces,
5800 as the precise format of these files is unclear, and
5801 nothing makes use of it.
5803 * A proxy daemon is now provided to proxy clients connecting
5804 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
5805 compatibility with classic D-Bus.
5807 * A bus driver implementation has been added that supports the
5808 classic D-Bus bus driver calls on kdbus, also for
5809 compatibility purposes.
5811 * A new API "sd-event.h" has been added that implements a
5812 minimal event loop API built around epoll. It provides a
5813 couple of features that direct epoll usage is lacking:
5814 prioritization of events, scales to large numbers of timer
5815 events, per-event timer slack (accuracy), system-wide
5816 coalescing of timer events, exit handlers, watchdog
5817 supervision support using systemd's sd_notify() API, child
5820 * A new API "sd-rntl.h" has been added that provides an API
5821 around the route netlink interface of the kernel, similar in
5822 style to "sd-bus.h".
5824 * A new API "sd-dhcp-client.h" has been added that provides a
5825 small DHCPv4 client-side implementation. This is used by
5828 * There is a new kernel command line option
5829 "systemd.restore_state=0|1". When set to "0", none of the
5830 systemd tools will restore saved runtime state to hardware
5831 devices. More specifically, the rfkill and backlight states
5834 * The FsckPassNo= compatibility option in mount/service units
5835 has been removed. The fstab generator will now add the
5836 necessary dependencies automatically, and does not require
5837 PID1's support for that anymore.
5839 * journalctl gained a new switch, --list-boots, that lists
5840 recent boots with their times and boot IDs.
5842 * The various tools like systemctl, loginctl, timedatectl,
5843 busctl, systemd-run, ... have gained a new switch "-M" to
5844 connect to a specific, local OS container (as direct
5845 connection, without requiring SSH). This works on any
5846 container that is registered with machined, such as those
5847 created by libvirt-lxc or nspawn.
5849 * systemd-run and systemd-analyze also gained support for "-H"
5850 to connect to remote hosts via SSH. This is particularly
5851 useful for systemd-run because it enables queuing of jobs
5852 onto remote systems.
5854 * machinectl gained a new command "login" to open a getty
5855 login in any local container. This works with any container
5856 that is registered with machined (such as those created by
5857 libvirt-lxc or nspawn), and which runs systemd inside.
5859 * machinectl gained a new "reboot" command that may be used to
5860 trigger a reboot on a specific container that is registered
5861 with machined. This works on any container that runs an init
5862 system of some kind.
5864 * systemctl gained a new "list-timers" command to print a nice
5865 listing of installed timer units with the times they elapse
5868 * Alternative reboot() parameters may now be specified on the
5869 "systemctl reboot" command line and are passed to the
5870 reboot() system call.
5872 * systemctl gained a new --job-mode= switch to configure the
5873 mode to queue a job with. This is a more generic version of
5874 --fail, --irreversible, and --ignore-dependencies, which are
5875 still available but not advertised anymore.
5877 * /etc/systemd/system.conf gained new settings to configure
5878 various default timeouts of units, as well as the default
5879 start limit interval and burst. These may still be overridden
5882 * PID1 will now export on the bus profile data of the security
5883 policy upload process (such as the SELinux policy upload to
5886 * journald: when forwarding logs to the console, include
5887 timestamps (following the setting in
5888 /sys/module/printk/parameters/time).
5890 * OnCalendar= in timer units now understands the special
5891 strings "yearly" and "annually". (Both are equivalent)
5893 * The accuracy of timer units is now configurable with the new
5894 AccuracySec= setting. It defaults to 1min.
5896 * A new dependency type JoinsNamespaceOf= has been added that
5897 allows running two services within the same /tmp and network
5898 namespace, if PrivateNetwork= or PrivateTmp= are used.
5900 * A new command "cat" has been added to systemctl. It outputs
5901 the original unit file of a unit, and concatenates the
5902 contents of additional "drop-in" unit file snippets, so that
5903 the full configuration is shown.
5905 * systemctl now supports globbing on the various "list-xyz"
5906 commands, like "list-units" or "list-sockets", as well as on
5907 those commands which take multiple unit names.
5909 * journalctl's --unit= switch gained support for globbing.
5911 * All systemd daemons now make use of the watchdog logic so
5912 that systemd automatically notices when they hang.
5914 * If the $container_ttys environment variable is set,
5915 getty-generator will automatically spawn a getty for each
5916 listed tty. This is useful for container managers to request
5917 login gettys to be spawned on as many ttys as needed.
5919 * %h, %s, %U specifier support is not available anymore when
5920 used in unit files for PID 1. This is because NSS calls are
5921 not safe from PID 1. They stay available for --user
5922 instances of systemd, and as special case for the root user.
5924 * loginctl gained a new "--no-legend" switch to turn off output
5927 * The "sd-login.h" API gained three new calls:
5928 sd_session_is_remote(), sd_session_get_remote_user(),
5929 sd_session_get_remote_host() to query information about
5932 * The udev hardware database now also carries vendor/product
5933 information of SDIO devices.
5935 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
5936 determine whether watchdog notifications are requested by
5939 * Socket-activated per-connection services now include a
5940 short description of the connection parameters in the
5943 * tmpfiles gained a new "--boot" option. When this is not used,
5944 only lines where the command character is not suffixed with
5945 "!" are executed. When this option is specified, those
5946 options are executed too. This partitions tmpfiles
5947 directives into those that can be safely executed at any
5948 time, and those which should be run only at boot (for
5949 example, a line that creates /run/nologin).
5951 * A new API "sd-resolve.h" has been added which provides a simple
5952 asynchronous wrapper around glibc NSS host name resolution
5953 calls, such as getaddrinfo(). In contrast to glibc's
5954 getaddrinfo_a(), it does not use signals. In contrast to most
5955 other asynchronous name resolution libraries, this one does
5956 not reimplement DNS, but reuses NSS, so that alternate
5957 host name resolution systems continue to work, such as mDNS,
5958 LDAP, etc. This API is based on libasyncns, but it has been
5959 cleaned up for inclusion in systemd.
5961 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
5962 "sd-daemon.h" are no longer found in individual libraries
5963 libsystemd-journal.so, libsystemd-login.so,
5964 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
5965 merged them into a single library, libsystemd.so, which
5966 provides all symbols. The reason for this is cyclic
5967 dependencies, as these libraries tend to use each other's
5968 symbols. So far, we have managed to workaround that by linking
5969 a copy of a good part of our code into each of these
5970 libraries again and again, which, however, makes certain
5971 things hard to do, like sharing static variables. Also, it
5972 substantially increases footprint. With this change, there
5973 is only one library for the basic APIs systemd
5974 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
5975 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
5976 library as well, however are subject to the --enable-kdbus
5977 switch (see below). Note that "sd-dhcp-client.h" is not part
5978 of this library (this is because it only consumes, never
5979 provides, services of/to other APIs). To make the transition
5980 easy from the separate libraries to the unified one, we
5981 provide the --enable-compat-libs compile-time switch which
5982 will generate stub libraries that are compatible with the
5983 old ones but redirect all calls to the new one.
5985 * All of the kdbus logic and the new APIs "sd-bus.h",
5986 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
5987 and "sd-utf8.h" are compile-time optional via the
5988 "--enable-kdbus" switch, and they are not compiled in by
5989 default. To make use of kdbus, you have to explicitly enable
5990 the switch. Note however, that neither the kernel nor the
5991 userspace API for all of this is considered stable yet. We
5992 want to maintain the freedom to still change the APIs for
5993 now. By specifying this build-time switch, you acknowledge
5994 that you are aware of the instability of the current
5997 * Also, note that while kdbus is pretty much complete,
5998 it lacks one thing: proper policy support. This means you
5999 can build a fully working system with all features; however,
6000 it will be highly insecure. Policy support will be added in
6001 one of the next releases, at the same time that we will
6002 declare the APIs stable.
6004 * When the kernel command line argument "kdbus" is specified,
6005 systemd will automatically load the kdbus.ko kernel module. At
6006 this stage of development, it is only useful for testing kdbus
6007 and should not be used in production. Note: if "--enable-kdbus"
6008 is specified, and the kdbus.ko kernel module is available, and
6009 "kdbus" is added to the kernel command line, the entire system
6010 runs with kdbus instead of dbus-daemon, with the above mentioned
6011 problem of missing the system policy enforcement. Also a future
6012 version of kdbus.ko or a newer systemd will not be compatible with
6013 each other, and will unlikely be able to boot the machine if only
6014 one of them is updated.
6016 * systemctl gained a new "import-environment" command which
6017 uploads the caller's environment (or parts thereof) into the
6018 service manager so that it is inherited by services started
6019 by the manager. This is useful to upload variables like
6020 $DISPLAY into the user service manager.
6022 * A new PrivateDevices= switch has been added to service units
6023 which allows running a service with a namespaced /dev
6024 directory that does not contain any device nodes for
6025 physical devices. More specifically, it only includes devices
6026 such as /dev/null, /dev/urandom, and /dev/zero which are API
6029 * logind has been extended to support behaviour like VT
6030 switching on seats that do not support a VT. This makes
6031 multi-session available on seats that are not the first seat
6032 (seat0), and on systems where kernel support for VTs has
6033 been disabled at compile-time.
6035 * If a process holds a delay lock for system sleep or shutdown
6036 and fails to release it in time, we will now log its
6037 identity. This makes it easier to identify processes that
6038 cause slow suspends or power-offs.
6040 * When parsing /etc/crypttab, support for a new key-slot=
6041 option as supported by Debian is added. It allows indicating
6042 which LUKS slot to use on disk, speeding up key loading.
6044 * The sd_journal_sendv() API call has been checked and
6045 officially declared to be async-signal-safe so that it may
6046 be invoked from signal handlers for logging purposes.
6048 * Boot-time status output is now enabled automatically after a
6049 short timeout if boot does not progress, in order to give
6050 the user an indication what she or he is waiting for.
6052 * The boot-time output has been improved to show how much time
6053 remains until jobs expire.
6055 * The KillMode= switch in service units gained a new possible
6056 value "mixed". If set, and the unit is shut down, then the
6057 initial SIGTERM signal is sent only to the main daemon
6058 process, while the following SIGKILL signal is sent to
6059 all remaining processes of the service.
6061 * When a scope unit is registered, a new property "Controller"
6062 may be set. If set to a valid bus name, systemd will send a
6063 RequestStop() signal to this name when it would like to shut
6064 down the scope. This may be used to hook manager logic into
6065 the shutdown logic of scope units. Also, scope units may now
6066 be put in a special "abandoned" state, in which case the
6067 manager process which created them takes no further
6068 responsibilities for it.
6070 * When reading unit files, systemd will now verify
6071 the access mode of these files, and warn about certain
6072 suspicious combinations. This has been added to make it
6073 easier to track down packaging bugs where unit files are
6074 marked executable or world-writable.
6076 * systemd-nspawn gained a new "--setenv=" switch to set
6077 container-wide environment variables. The similar option in
6078 systemd-activate was renamed from "--environment=" to
6079 "--setenv=" for consistency.
6081 * systemd-nspawn has been updated to create a new kdbus domain
6082 for each container that is invoked, thus allowing each
6083 container to have its own set of system and user buses,
6084 independent of the host.
6086 * systemd-nspawn gained a new --drop-capability= switch to run
6087 the container with less capabilities than the default. Both
6088 --drop-capability= and --capability= now take the special
6089 string "all" for dropping or keeping all capabilities.
6091 * systemd-nspawn gained new switches for executing containers
6092 with specific SELinux labels set.
6094 * systemd-nspawn gained a new --quiet switch to not generate
6095 any additional output but the container's own console
6098 * systemd-nspawn gained a new --share-system switch to run a
6099 container without PID namespacing enabled.
6101 * systemd-nspawn gained a new --register= switch to control
6102 whether the container is registered with systemd-machined or
6103 not. This is useful for containers that do not run full
6104 OS images, but only specific apps.
6106 * systemd-nspawn gained a new --keep-unit which may be used
6107 when invoked as the only program from a service unit, and
6108 results in registration of the unit service itself in
6109 systemd-machined, instead of a newly opened scope unit.
6111 * systemd-nspawn gained a new --network-interface= switch for
6112 moving arbitrary interfaces to the container. The new
6113 --network-veth switch creates a virtual Ethernet connection
6114 between host and container. The new --network-bridge=
6115 switch then allows assigning the host side of this virtual
6116 Ethernet connection to a bridge device.
6118 * systemd-nspawn gained a new --personality= switch for
6119 setting the kernel personality for the container. This is
6120 useful when running a 32-bit container on a 64-bit host. A
6121 similar option Personality= is now also available for service
6124 * logind will now also track a "Desktop" identifier for each
6125 session which encodes the desktop environment of it. This is
6126 useful for desktop environments that want to identify
6127 multiple running sessions of itself easily.
6129 * A new SELinuxContext= setting for service units has been
6130 added that allows setting a specific SELinux execution
6131 context for a service.
6133 * Most systemd client tools will now honour $SYSTEMD_LESS for
6134 settings of the "less" pager. By default, these tools will
6135 override $LESS to allow certain operations to work, such as
6136 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
6137 influence this logic.
6139 * systemd's "seccomp" hook-up has been changed to make use of
6140 the libseccomp library instead of using its own
6141 implementation. This has benefits for portability among
6144 * For usage together with SystemCallFilter=, a new
6145 SystemCallErrorNumber= setting has been introduced that
6146 allows configuration of a system error number to be returned
6147 on filtered system calls, instead of immediately killing the
6148 process. Also, SystemCallArchitectures= has been added to
6149 limit access to system calls of a particular architecture
6150 (in order to turn off support for unused secondary
6151 architectures). There is also a global
6152 SystemCallArchitectures= setting in system.conf now to turn
6153 off support for non-native system calls system-wide.
6155 * systemd requires a kernel with a working name_to_handle_at(),
6156 please see the kernel config requirements in the README file.
6158 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6159 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6160 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6161 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6162 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6163 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6164 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6165 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6166 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6167 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6168 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6169 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6170 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6171 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6172 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6173 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6174 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6175 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6176 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6177 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6178 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6179 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6180 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6181 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6183 — Berlin, 2014-02-20
6187 * logind has gained support for facilitating privileged input
6188 and drm device access for unprivileged clients. This work is
6189 useful to allow Wayland display servers (and similar
6190 programs, such as kmscon) to run under the user's ID and
6191 access input and drm devices which are normally
6192 protected. When this is used (and the kernel is new enough)
6193 logind will "mute" IO on the file descriptors passed to
6194 Wayland as long as it is in the background and "unmute" it
6195 if it returns into the foreground. This allows secure
6196 session switching without allowing background sessions to
6197 eavesdrop on input and display data. This also introduces
6198 session switching support if VT support is turned off in the
6199 kernel, and on seats that are not seat0.
6201 * A new kernel command line option luks.options= is understood
6202 now which allows specifying LUKS options for usage for LUKS
6203 encrypted partitions specified with luks.uuid=.
6205 * tmpfiles.d(5) snippets may now use specifier expansion in
6206 path names. More specifically %m, %b, %H, %v, are now
6207 replaced by the local machine id, boot id, hostname, and
6208 kernel version number.
6210 * A new tmpfiles.d(5) command "m" has been introduced which
6211 may be used to change the owner/group/access mode of a file
6212 or directory if it exists, but do nothing if it does not.
6214 * This release removes high-level support for the
6215 MemorySoftLimit= cgroup setting. The underlying kernel
6216 cgroup attribute memory.soft_limit= is currently badly
6217 designed and likely to be removed from the kernel API in its
6218 current form, hence we should not expose it for now.
6220 * The memory.use_hierarchy cgroup attribute is now enabled for
6221 all cgroups systemd creates in the memory cgroup
6222 hierarchy. This option is likely to be come the built-in
6223 default in the kernel anyway, and the non-hierarchical mode
6224 never made much sense in the intrinsically hierarchical
6227 * A new field _SYSTEMD_SLICE= is logged along with all journal
6228 messages containing the slice a message was generated
6229 from. This is useful to allow easy per-customer filtering of
6230 logs among other things.
6232 * systemd-journald will no longer adjust the group of journal
6233 files it creates to the "systemd-journal" group. Instead we
6234 rely on the journal directory to be owned by the
6235 "systemd-journal" group, and its setgid bit set, so that the
6236 kernel file system layer will automatically enforce that
6237 journal files inherit this group assignment. The reason for
6238 this change is that we cannot allow NSS look-ups from
6239 journald which would be necessary to resolve
6240 "systemd-journal" to a numeric GID, because this might
6241 create deadlocks if NSS involves synchronous queries to
6242 other daemons (such as nscd, or sssd) which in turn are
6243 logging clients of journald and might block on it, which
6244 would then dead lock. A tmpfiles.d(5) snippet included in
6245 systemd will make sure the setgid bit and group are
6246 properly set on the journal directory if it exists on every
6247 boot. However, we recommend adjusting it manually after
6248 upgrades too (or from RPM scriptlets), so that the change is
6249 not delayed until next reboot.
6251 * Backlight and random seed files in /var/lib/ have moved into
6252 the /var/lib/systemd/ directory, in order to centralize all
6253 systemd generated files in one directory.
6255 * Boot time performance measurements (as displayed by
6256 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6257 performance information if that's available to determine how
6258 much time BIOS and boot loader initialization required. With
6259 a sufficiently new BIOS you hence no longer need to boot
6260 with Gummiboot to get access to such information.
6262 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6263 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6264 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6265 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6266 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6267 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6268 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6270 — Berlin, 2013-10-02
6274 * The Restart= option for services now understands a new
6275 on-watchdog setting, which will restart the service
6276 automatically if the service stops sending out watchdog keep
6277 alive messages (as configured with WatchdogSec=).
6279 * The getty generator (which is responsible for bringing up a
6280 getty on configured serial consoles) will no longer only
6281 start a getty on the primary kernel console but on all
6282 others, too. This makes the order in which console= is
6283 specified on the kernel command line less important.
6285 * libsystemd-logind gained a new sd_session_get_vt() call to
6286 retrieve the VT number of a session.
6288 * If the option "tries=0" is set for an entry of /etc/crypttab
6289 its passphrase is queried indefinitely instead of any
6290 maximum number of tries.
6292 * If a service with a configure PID file terminates its PID
6293 file will now be removed automatically if it still exists
6294 afterwards. This should put an end to stale PID files.
6296 * systemd-run will now also take relative binary path names
6297 for execution and no longer insists on absolute paths.
6299 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6300 paths that are optionally prefixed with "-" to indicate that
6301 it should not be considered a failure if they do not exist.
6303 * journalctl -o (and similar commands) now understands a new
6304 output mode "short-precise", it is similar to "short" but
6305 shows timestamps with usec accuracy.
6307 * The option "discard" (as known from Debian) is now
6308 synonymous to "allow-discards" in /etc/crypttab. In fact,
6309 "discard" is preferred now (since it is easier to remember
6312 * Some licensing clean-ups were made, so that more code is now
6313 LGPL-2.1 licensed than before.
6315 * A minimal tool to save/restore the display backlight
6316 brightness across reboots has been added. It will store the
6317 backlight setting as late as possible at shutdown, and
6318 restore it as early as possible during reboot.
6320 * A logic to automatically discover and enable home and swap
6321 partitions on GPT disks has been added. With this in place
6322 /etc/fstab becomes optional for many setups as systemd can
6323 discover certain partitions located on the root disk
6324 automatically. Home partitions are recognized under their
6325 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6326 partitions are recognized under their GPT type ID
6327 0657fd6da4ab43c484e50933c84b4f4f.
6329 * systemd will no longer pass any environment from the kernel
6330 or initrd to system services. If you want to set an
6331 environment for all services, do so via the kernel command
6332 line systemd.setenv= assignment.
6334 * The systemd-sysctl tool no longer natively reads the file
6335 /etc/sysctl.conf. If desired, the file should be symlinked
6336 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6337 legacy support by a symlink rather than built-in code, it
6338 also makes the otherwise hidden order of application of the
6339 different files visible. (Note that this partly reverts to a
6340 pre-198 application order of sysctl knobs!)
6342 * The "systemctl set-log-level" and "systemctl dump" commands
6343 have been moved to systemd-analyze.
6345 * systemd-run learned the new --remain-after-exit switch,
6346 which causes the scope unit not to be cleaned up
6347 automatically after the process terminated.
6349 * tmpfiles learned a new --exclude-prefix= switch to exclude
6350 certain paths from operation.
6352 * journald will now automatically flush all messages to disk
6353 as soon as a message at the log level CRIT, ALERT or EMERG
6356 Contributions from: Andrew Cook, Brandon Philips, Christian
6357 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6358 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6359 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6360 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6361 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6362 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6363 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6364 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6365 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6366 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6367 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6368 William Giokas, Zbigniew Jędrzejewski-Szmek
6370 — Berlin, 2013-09-13
6374 * The documentation has been updated to cover the various new
6375 concepts introduced with 205.
6377 * Unit files now understand the new %v specifier which
6378 resolves to the kernel version string as returned by "uname
6381 * systemctl now supports filtering the unit list output by
6382 load state, active state and sub state, using the new
6385 * "systemctl status" will now show the results of the
6386 condition checks (like ConditionPathExists= and similar) of
6387 the last start attempts of the unit. They are also logged to
6390 * "journalctl -b" may now be used to look for boot output of a
6391 specific boot. Try "journalctl -b -1" for the previous boot,
6392 but the syntax is substantially more powerful.
6394 * "journalctl --show-cursor" has been added which prints the
6395 cursor string the last shown log line. This may then be used
6396 with the new "journalctl --after-cursor=" switch to continue
6397 browsing logs from that point on.
6399 * "journalctl --force" may now be used to force regeneration
6402 * Creation of "dead" device nodes has been moved from udev
6403 into kmod and tmpfiles. Previously, udev would read the kmod
6404 databases to pre-generate dead device nodes based on meta
6405 information contained in kernel modules, so that these would
6406 be auto-loaded on access rather then at boot. As this
6407 does not really have much to do with the exposing actual
6408 kernel devices to userspace this has always been slightly
6409 alien in the udev codebase. Following the new scheme kmod
6410 will now generate a runtime snippet for tmpfiles from the
6411 module meta information and it now is tmpfiles' job to the
6412 create the nodes. This also allows overriding access and
6413 other parameters for the nodes using the usual tmpfiles
6414 facilities. As side effect this allows us to remove the
6415 CAP_SYS_MKNOD capability bit from udevd entirely.
6417 * logind's device ACLs may now be applied to these "dead"
6418 devices nodes too, thus finally allowing managed access to
6419 devices such as /dev/snd/sequencer without loading the
6420 backing module right-away.
6422 * A new RPM macro has been added that may be used to apply
6423 tmpfiles configuration during package installation.
6425 * systemd-detect-virt and ConditionVirtualization= now can
6426 detect User-Mode-Linux machines (UML).
6428 * journald will now implicitly log the effective capabilities
6429 set of processes in the message metadata.
6431 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6433 * The initrd interface has been simplified (more specifically,
6434 support for passing performance data via environment
6435 variables and fsck results via files in /run has been
6436 removed). These features were non-essential, and are
6437 nowadays available in a much nicer way by having systemd in
6438 the initrd serialize its state and have the hosts systemd
6439 deserialize it again.
6441 * The udev "keymap" data files and tools to apply keyboard
6442 specific mappings of scan to key codes, and force-release
6443 scan code lists have been entirely replaced by a udev
6444 "keyboard" builtin and a hwdb data file.
6446 * systemd will now honour the kernel's "quiet" command line
6447 argument also during late shutdown, resulting in a
6448 completely silent shutdown when used.
6450 * There's now an option to control the SO_REUSEPORT socket
6451 option in .socket units.
6453 * Instance units will now automatically get a per-template
6454 subslice of system.slice unless something else is explicitly
6455 configured. For example, instances of sshd@.service will now
6456 implicitly be placed in system-sshd.slice rather than
6457 system.slice as before.
6459 * Test coverage support may now be enabled at build time.
6461 Contributions from: Dave Reisner, Frederic Crozat, Harald
6462 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6463 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6464 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6465 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6466 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6467 Giokas, Zbigniew Jędrzejewski-Szmek
6469 — Berlin, 2013-07-23
6473 * Two new unit types have been introduced:
6475 Scope units are very similar to service units, however, are
6476 created out of pre-existing processes — instead of PID 1
6477 forking off the processes. By using scope units it is
6478 possible for system services and applications to group their
6479 own child processes (worker processes) in a powerful way
6480 which then maybe used to organize them, or kill them
6481 together, or apply resource limits on them.
6483 Slice units may be used to partition system resources in an
6484 hierarchical fashion and then assign other units to them. By
6485 default there are now three slices: system.slice (for all
6486 system services), user.slice (for all user sessions),
6487 machine.slice (for VMs and containers).
6489 Slices and scopes have been introduced primarily in
6490 context of the work to move cgroup handling to a
6491 single-writer scheme, where only PID 1
6492 creates/removes/manages cgroups.
6494 * There's a new concept of "transient" units. In contrast to
6495 normal units these units are created via an API at runtime,
6496 not from configuration from disk. More specifically this
6497 means it is now possible to run arbitrary programs as
6498 independent services, with all execution parameters passed
6499 in via bus APIs rather than read from disk. Transient units
6500 make systemd substantially more dynamic then it ever was,
6501 and useful as a general batch manager.
6503 * logind has been updated to make use of scope and slice units
6504 for managing user sessions. As a user logs in he will get
6505 his own private slice unit, to which all sessions are added
6506 as scope units. We also added support for automatically
6507 adding an instance of user@.service for the user into the
6508 slice. Effectively logind will no longer create cgroup
6509 hierarchies on its own now, it will defer entirely to PID 1
6510 for this by means of scope, service and slice units. Since
6511 user sessions this way become entities managed by PID 1
6512 the output of "systemctl" is now a lot more comprehensive.
6514 * A new mini-daemon "systemd-machined" has been added which
6515 may be used by virtualization managers to register local
6516 VMs/containers. nspawn has been updated accordingly, and
6517 libvirt will be updated shortly. machined will collect a bit
6518 of meta information about the VMs/containers, and assign
6519 them their own scope unit (see above). The collected
6520 meta-data is then made available via the "machinectl" tool,
6521 and exposed in "ps" and similar tools. machined/machinectl
6522 is compile-time optional.
6524 * As discussed earlier, the low-level cgroup configuration
6525 options ControlGroup=, ControlGroupModify=,
6526 ControlGroupPersistent=, ControlGroupAttribute= have been
6527 removed. Please use high-level attribute settings instead as
6528 well as slice units.
6530 * A new bus call SetUnitProperties() has been added to alter
6531 various runtime parameters of a unit. This is primarily
6532 useful to alter cgroup parameters dynamically in a nice way,
6533 but will be extended later on to make more properties
6534 modifiable at runtime. systemctl gained a new set-properties
6535 command that wraps this call.
6537 * A new tool "systemd-run" has been added which can be used to
6538 run arbitrary command lines as transient services or scopes,
6539 while configuring a number of settings via the command
6540 line. This tool is currently very basic, however already
6541 very useful. We plan to extend this tool to even allow
6542 queuing of execution jobs with time triggers from the
6543 command line, similar in fashion to "at".
6545 * nspawn will now inform the user explicitly that kernels with
6546 audit enabled break containers, and suggest the user to turn
6549 * Support for detecting the IMA and AppArmor security
6550 frameworks with ConditionSecurity= has been added.
6552 * journalctl gained a new "-k" switch for showing only kernel
6553 messages, mimicking dmesg output; in addition to "--user"
6554 and "--system" switches for showing only user's own logs
6557 * systemd-delta can now show information about drop-in
6558 snippets extending unit files.
6560 * libsystemd-bus has been substantially updated but is still
6561 not available as public API.
6563 * systemd will now look for the "debug" argument on the kernel
6564 command line and enable debug logging, similar to what
6565 "systemd.log_level=debug" already did before.
6567 * "systemctl set-default", "systemctl get-default" has been
6568 added to configure the default.target symlink, which
6569 controls what to boot into by default.
6571 * "systemctl set-log-level" has been added as a convenient
6572 way to raise and lower systemd logging threshold.
6574 * "systemd-analyze plot" will now show the time the various
6575 generators needed for execution, as well as information
6576 about the unit file loading.
6578 * libsystemd-journal gained a new sd_journal_open_files() call
6579 for opening specific journal files. journactl also gained a
6580 new switch to expose this new functionality. Previously we
6581 only supported opening all files from a directory, or all
6582 files from the system, as opening individual files only is
6583 racy due to journal file rotation.
6585 * systemd gained the new DefaultEnvironment= setting in
6586 /etc/systemd/system.conf to set environment variables for
6589 * If a privileged process logs a journal message with the
6590 OBJECT_PID= field set, then journald will automatically
6591 augment this with additional OBJECT_UID=, OBJECT_GID=,
6592 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
6593 system services want to log events about specific client
6594 processes. journactl/systemctl has been updated to make use
6595 of this information if all log messages regarding a specific
6598 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
6599 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
6600 Reisner, David Coppa, David King, David Strauss, Eelco
6601 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
6602 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
6603 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
6604 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
6605 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
6606 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
6607 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
6608 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
6609 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
6610 Łukasz Stelmach, 장동준
6614 * The Python bindings gained some minimal support for the APIs
6615 exposed by libsystemd-logind.
6617 * ConditionSecurity= gained support for detecting SMACK. Since
6618 this condition already supports SELinux and AppArmor we only
6619 miss IMA for this. Patches welcome!
6621 Contributions from: Karol Lewandowski, Lennart Poettering,
6622 Zbigniew Jędrzejewski-Szmek
6626 * systemd-nspawn will now create /etc/resolv.conf if
6627 necessary, before bind-mounting the host's file onto it.
6629 * systemd-nspawn will now store meta information about a
6630 container on the container's cgroup as extended attribute
6631 fields, including the root directory.
6633 * The cgroup hierarchy has been reworked in many ways. All
6634 objects any of the components systemd creates in the cgroup
6635 tree are now suffixed. More specifically, user sessions are
6636 now placed in cgroups suffixed with ".session", users in
6637 cgroups suffixed with ".user", and nspawn containers in
6638 cgroups suffixed with ".nspawn". Furthermore, all cgroup
6639 names are now escaped in a simple scheme to avoid collision
6640 of userspace object names with kernel filenames. This work
6641 is preparation for making these objects relocatable in the
6642 cgroup tree, in order to allow easy resource partitioning of
6643 these objects without causing naming conflicts.
6645 * systemctl list-dependencies gained the new switches
6646 --plain, --reverse, --after and --before.
6648 * systemd-inhibit now shows the process name of processes that
6649 have taken an inhibitor lock.
6651 * nss-myhostname will now also resolve "localhost"
6652 implicitly. This makes /etc/hosts an optional file and
6653 nicely handles that on IPv6 ::1 maps to both "localhost" and
6656 * libsystemd-logind.so gained a new call
6657 sd_get_machine_names() to enumerate running containers and
6658 VMs (currently only supported by very new libvirt and
6659 nspawn). sd_login_monitor can now be used to watch
6660 VMs/containers coming and going.
6662 * .include is not allowed recursively anymore, and only in
6663 unit files. Usually it is better to use drop-in snippets in
6664 .d/*.conf anyway, as introduced with systemd 198.
6666 * systemd-analyze gained a new "critical-chain" command that
6667 determines the slowest chain of units run during system
6668 boot-up. It is very useful for tracking down where
6669 optimizing boot time is the most beneficial.
6671 * systemd will no longer allow manipulating service paths in
6672 the name=systemd:/system cgroup tree using ControlGroup= in
6673 units. (But is still fine with it in all other dirs.)
6675 * There's a new systemd-nspawn@.service service file that may
6676 be used to easily run nspawn containers as system
6677 services. With the container's root directory in
6678 /var/lib/container/foobar it is now sufficient to run
6679 "systemctl start systemd-nspawn@foobar.service" to boot it.
6681 * systemd-cgls gained a new parameter "--machine" to list only
6682 the processes within a certain container.
6684 * ConditionSecurity= now can check for "apparmor". We still
6685 are lacking checks for SMACK and IMA for this condition
6686 check though. Patches welcome!
6688 * A new configuration file /etc/systemd/sleep.conf has been
6689 added that may be used to configure which kernel operation
6690 systemd is supposed to execute when "suspend", "hibernate"
6691 or "hybrid-sleep" is requested. This makes the new kernel
6692 "freeze" state accessible to the user.
6694 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
6695 the passed argument if applicable.
6697 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
6698 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
6699 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
6700 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
6701 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
6702 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
6703 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
6708 * The output of 'systemctl list-jobs' got some polishing. The
6709 '--type=' argument may now be passed more than once. A new
6710 command 'systemctl list-sockets' has been added which shows
6711 a list of kernel sockets systemd is listening on with the
6712 socket units they belong to, plus the units these socket
6715 * The experimental libsystemd-bus library got substantial
6716 updates to work in conjunction with the (also experimental)
6717 kdbus kernel project. It works well enough to exchange
6718 messages with some sophistication. Note that kdbus is not
6719 ready yet, and the library is mostly an elaborate test case
6720 for now, and not installable.
6722 * systemd gained a new unit 'systemd-static-nodes.service'
6723 that generates static device nodes earlier during boot, and
6724 can run in conjunction with udev.
6726 * libsystemd-login gained a new call sd_pid_get_user_unit()
6727 to retrieve the user systemd unit a process is running
6728 in. This is useful for systems where systemd is used as
6731 * systemd-nspawn now places all containers in the new /machine
6732 top-level cgroup directory in the name=systemd
6733 hierarchy. libvirt will soon do the same, so that we get a
6734 uniform separation of /system, /user and /machine for system
6735 services, user processes and containers/virtual
6736 machines. This new cgroup hierarchy is also useful to stick
6737 stable names to specific container instances, which can be
6738 recognized later this way (this name may be controlled
6739 via systemd-nspawn's new -M switch). libsystemd-login also
6740 gained a new call sd_pid_get_machine_name() to retrieve the
6741 name of the container/VM a specific process belongs to.
6743 * bootchart can now store its data in the journal.
6745 * libsystemd-journal gained a new call
6746 sd_journal_add_conjunction() for AND expressions to the
6747 matching logic. This can be used to express more complex
6748 logical expressions.
6750 * journactl can now take multiple --unit= and --user-unit=
6753 * The cryptsetup logic now understands the "luks.key=" kernel
6754 command line switch for specifying a file to read the
6755 decryption key from. Also, if a configured key file is not
6756 found the tool will now automatically fall back to prompting
6759 * Python systemd.journal module was updated to wrap recently
6760 added functions from libsystemd-journal. The interface was
6761 changed to bring the low level interface in s.j._Reader
6762 closer to the C API, and the high level interface in
6763 s.j.Reader was updated to wrap and convert all data about
6766 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
6767 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
6768 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
6769 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
6770 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
6771 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6775 * journalctl --update-catalog now understands a new --root=
6776 option to operate on catalogs found in a different root
6779 * During shutdown after systemd has terminated all running
6780 services a final killing loop kills all remaining left-over
6781 processes. We will now print the name of these processes
6782 when we send SIGKILL to them, since this usually indicates a
6785 * If /etc/crypttab refers to password files stored on
6786 configured mount points automatic dependencies will now be
6787 generated to ensure the specific mount is established first
6788 before the key file is attempted to be read.
6790 * 'systemctl status' will now show information about the
6791 network sockets a socket unit is listening on.
6793 * 'systemctl status' will also shown information about any
6794 drop-in configuration file for units. (Drop-In configuration
6795 files in this context are files such as
6796 /etc/systemd/systemd/foobar.service.d/*.conf)
6798 * systemd-cgtop now optionally shows summed up CPU times of
6799 cgroups. Press '%' while running cgtop to switch between
6800 percentage and absolute mode. This is useful to determine
6801 which cgroups use up the most CPU time over the entire
6802 runtime of the system. systemd-cgtop has also been updated
6803 to be 'pipeable' for processing with further shell tools.
6805 * 'hostnamectl set-hostname' will now allow setting of FQDN
6808 * The formatting and parsing of time span values has been
6809 changed. The parser now understands fractional expressions
6810 such as "5.5h". The formatter will now output fractional
6811 expressions for all time spans under 1min, i.e. "5.123456s"
6812 rather than "5s 123ms 456us". For time spans under 1s
6813 millisecond values are shown, for those under 1ms
6814 microsecond values are shown. This should greatly improve
6815 all time-related output of systemd.
6817 * libsystemd-login and libsystemd-journal gained new
6818 functions for querying the poll() events mask and poll()
6819 timeout value for integration into arbitrary event
6822 * localectl gained the ability to list available X11 keymaps
6823 (models, layouts, variants, options).
6825 * 'systemd-analyze dot' gained the ability to filter for
6826 specific units via shell-style globs, to create smaller,
6827 more useful graphs. I.e. it is now possible to create simple
6828 graphs of all the dependencies between only target units, or
6829 of all units that Avahi has dependencies with.
6831 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
6832 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
6833 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
6834 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
6835 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
6836 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
6837 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
6841 * The boot-time readahead implementation for rotating media
6842 will now read the read-ahead data in multiple passes which
6843 consist of all read requests made in equidistant time
6844 intervals. This means instead of strictly reading read-ahead
6845 data in its physical order on disk we now try to find a
6846 middle ground between physical and access time order.
6848 * /etc/os-release files gained a new BUILD_ID= field for usage
6849 on operating systems that provide continuous builds of OS
6852 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
6853 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
6854 William Douglas, Zbigniew Jędrzejewski-Szmek
6858 * systemd-python gained an API exposing libsystemd-daemon.
6860 * The SMACK setup logic gained support for uploading CIPSO
6863 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
6864 ReadOnlyDirectories= and InaccessibleDirectories= has
6865 changed. The private /tmp and /var/tmp directories are now
6866 shared by all processes of a service (which means
6867 ExecStartPre= may now leave data in /tmp that ExecStart= of
6868 the same service can still access). When a service is
6869 stopped its temporary directories are immediately deleted
6870 (normal clean-up with tmpfiles is still done in addition to
6873 * By default, systemd will now set a couple of sysctl
6874 variables in the kernel: the safe sysrq options are turned
6875 on, IP route verification is turned on, and source routing
6876 disabled. The recently added hardlink and softlink
6877 protection of the kernel is turned on. These settings should
6878 be reasonably safe, and good defaults for all new systems.
6880 * The predictable network naming logic may now be turned off
6881 with a new kernel command line switch: net.ifnames=0.
6883 * A new libsystemd-bus module has been added that implements a
6884 pretty complete D-Bus client library. For details see:
6886 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
6888 * journald will now explicitly flush the journal files to disk
6889 at the latest 5min after each write. The file will then also
6890 be marked offline until the next write. This should increase
6891 reliability in case of a crash. The synchronization delay
6892 can be configured via SyncIntervalSec= in journald.conf.
6894 * There's a new remote-fs-setup.target unit that can be used
6895 to pull in specific services when at least one remote file
6896 system is to be mounted.
6898 * There are new targets timers.target and paths.target as
6899 canonical targets to pull user timer and path units in
6900 from. This complements sockets.target with a similar
6901 purpose for socket units.
6903 * libudev gained a new call udev_device_set_attribute_value()
6904 to set sysfs attributes of a device.
6906 * The udev daemon now sets the default number of worker
6907 processes executed in parallel based on the number of available
6908 CPUs instead of the amount of available RAM. This is supposed
6909 to provide a more reliable default and limit a too aggressive
6910 parallelism for setups with 1000s of devices connected.
6912 Contributions from: Auke Kok, Colin Walters, Cristian
6913 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
6914 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
6915 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
6916 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
6917 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
6918 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
6919 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
6920 Zbigniew Jędrzejewski-Szmek
6924 * Configuration of unit files may now be extended via drop-in
6925 files without having to edit/override the unit files
6926 themselves. More specifically, if the administrator wants to
6927 change one value for a service file foobar.service he can
6928 now do so by dropping in a configuration snippet into
6929 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
6930 will load all these snippets and apply them on top of the
6931 main unit configuration file, possibly extending or
6932 overriding its settings. Using these drop-in snippets is
6933 generally nicer than the two earlier options for changing
6934 unit files locally: copying the files from
6935 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
6936 them there; or creating a new file in /etc/systemd/system/
6937 that incorporates the original one via ".include". Drop-in
6938 snippets into these .d/ directories can be placed in any
6939 directory systemd looks for units in, and the usual
6940 overriding semantics between /usr/lib, /etc and /run apply
6943 * Most unit file settings which take lists of items can now be
6944 reset by assigning the empty string to them. For example,
6945 normally, settings such as Environment=FOO=BAR append a new
6946 environment variable assignment to the environment block,
6947 each time they are used. By assigning Environment= the empty
6948 string the environment block can be reset to empty. This is
6949 particularly useful with the .d/*.conf drop-in snippets
6950 mentioned above, since this adds the ability to reset list
6951 settings from vendor unit files via these drop-ins.
6953 * systemctl gained a new "list-dependencies" command for
6954 listing the dependencies of a unit recursively.
6956 * Inhibitors are now honored and listed by "systemctl
6957 suspend", "systemctl poweroff" (and similar) too, not only
6958 GNOME. These commands will also list active sessions by
6961 * Resource limits (as exposed by the various control group
6962 controllers) can now be controlled dynamically at runtime
6963 for all units. More specifically, you can now use a command
6964 like "systemctl set-cgroup-attr foobar.service cpu.shares
6965 2000" to alter the CPU shares a specific service gets. These
6966 settings are stored persistently on disk, and thus allow the
6967 administrator to easily adjust the resource usage of
6968 services with a few simple commands. This dynamic resource
6969 management logic is also available to other programs via the
6970 bus. Almost any kernel cgroup attribute and controller is
6973 * systemd-vconsole-setup will now copy all font settings to
6974 all allocated VTs, where it previously applied them only to
6977 * libsystemd-login gained the new sd_session_get_tty() API
6980 * This release drops support for a few legacy or
6981 distribution-specific LSB facility names when parsing init
6982 scripts: $x-display-manager, $mail-transfer-agent,
6983 $mail-transport-agent, $mail-transfer-agent, $smtp,
6984 $null. Also, the mail-transfer-agent.target unit backing
6985 this has been removed. Distributions which want to retain
6986 compatibility with this should carry the burden for
6987 supporting this themselves and patch support for these back
6988 in, if they really need to. Also, the facilities $syslog and
6989 $local_fs are now ignored, since systemd does not support
6990 early-boot LSB init scripts anymore, and these facilities
6991 are implied anyway for normal services. syslog.target has
6994 * There are new bus calls on PID1's Manager object for
6995 cancelling jobs, and removing snapshot units. Previously,
6996 both calls were only available on the Job and Snapshot
6999 * systemd-journal-gatewayd gained SSL support.
7001 * The various "environment" files, such as /etc/locale.conf
7002 now support continuation lines with a backslash ("\") as
7003 last character in the line, similarly in style (but different)
7004 to how this is supported in shells.
7006 * For normal user processes the _SYSTEMD_USER_UNIT= field is
7007 now implicitly appended to every log entry logged. systemctl
7008 has been updated to filter by this field when operating on a
7009 user systemd instance.
7011 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
7012 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7013 the container. This makes it easier to boot unmodified
7014 Fedora systems in a container, which however still requires
7015 audit=0 to be passed on the kernel command line. Auditing in
7016 kernel and userspace is unfortunately still too broken in
7017 context of containers, hence we recommend compiling it out
7018 of the kernel or using audit=0. Hopefully this will be fixed
7019 one day for good in the kernel.
7021 * nspawn gained the new --bind= and --bind-ro= parameters to
7022 bind mount specific directories from the host into the
7025 * nspawn will now mount its own devpts file system instance
7026 into the container, in order not to leak pty devices from
7027 the host into the container.
7029 * systemd will now read the firmware boot time performance
7030 information from the EFI variables, if the used boot loader
7031 supports this, and takes it into account for boot performance
7032 analysis via "systemd-analyze". This is currently supported
7033 only in conjunction with Gummiboot, but could be supported
7034 by other boot loaders too. For details see:
7036 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
7038 * A new generator has been added that automatically mounts the
7039 EFI System Partition (ESP) to /boot, if that directory
7040 exists, is empty, and no other file system has been
7041 configured to be mounted there.
7043 * logind will now send out PrepareForSleep(false) out
7044 unconditionally, after coming back from suspend. This may be
7045 used by applications as asynchronous notification for
7046 system resume events.
7048 * "systemctl unlock-sessions" has been added, that allows
7049 unlocking the screens of all user sessions at once, similar
7050 to how "systemctl lock-sessions" already locked all users
7051 sessions. This is backed by a new D-Bus call UnlockSessions().
7053 * "loginctl seat-status" will now show the master device of a
7054 seat. (i.e. the device of a seat that needs to be around for
7055 the seat to be considered available, usually the graphics
7058 * tmpfiles gained a new "X" line type, that allows
7059 configuration of files and directories (with wildcards) that
7060 shall be excluded from automatic cleanup ("aging").
7062 * udev default rules set the device node permissions now only
7063 at "add" events, and do not change them any longer with a
7064 later "change" event.
7066 * The log messages for lid events and power/sleep keypresses
7067 now carry a message ID.
7069 * We now have a substantially larger unit test suite, but this
7070 continues to be work in progress.
7072 * udevadm hwdb gained a new --root= parameter to change the
7073 root directory to operate relative to.
7075 * logind will now issue a background sync() request to the kernel
7076 early at shutdown, so that dirty buffers are flushed to disk early
7077 instead of at the last moment, in order to optimize shutdown
7080 * A new bootctl tool has been added that is an interface for
7081 certain boot loader operations. This is currently a preview
7082 and is likely to be extended into a small mechanism daemon
7083 like timedated, localed, hostnamed, and can be used by
7084 graphical UIs to enumerate available boot options, and
7085 request boot into firmware operations.
7087 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7088 the rest of the package. It also has been updated to work
7089 correctly in initrds.
7091 * Policykit previously has been runtime optional, and is now
7092 also compile time optional via a configure switch.
7094 * systemd-analyze has been reimplemented in C. Also "systemctl
7095 dot" has moved into systemd-analyze.
7097 * "systemctl status" with no further parameters will now print
7098 the status of all active or failed units.
7100 * Operations such as "systemctl start" can now be executed
7101 with a new mode "--irreversible" which may be used to queue
7102 operations that cannot accidentally be reversed by a later
7103 job queuing. This is by default used to make shutdown
7104 requests more robust.
7106 * The Python API of systemd now gained a new module for
7107 reading journal files.
7109 * A new tool kernel-install has been added that can install
7110 kernel images according to the Boot Loader Specification:
7112 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
7114 * Boot time console output has been improved to provide
7115 animated boot time output for hanging jobs.
7117 * A new tool systemd-activate has been added which can be used
7118 to test socket activation with, directly from the command
7119 line. This should make it much easier to test and debug
7120 socket activation in daemons.
7122 * journalctl gained a new "--reverse" (or -r) option to show
7123 journal output in reverse order (i.e. newest line first).
7125 * journalctl gained a new "--pager-end" (or -e) option to jump
7126 to immediately jump to the end of the journal in the
7127 pager. This is only supported in conjunction with "less".
7129 * journalctl gained a new "--user-unit=" option, that works
7130 similarly to "--unit=" but filters for user units rather than
7133 * A number of unit files to ease adoption of systemd in
7134 initrds has been added. This moves some minimal logic from
7135 the various initrd implementations into systemd proper.
7137 * The journal files are now owned by a new group
7138 "systemd-journal", which exists specifically to allow access
7139 to the journal, and nothing else. Previously, we used the
7140 "adm" group for that, which however possibly covers more
7141 than just journal/log file access. This new group is now
7142 already used by systemd-journal-gatewayd to ensure this
7143 daemon gets access to the journal files and as little else
7144 as possible. Note that "make install" will also set FS ACLs
7145 up for /var/log/journal to give "adm" and "wheel" read
7146 access to it, in addition to "systemd-journal" which owns
7147 the journal files. We recommend that packaging scripts also
7148 add read access to "adm" + "wheel" to /var/log/journal, and
7149 all existing/future journal files. To normal users and
7150 administrators little changes, however packagers need to
7151 ensure to create the "systemd-journal" system group at
7152 package installation time.
7154 * The systemd-journal-gatewayd now runs as unprivileged user
7155 systemd-journal-gateway:systemd-journal-gateway. Packaging
7156 scripts need to create these system user/group at
7159 * timedated now exposes a new boolean property CanNTP that
7160 indicates whether a local NTP service is available or not.
7162 * systemd-detect-virt will now also detect xen PVs
7164 * The pstore file system is now mounted by default, if it is
7167 * In addition to the SELinux and IMA policies we will now also
7168 load SMACK policies at early boot.
7170 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7171 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7172 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7173 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7174 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7175 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7176 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7177 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7178 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7179 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7180 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7181 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7182 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7183 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7187 * Timer units now support calendar time events in addition to
7188 monotonic time events. That means you can now trigger a unit
7189 based on a calendar time specification such as "Thu,Fri
7190 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7191 or fifth day of any month of the year 2013, given that it is
7192 a thursday or friday. This brings timer event support
7193 considerably closer to cron's capabilities. For details on
7194 the supported calendar time specification language see
7197 * udev now supports a number of different naming policies for
7198 network interfaces for predictable names, and a combination
7199 of these policies is now the default. Please see this wiki
7200 document for details:
7202 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
7204 * Auke Kok's bootchart implementation has been added to the
7205 systemd tree. It is an optional component that can graph the
7206 boot in quite some detail. It is one of the best bootchart
7207 implementations around and minimal in its code and
7210 * nss-myhostname has been integrated into the systemd source
7211 tree. nss-myhostname guarantees that the local hostname
7212 always stays resolvable via NSS. It has been a weak
7213 requirement of systemd-hostnamed since a long time, and
7214 since its code is actually trivial we decided to just
7215 include it in systemd's source tree. It can be turned off
7216 with a configure switch.
7218 * The read-ahead logic is now capable of properly detecting
7219 whether a btrfs file system is on SSD or rotating media, in
7220 order to optimize the read-ahead scheme. Previously, it was
7221 only capable of detecting this on traditional file systems
7224 * In udev, additional device properties are now read from the
7225 IAB in addition to the OUI database. Also, Bluetooth company
7226 identities are attached to the devices as well.
7228 * In service files %U may be used as specifier that is
7229 replaced by the configured user name of the service.
7231 * nspawn may now be invoked without a controlling TTY. This
7232 makes it suitable for invocation as its own service. This
7233 may be used to set up a simple containerized server system
7234 using only core OS tools.
7236 * systemd and nspawn can now accept socket file descriptors
7237 when they are started for socket activation. This enables
7238 implementation of socket activated nspawn
7239 containers. i.e. think about autospawning an entire OS image
7240 when the first SSH or HTTP connection is received. We expect
7241 that similar functionality will also be added to libvirt-lxc
7244 * journalctl will now suppress ANSI color codes when
7245 presenting log data.
7247 * systemctl will no longer show control group information for
7248 a unit if the control group is empty anyway.
7250 * logind can now automatically suspend/hibernate/shutdown the
7253 * /etc/machine-info and hostnamed now also expose the chassis
7254 type of the system. This can be used to determine whether
7255 the local system is a laptop, desktop, handset or
7256 tablet. This information may either be configured by the
7257 user/vendor or is automatically determined from ACPI and DMI
7258 information if possible.
7260 * A number of PolicyKit actions are now bound together with
7261 "imply" rules. This should simplify creating UIs because
7262 many actions will now authenticate similar ones as well.
7264 * Unit files learnt a new condition ConditionACPower= which
7265 may be used to conditionalize a unit depending on whether an
7266 AC power source is connected or not, of whether the system
7267 is running on battery power.
7269 * systemctl gained a new "is-failed" verb that may be used in
7270 shell scripts and suchlike to check whether a specific unit
7271 is in the "failed" state.
7273 * The EnvironmentFile= setting in unit files now supports file
7274 globbing, and can hence be used to easily read a number of
7275 environment files at once.
7277 * systemd will no longer detect and recognize specific
7278 distributions. All distribution-specific #ifdeffery has been
7279 removed, systemd is now fully generic and
7280 distribution-agnostic. Effectively, not too much is lost as
7281 a lot of the code is still accessible via explicit configure
7282 switches. However, support for some distribution specific
7283 legacy configuration file formats has been dropped. We
7284 recommend distributions to simply adopt the configuration
7285 files everybody else uses now and convert the old
7286 configuration from packaging scripts. Most distributions
7287 already did that. If that's not possible or desirable,
7288 distributions are welcome to forward port the specific
7289 pieces of code locally from the git history.
7291 * When logging a message about a unit systemd will now always
7292 log the unit name in the message meta data.
7294 * localectl will now also discover system locale data that is
7295 not stored in locale archives, but directly unpacked.
7297 * logind will no longer unconditionally use framebuffer
7298 devices as seat masters, i.e. as devices that are required
7299 to be existing before a seat is considered preset. Instead,
7300 it will now look for all devices that are tagged as
7301 "seat-master" in udev. By default, framebuffer devices will
7302 be marked as such, but depending on local systems, other
7303 devices might be marked as well. This may be used to
7304 integrate graphics cards using closed source drivers (such
7305 as NVidia ones) more nicely into logind. Note however, that
7306 we recommend using the open source NVidia drivers instead,
7307 and no udev rules for the closed-source drivers will be
7308 shipped from us upstream.
7310 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7311 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7312 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7313 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7314 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7315 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7316 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7317 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7318 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7319 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7320 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7321 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7326 * udev gained support for loading additional device properties
7327 from an indexed database that is keyed by vendor/product IDs
7328 and similar device identifiers. For the beginning this
7329 "hwdb" is populated with data from the well-known PCI and
7330 USB database, but also includes PNP, ACPI and OID data. In
7331 the longer run this indexed database shall grow into
7332 becoming the one central database for non-essential
7333 userspace device metadata. Previously, data from the PCI/USB
7334 database was only attached to select devices, since the
7335 lookup was a relatively expensive operation due to O(n) time
7336 complexity (with n being the number of entries in the
7337 database). Since this is now O(1), we decided to add in this
7338 data for all devices where this is available, by
7339 default. Note that the indexed database needs to be rebuilt
7340 when new data files are installed. To achieve this you need
7341 to update your packaging scripts to invoke "udevadm hwdb
7342 --update" after installation of hwdb data files. For
7343 RPM-based distributions we introduced the new
7344 %udev_hwdb_update macro for this purpose.
7346 * The Journal gained support for the "Message Catalog", an
7347 indexed database to link up additional information with
7348 journal entries. For further details please check:
7350 https://www.freedesktop.org/wiki/Software/systemd/catalog
7352 The indexed message catalog database also needs to be
7353 rebuilt after installation of message catalog files. Use
7354 "journalctl --update-catalog" for this. For RPM-based
7355 distributions we introduced the %journal_catalog_update
7356 macro for this purpose.
7358 * The Python Journal bindings gained support for the standard
7359 Python logging framework.
7361 * The Journal API gained new functions for checking whether
7362 the underlying file system of a journal file is capable of
7363 properly reporting file change notifications, or whether
7364 applications that want to reflect journal changes "live"
7365 need to recheck journal files continuously in appropriate
7368 * It is now possible to set the "age" field for tmpfiles
7369 entries to 0, indicating that files matching this entry
7370 shall always be removed when the directories are cleaned up.
7372 * coredumpctl gained a new "gdb" verb which invokes gdb
7373 right-away on the selected coredump.
7375 * There's now support for "hybrid sleep" on kernels that
7376 support this, in addition to "suspend" and "hibernate". Use
7377 "systemctl hybrid-sleep" to make use of this.
7379 * logind's HandleSuspendKey= setting (and related settings)
7380 now gained support for a new "lock" setting to simply
7381 request the screen lock on all local sessions, instead of
7382 actually executing a suspend or hibernation.
7384 * systemd will now mount the EFI variables file system by
7387 * Socket units now gained support for configuration of the
7388 SMACK security label.
7390 * timedatectl will now output the time of the last and next
7391 daylight saving change.
7393 * We dropped support for various legacy and distro-specific
7394 concepts, such as insserv, early-boot SysV services
7395 (i.e. those for non-standard runlevels such as 'b' or 'S')
7396 or ArchLinux /etc/rc.conf support. We recommend the
7397 distributions who still need support this to either continue
7398 to maintain the necessary patches downstream, or find a
7399 different solution. (Talk to us if you have questions!)
7401 * Various systemd components will now bypass PolicyKit checks
7402 for root and otherwise handle properly if PolicyKit is not
7403 found to be around. This should fix most issues for
7404 PolicyKit-less systems. Quite frankly this should have been
7405 this way since day one. It is absolutely our intention to
7406 make systemd work fine on PolicyKit-less systems, and we
7407 consider it a bug if something does not work as it should if
7408 PolicyKit is not around.
7410 * For embedded systems it is now possible to build udev and
7411 systemd without blkid and/or kmod support.
7413 * "systemctl switch-root" is now capable of switching root
7414 more than once. I.e. in addition to transitions from the
7415 initrd to the host OS it is now possible to transition to
7416 further OS images from the host. This is useful to implement
7417 offline updating tools.
7419 * Various other additions have been made to the RPM macros
7420 shipped with systemd. Use %udev_rules_update() after
7421 installing new udev rules files. %_udevhwdbdir,
7422 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7423 %_sysctldir are now available which resolve to the right
7424 directories for packages to place various data files in.
7426 * journalctl gained the new --full switch (in addition to
7427 --all, to disable ellipsation for long messages.
7429 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7430 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7431 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7432 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7433 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7434 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7435 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7436 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7437 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7441 * journalctl gained new --since= and --until= switches to
7442 filter by time. It also now supports nice filtering for
7443 units via --unit=/-u.
7445 * Type=oneshot services may use ExecReload= and do the
7448 * The journal daemon now supports time-based rotation and
7449 vacuuming, in addition to the usual disk-space based
7452 * The journal will now index the available field values for
7453 each field name. This enables clients to show pretty drop
7454 downs of available match values when filtering. The bash
7455 completion of journalctl has been updated
7456 accordingly. journalctl gained a new switch -F to list all
7457 values a certain field takes in the journal database.
7459 * More service events are now written as structured messages
7460 to the journal, and made recognizable via message IDs.
7462 * The timedated, localed and hostnamed mini-services which
7463 previously only provided support for changing time, locale
7464 and hostname settings from graphical DEs such as GNOME now
7465 also have a minimal (but very useful) text-based client
7466 utility each. This is probably the nicest way to changing
7467 these settings from the command line now, especially since
7468 it lists available options and is fully integrated with bash
7471 * There's now a new tool "systemd-coredumpctl" to list and
7472 extract coredumps from the journal.
7474 * We now install a README each in /var/log/ and
7475 /etc/rc.d/init.d explaining where the system logs and init
7476 scripts went. This hopefully should help folks who go to
7477 that dirs and look into the otherwise now empty void and
7478 scratch their heads.
7480 * When user-services are invoked (by systemd --user) the
7481 $MANAGERPID env var is set to the PID of systemd.
7483 * SIGRTMIN+24 when sent to a --user instance will now result
7484 in immediate termination of systemd.
7486 * gatewayd received numerous feature additions such as a
7487 "follow" mode, for live syncing and filtering.
7489 * browse.html now allows filtering and showing detailed
7490 information on specific entries. Keyboard navigation and
7491 mouse screen support has been added.
7493 * gatewayd/journalctl now supports HTML5/JSON
7494 Server-Sent-Events as output.
7496 * The SysV init script compatibility logic will now
7497 heuristically determine whether a script supports the
7498 "reload" verb, and only then make this available as
7501 * "systemctl status --follow" has been removed, use "journalctl
7504 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
7505 have been removed since they are hardly useful to be
7508 * And I'd like to take the opportunity to specifically mention
7509 Zbigniew for his great contributions. Zbigniew, you rock!
7511 Contributions from: Andrew Eikum, Christian Hesse, Colin
7512 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
7513 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
7514 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
7515 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
7516 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
7517 Jędrzejewski-Szmek, Сковорода Никита Андреевич
7521 * If /etc/vconsole.conf is non-existent or empty we will no
7522 longer load any console font or key map at boot by
7523 default. Instead the kernel defaults will be left
7524 intact. This is definitely the right thing to do, as no
7525 configuration should mean no configuration, and hard-coding
7526 font names that are different on all archs is probably a bad
7527 idea. Also, the kernel default key map and font should be
7528 good enough for most cases anyway, and mostly identical to
7529 the userspace fonts/key maps we previously overloaded them
7530 with. If distributions want to continue to default to a
7531 non-kernel font or key map they should ship a default
7532 /etc/vconsole.conf with the appropriate contents.
7534 Contributions from: Colin Walters, Daniel J Walsh, Dave
7535 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
7536 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7540 * journalctl gained a new --cursor= switch to show entries
7541 starting from the specified location in the journal.
7543 * We now enforce a size limit on journal entry fields exported
7544 with "-o json" in journalctl. Fields larger than 4K will be
7545 assigned null. This can be turned off with --all.
7547 * An (optional) journal gateway daemon is now available as
7548 "systemd-journal-gatewayd.service". This service provides
7549 access to the journal via HTTP and JSON. This functionality
7550 will be used to implement live log synchronization in both
7551 pull and push modes, but has various other users too, such
7552 as easy log access for debugging of embedded devices. Right
7553 now it is already useful to retrieve the journal via HTTP:
7555 # systemctl start systemd-journal-gatewayd.service
7556 # wget http://localhost:19531/entries
7558 This will download the journal contents in a
7559 /var/log/messages compatible format. The same as JSON:
7561 # curl -H"Accept: application/json" http://localhost:19531/entries
7563 This service is also accessible via a web browser where a
7564 single static HTML5 app is served that uses the JSON logic
7565 to enable the user to do some basic browsing of the
7566 journal. This will be extended later on. Here's an example
7567 screenshot of this app in its current state:
7569 http://0pointer.de/public/journal-gatewayd
7571 Contributions from: Kay Sievers, Lennart Poettering, Robert
7572 Milasan, Tom Gundersen
7576 * The bash completion logic is now available for journalctl
7579 * We do not mount the "cpuset" controller anymore together with
7580 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
7581 started if no parameters are assigned to it. "cpuset" hence
7582 broke code that assumed it could create "cpu" groups and
7585 * journalctl -f will now subscribe to terminal size changes,
7586 and line break accordingly.
7588 Contributions from: Dave Reisner, Kay Sievers, Lennart
7589 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
7593 * nspawn will now create a symlink /etc/localtime in the
7594 container environment, copying the host's timezone
7595 setting. Previously this has been done via a bind mount, but
7596 since symlinks cannot be bind mounted this has now been
7597 changed to create/update the appropriate symlink.
7599 * journalctl -n's line number argument is now optional, and
7600 will default to 10 if omitted.
7602 * journald will now log the maximum size the journal files may
7603 take up on disk. This is particularly useful if the default
7604 built-in logic of determining this parameter from the file
7605 system size is used. Use "systemctl status
7606 systemd-journald.service" to see this information.
7608 * The multi-seat X wrapper tool has been stripped down. As X
7609 is now capable of enumerating graphics devices via udev in a
7610 seat-aware way the wrapper is not strictly necessary
7611 anymore. A stripped down temporary stop-gap is still shipped
7612 until the upstream display managers have been updated to
7613 fully support the new X logic. Expect this wrapper to be
7614 removed entirely in one of the next releases.
7616 * HandleSleepKey= in logind.conf has been split up into
7617 HandleSuspendKey= and HandleHibernateKey=. The old setting
7618 is not available anymore. X11 and the kernel are
7619 distinguishing between these keys and we should too. This
7620 also means the inhibition lock for these keys has been split
7623 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
7624 Poettering, Lukas Nykryn, Václav Pavlín
7628 * Whenever a unit changes state we will now log this to the
7629 journal and show along the unit's own log output in
7632 * ConditionPathIsMountPoint= can now properly detect bind
7633 mount points too. (Previously, a bind mount of one file
7634 system to another place in the same file system could not be
7635 detected as mount, since they shared struct stat's st_dev
7638 * We will now mount the cgroup controllers cpu, cpuacct,
7639 cpuset and the controllers net_cls, net_prio together by
7642 * nspawn containers will now have a virtualized boot
7643 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
7644 over with a randomized ID at container initialization). This
7645 has the effect of making "journalctl -b" do the right thing
7648 * The JSON output journal serialization has been updated not
7649 to generate "endless" list objects anymore, but rather one
7650 JSON object per line. This is more in line how most JSON
7651 parsers expect JSON objects. The new output mode
7652 "json-pretty" has been added to provide similar output, but
7653 neatly aligned for readability by humans.
7655 * We dropped all explicit sync() invocations in the shutdown
7656 code. The kernel does this implicitly anyway in the kernel
7657 reboot() syscall. halt(8)'s -n option is now a compatibility
7660 * We now support virtualized reboot() in containers, as
7661 supported by newer kernels. We will fall back to exit() if
7662 CAP_SYS_REBOOT is not available to the container. Also,
7663 nspawn makes use of this now and will actually reboot the
7664 container if the containerized OS asks for that.
7666 * journalctl will only show local log output by default
7667 now. Use --merge (-m) to show remote log output, too.
7669 * libsystemd-journal gained the new sd_journal_get_usage()
7670 call to determine the current disk usage of all journal
7671 files. This is exposed in the new "journalctl --disk-usage"
7674 * journald gained a new configuration setting SplitMode= in
7675 journald.conf which may be used to control how user journals
7676 are split off. See journald.conf(5) for details.
7678 * A new condition type ConditionFileNotEmpty= has been added.
7680 * tmpfiles' "w" lines now support file globbing, to write
7681 multiple files at once.
7683 * We added Python bindings for the journal submission
7684 APIs. More Python APIs for a number of selected APIs will
7685 likely follow. Note that we intend to add native bindings
7686 only for the Python language, as we consider it common
7687 enough to deserve bindings shipped within systemd. There are
7688 various projects outside of systemd that provide bindings
7689 for languages such as PHP or Lua.
7691 * Many conditions will now resolve specifiers such as %i. In
7692 addition, PathChanged= and related directives of .path units
7693 now support specifiers as well.
7695 * There's now a new RPM macro definition for the system preset
7698 * journald will now warn if it ca not forward a message to the
7699 syslog daemon because its socket is full.
7701 * timedated will no longer write or process /etc/timezone,
7702 except on Debian. As we do not support late mounted /usr
7703 anymore /etc/localtime always being a symlink is now safe,
7704 and hence the information in /etc/timezone is not necessary
7707 * logind will now always reserve one VT for a text getty (VT6
7708 by default). Previously if more than 6 X sessions where
7709 started they took up all the VTs with auto-spawned gettys,
7710 so that no text gettys were available anymore.
7712 * udev will now automatically inform the btrfs kernel logic
7713 about btrfs RAID components showing up. This should make
7714 simple hotplug based btrfs RAID assembly work.
7716 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
7717 (but not for its children which will stay at the kernel
7718 default). This should allow setups with a lot more listening
7721 * systemd will now always pass the configured timezone to the
7722 kernel at boot. timedated will do the same when the timezone
7725 * logind's inhibition logic has been updated. By default,
7726 logind will now handle the lid switch, the power and sleep
7727 keys all the time, even in graphical sessions. If DEs want
7728 to handle these events on their own they should take the new
7729 handle-power-key, handle-sleep-key and handle-lid-switch
7730 inhibitors during their runtime. A simple way to achieve
7731 that is to invoke the DE wrapped in an invocation of:
7733 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
7735 * Access to unit operations is now checked via SELinux taking
7736 the unit file label and client process label into account.
7738 * systemd will now notify the administrator in the journal
7739 when he over-mounts a non-empty directory.
7741 * There are new specifiers that are resolved in unit files,
7742 for the host name (%H), the machine ID (%m) and the boot ID
7745 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
7746 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
7747 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
7748 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
7749 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
7750 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
7751 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
7755 * Support for reading structured kernel messages from
7756 /dev/kmsg has now been added and is enabled by default.
7758 * Support for reading kernel messages from /proc/kmsg has now
7759 been removed. If you want kernel messages in the journal
7760 make sure to run a recent kernel (>= 3.5) that supports
7761 reading structured messages from /dev/kmsg (see
7762 above). /proc/kmsg is now exclusive property of classic
7763 syslog daemons again.
7765 * The libudev API gained the new
7766 udev_device_new_from_device_id() call.
7768 * The logic for file system namespace (ReadOnlyDirectory=,
7769 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
7770 require pivot_root() anymore. This means fewer temporary
7771 directories are created below /tmp for this feature.
7773 * nspawn containers will now see and receive all submounts
7774 made on the host OS below the root file system of the
7777 * Forward Secure Sealing is now supported for Journal files,
7778 which provide cryptographical sealing of journal files so
7779 that attackers cannot alter log history anymore without this
7780 being detectable. Lennart will soon post a blog story about
7781 this explaining it in more detail.
7783 * There are two new service settings RestartPreventExitStatus=
7784 and SuccessExitStatus= which allow configuration of exit
7785 status (exit code or signal) which will be excepted from the
7786 restart logic, resp. consider successful.
7788 * journalctl gained the new --verify switch that can be used
7789 to check the integrity of the structure of journal files and
7790 (if Forward Secure Sealing is enabled) the contents of
7793 * nspawn containers will now be run with /dev/stdin, /dev/fd/
7794 and similar symlinks pre-created. This makes running shells
7795 as container init process a lot more fun.
7797 * The fstab support can now handle PARTUUID= and PARTLABEL=
7800 * A new ConditionHost= condition has been added to match
7801 against the hostname (with globs) and machine ID. This is
7802 useful for clusters where a single OS image is used to
7803 provision a large number of hosts which shall run slightly
7804 different sets of services.
7806 * Services which hit the restart limit will now be placed in a
7809 Contributions from: Bertram Poettering, Dave Reisner, Huang
7810 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
7811 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
7815 * When running in --user mode systemd will now become a
7816 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
7817 tree a lot more organized.
7819 * A new PartOf= unit dependency type has been introduced that
7820 may be used to group services in a natural way.
7822 * "systemctl enable" may now be used to enable instances of
7825 * journalctl now prints error log levels in red, and
7826 warning/notice log levels in bright white. It also supports
7827 filtering by log level now.
7829 * cgtop gained a new -n switch (similar to top), to configure
7830 the maximum number of iterations to run for. It also gained
7831 -b, to run in batch mode (accepting no input).
7833 * The suffix ".service" may now be omitted on most systemctl
7834 command lines involving service unit names.
7836 * There's a new bus call in logind to lock all sessions, as
7837 well as a loginctl verb for it "lock-sessions".
7839 * libsystemd-logind.so gained a new call sd_journal_perror()
7840 that works similar to libc perror() but logs to the journal
7841 and encodes structured information about the error number.
7843 * /etc/crypttab entries now understand the new keyfile-size=
7846 * shutdown(8) now can send a (configurable) wall message when
7847 a shutdown is cancelled.
7849 * The mount propagation mode for the root file system will now
7850 default to "shared", which is useful to make containers work
7851 nicely out-of-the-box so that they receive new mounts from
7852 the host. This can be undone locally by running "mount
7853 --make-rprivate /" if needed.
7855 * The prefdm.service file has been removed. Distributions
7856 should maintain this unit downstream if they intend to keep
7857 it around. However, we recommend writing normal unit files
7858 for display managers instead.
7860 * Since systemd is a crucial part of the OS we will now
7861 default to a number of compiler switches that improve
7862 security (hardening) such as read-only relocations, stack
7863 protection, and suchlike.
7865 * The TimeoutSec= setting for services is now split into
7866 TimeoutStartSec= and TimeoutStopSec= to allow configuration
7867 of individual time outs for the start and the stop phase of
7870 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
7871 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
7872 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
7873 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
7874 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
7875 Gundersen, Zbigniew Jędrzejewski-Szmek
7879 * The journal and id128 C APIs are now fully documented as man
7882 * Extra safety checks have been added when transitioning from
7883 the initial RAM disk to the main system to avoid accidental
7886 * /etc/crypttab entries now understand the new keyfile-offset=
7889 * systemctl -t can now be used to filter by unit load state.
7891 * The journal C API gained the new sd_journal_wait() call to
7892 make writing synchronous journal clients easier.
7894 * journalctl gained the new -D switch to show journals from a
7897 * journalctl now displays a special marker between log
7898 messages of two different boots.
7900 * The journal is now explicitly flushed to /var via a service
7901 systemd-journal-flush.service, rather than implicitly simply
7902 by seeing /var/log/journal to be writable.
7904 * journalctl (and the journal C APIs) can now match for much
7905 more complex expressions, with alternatives and
7908 * When transitioning from the initial RAM disk to the main
7909 system we will now kill all processes in a killing spree to
7910 ensure no processes stay around by accident.
7912 * Three new specifiers may be used in unit files: %u, %h, %s
7913 resolve to the user name, user home directory resp. user
7914 shell. This is useful for running systemd user instances.
7916 * We now automatically rotate journal files if their data
7917 object hash table gets a fill level > 75%. We also size the
7918 hash table based on the configured maximum file size. This
7919 together should lower hash collisions drastically and thus
7920 speed things up a bit.
7922 * journalctl gained the new "--header" switch to introspect
7923 header data of journal files.
7925 * A new setting SystemCallFilters= has been added to services
7926 which may be used to apply blacklists or whitelists to
7927 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
7929 * nspawn gained a new --link-journal= switch (and quicker: -j)
7930 to link the container journal with the host. This makes it
7931 very easy to centralize log viewing on the host for all
7932 guests while still keeping the journal files separated.
7934 * Many bugfixes and optimizations
7936 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
7937 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
7938 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
7943 * Several tools now understand kernel command line arguments,
7944 which are only read when run in an initial RAM disk. They
7945 usually follow closely their normal counterparts, but are
7948 * There's a new tool to analyze the readahead files that are
7949 automatically generated at boot. Use:
7951 /usr/lib/systemd/systemd-readahead analyze /.readahead
7953 * We now provide an early debug shell on tty9 if this enabled. Use:
7955 systemctl enable debug-shell.service
7957 * All plymouth related units have been moved into the Plymouth
7958 package. Please make sure to upgrade your Plymouth version
7961 * systemd-tmpfiles now supports getting passed the basename of
7962 a configuration file only, in which case it will look for it
7963 in all appropriate directories automatically.
7965 * udevadm info now takes a /dev or /sys path as argument, and
7966 does the right thing. Example:
7968 udevadm info /dev/sda
7969 udevadm info /sys/class/block/sda
7971 * systemctl now prints a warning if a unit is stopped but a
7972 unit that might trigger it continues to run. Example: a
7973 service is stopped but the socket that activates it is left
7976 * "systemctl status" will now mention if the log output was
7977 shortened due to rotation since a service has been started.
7979 * The journal API now exposes functions to determine the
7980 "cutoff" times due to rotation.
7982 * journald now understands SIGUSR1 and SIGUSR2 for triggering
7983 immediately flushing of runtime logs to /var if possible,
7984 resp. for triggering immediate rotation of the journal
7987 * It is now considered an error if a service is attempted to
7988 be stopped that is not loaded.
7990 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
7992 * systemd-analyze now supports Python 3
7994 * tmpfiles now supports cleaning up directories via aging
7995 where the first level dirs are always kept around but
7996 directories beneath it automatically aged. This is enabled
7997 by prefixing the age field with '~'.
7999 * Seat objects now expose CanGraphical, CanTTY properties
8000 which is required to deal with very fast bootups where the
8001 display manager might be running before the graphics drivers
8002 completed initialization.
8004 * Seat objects now expose a State property.
8006 * We now include RPM macros for service enabling/disabling
8007 based on the preset logic. We recommend RPM based
8008 distributions to make use of these macros if possible. This
8009 makes it simpler to reuse RPM spec files across
8012 * We now make sure that the collected systemd unit name is
8013 always valid when services log to the journal via
8016 * There's a new man page kernel-command-line(7) detailing all
8017 command line options we understand.
8019 * The fstab generator may now be disabled at boot by passing
8020 fstab=0 on the kernel command line.
8022 * A new kernel command line option modules-load= is now understood
8023 to load a specific kernel module statically, early at boot.
8025 * Unit names specified on the systemctl command line are now
8026 automatically escaped as needed. Also, if file system or
8027 device paths are specified they are automatically turned
8028 into the appropriate mount or device unit names. Example:
8030 systemctl status /home
8031 systemctl status /dev/sda
8033 * The SysVConsole= configuration option has been removed from
8034 system.conf parsing.
8036 * The SysV search path is no longer exported on the D-Bus
8039 * The Names= option has been removed from unit file parsing.
8041 * There's a new man page bootup(7) detailing the boot process.
8043 * Every unit and every generator we ship with systemd now
8044 comes with full documentation. The self-explanatory boot is
8047 * A couple of services gained "systemd-" prefixes in their
8048 name if they wrap systemd code, rather than only external
8049 code. Among them fsck@.service which is now
8050 systemd-fsck@.service.
8052 * The HaveWatchdog property has been removed from the D-Bus
8055 * systemd.confirm_spawn= on the kernel command line should now
8058 * There's a new man page crypttab(5) which details all options
8059 we actually understand.
8061 * systemd-nspawn gained a new --capability= switch to pass
8062 additional capabilities to the container.
8064 * timedated will now read known NTP implementation unit names
8065 from /usr/lib/systemd/ntp-units.d/*.list,
8066 systemd-timedated-ntp.target has been removed.
8068 * journalctl gained a new switch "-b" that lists log data of
8069 the current boot only.
8071 * The notify socket is in the abstract namespace again, in
8072 order to support daemons which chroot() at start-up.
8074 * There is a new Storage= configuration option for journald
8075 which allows configuration of where log data should go. This
8076 also provides a way to disable journal logging entirely, so
8077 that data collected is only forwarded to the console, the
8078 kernel log buffer or another syslog implementation.
8080 * Many bugfixes and optimizations
8082 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8083 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8084 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8085 Shawn Landden, Tom Gundersen
8089 * "systemctl help <unit>" now shows the man page if one is
8092 * Several new man pages have been added.
8094 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8095 MaxLevelConsole= can now be specified in
8096 journald.conf. These options allow reducing the amount of
8097 data stored on disk or forwarded by the log level.
8099 * TimerSlackNSec= can now be specified in system.conf for
8100 PID1. This allows system-wide power savings.
8102 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8103 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8108 * logind is now capable of (optionally) handling power and
8109 sleep keys as well as the lid switch.
8111 * journalctl now understands the syntax "journalctl
8112 /usr/bin/avahi-daemon" to get all log output of a specific
8115 * CapabilityBoundingSet= in system.conf now also influences
8116 the capability bound set of usermode helpers of the kernel.
8118 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8119 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8120 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8121 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8125 * Note that we skipped 139 releases here in order to set the
8126 new version to something that is greater than both udev's
8127 and systemd's most recent version number.
8129 * udev: all udev sources are merged into the systemd source tree now.
8130 All future udev development will happen in the systemd tree. It
8131 is still fully supported to use the udev daemon and tools without
8132 systemd running, like in initramfs or other init systems. Building
8133 udev though, will require the *build* of the systemd tree, but
8134 udev can be properly *run* without systemd.
8136 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
8137 should be used to create dead device nodes as workarounds for broken
8140 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8141 no longer supported. udev_monitor_new_from_netlink() needs to be
8142 used to subscribe to events.
8144 * udev: when udevd is started by systemd, processes which are left
8145 behind by forking them off of udev rules, are unconditionally cleaned
8146 up and killed now after the event handling has finished. Services or
8147 daemons must be started as systemd services. Services can be
8148 pulled-in by udev to get started, but they can no longer be directly
8149 forked by udev rules.
8151 * udev: the daemon binary is called systemd-udevd now and installed
8152 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8153 to adapt to that, create symlink, or rename the binary after building
8156 * libudev no longer provides these symbols:
8157 udev_monitor_from_socket()
8158 udev_queue_get_failed_list_entry()
8159 udev_get_{dev,sys,run}_path()
8160 The versions number was bumped and symbol versioning introduced.
8162 * systemd-loginctl and systemd-journalctl have been renamed
8163 to loginctl and journalctl to match systemctl.
8165 * The config files: /etc/systemd/systemd-logind.conf and
8166 /etc/systemd/systemd-journald.conf have been renamed to
8167 logind.conf and journald.conf. Package updates should rename
8168 the files to the new names on upgrade.
8170 * For almost all files the license is now LGPL2.1+, changed
8171 from the previous GPL2.0+. Exceptions are some minor stuff
8172 of udev (which will be changed to LGPL2.1 eventually, too),
8173 and the MIT licensed sd-daemon.[ch] library that is suitable
8174 to be used as drop-in files.
8176 * systemd and logind now handle system sleep states, in
8177 particular suspending and hibernating.
8179 * logind now implements a sleep/shutdown/idle inhibiting logic
8180 suitable for a variety of uses. Soonishly Lennart will blog
8181 about this in more detail.
8183 * var-run.mount and var-lock.mount are no longer provided
8184 (which previously bind mounted these directories to their new
8185 places). Distributions which have not converted these
8186 directories to symlinks should consider stealing these files
8187 from git history and add them downstream.
8189 * We introduced the Documentation= field for units and added
8190 this to all our shipped units. This is useful to make it
8191 easier to explore the boot and the purpose of the various
8194 * All smaller setup units (such as
8195 systemd-vconsole-setup.service) now detect properly if they
8196 are run in a container and are skipped when
8197 appropriate. This guarantees an entirely noise-free boot in
8198 Linux container environments such as systemd-nspawn.
8200 * A framework for implementing offline system updates is now
8201 integrated, for details see:
8202 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
8204 * A new service type Type=idle is available now which helps us
8205 avoiding ugly interleaving of getty output and boot status
8208 * There's now a system-wide CapabilityBoundingSet= option to
8209 globally reduce the set of capabilities for the
8210 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8211 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8212 even CAP_NET_ADMIN system-wide for secure systems.
8214 * There are now system-wide DefaultLimitXXX= options to
8215 globally change the defaults of the various resource limits
8216 for all units started by PID 1.
8218 * Harald Hoyer's systemd test suite has been integrated into
8219 systemd which allows easy testing of systemd builds in qemu
8220 and nspawn. (This is really awesome! Ask us for details!)
8222 * The fstab parser is now implemented as generator, not inside
8225 * systemctl will now warn you if .mount units generated from
8226 /etc/fstab are out of date due to changes in fstab that
8227 have not been read by systemd yet.
8229 * systemd is now suitable for usage in initrds. Dracut has
8230 already been updated to make use of this. With this in place
8231 initrds get a slight bit faster but primarily are much
8232 easier to introspect and debug since "systemctl status" in
8233 the host system can be used to introspect initrd services,
8234 and the journal from the initrd is kept around too.
8236 * systemd-delta has been added, a tool to explore differences
8237 between user/admin configuration and vendor defaults.
8239 * PrivateTmp= now affects both /tmp and /var/tmp.
8241 * Boot time status messages are now much prettier and feature
8242 proper english language. Booting up systemd has never been
8245 * Read-ahead pack files now include the inode number of all
8246 files to pre-cache. When the inode changes the pre-caching
8247 is not attempted. This should be nicer to deal with updated
8248 packages which might result in changes of read-ahead
8251 * We now temporaritly lower the kernel's read_ahead_kb variable
8252 when collecting read-ahead data to ensure the kernel's
8253 built-in read-ahead does not add noise to our measurements
8254 of necessary blocks to pre-cache.
8256 * There's now RequiresMountsFor= to add automatic dependencies
8257 for all mounts necessary for a specific file system path.
8259 * MountAuto= and SwapAuto= have been removed from
8260 system.conf. Mounting file systems at boot has to take place
8263 * nspawn now learned a new switch --uuid= to set the machine
8264 ID on the command line.
8266 * nspawn now learned the -b switch to automatically search
8269 * vt102 is now the default TERM for serial TTYs, upgraded from
8272 * systemd-logind now works on VT-less systems.
8274 * The build tree has been reorganized. The individual
8275 components now have directories of their own.
8277 * A new condition type ConditionPathIsReadWrite= is now available.
8279 * nspawn learned the new -C switch to create cgroups for the
8280 container in other hierarchies.
8282 * We now have support for hardware watchdogs, configurable in
8285 * The scheduled shutdown logic now has a public API.
8287 * We now mount /tmp as tmpfs by default, but this can be
8288 masked and /etc/fstab can override it.
8290 * Since udisks does not make use of /media anymore we are not
8291 mounting a tmpfs on it anymore.
8293 * journalctl gained a new --local switch to only interleave
8294 locally generated journal files.
8296 * We can now load the IMA policy at boot automatically.
8298 * The GTK tools have been split off into a systemd-ui.
8300 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8301 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8302 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8303 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8304 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8305 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8306 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8307 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8308 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8313 * This is mostly a bugfix release
8315 * Support optional initialization of the machine ID from the
8316 KVM or container configured UUID.
8318 * Support immediate reboots with "systemctl reboot -ff"
8320 * Show /etc/os-release data in systemd-analyze output
8322 * Many bugfixes for the journal, including endianness fixes and
8323 ensuring that disk space enforcement works
8325 * sd-login.h is C++ compatible again
8327 * Extend the /etc/os-release format on request of the Debian
8330 * We now refuse non-UTF8 strings used in various configuration
8331 and unit files. This is done to ensure we do not pass invalid
8332 data over D-Bus or expose it elsewhere.
8334 * Register Mimo USB Screens as suitable for automatic seat
8337 * Read SELinux client context from journal clients in a race
8340 * Reorder configuration file lookup order. /etc now always
8341 overrides /run in order to allow the administrator to always
8342 and unconditionally override vendor-supplied or
8343 automatically generated data.
8345 * The various user visible bits of the journal now have man
8346 pages. We still lack man pages for the journal API calls
8349 * We now ship all man pages in HTML format again in the
8352 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8353 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8354 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8359 * This is mostly a bugfix release
8361 * systems lacking /etc/os-release are no longer supported.
8363 * Various functionality updates to libsystemd-login.so
8365 * Track class of PAM logins to distinguish greeters from
8368 Contributions from: Kay Sievers, Lennart Poettering, Michael
8373 * This is an important bugfix release for v41.
8375 * Building man pages is now optional which should be useful
8376 for those building systemd from git but unwilling to install
8379 * Watchdog support for supervising services is now usable. In
8380 a future release support for hardware watchdogs
8381 (i.e. /dev/watchdog) will be added building on this.
8383 * Service start rate limiting is now configurable and can be
8384 turned off per service. When a start rate limit is hit a
8385 reboot can automatically be triggered.
8387 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8389 Contributions from: Benjamin Franzke, Bill Nottingham,
8390 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8391 Schmidt, Michał Górny, Piotr Drąg
8395 * The systemd binary is installed /usr/lib/systemd/systemd now;
8396 An existing /sbin/init symlink needs to be adapted with the
8399 * The code that loads kernel modules has been ported to invoke
8400 libkmod directly, instead of modprobe. This means we do not
8401 support systems with module-init-tools anymore.
8403 * Watchdog support is now already useful, but still not
8406 * A new kernel command line option systemd.setenv= is
8407 understood to set system wide environment variables
8408 dynamically at boot.
8410 * We now limit the set of capabilities of systemd-journald.
8412 * We now set SIGPIPE to ignore by default, since it only is
8413 useful in shell pipelines, and has little use in general
8414 code. This can be disabled with IgnoreSIPIPE=no in unit
8417 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8418 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8423 * This is mostly a bugfix release
8425 * We now expose the reason why a service failed in the
8426 "Result" D-Bus property.
8428 * Rudimentary service watchdog support (will be completed over
8429 the next few releases.)
8431 * When systemd forks off in order execute some service we will
8432 now immediately changes its argv[0] to reflect which process
8433 it will execute. This is useful to minimize the time window
8434 with a generic argv[0], which makes bootcharts more useful
8436 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8437 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8438 Mike Kazantsev, Ray Strode
8442 * This is mostly a test release, but incorporates many
8445 * New systemd-cgtop tool to show control groups by their
8448 * Linking against libacl for ACLs is optional again. If
8449 disabled, support tracking device access for active logins
8450 goes becomes unavailable, and so does access to the user
8451 journals by the respective users.
8453 * If a group "adm" exists, journal files are automatically
8454 owned by them, thus allow members of this group full access
8455 to the system journal as well as all user journals.
8457 * The journal now stores the SELinux context of the logging
8458 client for all entries.
8460 * Add C++ inclusion guards to all public headers
8462 * New output mode "cat" in the journal to print only text
8463 messages, without any meta data like date or time.
8465 * Include tiny X server wrapper as a temporary stop-gap to
8466 teach XOrg udev display enumeration. This is used by display
8467 managers such as gdm, and will go away as soon as XOrg
8468 learned native udev hotplugging for display devices.
8470 * Add new systemd-cat tool for executing arbitrary programs
8471 with STDERR/STDOUT connected to the journal. Can also act as
8472 BSD logger replacement, and does so by default.
8474 * Optionally store all locally generated coredumps in the
8475 journal along with meta data.
8477 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8478 writing short strings to files (for usage for /sys), and for
8479 creating symlinks, character and block device nodes.
8481 * New unit file option ControlGroupPersistent= to make cgroups
8482 persistent, following the mechanisms outlined in
8483 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
8485 * Support multiple local RTCs in a sane way
8487 * No longer monopolize IO when replaying readahead data on
8488 rotating disks, since we might starve non-file-system IO to
8489 death, since fanotify() will not see accesses done by blkid,
8492 * Do not show kernel threads in systemd-cgls anymore, unless
8493 requested with new -k switch.
8495 Contributions from: Dan Horák, Kay Sievers, Lennart
8496 Poettering, Michal Schmidt
8500 * This is mostly a test release, but incorporates many
8503 * The git repository moved to:
8504 git://anongit.freedesktop.org/systemd/systemd
8505 ssh://git.freedesktop.org/git/systemd/systemd
8507 * First release with the journal
8508 http://0pointer.de/blog/projects/the-journal.html
8510 * The journal replaces both systemd-kmsg-syslogd and
8511 systemd-stdout-bridge.
8513 * New sd_pid_get_unit() API call in libsystemd-logind
8515 * Many systemadm clean-ups
8517 * Introduce remote-fs-pre.target which is ordered before all
8518 remote mounts and may be used to start services before all
8521 * Added Mageia support
8523 * Add bash completion for systemd-loginctl
8525 * Actively monitor PID file creation for daemons which exit in
8526 the parent process before having finished writing the PID
8527 file in the daemon process. Daemons which do this need to be
8528 fixed (i.e. PID file creation must have finished before the
8529 parent exits), but we now react a bit more gracefully to them.
8531 * Add colourful boot output, mimicking the well-known output
8532 of existing distributions.
8534 * New option PassCredentials= for socket units, for
8535 compatibility with a recent kernel ABI breakage.
8537 * /etc/rc.local is now hooked in via a generator binary, and
8538 thus will no longer act as synchronization point during
8541 * systemctl list-unit-files now supports --root=.
8543 * systemd-tmpfiles now understands two new commands: z, Z for
8544 relabelling files according to the SELinux database. This is
8545 useful to apply SELinux labels to specific files in /sys,
8548 * Output of SysV services is now forwarded to both the console
8549 and the journal by default, not only just the console.
8551 * New man pages for all APIs from libsystemd-login.
8553 * The build tree got reorganized and the build system is a
8554 lot more modular allowing embedded setups to specifically
8555 select the components of systemd they are interested in.
8557 * Support for Linux systems lacking the kernel VT subsystem is
8560 * configure's --with-rootdir= got renamed to
8561 --with-rootprefix= to follow the naming used by udev and
8564 * Unless specified otherwise we will now install to /usr instead
8565 of /usr/local by default.
8567 * Processes with '@' in argv[0][0] are now excluded from the
8568 final shut-down killing spree, following the logic explained
8570 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
8572 * All processes remaining in a service cgroup when we enter
8573 the START or START_PRE states are now killed with
8574 SIGKILL. That means it is no longer possible to spawn
8575 background processes from ExecStart= lines (which was never
8576 supported anyway, and bad style).
8578 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
8579 reloading of units together.
8581 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
8582 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
8583 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
8584 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
8585 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
~ianmdlvl / elogind.git / blob