1 systemd System and Service Manager
5 * The output of 'systemctl list-jobs' got some polishing. The
6 '--type=' argument may now be passed more than once. A new
7 command 'systemctl list-sockets' has been added which shows
8 a list of kernel sockets systemd is listening on with the
9 socket units they belong to, plus the units these socket
12 * The experimental libsystemd-bus library got substantial
13 updates to work in conjunction with the (also experimental)
14 kdbus kernel project. It works well enough to exchange
15 messages with some sophistication. Note that kdbus is not
16 ready yet, and the library is mostly an elaborate test case
17 for now, and not installable.
19 * systemd gained a new unit 'systemd-static-nodes.service'
20 that generates static device nodes earlier during boot, and
21 can run in conjunction with udev.
23 * libsystemd-login gained a new call sd_pid_get_user_unit()
24 to retrieve the user systemd unit a process is running
25 in. This is useful for systems where systemd is used as
28 * systemd-nspawn now places all containers in the new /machine
29 top-level cgroup directory in the name=systemd
30 hierarchy. libvirt will soon do the same, so that we get a
31 uniform separation of /system, /user and /machine for system
32 services, user processes and containers/virtual
33 machines. This new cgroup hierarchy is also useful to stick
34 stable names to specific container instances, which can be
35 recognized later this way (this name may be controlled
36 via systemd-nspawn's new -M switch). libsystemd-login also
37 gained a new call sd_pid_get_machine_name() to retrieve the
38 name of the container/VM a specific process belongs to.
40 * bootchart can now store its data in the journal.
42 * libsystemd-journal gained a new call
43 sd_journal_add_conjunction() for AND expressions to the
44 matching logic. This can be used to express more complex
47 * journactl can now take multiple --unit= and --user-unit=
50 * The cryptsetup logic now understands the "luks.key=" kernel
51 command line switch for specifying a file to read the
52 decryption key from. Also, if a configured key file is not
53 found the tool will now automatically fall back to prompting
56 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
57 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
58 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
59 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
60 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
61 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
65 * journalctl --update-catalog now understands a new --root=
66 option to operate on catalogs found in a different root
69 * During shutdown after systemd has terminated all running
70 services a final killing loop kills all remaining left-over
71 processes. We will now print the name of these processes
72 when we send SIGKILL to them, since this usually indicates a
75 * If /etc/crypttab refers to password files stored on
76 configured mount points automatic dependencies will now be
77 generated to ensure the specific mount is established first
78 before the key file is attempted to be read.
80 * 'systemctl status' will now show information about the
81 network sockets a socket unit is listening on.
83 * 'systemctl status' will also shown information about any
84 drop-in configuration file for units. (Drop-In configuration
85 files in this context are files such as
86 /etc/systemd/systemd/foobar.service.d/*.conf)
88 * systemd-cgtop now optionally shows summed up CPU times of
89 cgroups. Press '%' while running cgtop to switch between
90 percentage and absolute mode. This is useful to determine
91 which cgroups use up the most CPU time over the entire
92 runtime of the system. systemd-cgtop has also been updated
93 to be 'pipeable' for processing with further shell tools.
95 * 'hostnamectl set-hostname' will now allow setting of FQDN
98 * The formatting and parsing of time span values has been
99 changed. The parser now understands fractional expressions
100 such as "5.5h". The formatter will now output fractional
101 expressions for all time spans under 1min, i.e. "5.123456s"
102 rather than "5s 123ms 456us". For time spans under 1s
103 millisecond values are shown, for those under 1ms
104 microsecond values are shown. This should greatly improve
105 all time-related output of systemd.
107 * libsystemd-login and libsystemd-journal gained new
108 functions for querying the poll() events mask and poll()
109 timeout value for integration into arbitrary event
112 * localectl gained the ability to list available X11 keymaps
113 (models, layouts, variants, options).
115 * 'systemd-analyze dot' gained the ability to filter for
116 specific units via shell-style globs, to create smaller,
117 more useful graphs. I.e. it's now possible to create simple
118 graphs of all the dependencies between only target units, or
119 of all units that Avahi has dependencies with.
121 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
122 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
123 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
124 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
125 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
126 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
127 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
131 * The boot-time readahead implementation for rotating media
132 will now read the read-ahead data in multiple passes which
133 consist of all read requests made in equidistant time
134 intervals. This means instead of strictly reading read-ahead
135 data in its physical order on disk we now try to find a
136 middle ground between physical and access time order.
138 * /etc/os-release files gained a new BUILD_ID= field for usage
139 on operating systems that provide continuous builds of OS
142 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
143 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
144 William Douglas, Zbigniew Jędrzejewski-Szmek
148 * systemd-python gained an API exposing libsystemd-daemon.
150 * The SMACK setup logic gained support for uploading CIPSO
153 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
154 ReadOnlyDirectories= and InaccessibleDirectories= has
155 changed. The private /tmp and /var/tmp directories are now
156 shared by all processes of a service (which means
157 ExecStartPre= may now leave data in /tmp that ExecStart= of
158 the same service can still access). When a service is
159 stopped its temporary directories are immediately deleted
160 (normal clean-up with tmpfiles is still done in addition to
163 * By default, systemd will now set a couple of sysctl
164 variables in the kernel: the safe sysrq options are turned
165 on, IP route verification is turned on, and source routing
166 disabled. The recently added hardlink and softlink
167 protection of the kernel is turned on. These settings should
168 be reasonably safe, and good defaults for all new systems.
170 * The predictable network naming logic may now be turned off
171 with a new kernel command line switch: net.ifnames=0.
173 * A new libsystemd-bus module has been added that implements a
174 pretty complete D-Bus client library. For details see:
176 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
178 * journald will now explicitly flush the journal files to disk
179 at the latest 5min after each write. The file will then also
180 be marked offline until the next write. This should increase
181 reliability in case of a crash. The synchronization delay
182 can be configured via SyncIntervalSec= in journald.conf.
184 * There's a new remote-fs-setup.target unit that can be used
185 to pull in specific services when at least one remote file
186 system is to be mounted.
188 * There are new targets timers.target and paths.target as
189 canonical targets to pull user timer and path units in
190 from. This complements sockets.target with a similar
191 purpose for socket units.
193 * libudev gained a new call udev_device_set_attribute_value()
194 to set sysfs attributes of a device.
196 * The udev daemon now sets the default number of worker
197 processes executed in parallel based on the number of available
198 CPUs instead of the amount of available RAM. This is supposed
199 to provide a more reliable default and limit a too aggressive
200 paralellism for setups with 1000s of devices connected.
202 Contributions from: Auke Kok, Colin Walters, Cristian
203 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
204 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
205 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
206 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
207 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
208 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
209 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
210 Zbigniew Jędrzejewski-Szmek
214 * Configuration of unit files may now be extended via drop-in
215 files without having to edit/override the unit files
216 themselves. More specifically, if the administrator wants to
217 change one value for a service file foobar.service he can
218 now do so by dropping in a configuration snippet into
219 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
220 will load all these snippets and apply them on top of the
221 main unit configuration file, possibly extending or
222 overriding its settings. Using these drop-in snippets is
223 generally nicer than the two earlier options for changing
224 unit files locally: copying the files from
225 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
226 them there; or creating a new file in /etc/systemd/system/
227 that incorporates the original one via ".include". Drop-in
228 snippets into these .d/ directories can be placed in any
229 directory systemd looks for units in, and the usual
230 overriding semantics between /usr/lib, /etc and /run apply
233 * Most unit file settings which take lists of items can now be
234 reset by assigning the empty string to them. For example,
235 normally, settings such as Environment=FOO=BAR append a new
236 environment variable assignment to the environment block,
237 each time they are used. By assigning Environment= the empty
238 string the environment block can be reset to empty. This is
239 particularly useful with the .d/*.conf drop-in snippets
240 mentioned above, since this adds the ability to reset list
241 settings from vendor unit files via these drop-ins.
243 * systemctl gained a new "list-dependencies" command for
244 listing the dependencies of a unit recursively.
246 * Inhibitors are now honored and listed by "systemctl
247 suspend", "systemctl poweroff" (and similar) too, not only
248 GNOME. These commands will also list active sessions by
251 * Resource limits (as exposed by the various control group
252 controllers) can now be controlled dynamically at runtime
253 for all units. More specifically, you can now use a command
254 like "systemctl set-cgroup-attr foobar.service cpu.shares
255 2000" to alter the CPU shares a specific service gets. These
256 settings are stored persistently on disk, and thus allow the
257 administrator to easily adjust the resource usage of
258 services with a few simple commands. This dynamic resource
259 management logic is also available to other programs via the
260 bus. Almost any kernel cgroup attribute and controller is
263 * systemd-vconsole-setup will now copy all font settings to
264 all allocated VTs, where it previously applied them only to
267 * libsystemd-login gained the new sd_session_get_tty() API
270 * This release drops support for a few legacy or
271 distribution-specific LSB facility names when parsing init
272 scripts: $x-display-manager, $mail-transfer-agent,
273 $mail-transport-agent, $mail-transfer-agent, $smtp,
274 $null. Also, the mail-transfer-agent.target unit backing
275 this has been removed. Distributions which want to retain
276 compatibility with this should carry the burden for
277 supporting this themselves and patch support for these back
278 in, if they really need to. Also, the facilities $syslog and
279 $local_fs are now ignored, since systemd does not support
280 early-boot LSB init scripts anymore, and these facilities
281 are implied anyway for normal services. syslog.target has
284 * There are new bus calls on PID1's Manager object for
285 cancelling jobs, and removing snapshot units. Previously,
286 both calls were only available on the Job and Snapshot
289 * systemd-journal-gatewayd gained SSL support.
291 * The various "environment" files, such as /etc/locale.conf
292 now support continuation lines with a backslash ("\") as
293 last character in the line, similar in style (but different)
294 to how this is supported in shells.
296 * For normal user processes the _SYSTEMD_USER_UNIT= field is
297 now implicitly appended to every log entry logged. systemctl
298 has been updated to filter by this field when operating on a
299 user systemd instance.
301 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
302 CAP_AUDIT_CONTROL capabilities to the capabilities set for
303 the container. This makes it easier to boot unmodified
304 Fedora systems in a container, which however still requires
305 audit=0 to be passed on the kernel command line. Auditing in
306 kernel and userspace is unfortunately still too broken in
307 context of containers, hence we recommend compiling it out
308 of the kernel or using audit=0. Hopefully this will be fixed
309 one day for good in the kernel.
311 * nspawn gained the new --bind= and --bind-ro= parameters to
312 bind mount specific directories from the host into the
315 * nspawn will now mount its own devpts file system instance
316 into the container, in order not to leak pty devices from
317 the host into the container.
319 * systemd will now read the firmware boot time performance
320 information from the EFI variables, if the used boot loader
321 supports this, and takes it into account for boot performance
322 analysis via "systemd-analyze". This is currently supported
323 only in conjunction with Gummiboot, but could be supported
324 by other boot loaders too. For details see:
326 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
328 * A new generator has been added that automatically mounts the
329 EFI System Partition (ESP) to /boot, if that directory
330 exists, is empty, and no other file system has been
331 configured to be mounted there.
333 * logind will now send out PrepareForSleep(false) out
334 unconditionally, after coming back from suspend. This may be
335 used by applications as asynchronous notification for
336 system resume events.
338 * "systemctl unlock-sessions" has been added, that allows
339 unlocking the screens of all user sessions at once, similar
340 how "systemctl lock-sessions" already locked all users
341 sessions. This is backed by a new D-Bus call UnlockSessions().
343 * "loginctl seat-status" will now show the master device of a
344 seat. (i.e. the device of a seat that needs to be around for
345 the seat to be considered available, usually the graphics
348 * tmpfiles gained a new "X" line type, that allows
349 configuration of files and directories (with wildcards) that
350 shall be excluded from automatic cleanup ("aging").
352 * udev default rules set the device node permissions now only
353 at "add" events, and do not change them any longer with a
354 later "change" event.
356 * The log messages for lid events and power/sleep keypresses
357 now carry a message ID.
359 * We now have a substantially larger unit test suite, but this
360 continues to be work in progress.
362 * udevadm hwdb gained a new --root= parameter to change the
363 root directory to operate relative to.
365 * logind will now issue a background sync() request to the kernel
366 early at shutdown, so that dirty buffers are flushed to disk early
367 instead of at the last moment, in order to optimize shutdown
370 * A new bootctl tool has been added that is an interface for
371 certain boot loader operations. This is currently a preview
372 and is likely to be extended into a small mechanism daemon
373 like timedated, localed, hostnamed, and can be used by
374 graphical UIs to enumerate available boot options, and
375 request boot into firmware operations.
377 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
378 the rest of the package. It also has been updated to work
379 correctly in initrds.
381 * Policykit previously has been runtime optional, and is now
382 also compile time optional via a configure switch.
384 * systemd-analyze has been reimplemented in C. Also "systemctl
385 dot" has moved into systemd-analyze.
387 * "systemctl status" with no further parameters will now print
388 the status of all active or failed units.
390 * Operations such as "systemctl start" can now be executed
391 with a new mode "--irreversible" which may be used to queue
392 operations that cannot accidentally be reversed by a later
393 job queuing. This is by default used to make shutdown
394 requests more robust.
396 * The Python API of systemd now gained a new module for
397 reading journal files.
399 * A new tool kernel-install has been added that can install
400 kernel images according to the Boot Loader Specification:
402 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
404 * Boot time console output has been improved to provide
405 animated boot time output for hanging jobs.
407 * A new tool systemd-activate has been added which can be used
408 to test socket activation with, directly from the command
409 line. This should make it much easier to test and debug
410 socket activation in daemons.
412 * journalctl gained a new "--reverse" (or -r) option to show
413 journal output in reverse order (i.e. newest line first).
415 * journalctl gained a new "--pager-end" (or -e) option to jump
416 to immediately jump to the end of the journal in the
417 pager. This is only supported in conjunction with "less".
419 * journalctl gained a new "--user-unit=" option, that works
420 similar to "--unit=" but filters for user units rather than
423 * A number of unit files to ease adoption of systemd in
424 initrds has been added. This moves some minimal logic from
425 the various initrd implementations into systemd proper.
427 * The journal files are now owned by a new group
428 "systemd-journal", which exists specifically to allow access
429 to the journal, and nothing else. Previously, we used the
430 "adm" group for that, which however possibly covers more
431 than just journal/log file access. This new group is now
432 already used by systemd-journal-gatewayd to ensure this
433 daemon gets access to the journal files and as little else
434 as possible. Note that "make install" will also set FS ACLs
435 up for /var/log/journal to give "adm" and "wheel" read
436 access to it, in addition to "systemd-journal" which owns
437 the journal files. We recommend that packaging scripts also
438 add read access to "adm" + "wheel" to /var/log/journal, and
439 all existing/future journal files. To normal users and
440 administrators little changes, however packagers need to
441 ensure to create the "systemd-journal" system group at
442 package installation time.
444 * The systemd-journal-gatewayd now runs as unprivileged user
445 systemd-journal-gateway:systemd-journal-gateway. Packaging
446 scripts need to create these system user/group at
449 * timedated now exposes a new boolean property CanNTP that
450 indicates whether a local NTP service is available or not.
452 * systemd-detect-virt will now also detect xen PVs
454 * The pstore file system is now mounted by default, if it is
457 * In addition to the SELinux and IMA policies we will now also
458 load SMACK policies at early boot.
460 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
461 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
462 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
463 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
464 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
465 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
466 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
467 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
468 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
469 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
470 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
471 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
472 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
473 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
477 * Timer units now support calendar time events in addition to
478 monotonic time events. That means you can now trigger a unit
479 based on a calendar time specification such as "Thu,Fri
480 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
481 or fifth day of any month of the year 2013, given that it is
482 a thursday or friday. This brings timer event support
483 considerably closer to cron's capabilities. For details on
484 the supported calendar time specification language see
487 * udev now supports a number of different naming policies for
488 network interfaces for predictable names, and a combination
489 of these policies is now the default. Please see this wiki
490 document for details:
492 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
494 * Auke Kok's bootchart implementation has been added to the
495 systemd tree. It's an optional component that can graph the
496 boot in quite some detail. It's one of the best bootchart
497 implementations around and minimal in its code and
500 * nss-myhostname has been integrated into the systemd source
501 tree. nss-myhostname guarantees that the local hostname
502 always stays resolvable via NSS. It has been a weak
503 requirement of systemd-hostnamed since a long time, and
504 since its code is actually trivial we decided to just
505 include it in systemd's source tree. It can be turned off
506 with a configure switch.
508 * The read-ahead logic is now capable of properly detecting
509 whether a btrfs file system is on SSD or rotating media, in
510 order to optimize the read-ahead scheme. Previously, it was
511 only capable of detecting this on traditional file systems
514 * In udev, additional device properties are now read from the
515 IAB in addition to the OUI database. Also, Bluetooth company
516 identities are attached to the devices as well.
518 * In service files %U may be used as specifier that is
519 replaced by the configured user name of the service.
521 * nspawn may now be invoked without a controlling TTY. This
522 makes it suitable for invocation as its own service. This
523 may be used to set up a simple containerized server system
524 using only core OS tools.
526 * systemd and nspawn can now accept socket file descriptors
527 when they are started for socket activation. This enables
528 implementation of socket activated nspawn
529 containers. i.e. think about autospawning an entire OS image
530 when the first SSH or HTTP connection is received. We expect
531 that similar functionality will also be added to libvirt-lxc
534 * journalctl will now suppress ANSI color codes when
537 * systemctl will no longer show control group information for
538 a unit if a the control group is empty anyway.
540 * logind can now automatically suspend/hibernate/shutdown the
543 * /etc/machine-info and hostnamed now also expose the chassis
544 type of the system. This can be used to determine whether
545 the local system is a laptop, desktop, handset or
546 tablet. This information may either be configured by the
547 user/vendor or is automatically determined from ACPI and DMI
548 information if possible.
550 * A number of PolicyKit actions are now bound together with
551 "imply" rules. This should simplify creating UIs because
552 many actions will now authenticate similar ones as well.
554 * Unit files learnt a new condition ConditionACPower= which
555 may be used to conditionalize a unit depending on whether an
556 AC power source is connected or not, of whether the system
557 is running on battery power.
559 * systemctl gained a new "is-failed" verb that may be used in
560 shell scripts and suchlike to check whether a specific unit
561 is in the "failed" state.
563 * The EnvironmentFile= setting in unit files now supports file
564 globbing, and can hence be used to easily read a number of
565 environment files at once.
567 * systemd will no longer detect and recognize specific
568 distributions. All distribution-specific #ifdeffery has been
569 removed, systemd is now fully generic and
570 distribution-agnostic. Effectively, not too much is lost as
571 a lot of the code is still accessible via explicit configure
572 switches. However, support for some distribution specific
573 legacy configuration file formats has been dropped. We
574 recommend distributions to simply adopt the configuration
575 files everybody else uses now and convert the old
576 configuration from packaging scripts. Most distributions
577 already did that. If that's not possible or desirable,
578 distributions are welcome to forward port the specific
579 pieces of code locally from the git history.
581 * When logging a message about a unit systemd will now always
582 log the unit name in the message meta data.
584 * localectl will now also discover system locale data that is
585 not stored in locale archives, but directly unpacked.
587 * logind will no longer unconditionally use framebuffer
588 devices as seat masters, i.e. as devices that are required
589 to be existing before a seat is considered preset. Instead,
590 it will now look for all devices that are tagged as
591 "seat-master" in udev. By default framebuffer devices will
592 be marked as such, but depending on local systems other
593 devices might be marked as well. This may be used to
594 integrate graphics cards using closed source drivers (such
595 as NVidia ones) more nicely into logind. Note however, that
596 we recommend using the open source NVidia drivers instead,
597 and no udev rules for the closed-source drivers will be
598 shipped from us upstream.
600 Contributions from: Adam Williamson, Alessandro Crismani, Auke
601 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
602 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
603 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
604 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
605 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
606 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
607 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
608 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
609 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
610 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
611 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
616 * udev gained support for loading additional device properties
617 from an indexed database that is keyed by vendor/product IDs
618 and similar device identifiers. For the beginning this
619 "hwdb" is populated with data from the well-known PCI and
620 USB database, but also includes PNP, ACPI and OID data. In
621 the longer run this indexed database shall grow into
622 becoming the one central database for non-essential
623 userspace device metadata. Previously, data from the PCI/USB
624 database was only attached to select devices, since the
625 lookup was a relatively expensive operation due to O(n) time
626 complexity (with n being the number of entries in the
627 database). Since this is now O(1), we decided to add in this
628 data for all devices where this is available, by
629 default. Note that the indexed database needs to be rebuilt
630 when new data files are installed. To achieve this you need
631 to update your packaging scripts to invoke "udevadm hwdb
632 --update" after installation of hwdb data files. For
633 RPM-based distributions we introduced the new
634 %udev_hwdb_update macro for this purpose.
636 * The Journal gained support for the "Message Catalog", an
637 indexed database to link up additional information with
638 journal entries. For further details please check:
640 http://www.freedesktop.org/wiki/Software/systemd/catalog
642 The indexed message catalog database also needs to be
643 rebuilt after installation of message catalog files. Use
644 "journalctl --update-catalog" for this. For RPM-based
645 distributions we introduced the %journal_catalog_update
646 macro for this purpose.
648 * The Python Journal bindings gained support for the standard
649 Python logging framework.
651 * The Journal API gained new functions for checking whether
652 the underlying file system of a journal file is capable of
653 properly reporting file change notifications, or whether
654 applications that want to reflect journal changes "live"
655 need to recheck journal files continuously in appropriate
658 * It is now possible to set the "age" field for tmpfiles
659 entries to 0, indicating that files matching this entry
660 shall always be removed when the directories are cleaned up.
662 * coredumpctl gained a new "gdb" verb which invokes gdb
663 right-away on the selected coredump.
665 * There's now support for "hybrid sleep" on kernels that
666 support this, in addition to "suspend" and "hibernate". Use
667 "systemctl hybrid-sleep" to make use of this.
669 * logind's HandleSuspendKey= setting (and related settings)
670 now gained support for a new "lock" setting to simply
671 request the screen lock on all local sessions, instead of
672 actually executing a suspend or hibernation.
674 * systemd will now mount the EFI variables file system by
677 * Socket units now gained support for configuration of the
678 SMACK security label.
680 * timedatectl will now output the time of the last and next
681 daylight saving change.
683 * We dropped support for various legacy and distro-specific
684 concepts, such as insserv, early-boot SysV services
685 (i.e. those for non-standard runlevels such as 'b' or 'S')
686 or ArchLinux /etc/rc.conf support. We recommend the
687 distributions who still need support this to either continue
688 to maintain the necessary patches downstream, or find a
689 different solution. (Talk to us if you have questions!)
691 * Various systemd components will now bypass PolicyKit checks
692 for root and otherwise handle properly if PolicyKit is not
693 found to be around. This should fix most issues for
694 PolicyKit-less systems. Quite frankly this should have been
695 this way since day one. It is absolutely our intention to
696 make systemd work fine on PolicyKit-less systems, and we
697 consider it a bug if something doesn't work as it should if
698 PolicyKit is not around.
700 * For embedded systems it is now possible to build udev and
701 systemd without blkid and/or kmod support.
703 * "systemctl switch-root" is now capable of switching root
704 more than once. I.e. in addition to transitions from the
705 initrd to the host OS it is now possible to transition to
706 further OS images from the host. This is useful to implement
707 offline updating tools.
709 * Various other additions have been made to the RPM macros
710 shipped with systemd. Use %udev_rules_update() after
711 installing new udev rules files. %_udevhwdbdir,
712 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
713 %_sysctldir are now available which resolve to the right
714 directories for packages to place various data files in.
716 * journalctl gained the new --full switch (in addition to
717 --all, to disable ellipsation for long messages.
719 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
720 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
721 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
722 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
723 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
724 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
725 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
726 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
727 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
731 * journalctl gained new --since= and --until= switches to
732 filter by time. It also now supports nice filtering for
733 units via --unit=/-u.
735 * Type=oneshot services may use ExecReload= and do the
738 * The journal daemon now supports time-based rotation and
739 vacuuming, in addition to the usual disk-space based
742 * The journal will now index the available field values for
743 each field name. This enables clients to show pretty drop
744 downs of available match values when filtering. The bash
745 completion of journalctl has been updated
746 accordingly. journalctl gained a new switch -F to list all
747 values a certain field takes in the journal database.
749 * More service events are now written as structured messages
750 to the journal, and made recognizable via message IDs.
752 * The timedated, localed and hostnamed mini-services which
753 previously only provided support for changing time, locale
754 and hostname settings from graphical DEs such as GNOME now
755 also have a minimal (but very useful) text-based client
756 utility each. This is probably the nicest way to changing
757 these settings from the command line now, especially since
758 it lists available options and is fully integrated with bash
761 * There's now a new tool "systemd-coredumpctl" to list and
762 extract coredumps from the journal.
764 * We now install a README each in /var/log/ and
765 /etc/rc.d/init.d explaining where the system logs and init
766 scripts went. This hopefully should help folks who go to
767 that dirs and look into the otherwise now empty void and
770 * When user-services are invoked (by systemd --user) the
771 $MANAGERPID env var is set to the PID of systemd.
773 * SIGRTMIN+24 when sent to a --user instance will now result
774 in immediate termination of systemd.
776 * gatewayd received numerous feature additions such as a
777 "follow" mode, for live syncing and filtering.
779 * browse.html now allows filtering and showing detailed
780 information on specific entries. Keyboard navigation and
781 mouse screen support has been added.
783 * gatewayd/journalctl now supports HTML5/JSON
784 Server-Sent-Events as output.
786 * The SysV init script compatibility logic will now
787 heuristically determine whether a script supports the
788 "reload" verb, and only then make this available as
791 * "systemctl status --follow" has been removed, use "journalctl
794 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
795 have been removed since they are hardly useful to be
798 * And I'd like to take the opportunity to specifically mention
799 Zbigniew for his great contributions. Zbigniew, you rock!
801 Contributions from: Andrew Eikum, Christian Hesse, Colin
802 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
803 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
804 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
805 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
806 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
807 Jędrzejewski-Szmek, Сковорода Никита Андреевич
811 * If /etc/vconsole.conf is non-existent or empty we will no
812 longer load any console font or key map at boot by
813 default. Instead the kernel defaults will be left
814 intact. This is definitely the right thing to do, as no
815 configuration should mean no configuration, and hard-coding
816 font names that are different on all archs is probably a bad
817 idea. Also, the kernel default key map and font should be
818 good enough for most cases anyway, and mostly identical to
819 the userspace fonts/key maps we previously overloaded them
820 with. If distributions want to continue to default to a
821 non-kernel font or key map they should ship a default
822 /etc/vconsole.conf with the appropriate contents.
824 Contributions from: Colin Walters, Daniel J Walsh, Dave
825 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
826 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
830 * journalctl gained a new --cursor= switch to show entries
831 starting from the specified location in the journal.
833 * We now enforce a size limit on journal entry fields exported
834 with "-o json" in journalctl. Fields larger than 4K will be
835 assigned null. This can be turned off with --all.
837 * An (optional) journal gateway daemon is now available as
838 "systemd-journal-gatewayd.service". This service provides
839 access to the journal via HTTP and JSON. This functionality
840 will be used to implement live log synchronization in both
841 pull and push modes, but has various other users too, such
842 as easy log access for debugging of embedded devices. Right
843 now it is already useful to retrieve the journal via HTTP:
845 # systemctl start systemd-journal-gatewayd.service
846 # wget http://localhost:19531/entries
848 This will download the journal contents in a
849 /var/log/messages compatible format. The same as JSON:
851 # curl -H"Accept: application/json" http://localhost:19531/entries
853 This service is also accessible via a web browser where a
854 single static HTML5 app is served that uses the JSON logic
855 to enable the user to do some basic browsing of the
856 journal. This will be extended later on. Here's an example
857 screenshot of this app in its current state:
859 http://0pointer.de/public/journal-gatewayd
861 Contributions from: Kay Sievers, Lennart Poettering, Robert
862 Milasan, Tom Gundersen
866 * The bash completion logic is now available for journalctl
869 * We don't mount the "cpuset" controller anymore together with
870 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
871 started if no parameters are assigned to it. "cpuset" hence
872 broke code that assumed it it could create "cpu" groups and
875 * journalctl -f will now subscribe to terminal size changes,
876 and line break accordingly.
878 Contributions from: Dave Reisner, Kay Sievers, Lennart
879 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
883 * nspawn will now create a symlink /etc/localtime in the
884 container environment, copying the host's timezone
885 setting. Previously this has been done via a bind mount, but
886 since symlinks cannot be bind mounted this has now been
887 changed to create/update the appropriate symlink.
889 * journalctl -n's line number argument is now optional, and
890 will default to 10 if omitted.
892 * journald will now log the maximum size the journal files may
893 take up on disk. This is particularly useful if the default
894 built-in logic of determining this parameter from the file
895 system size is used. Use "systemctl status
896 systemd-journald.service" to see this information.
898 * The multi-seat X wrapper tool has been stripped down. As X
899 is now capable of enumerating graphics devices via udev in a
900 seat-aware way the wrapper is not strictly necessary
901 anymore. A stripped down temporary stop-gap is still shipped
902 until the upstream display managers have been updated to
903 fully support the new X logic. Expect this wrapper to be
904 removed entirely in one of the next releases.
906 * HandleSleepKey= in logind.conf has been split up into
907 HandleSuspendKey= and HandleHibernateKey=. The old setting
908 is not available anymore. X11 and the kernel are
909 distuingishing between these keys and we should too. This
910 also means the inhibition lock for these keys has been split
913 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
914 Poettering, Lukas Nykryn, Václav Pavlín
918 * Whenever a unit changes state we'll now log this to the
919 journal and show along the unit's own log output in
922 * ConditionPathIsMountPoint= can now properly detect bind
923 mount points too. (Previously, a bind mount of one file
924 system to another place in the same file system could not be
925 detected as mount, since they shared struct stat's st_dev
928 * We will now mount the cgroup controllers cpu, cpuacct,
929 cpuset and the controllers net_cls, net_prio together by
932 * nspawn containers will now have a virtualized boot
933 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
934 over with a randomized ID at container initialization). This
935 has the effect of making "journalctl -b" do the right thing
938 * The JSON output journal serialization has been updated not
939 to generate "endless" list objects anymore, but rather one
940 JSON object per line. This is more in line how most JSON
941 parsers expect JSON objects. The new output mode
942 "json-pretty" has been added to provide similar output, but
943 neatly aligned for readability by humans.
945 * We dropped all explicit sync() invocations in the shutdown
946 code. The kernel does this implicitly anyway in the kernel
947 reboot() syscall. halt(8)'s -n option is now a compatibility
950 * We now support virtualized reboot() in containers, as
951 supported by newer kernels. We will fall back to exit() if
952 CAP_SYS_REBOOT is not available to the container. Also,
953 nspawn makes use of this now and will actually reboot the
954 container if the containerized OS asks for that.
956 * journalctl will only show local log output by default
957 now. Use --merge (-m) to show remote log output, too.
959 * libsystemd-journal gained the new sd_journal_get_usage()
960 call to determine the current disk usage of all journal
961 files. This is exposed in the new "journalctl --disk-usage"
964 * journald gained a new configuration setting SplitMode= in
965 journald.conf which may be used to control how user journals
966 are split off. See journald.conf(5) for details.
968 * A new condition type ConditionFileNotEmpty= has been added.
970 * tmpfiles' "w" lines now support file globbing, to write
971 multiple files at once.
973 * We added Python bindings for the journal submission
974 APIs. More Python APIs for a number of selected APIs will
975 likely follow. Note that we intend to add native bindings
976 only for the Python language, as we consider it common
977 enough to deserve bindings shipped within systemd. There are
978 various projects outside of systemd that provide bindings
979 for languages such as PHP or Lua.
981 * Many conditions will now resolve specifiers such as %i. In
982 addition, PathChanged= and related directives of .path units
983 now support specifiers as well.
985 * There's now a new RPM macro definition for the system preset
988 * journald will now warn if it can't forward a message to the
989 syslog daemon because it's socket is full.
991 * timedated will no longer write or process /etc/timezone,
992 except on Debian. As we do not support late mounted /usr
993 anymore /etc/localtime always being a symlink is now safe,
994 and hence the information in /etc/timezone is not necessary
997 * logind will now always reserve one VT for a text getty (VT6
998 by default). Previously if more than 6 X sessions where
999 started they took up all the VTs with auto-spawned gettys,
1000 so that no text gettys were available anymore.
1002 * udev will now automatically inform the btrfs kernel logic
1003 about btrfs RAID components showing up. This should make
1004 simple hotplug based btrfs RAID assembly work.
1006 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
1007 (but not for its children which will stay at the kernel
1008 default). This should allow setups with a lot more listening
1011 * systemd will now always pass the configured timezone to the
1012 kernel at boot. timedated will do the same when the timezone
1015 * logind's inhibition logic has been updated. By default,
1016 logind will now handle the lid switch, the power and sleep
1017 keys all the time, even in graphical sessions. If DEs want
1018 to handle these events on their own they should take the new
1019 handle-power-key, handle-sleep-key and handle-lid-switch
1020 inhibitors during their runtime. A simple way to achiveve
1021 that is to invoke the DE wrapped in an invocation of:
1023 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
1025 * Access to unit operations is now checked via SELinux taking
1026 the unit file label and client process label into account.
1028 * systemd will now notify the administrator in the journal
1029 when he over-mounts a non-empty directory.
1031 * There are new specifiers that are resolved in unit files,
1032 for the host name (%H), the machine ID (%m) and the boot ID
1035 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
1036 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
1037 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
1038 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
1039 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
1040 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
1041 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
1045 * Support for reading structured kernel messages from
1046 /dev/kmsg has now been added and is enabled by default.
1048 * Support for reading kernel messages from /proc/kmsg has now
1049 been removed. If you want kernel messages in the journal
1050 make sure to run a recent kernel (>= 3.5) that supports
1051 reading structured messages from /dev/kmsg (see
1052 above). /proc/kmsg is now exclusive property of classic
1053 syslog daemons again.
1055 * The libudev API gained the new
1056 udev_device_new_from_device_id() call.
1058 * The logic for file system namespace (ReadOnlyDirectory=,
1059 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
1060 require pivot_root() anymore. This means fewer temporary
1061 directories are created below /tmp for this feature.
1063 * nspawn containers will now see and receive all submounts
1064 made on the host OS below the root file system of the
1067 * Forward Secure Sealing is now supported for Journal files,
1068 which provide cryptographical sealing of journal files so
1069 that attackers cannot alter log history anymore without this
1070 being detectable. Lennart will soon post a blog story about
1071 this explaining it in more detail.
1073 * There are two new service settings RestartPreventExitStatus=
1074 and SuccessExitStatus= which allow configuration of exit
1075 status (exit code or signal) which will be excepted from the
1076 restart logic, resp. consider successful.
1078 * journalctl gained the new --verify switch that can be used
1079 to check the integrity of the structure of journal files and
1080 (if Forward Secure Sealing is enabled) the contents of
1083 * nspawn containers will now be run with /dev/stdin, /dev/fd/
1084 and similar symlinks pre-created. This makes running shells
1085 as container init process a lot more fun.
1087 * The fstab support can now handle PARTUUID= and PARTLABEL=
1090 * A new ConditionHost= condition has been added to match
1091 against the hostname (with globs) and machine ID. This is
1092 useful for clusters where a single OS image is used to
1093 provision a large number of hosts which shall run slightly
1094 different sets of services.
1096 * Services which hit the restart limit will now be placed in a
1099 Contributions from: Bertram Poettering, Dave Reisner, Huang
1100 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
1101 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
1105 * When running in --user mode systemd will now become a
1106 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
1107 tree a lot more organized.
1109 * A new PartOf= unit dependency type has been introduced that
1110 may be used to group services in a natural way.
1112 * "systemctl enable" may now be used to enable instances of
1115 * journalctl now prints error log levels in red, and
1116 warning/notice log levels in bright white. It also supports
1117 filtering by log level now.
1119 * cgtop gained a new -n switch (similar to top), to configure
1120 the maximum number of iterations to run for. It also gained
1121 -b, to run in batch mode (accepting no input).
1123 * The suffix ".service" may now be omitted on most systemctl
1124 command lines involving service unit names.
1126 * There's a new bus call in logind to lock all sessions, as
1127 well as a loginctl verb for it "lock-sessions".
1129 * libsystemd-logind.so gained a new call sd_journal_perror()
1130 that works similar to libc perror() but logs to the journal
1131 and encodes structured information about the error number.
1133 * /etc/crypttab entries now understand the new keyfile-size=
1136 * shutdown(8) now can send a (configurable) wall message when
1137 a shutdown is cancelled.
1139 * The mount propagation mode for the root file system will now
1140 default to "shared", which is useful to make containers work
1141 nicely out-of-the-box so that they receive new mounts from
1142 the host. This can be undone locally by running "mount
1143 --make-rprivate /" if needed.
1145 * The prefdm.service file has been removed. Distributions
1146 should maintain this unit downstream if they intend to keep
1147 it around. However, we recommend writing normal unit files
1148 for display managers instead.
1150 * Since systemd is a crucial part of the OS we will now
1151 default to a number of compiler switches that improve
1152 security (hardening) such as read-only relocations, stack
1153 protection, and suchlike.
1155 * The TimeoutSec= setting for services is now split into
1156 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1157 of individual time outs for the start and the stop phase of
1160 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1161 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1162 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1163 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1164 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1165 Gundersen, Zbigniew Jędrzejewski-Szmek
1169 * The journal and id128 C APIs are now fully documented as man
1172 * Extra safety checks have been added when transitioning from
1173 the initial RAM disk to the main system to avoid accidental
1176 * /etc/crypttab entries now understand the new keyfile-offset=
1179 * systemctl -t can now be used to filter by unit load state.
1181 * The journal C API gained the new sd_journal_wait() call to
1182 make writing synchronous journal clients easier.
1184 * journalctl gained the new -D switch to show journals from a
1187 * journalctl now displays a special marker between log
1188 messages of two different boots.
1190 * The journal is now explicitly flushed to /var via a service
1191 systemd-journal-flush.service, rather than implicitly simply
1192 by seeing /var/log/journal to be writable.
1194 * journalctl (and the journal C APIs) can now match for much
1195 more complex expressions, with alternatives and
1198 * When transitioning from the initial RAM disk to the main
1199 system we will now kill all processes in a killing spree to
1200 ensure no processes stay around by accident.
1202 * Three new specifiers may be used in unit files: %u, %h, %s
1203 resolve to the user name, user home directory resp. user
1204 shell. This is useful for running systemd user instances.
1206 * We now automatically rotate journal files if their data
1207 object hash table gets a fill level > 75%. We also size the
1208 hash table based on the configured maximum file size. This
1209 together should lower hash collisions drastically and thus
1210 speed things up a bit.
1212 * journalctl gained the new "--header" switch to introspect
1213 header data of journal files.
1215 * A new setting SystemCallFilters= has been added to services
1216 which may be used to apply blacklists or whitelists to
1217 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1219 * nspawn gained a new --link-journal= switch (and quicker: -j)
1220 to link the container journal with the host. This makes it
1221 very easy to centralize log viewing on the host for all
1222 guests while still keeping the journal files separated.
1224 * Many bugfixes and optimizations
1226 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1227 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1228 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1233 * Several tools now understand kernel command line arguments,
1234 which are only read when run in an initial RAM disk. They
1235 usually follow closely their normal counterparts, but are
1238 * There's a new tool to analyze the readahead files that are
1239 automatically generated at boot. Use:
1241 /usr/lib/systemd/systemd-readahead analyze /.readahead
1243 * We now provide an early debug shell on tty9 if this enabled. Use:
1245 systemctl enable debug-shell.service
1247 * All plymouth related units have been moved into the Plymouth
1248 package. Please make sure to upgrade your Plymouth version
1251 * systemd-tmpfiles now supports getting passed the basename of
1252 a configuration file only, in which case it will look for it
1253 in all appropriate directories automatically.
1255 * udevadm info now takes a /dev or /sys path as argument, and
1256 does the right thing. Example:
1258 udevadm info /dev/sda
1259 udevadm info /sys/class/block/sda
1261 * systemctl now prints a warning if a unit is stopped but a
1262 unit that might trigger it continues to run. Example: a
1263 service is stopped but the socket that activates it is left
1266 * "systemctl status" will now mention if the log output was
1267 shortened due to rotation since a service has been started.
1269 * The journal API now exposes functions to determine the
1270 "cutoff" times due to rotation.
1272 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1273 immediately flushing of runtime logs to /var if possible,
1274 resp. for triggering immediate rotation of the journal
1277 * It is now considered an error if a service is attempted to
1278 be stopped that is not loaded.
1280 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1282 * systemd-analyze now supports Python 3
1284 * tmpfiles now supports cleaning up directories via aging
1285 where the first level dirs are always kept around but
1286 directories beneath it automatically aged. This is enabled
1287 by prefixing the age field with '~'.
1289 * Seat objects now expose CanGraphical, CanTTY properties
1290 which is required to deal with very fast bootups where the
1291 display manager might be running before the graphics drivers
1292 completed initialization.
1294 * Seat objects now expose a State property.
1296 * We now include RPM macros for service enabling/disabling
1297 based on the preset logic. We recommend RPM based
1298 distributions to make use of these macros if possible. This
1299 makes it simpler to reuse RPM spec files across
1302 * We now make sure that the collected systemd unit name is
1303 always valid when services log to the journal via
1306 * There's a new man page kernel-command-line(7) detailing all
1307 command line options we understand.
1309 * The fstab generator may now be disabled at boot by passing
1310 fstab=0 on the kernel command line.
1312 * A new kernel command line option modules-load= is now understood
1313 to load a specific kernel module statically, early at boot.
1315 * Unit names specified on the systemctl command line are now
1316 automatically escaped as needed. Also, if file system or
1317 device paths are specified they are automatically turned
1318 into the appropriate mount or device unit names. Example:
1320 systemctl status /home
1321 systemctl status /dev/sda
1323 * The SysVConsole= configuration option has been removed from
1324 system.conf parsing.
1326 * The SysV search path is no longer exported on the D-Bus
1329 * The Names= option is been removed from unit file parsing.
1331 * There's a new man page bootup(7) detailing the boot process.
1333 * Every unit and every generator we ship with systemd now
1334 comes with full documentation. The self-explanatory boot is
1337 * A couple of services gained "systemd-" prefixes in their
1338 name if they wrap systemd code, rather than only external
1339 code. Among them fsck@.service which is now
1340 systemd-fsck@.service.
1342 * The HaveWatchdog property has been removed from the D-Bus
1345 * systemd.confirm_spawn= on the kernel command line should now
1348 * There's a new man page crypttab(5) which details all options
1349 we actually understand.
1351 * systemd-nspawn gained a new --capability= switch to pass
1352 additional capabilities to the container.
1354 * timedated will now read known NTP implementation unit names
1355 from /usr/lib/systemd/ntp-units.d/*.list,
1356 systemd-timedated-ntp.target has been removed.
1358 * journalctl gained a new switch "-b" that lists log data of
1359 the current boot only.
1361 * The notify socket is in the abstract namespace again, in
1362 order to support daemons which chroot() at start-up.
1364 * There is a new Storage= configuration option for journald
1365 which allows configuration of where log data should go. This
1366 also provides a way to disable journal logging entirely, so
1367 that data collected is only forwarded to the console, the
1368 kernel log buffer or another syslog implementation.
1370 * Many bugfixes and optimizations
1372 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1373 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1374 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1375 Shawn Landden, Tom Gundersen
1379 * "systemctl help <unit>" now shows the man page if one is
1382 * Several new man pages have been added.
1384 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1385 MaxLevelConsole= can now be specified in
1386 journald.conf. These options allow reducing the amount of
1387 data stored on disk or forwarded by the log level.
1389 * TimerSlackNSec= can now be specified in system.conf for
1390 PID1. This allows system-wide power savings.
1392 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1393 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1398 * logind is now capable of (optionally) handling power and
1399 sleep keys as well as the lid switch.
1401 * journalctl now understands the syntax "journalctl
1402 /usr/bin/avahi-daemon" to get all log output of a specific
1405 * CapabilityBoundingSet= in system.conf now also influences
1406 the capability bound set of usermode helpers of the kernel.
1408 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1409 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1410 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1411 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1415 * Note that we skipped 139 releases here in order to set the
1416 new version to something that is greater than both udev's
1417 and systemd's most recent version number.
1419 * udev: all udev sources are merged into the systemd source tree now.
1420 All future udev development will happen in the systemd tree. It
1421 is still fully supported to use the udev daemon and tools without
1422 systemd running, like in initramfs or other init systems. Building
1423 udev though, will require the *build* of the systemd tree, but
1424 udev can be properly *run* without systemd.
1426 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1427 should be used to create dead device nodes as workarounds for broken
1430 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1431 no longer supported. udev_monitor_new_from_netlink() needs to be
1432 used to subscribe to events.
1434 * udev: when udevd is started by systemd, processes which are left
1435 behind by forking them off of udev rules, are unconditionally cleaned
1436 up and killed now after the event handling has finished. Services or
1437 daemons must be started as systemd services. Services can be
1438 pulled-in by udev to get started, but they can no longer be directly
1439 forked by udev rules.
1441 * udev: the daemon binary is called systemd-udevd now and installed
1442 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1443 to adapt to that, create symlink, or rename the binary after building
1446 * libudev no longer provides these symbols:
1447 udev_monitor_from_socket()
1448 udev_queue_get_failed_list_entry()
1449 udev_get_{dev,sys,run}_path()
1450 The versions number was bumped and symbol versioning introduced.
1452 * systemd-loginctl and systemd-journalctl have been renamed
1453 to loginctl and journalctl to match systemctl.
1455 * The config files: /etc/systemd/systemd-logind.conf and
1456 /etc/systemd/systemd-journald.conf have been renamed to
1457 logind.conf and journald.conf. Package updates should rename
1458 the files to the new names on upgrade.
1460 * For almost all files the license is now LGPL2.1+, changed
1461 from the previous GPL2.0+. Exceptions are some minor stuff
1462 of udev (which will be changed to LGPL2.1 eventually, too),
1463 and the MIT licensed sd-daemon.[ch] library that is suitable
1464 to be used as drop-in files.
1466 * systemd and logind now handle system sleep states, in
1467 particular suspending and hibernating.
1469 * logind now implements a sleep/shutdown/idle inhibiting logic
1470 suitable for a variety of uses. Soonishly Lennart will blog
1471 about this in more detail.
1473 * var-run.mount and var-lock.mount are no longer provided
1474 (which prevously bind mounted these directories to their new
1475 places). Distributions which have not converted these
1476 directories to symlinks should consider stealing these files
1477 from git history and add them downstream.
1479 * We introduced the Documentation= field for units and added
1480 this to all our shipped units. This is useful to make it
1481 easier to explore the boot and the purpose of the various
1484 * All smaller setup units (such as
1485 systemd-vconsole-setup.service) now detect properly if they
1486 are run in a container and are skipped when
1487 appropriate. This guarantees an entirely noise-free boot in
1488 Linux container environments such as systemd-nspawn.
1490 * A framework for implementing offline system updates is now
1491 integrated, for details see:
1492 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1494 * A new service type Type=idle is available now which helps us
1495 avoiding ugly interleaving of getty output and boot status
1498 * There's now a system-wide CapabilityBoundingSet= option to
1499 globally reduce the set of capabilities for the
1500 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1501 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1502 even CAP_NET_ADMIN system-wide for secure systems.
1504 * There are now system-wide DefaultLimitXXX= options to
1505 globally change the defaults of the various resource limits
1506 for all units started by PID 1.
1508 * Harald Hoyer's systemd test suite has been integrated into
1509 systemd which allows easy testing of systemd builds in qemu
1510 and nspawn. (This is really awesome! Ask us for details!)
1512 * The fstab parser is now implemented as generator, not inside
1515 * systemctl will now warn you if .mount units generated from
1516 /etc/fstab are out of date due to changes in fstab that
1517 haven't been read by systemd yet.
1519 * systemd is now suitable for usage in initrds. Dracut has
1520 already been updated to make use of this. With this in place
1521 initrds get a slight bit faster but primarily are much
1522 easier to introspect and debug since "systemctl status" in
1523 the host system can be used to introspect initrd services,
1524 and the journal from the initrd is kept around too.
1526 * systemd-delta has been added, a tool to explore differences
1527 between user/admin configuration and vendor defaults.
1529 * PrivateTmp= now affects both /tmp and /var/tmp.
1531 * Boot time status messages are now much prettier and feature
1532 proper english language. Booting up systemd has never been
1535 * Read-ahead pack files now include the inode number of all
1536 files to pre-cache. When the inode changes the pre-caching
1537 is not attempted. This should be nicer to deal with updated
1538 packages which might result in changes of read-ahead
1541 * We now temporaritly lower the kernel's read_ahead_kb variable
1542 when collecting read-ahead data to ensure the kernel's
1543 built-in read-ahead does not add noise to our measurements
1544 of necessary blocks to pre-cache.
1546 * There's now RequiresMountsFor= to add automatic dependencies
1547 for all mounts necessary for a specific file system path.
1549 * MountAuto= and SwapAuto= have been removed from
1550 system.conf. Mounting file systems at boot has to take place
1553 * nspawn now learned a new switch --uuid= to set the machine
1554 ID on the command line.
1556 * nspawn now learned the -b switch to automatically search
1559 * vt102 is now the default TERM for serial TTYs, upgraded from
1562 * systemd-logind now works on VT-less systems.
1564 * The build tree has been reorganized. The individual
1565 components now have directories of their own.
1567 * A new condition type ConditionPathIsReadWrite= is now available.
1569 * nspawn learned the new -C switch to create cgroups for the
1570 container in other hierarchies.
1572 * We now have support for hardware watchdogs, configurable in
1575 * The scheduled shutdown logic now has a public API.
1577 * We now mount /tmp as tmpfs by default, but this can be
1578 masked and /etc/fstab can override it.
1580 * Since udisks doesn't make use of /media anymore we are not
1581 mounting a tmpfs on it anymore.
1583 * journalctl gained a new --local switch to only interleave
1584 locally generated journal files.
1586 * We can now load the IMA policy at boot automatically.
1588 * The GTK tools have been split off into a systemd-ui.
1590 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1591 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1592 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1593 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1594 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1595 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1596 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1597 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1598 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1603 * This is mostly a bugfix release
1605 * Support optional initialization of the machine ID from the
1606 KVM or container configured UUID.
1608 * Support immediate reboots with "systemctl reboot -ff"
1610 * Show /etc/os-release data in systemd-analyze output
1612 * Many bugfixes for the journal, including endianness fixes and
1613 ensuring that disk space enforcement works
1615 * sd-login.h is C++ comptaible again
1617 * Extend the /etc/os-release format on request of the Debian
1620 * We now refuse non-UTF8 strings used in various configuration
1621 and unit files. This is done to ensure we don't pass invalid
1622 data over D-Bus or expose it elsewhere.
1624 * Register Mimo USB Screens as suitable for automatic seat
1627 * Read SELinux client context from journal clients in a race
1630 * Reorder configuration file lookup order. /etc now always
1631 overrides /run in order to allow the administrator to always
1632 and unconditionally override vendor supplied or
1633 automatically generated data.
1635 * The various user visible bits of the journal now have man
1636 pages. We still lack man pages for the journal API calls
1639 * We now ship all man pages in HTML format again in the
1642 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1643 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1644 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1649 * This is mostly a bugfix release
1651 * systems lacking /etc/os-release are no longer supported.
1653 * Various functionality updates to libsystemd-login.so
1655 * Track class of PAM logins to distuingish greeters from
1658 Contributions from: Kay Sievers, Lennart Poettering, Michael
1663 * This is an important bugfix release for v41.
1665 * Building man pages is now optional which should be useful
1666 for those building systemd from git but unwilling to install
1669 * Watchdog support for supervising services is now usable. In
1670 a future release support for hardware watchdogs
1671 (i.e. /dev/watchdog) will be added building on this.
1673 * Service start rate limiting is now configurable and can be
1674 turned off per service. When a start rate limit is hit a
1675 reboot can automatically be triggered.
1677 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1679 Contributions from: Benjamin Franzke, Bill Nottingham,
1680 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1681 Schmidt, Michał Górny, Piotr Drąg
1685 * The systemd binary is installed /usr/lib/systemd/systemd now;
1686 An existing /sbin/init symlink needs to be adapted with the
1689 * The code that loads kernel modules has been ported to invoke
1690 libkmod directly, instead of modprobe. This means we do not
1691 support systems with module-init-tools anymore.
1693 * Watchdog support is now already useful, but still not
1696 * A new kernel command line option systemd.setenv= is
1697 understood to set system wide environment variables
1698 dynamically at boot.
1700 * We now limit the set of capabilities of systemd-journald.
1702 * We now set SIGPIPE to ignore by default, since it only is
1703 useful in shell pipelines, and has little use in general
1704 code. This can be disabled with IgnoreSIPIPE=no in unit
1707 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1708 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1713 * This is mostly a bugfix release
1715 * We now expose the reason why a service failed in the
1716 "Result" D-Bus property.
1718 * Rudimentary service watchdog support (will be completed over
1719 the next few releases.)
1721 * When systemd forks off in order execute some service we will
1722 now immediately changes its argv[0] to reflect which process
1723 it will execute. This is useful to minimize the time window
1724 with a generic argv[0], which makes bootcharts more useful
1726 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1727 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1728 Mike Kazantsev, Ray Strode
1732 * This is mostly a test release, but incorporates many
1735 * New systemd-cgtop tool to show control groups by their
1738 * Linking against libacl for ACLs is optional again. If
1739 disabled, support tracking device access for active logins
1740 goes becomes unavailable, and so does access to the user
1741 journals by the respective users.
1743 * If a group "adm" exists, journal files are automatically
1744 owned by them, thus allow members of this group full access
1745 to the system journal as well as all user journals.
1747 * The journal now stores the SELinux context of the logging
1748 client for all entries.
1750 * Add C++ inclusion guards to all public headers
1752 * New output mode "cat" in the journal to print only text
1753 messages, without any meta data like date or time.
1755 * Include tiny X server wrapper as a temporary stop-gap to
1756 teach XOrg udev display enumeration. This is used by display
1757 managers such as gdm, and will go away as soon as XOrg
1758 learned native udev hotplugging for display devices.
1760 * Add new systemd-cat tool for executing arbitrary programs
1761 with STDERR/STDOUT connected to the journal. Can also act as
1762 BSD logger replacement, and does so by default.
1764 * Optionally store all locally generated coredumps in the
1765 journal along with meta data.
1767 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
1768 writing short strings to files (for usage for /sys), and for
1769 creating symlinks, character and block device nodes.
1771 * New unit file option ControlGroupPersistent= to make cgroups
1772 persistent, following the mechanisms outlined in
1773 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
1775 * Support multiple local RTCs in a sane way
1777 * No longer monopolize IO when replaying readahead data on
1778 rotating disks, since we might starve non-file-system IO to
1779 death, since fanotify() will not see accesses done by blkid,
1782 * Don't show kernel threads in systemd-cgls anymore, unless
1783 requested with new -k switch.
1785 Contributions from: Dan Horák, Kay Sievers, Lennart
1786 Poettering, Michal Schmidt
1790 * This is mostly a test release, but incorporates many
1793 * The git repository moved to:
1794 git://anongit.freedesktop.org/systemd/systemd
1795 ssh://git.freedesktop.org/git/systemd/systemd
1797 * First release with the journal
1798 http://0pointer.de/blog/projects/the-journal.html
1800 * The journal replaces both systemd-kmsg-syslogd and
1801 systemd-stdout-bridge.
1803 * New sd_pid_get_unit() API call in libsystemd-logind
1805 * Many systemadm clean-ups
1807 * Introduce remote-fs-pre.target which is ordered before all
1808 remote mounts and may be used to start services before all
1811 * Added Mageia support
1813 * Add bash completion for systemd-loginctl
1815 * Actively monitor PID file creation for daemons which exit in
1816 the parent process before having finished writing the PID
1817 file in the daemon process. Daemons which do this need to be
1818 fixed (i.e. PID file creation must have finished before the
1819 parent exits), but we now react a bit more gracefully to them.
1821 * Add colourful boot output, mimicking the well-known output
1822 of existing distributions.
1824 * New option PassCredentials= for socket units, for
1825 compatibility with a recent kernel ABI breakage.
1827 * /etc/rc.local is now hooked in via a generator binary, and
1828 thus will no longer act as synchronization point during
1831 * systemctl list-unit-files now supports --root=.
1833 * systemd-tmpfiles now understands two new commands: z, Z for
1834 relabelling files according to the SELinux database. This is
1835 useful to apply SELinux labels to specific files in /sys,
1838 * Output of SysV services is now forwarded to both the console
1839 and the journal by default, not only just the console.
1841 * New man pages for all APIs from libsystemd-login.
1843 * The build tree got reorganized and a the build system is a
1844 lot more modular allowing embedded setups to specifically
1845 select the components of systemd they are interested in.
1847 * Support for Linux systems lacking the kernel VT subsystem is
1850 * configure's --with-rootdir= got renamed to
1851 --with-rootprefix= to follow the naming used by udev and
1854 * Unless specified otherwise we'll now install to /usr instead
1855 of /usr/local by default.
1857 * Processes with '@' in argv[0][0] are now excluded from the
1858 final shut-down killing spree, following the logic explained
1860 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1862 * All processes remaining in a service cgroup when we enter
1863 the START or START_PRE states are now killed with
1864 SIGKILL. That means it is no longer possible to spawn
1865 background processes from ExecStart= lines (which was never
1866 supported anyway, and bad style).
1868 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1869 reloading of units together.
1871 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1872 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1873 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1874 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1875 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek