1 systemd System and Service Manager
5 * systemd-python gained an API exposing libsystemd-daemon.
7 * The SMACK setup logic gained support for uploading CIPSO
10 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
11 ReadOnlyDirectories= and InaccessibleDirectories= has
12 changed. The private /tmp and /var/tmp directories are now
13 shared by all processes of a service (which means
14 ExecStartPre= may now leave data in /tmp that ExecStart= of
15 the same service can still access). When a service is
16 stopped its temporary directories are immediately deleted
17 (normal clean-up with tmpfiles is still done in addition to
20 * By default, systemd will now set a couple of sysctl
21 variables in the kernel: the safe sysrq options are turned
22 on, IP route verification is turned on, and source routing
23 disabled. The recently added hardlink and softlink
24 protection of the kernel is turned on. These settings should
25 be reasonably safe, and good defaults for all new systems.
27 * The predictable network naming logic may now be turned off
28 with a new kernel command line switch: net.ifnames=0.
30 * A new libsystemd-bus module has been added that implements a
31 pretty complete D-Bus client library. For details see:
33 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
35 * journald will now explicitly flush the journal files to disk
36 at the latest 5min after each write. The file will then also
37 be marked offline until the next write. This should increase
38 reliability in case of a crash. The synchronization delay
39 can be configured via SyncIntervalSec= in journald.conf.
41 * There's a new remote-fs-setup.target unit that can be used
42 to pull in specific services when at least one remote file
43 system is to be mounted.
45 * There are new targets timers.target and paths.target as
46 canonical targets to pull user timer and path units in
47 from. This complements sockets.target with a similar
48 purpose for socket units.
50 * libudev gained a new call udev_device_set_attribute_value()
51 to set sysfs attributes of a device.
53 * The udev daemon now sets the default number of worker
54 processes executed in parallel based on the number of available
55 CPUs instead of the amount of available RAM. This is supposed
56 to provide a more reliable default and limit a too agressive
57 paralellism for setups with 1000s of devices connected.
59 Contributions from: Auke Kok, Colin Walters, Cristian
60 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
61 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
62 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
63 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
64 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
65 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
66 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
67 Zbigniew Jędrzejewski-Szmek
71 * Configuration of unit files may now be extended via drop-in
72 files without having to edit/override the unit files
73 themselves. More specifically, if the administrator wants to
74 change one value for a service file foobar.service he can
75 now do so by dropping in a configuration snippet into
76 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
77 will load all these snippets and apply them on top of the
78 main unit configuration file, possibly extending or
79 overriding its settings. Using these drop-in snippets is
80 generally nicer than the two earlier options for changing
81 unit files locally: copying the files from
82 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
83 them there; or creating a new file in /etc/systemd/system/
84 that incorporates the original one via ".include". Drop-in
85 snippets into these .d/ directories can be placed in any
86 directory systemd looks for units in, and the usual
87 overriding semantics between /usr/lib, /etc and /run apply
90 * Most unit file settings which take lists of items can now be
91 reset by assigning the empty string to them. For example,
92 normally, settings such as Environment=FOO=BAR append a new
93 environment variable assignment to the environment block,
94 each time they are used. By assigning Environment= the empty
95 string the environment block can be reset to empty. This is
96 particularly useful with the .d/*.conf drop-in snippets
97 mentioned above, since this adds the ability to reset list
98 settings from vendor unit files via these drop-ins.
100 * systemctl gained a new "list-dependencies" command for
101 listing the dependencies of a unit recursively.
103 * Inhibitors are now honored and listed by "systemctl
104 suspend", "systemctl poweroff" (and similar) too, not only
105 GNOME. These commands will also list active sessions by
108 * Resource limits (as exposed by the various control group
109 controllers) can now be controlled dynamically at runtime
110 for all units. More specifically, you can now use a command
111 like "systemctl set-cgroup-attr foobar.service cpu.shares
112 2000" to alter the CPU shares a specific service gets. These
113 settings are stored persistently on disk, and thus allow the
114 administrator to easily adjust the resource usage of
115 services with a few simple commands. This dynamic resource
116 management logic is also available to other programs via the
117 bus. Almost any kernel cgroup attribute and controller is
120 * systemd-vconsole-setup will now copy all font settings to
121 all allocated VTs, where it previously applied them only to
124 * libsystemd-login gained the new sd_session_get_tty() API
127 * This release drops support for a few legacy or
128 distribution-specific LSB facility names when parsing init
129 scripts: $x-display-manager, $mail-transfer-agent,
130 $mail-transport-agent, $mail-transfer-agent, $smtp,
131 $null. Also, the mail-transfer-agent.target unit backing
132 this has been removed. Distributions which want to retain
133 compatibility with this should carry the burden for
134 supporting this themselves and patch support for these back
135 in, if they really need to. Also, the facilities $syslog and
136 $local_fs are now ignored, since systemd does not support
137 early-boot LSB init scripts anymore, and these facilities
138 are implied anyway for normal services. syslog.target has
141 * There are new bus calls on PID1's Manager object for
142 cancelling jobs, and removing snapshot units. Previously,
143 both calls were only available on the Job and Snapshot
146 * systemd-journal-gatewayd gained SSL support.
148 * The various "environment" files, such as /etc/locale.conf
149 now support continuation lines with a backslash ("\") as
150 last character in the line, similar in style (but different)
151 to how this is supported in shells.
153 * For normal user processes the _SYSTEMD_USER_UNIT= field is
154 now implicitly appended to every log entry logged. systemctl
155 has been updated to filter by this field when operating on a
156 user systemd instance.
158 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
159 CAP_AUDIT_CONTROL capabilities to the capabilities set for
160 the container. This makes it easier to boot unmodified
161 Fedora systems in a container, which however still requires
162 audit=0 to be passed on the kernel command line. Auditing in
163 kernel and userspace is unfortunately still too broken in
164 context of containers, hence we recommend compiling it out
165 of the kernel or using audit=0. Hopefully this will be fixed
166 one day for good in the kernel.
168 * nspawn gained the new --bind= and --bind-ro= parameters to
169 bind mount specific directories from the host into the
172 * nspawn will now mount its own devpts file system instance
173 into the container, in order not to leak pty devices from
174 the host into the container.
176 * systemd will now read the firmware boot time performance
177 information from the EFI variables, if the used boot loader
178 supports this, and takes it into account for boot performance
179 analysis via "systemd-analyze". This is currently supported
180 only in conjunction with Gummiboot, but could be supported
181 by other boot loaders too. For details see:
183 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
185 * A new generator has been added that automatically mounts the
186 EFI System Partition (ESP) to /boot, if that directory
187 exists, is empty, and no other file system has been
188 configured to be mounted there.
190 * logind will now send out PrepareForSleep(false) out
191 unconditionally, after coming back from suspend. This may be
192 used by applications as asynchronous notification for
193 system resume events.
195 * "systemctl unlock-sessions" has been added, that allows
196 unlocking the screens of all user sessions at once, similar
197 how "systemctl lock-sessions" already locked all users
198 sessions. This is backed by a new D-Bus call UnlockSessions().
200 * "loginctl seat-status" will now show the master device of a
201 seat. (i.e. the device of a seat that needs to be around for
202 the seat to be considered available, usually the graphics
205 * tmpfiles gained a new "X" line type, that allows
206 configuration of files and directories (with wildcards) that
207 shall be excluded from automatic cleanup ("aging").
209 * udev default rules set the device node permissions now only
210 at "add" events, and do not change them any longer with a
211 later "change" event.
213 * The log messages for lid events and power/sleep keypresses
214 now carry a message ID.
216 * We now have a substantially larger unit test suite, but this
217 continues to be work in progress.
219 * udevadm hwdb gained a new --root= parameter to change the
220 root directory to operate relative to.
222 * logind will now issue a background sync() request to the kernel
223 early at shutdown, so that dirty buffers are flushed to disk early
224 instead of at the last moment, in order to optimize shutdown
227 * A new bootctl tool has been added that is an interface for
228 certain boot loader operations. This is currently a preview
229 and is likely to be extended into a small mechanism daemon
230 like timedated, localed, hostnamed, and can be used by
231 graphical UIs to enumerate available boot options, and
232 request boot into firmware operations.
234 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
235 the rest of the package. It also has been updated to work
236 correctly in initrds.
238 * Policykit previously has been runtime optional, and is now
239 also compile time optional via a configure switch.
241 * systemd-analyze has been reimplemented in C. Also "systemctl
242 dot" has moved into systemd-analyze.
244 * "systemctl status" with no further parameters will now print
245 the status of all active or failed units.
247 * Operations such as "systemctl start" can now be executed
248 with a new mode "--irreversible" which may be used to queue
249 operations that cannot accidentally be reversed by a later
250 job queuing. This is by default used to make shutdown
251 requests more robust.
253 * The Python API of systemd now gained a new module for
254 reading journal files.
256 * A new tool kernel-install has been added that can install
257 kernel images according to the Boot Loader Specification:
259 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
261 * Boot time console output has been improved to provide
262 animated boot time output for hanging jobs.
264 * A new tool systemd-activate has been added which can be used
265 to test socket activation with, directly from the command
266 line. This should make it much easier to test and debug
267 socket activation in daemons.
269 * journalctl gained a new "--reverse" (or -r) option to show
270 journal output in reverse order (i.e. newest line first).
272 * journalctl gained a new "--pager-end" (or -e) option to jump
273 to immediately jump to the end of the journal in the
274 pager. This is only supported in conjunction with "less".
276 * journalctl gained a new "--user-unit=" option, that works
277 similar to "--unit=" but filters for user units rather than
280 * A number of unit files to ease adoption of systemd in
281 initrds has been added. This moves some minimal logic from
282 the various initrd implementations into systemd proper.
284 * The journal files are now owned by a new group
285 "systemd-journal", which exists specifically to allow access
286 to the journal, and nothing else. Previously, we used the
287 "adm" group for that, which however possibly covers more
288 than just journal/log file access. This new group is now
289 already used by systemd-journal-gatewayd to ensure this
290 daemon gets access to the journal files and as little else
291 as possible. Note that "make install" will also set FS ACLs
292 up for /var/log/journal to give "adm" and "wheel" read
293 access to it, in addition to "systemd-journal" which owns
294 the journal files. We recommend that packaging scripts also
295 add read access to "adm" + "wheel" to /var/log/journal, and
296 all existing/future journal files. To normal users and
297 administrators little changes, however packagers need to
298 ensure to create the "systemd-journal" system group at
299 package installation time.
301 * The systemd-journal-gatewayd now runs as unprivileged user
302 systemd-journal-gateway:systemd-journal-gateway. Packaging
303 scripts need to create these system user/group at
306 * timedated now exposes a new boolean property CanNTP that
307 indicates whether a local NTP service is available or not.
309 * systemd-detect-virt will now also detect xen PVs
311 * The pstore file system is now mounted by default, if it is
314 * In addition to the SELinux and IMA policies we will now also
315 load SMACK policies at early boot.
317 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
318 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
319 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
320 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
321 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
322 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
323 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
324 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
325 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
326 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
327 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
328 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
329 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
330 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
334 * Timer units now support calendar time events in addition to
335 monotonic time events. That means you can now trigger a unit
336 based on a calendar time specification such as "Thu,Fri
337 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
338 or fifth day of any month of the year 2013, given that it is
339 a thursday or friday. This brings timer event support
340 considerably closer to cron's capabilities. For details on
341 the supported calendar time specification language see
344 * udev now supports a number of different naming policies for
345 network interfaces for predictable names, and a combination
346 of these policies is now the default. Please see this wiki
347 document for details:
349 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
351 * Auke Kok's bootchart implementation has been added to the
352 systemd tree. It's an optional component that can graph the
353 boot in quite some detail. It's one of the best bootchart
354 implementations around and minimal in its code and
357 * nss-myhostname has been integrated into the systemd source
358 tree. nss-myhostname guarantees that the local hostname
359 always stays resolvable via NSS. It has been a weak
360 requirement of systemd-hostnamed since a long time, and
361 since its code is actually trivial we decided to just
362 include it in systemd's source tree. It can be turned off
363 with a configure switch.
365 * The read-ahead logic is now capable of properly detecting
366 whether a btrfs file system is on SSD or rotating media, in
367 order to optimize the read-ahead scheme. Previously, it was
368 only capable of detecting this on traditional file systems
371 * In udev, additional device properties are now read from the
372 IAB in addition to the OUI database. Also, Bluetooth company
373 identities are attached to the devices as well.
375 * In service files %U may be used as specifier that is
376 replaced by the configured user name of the service.
378 * nspawn may now be invoked without a controlling TTY. This
379 makes it suitable for invocation as its own service. This
380 may be used to set up a simple containerized server system
381 using only core OS tools.
383 * systemd and nspawn can now accept socket file descriptors
384 when they are started for socket activation. This enables
385 implementation of socket activated nspawn
386 containers. i.e. think about autospawning an entire OS image
387 when the first SSH or HTTP connection is received. We expect
388 that similar functionality will also be added to libvirt-lxc
391 * journalctl will now suppress ANSI color codes when
394 * systemctl will no longer show control group information for
395 a unit if a the control group is empty anyway.
397 * logind can now automatically suspend/hibernate/shutdown the
400 * /etc/machine-info and hostnamed now also expose the chassis
401 type of the system. This can be used to determine whether
402 the local system is a laptop, desktop, handset or
403 tablet. This information may either be configured by the
404 user/vendor or is automatically determined from ACPI and DMI
405 information if possible.
407 * A number of PolicyKit actions are now bound together with
408 "imply" rules. This should simplify creating UIs because
409 many actions will now authenticate similar ones as well.
411 * Unit files learnt a new condition ConditionACPower= which
412 may be used to conditionalize a unit depending on whether an
413 AC power source is connected or not, of whether the system
414 is running on battery power.
416 * systemctl gained a new "is-failed" verb that may be used in
417 shell scripts and suchlike to check whether a specific unit
418 is in the "failed" state.
420 * The EnvironmentFile= setting in unit files now supports file
421 globbing, and can hence be used to easily read a number of
422 environment files at once.
424 * systemd will no longer detect and recognize specific
425 distributions. All distribution-specific #ifdeffery has been
426 removed, systemd is now fully generic and
427 distribution-agnostic. Effectively, not too much is lost as
428 a lot of the code is still accessible via explicit configure
429 switches. However, support for some distribution specific
430 legacy configuration file formats has been dropped. We
431 recommend distributions to simply adopt the configuration
432 files everybody else uses now and convert the old
433 configuration from packaging scripts. Most distributions
434 already did that. If that's not possible or desirable,
435 distributions are welcome to forward port the specific
436 pieces of code locally from the git history.
438 * When logging a message about a unit systemd will now always
439 log the unit name in the message meta data.
441 * localectl will now also discover system locale data that is
442 not stored in locale archives, but directly unpacked.
444 * logind will no longer unconditionally use framebuffer
445 devices as seat masters, i.e. as devices that are required
446 to be existing before a seat is considered preset. Instead,
447 it will now look for all devices that are tagged as
448 "seat-master" in udev. By default framebuffer devices will
449 be marked as such, but depending on local systems other
450 devices might be marked as well. This may be used to
451 integrate graphics cards using closed source drivers (such
452 as NVidia ones) more nicely into logind. Note however, that
453 we recommend using the open source NVidia drivers instead,
454 and no udev rules for the closed-source drivers will be
455 shipped from us upstream.
457 Contributions from: Adam Williamson, Alessandro Crismani, Auke
458 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
459 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
460 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
461 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
462 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
463 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
464 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
465 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
466 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
467 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
468 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
473 * udev gained support for loading additional device properties
474 from an indexed database that is keyed by vendor/product IDs
475 and similar device identifiers. For the beginning this
476 "hwdb" is populated with data from the well-known PCI and
477 USB database, but also includes PNP, ACPI and OID data. In
478 the longer run this indexed database shall grow into
479 becoming the one central database for non-essential
480 userspace device metadata. Previously, data from the PCI/USB
481 database was only attached to select devices, since the
482 lookup was a relatively expensive operation due to O(n) time
483 complexity (with n being the number of entries in the
484 database). Since this is now O(1), we decided to add in this
485 data for all devices where this is available, by
486 default. Note that the indexed database needs to be rebuilt
487 when new data files are installed. To achieve this you need
488 to update your packaging scripts to invoke "udevadm hwdb
489 --update" after installation of hwdb data files. For
490 RPM-based distributions we introduced the new
491 %udev_hwdb_update macro for this purpose.
493 * The Journal gained support for the "Message Catalog", an
494 indexed database to link up additional information with
495 journal entries. For further details please check:
497 http://www.freedesktop.org/wiki/Software/systemd/catalog
499 The indexed message catalog database also needs to be
500 rebuilt after installation of message catalog files. Use
501 "journalctl --update-catalog" for this. For RPM-based
502 distributions we introduced the %journal_catalog_update
503 macro for this purpose.
505 * The Python Journal bindings gained support for the standard
506 Python logging framework.
508 * The Journal API gained new functions for checking whether
509 the underlying file system of a journal file is capable of
510 properly reporting file change notifications, or whether
511 applications that want to reflect journal changes "live"
512 need to recheck journal files continously in appropriate
515 * It is now possible to set the "age" field for tmpfiles
516 entries to 0, indicating that files matching this entry
517 shall always be removed when the directories are cleaned up.
519 * coredumpctl gained a new "gdb" verb which invokes gdb
520 right-away on the selected coredump.
522 * There's now support for "hybrid sleep" on kernels that
523 support this, in addition to "suspend" and "hibernate". Use
524 "systemctl hybrid-sleep" to make use of this.
526 * logind's HandleSuspendKey= setting (and related settings)
527 now gained support for a new "lock" setting to simply
528 request the screen lock on all local sessions, instead of
529 actually executing a suspend or hibernation.
531 * systemd will now mount the EFI variables file system by
534 * Socket units now gained support for configuration of the
535 SMACK security label.
537 * timedatectl will now output the time of the last and next
538 daylight saving change.
540 * We dropped support for various legacy and distro-specific
541 concepts, such as insserv, early-boot SysV services
542 (i.e. those for non-standard runlevels such as 'b' or 'S')
543 or ArchLinux /etc/rc.conf support. We recommend the
544 distributions who still need support this to either continue
545 to maintain the necessary patches downstream, or find a
546 different solution. (Talk to us if you have questions!)
548 * Various systemd components will now bypass PolicyKit checks
549 for root and otherwise handle properly if PolicyKit is not
550 found to be around. This should fix most issues for
551 PolicyKit-less systems. Quite frankly this should have been
552 this way since day one. It is absolutely our intention to
553 make systemd work fine on PolicyKit-less systems, and we
554 consider it a bug if something doesn't work as it should if
555 PolicyKit is not around.
557 * For embedded systems it is now possible to build udev and
558 systemd without blkid and/or kmod support.
560 * "systemctl switch-root" is now capable of switching root
561 more than once. I.e. in addition to transitions from the
562 initrd to the host OS it is now possible to transition to
563 further OS images from the host. This is useful to implement
564 offline updating tools.
566 * Various other additions have been made to the RPM macros
567 shipped with systemd. Use %udev_rules_update() after
568 installing new udev rules files. %_udevhwdbdir,
569 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
570 %_sysctldir are now available which resolve to the right
571 directories for packages to place various data files in.
573 * journalctl gained the new --full switch (in addition to
574 --all, to disable ellipsation for long messages.
576 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
577 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
578 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
579 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
580 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
581 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
582 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
583 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
584 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
588 * journalctl gained new --since= and --until= switches to
589 filter by time. It also now supports nice filtering for
590 units via --unit=/-u.
592 * Type=oneshot services may use ExecReload= and do the
595 * The journal daemon now supports time-based rotation and
596 vacuuming, in addition to the usual disk-space based
599 * The journal will now index the available field values for
600 each field name. This enables clients to show pretty drop
601 downs of available match values when filtering. The bash
602 completion of journalctl has been updated
603 accordingly. journalctl gained a new switch -F to list all
604 values a certain field takes in the journal database.
606 * More service events are now written as structured messages
607 to the journal, and made recognizable via message IDs.
609 * The timedated, localed and hostnamed mini-services which
610 previously only provided support for changing time, locale
611 and hostname settings from graphical DEs such as GNOME now
612 also have a minimal (but very useful) text-based client
613 utility each. This is probably the nicest way to changing
614 these settings from the command line now, especially since
615 it lists available options and is fully integrated with bash
618 * There's now a new tool "systemd-coredumpctl" to list and
619 extract coredumps from the journal.
621 * We now install a README each in /var/log/ and
622 /etc/rc.d/init.d explaining where the system logs and init
623 scripts went. This hopefully should help folks who go to
624 that dirs and look into the otherwise now empty void and
627 * When user-services are invoked (by systemd --user) the
628 $MANAGERPID env var is set to the PID of systemd.
630 * SIGRTMIN+24 when sent to a --user instance will now result
631 in immediate termination of systemd.
633 * gatewayd received numerous feature additions such as a
634 "follow" mode, for live syncing and filtering.
636 * browse.html now allows filtering and showing detailed
637 information on specific entries. Keyboard navigation and
638 mouse screen support has been added.
640 * gatewayd/journalctl now supports HTML5/JSON
641 Server-Sent-Events as output.
643 * The SysV init script compatibility logic will now
644 heuristically determine whether a script supports the
645 "reload" verb, and only then make this available as
648 * "systemctl status --follow" has been removed, use "journalctl
651 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
652 have been removed since they are hardly useful to be
655 * And I'd like to take the opportunity to specifically mention
656 Zbigniew for his great contributions. Zbigniew, you rock!
658 Contributions from: Andrew Eikum, Christian Hesse, Colin
659 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
660 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
661 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
662 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
663 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
664 Jędrzejewski-Szmek, Сковорода Никита Андреевич
668 * If /etc/vconsole.conf is non-existent or empty we will no
669 longer load any console font or key map at boot by
670 default. Instead the kernel defaults will be left
671 intact. This is definitely the right thing to do, as no
672 configuration should mean no configuration, and hard-coding
673 font names that are different on all archs is probably a bad
674 idea. Also, the kernel default key map and font should be
675 good enough for most cases anyway, and mostly identical to
676 the userspace fonts/key maps we previously overloaded them
677 with. If distributions want to continue to default to a
678 non-kernel font or key map they should ship a default
679 /etc/vconsole.conf with the appropriate contents.
681 Contributions from: Colin Walters, Daniel J Walsh, Dave
682 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
683 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
687 * journalctl gained a new --cursor= switch to show entries
688 starting from the specified location in the journal.
690 * We now enforce a size limit on journal entry fields exported
691 with "-o json" in journalctl. Fields larger than 4K will be
692 assigned null. This can be turned off with --all.
694 * An (optional) journal gateway daemon is now available as
695 "systemd-journal-gatewayd.service". This service provides
696 access to the journal via HTTP and JSON. This functionality
697 will be used to implement live log synchronization in both
698 pull and push modes, but has various other users too, such
699 as easy log access for debugging of embedded devices. Right
700 now it is already useful to retrieve the journal via HTTP:
702 # systemctl start systemd-journal-gatewayd.service
703 # wget http://localhost:19531/entries
705 This will download the journal contents in a
706 /var/log/messages compatible format. The same as JSON:
708 # curl -H"Accept: application/json" http://localhost:19531/entries
710 This service is also accessible via a web browser where a
711 single static HTML5 app is served that uses the JSON logic
712 to enable the user to do some basic browsing of the
713 journal. This will be extended later on. Here's an example
714 screenshot of this app in its current state:
716 http://0pointer.de/public/journal-gatewayd
718 Contributions from: Kay Sievers, Lennart Poettering, Robert
719 Milasan, Tom Gundersen
723 * The bash completion logic is now available for journalctl
726 * We don't mount the "cpuset" controller anymore together with
727 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
728 started if no parameters are assigned to it. "cpuset" hence
729 broke code that assumed it it could create "cpu" groups and
732 * journalctl -f will now subscribe to terminal size changes,
733 and line break accordingly.
735 Contributions from: Dave Reisner, Kay Sievers, Lennart
736 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
740 * nspawn will now create a symlink /etc/localtime in the
741 container environment, copying the host's timezone
742 setting. Previously this has been done via a bind mount, but
743 since symlinks cannot be bind mounted this has now been
744 changed to create/update the appropriate symlink.
746 * journalctl -n's line number argument is now optional, and
747 will default to 10 if omitted.
749 * journald will now log the maximum size the journal files may
750 take up on disk. This is particularly useful if the default
751 built-in logic of determining this parameter from the file
752 system size is used. Use "systemctl status
753 systemd-journald.service" to see this information.
755 * The multi-seat X wrapper tool has been stripped down. As X
756 is now capable of enumerating graphics devices via udev in a
757 seat-aware way the wrapper is not strictly necessary
758 anymore. A stripped down temporary stop-gap is still shipped
759 until the upstream display managers have been updated to
760 fully support the new X logic. Expect this wrapper to be
761 removed entirely in one of the next releases.
763 * HandleSleepKey= in logind.conf has been split up into
764 HandleSuspendKey= and HandleHibernateKey=. The old setting
765 is not available anymore. X11 and the kernel are
766 distuingishing between these keys and we should too. This
767 also means the inhibition lock for these keys has been split
770 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
771 Poettering, Lukas Nykryn, Václav Pavlín
775 * Whenever a unit changes state we'll now log this to the
776 journal and show along the unit's own log output in
779 * ConditionPathIsMountPoint= can now properly detect bind
780 mount points too. (Previously, a bind mount of one file
781 system to another place in the same file system could not be
782 detected as mount, since they shared struct stat's st_dev
785 * We will now mount the cgroup controllers cpu, cpuacct,
786 cpuset and the controllers net_cls, net_prio together by
789 * nspawn containers will now have a virtualized boot
790 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
791 over with a randomized ID at container initialization). This
792 has the effect of making "journalctl -b" do the right thing
795 * The JSON output journal serialization has been updated not
796 to generate "endless" list objects anymore, but rather one
797 JSON object per line. This is more in line how most JSON
798 parsers expect JSON objects. The new output mode
799 "json-pretty" has been added to provide similar output, but
800 neatly aligned for readability by humans.
802 * We dropped all explicit sync() invocations in the shutdown
803 code. The kernel does this implicitly anyway in the kernel
804 reboot() syscall. halt(8)'s -n option is now a compatibility
807 * We now support virtualized reboot() in containers, as
808 supported by newer kernels. We will fall back to exit() if
809 CAP_SYS_REBOOT is not available to the container. Also,
810 nspawn makes use of this now and will actually reboot the
811 container if the containerized OS asks for that.
813 * journalctl will only show local log output by default
814 now. Use --merge (-m) to show remote log output, too.
816 * libsystemd-journal gained the new sd_journal_get_usage()
817 call to determine the current disk usage of all journal
818 files. This is exposed in the new "journalctl --disk-usage"
821 * journald gained a new configuration setting SplitMode= in
822 journald.conf which may be used to control how user journals
823 are split off. See journald.conf(5) for details.
825 * A new condition type ConditionFileNotEmpty= has been added.
827 * tmpfiles' "w" lines now support file globbing, to write
828 multiple files at once.
830 * We added Python bindings for the journal submission
831 APIs. More Python APIs for a number of selected APIs will
832 likely follow. Note that we intend to add native bindings
833 only for the Python language, as we consider it common
834 enough to deserve bindings shipped within systemd. There are
835 various projects outside of systemd that provide bindings
836 for languages such as PHP or Lua.
838 * Many conditions will now resolve specifiers such as %i. In
839 addition, PathChanged= and related directives of .path units
840 now support specifiers as well.
842 * There's now a new RPM macro definition for the system preset
845 * journald will now warn if it can't foward a message to the
846 syslog daemon because it's socket is full.
848 * timedated will no longer write or process /etc/timezone,
849 except on Debian. As we do not support late mounted /usr
850 anymore /etc/localtime always being a symlink is now safe,
851 and hence the information in /etc/timezone is not necessary
854 * logind will now always reserve one VT for a text getty (VT6
855 by default). Previously if more than 6 X sessions where
856 started they took up all the VTs with auto-spawned gettys,
857 so that no text gettys were available anymore.
859 * udev will now automatically inform the btrfs kernel logic
860 about btrfs RAID components showing up. This should make
861 simple hotplug based btrfs RAID assembly work.
863 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
864 (but not for its children which will stay at the kernel
865 default). This should allow setups with a lot more listening
868 * systemd will now always pass the configured timezone to the
869 kernel at boot. timedated will do the same when the timezone
872 * logind's inhibition logic has been updated. By default,
873 logind will now handle the lid switch, the power and sleep
874 keys all the time, even in graphical sessions. If DEs want
875 to handle these events on their own they should take the new
876 handle-power-key, handle-sleep-key and handle-lid-switch
877 inhibitors during their runtime. A simple way to achiveve
878 that is to invoke the DE wrapped in an invocation of:
880 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
882 * Access to unit operations is now checked via SELinux taking
883 the unit file label and client process label into account.
885 * systemd will now notify the administrator in the journal
886 when he over-mounts a non-empty directory.
888 * There are new specifiers that are resolved in unit files,
889 for the host name (%H), the machine ID (%m) and the boot ID
892 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
893 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
894 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
895 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
896 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
897 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
898 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
902 * Support for reading structured kernel messages from
903 /dev/kmsg has now been added and is enabled by default.
905 * Support for reading kernel messages from /proc/kmsg has now
906 been removed. If you want kernel messages in the journal
907 make sure to run a recent kernel (>= 3.5) that supports
908 reading structured messages from /dev/kmsg (see
909 above). /proc/kmsg is now exclusive property of classic
910 syslog daemons again.
912 * The libudev API gained the new
913 udev_device_new_from_device_id() call.
915 * The logic for file system namespace (ReadOnlyDirectory=,
916 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
917 require pivot_root() anymore. This means fewer temporary
918 directories are created below /tmp for this feature.
920 * nspawn containers will now see and receive all submounts
921 made on the host OS below the root file system of the
924 * Forward Secure Sealing is now supported for Journal files,
925 which provide cryptographical sealing of journal files so
926 that attackers cannot alter log history anymore without this
927 being detectable. Lennart will soon post a blog story about
928 this explaining it in more detail.
930 * There are two new service settings RestartPreventExitStatus=
931 and SuccessExitStatus= which allow configuration of exit
932 status (exit code or signal) which will be excepted from the
933 restart logic, resp. consider successful.
935 * journalctl gained the new --verify switch that can be used
936 to check the integrity of the structure of journal files and
937 (if Forward Secure Sealing is enabled) the contents of
940 * nspawn containers will now be run with /dev/stdin, /dev/fd/
941 and similar symlinks pre-created. This makes running shells
942 as container init process a lot more fun.
944 * The fstab support can now handle PARTUUID= and PARTLABEL=
947 * A new ConditionHost= condition has been added to match
948 against the hostname (with globs) and machine ID. This is
949 useful for clusters where a single OS image is used to
950 provision a large number of hosts which shall run slightly
951 different sets of services.
953 * Services which hit the restart limit will now be placed in a
956 Contributions from: Bertram Poettering, Dave Reisner, Huang
957 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
958 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
962 * When running in --user mode systemd will now become a
963 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
964 tree a lot more organized.
966 * A new PartOf= unit dependency type has been introduced that
967 may be used to group services in a natural way.
969 * "systemctl enable" may now be used to enable instances of
972 * journalctl now prints error log levels in red, and
973 warning/notice log levels in bright white. It also supports
974 filtering by log level now.
976 * cgtop gained a new -n switch (similar to top), to configure
977 the maximum number of iterations to run for. It also gained
978 -b, to run in batch mode (accepting no input).
980 * The suffix ".service" may now be ommited on most systemctl
981 command lines involving service unit names.
983 * There's a new bus call in logind to lock all sessions, as
984 well as a loginctl verb for it "lock-sessions".
986 * libsystemd-logind.so gained a new call sd_journal_perror()
987 that works similar to libc perror() but logs to the journal
988 and encodes structured information about the error number.
990 * /etc/crypttab entries now understand the new keyfile-size=
993 * shutdown(8) now can send a (configurable) wall message when
994 a shutdown is cancelled.
996 * The mount propagation mode for the root file system will now
997 default to "shared", which is useful to make containers work
998 nicely out-of-the-box so that they receive new mounts from
999 the host. This can be undone locally by running "mount
1000 --make-rprivate /" if needed.
1002 * The prefdm.service file has been removed. Distributions
1003 should maintain this unit downstream if they intend to keep
1004 it around. However, we recommend writing normal unit files
1005 for display managers instead.
1007 * Since systemd is a crucial part of the OS we will now
1008 default to a number of compiler switches that improve
1009 security (hardening) such as read-only relocations, stack
1010 protection, and suchlike.
1012 * The TimeoutSec= setting for services is now split into
1013 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1014 of individual time outs for the start and the stop phase of
1017 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1018 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1019 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1020 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1021 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1022 Gundersen, Zbigniew Jędrzejewski-Szmek
1026 * The journal and id128 C APIs are now fully documented as man
1029 * Extra safety checks have been added when transitioning from
1030 the initial RAM disk to the main system to avoid accidental
1033 * /etc/crypttab entries now understand the new keyfile-offset=
1036 * systemctl -t can now be used to filter by unit load state.
1038 * The journal C API gained the new sd_journal_wait() call to
1039 make writing synchronous journal clients easier.
1041 * journalctl gained the new -D switch to show journals from a
1044 * journalctl now displays a special marker between log
1045 messages of two different boots.
1047 * The journal is now explicitly flushed to /var via a service
1048 systemd-journal-flush.service, rather than implicitly simply
1049 by seeing /var/log/journal to be writable.
1051 * journalctl (and the journal C APIs) can now match for much
1052 more complex expressions, with alternatives and
1055 * When transitioning from the initial RAM disk to the main
1056 system we will now kill all processes in a killing spree to
1057 ensure no processes stay around by accident.
1059 * Three new specifiers may be used in unit files: %u, %h, %s
1060 resolve to the user name, user home directory resp. user
1061 shell. This is useful for running systemd user instances.
1063 * We now automatically rotate journal files if their data
1064 object hash table gets a fill level > 75%. We also size the
1065 hash table based on the configured maximum file size. This
1066 together should lower hash collisions drastically and thus
1067 speed things up a bit.
1069 * journalctl gained the new "--header" switch to introspect
1070 header data of journal files.
1072 * A new setting SystemCallFilters= has been added to services
1073 which may be used to apply blacklists or whitelists to
1074 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1076 * nspawn gained a new --link-journal= switch (and quicker: -j)
1077 to link the container journal with the host. This makes it
1078 very easy to centralize log viewing on the host for all
1079 guests while still keeping the journal files separated.
1081 * Many bugfixes and optimizations
1083 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1084 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1085 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1090 * Several tools now understand kernel command line arguments,
1091 which are only read when run in an initial RAM disk. They
1092 usually follow closely their normal counterparts, but are
1095 * There's a new tool to analyze the readahead files that are
1096 automatically generated at boot. Use:
1098 /usr/lib/systemd/systemd-readahead analyze /.readahead
1100 * We now provide an early debug shell on tty9 if this enabled. Use:
1102 systemctl enable debug-shell.service
1104 * All plymouth related units have been moved into the Plymouth
1105 package. Please make sure to upgrade your Plymouth version
1108 * systemd-tmpfiles now supports getting passed the basename of
1109 a configuration file only, in which case it will look for it
1110 in all appropriate directories automatically.
1112 * udevadm info now takes a /dev or /sys path as argument, and
1113 does the right thing. Example:
1115 udevadm info /dev/sda
1116 udevadm info /sys/class/block/sda
1118 * systemctl now prints a warning if a unit is stopped but a
1119 unit that might trigger it continues to run. Example: a
1120 service is stopped but the socket that activates it is left
1123 * "systemctl status" will now mention if the log output was
1124 shortened due to rotation since a service has been started.
1126 * The journal API now exposes functions to determine the
1127 "cutoff" times due to rotation.
1129 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1130 immediately flushing of runtime logs to /var if possible,
1131 resp. for triggering immediate rotation of the journal
1134 * It is now considered an error if a service is attempted to
1135 be stopped that is not loaded.
1137 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1139 * systemd-analyze now supports Python 3
1141 * tmpfiles now supports cleaning up directories via aging
1142 where the first level dirs are always kept around but
1143 directories beneath it automatically aged. This is enabled
1144 by prefixing the age field with '~'.
1146 * Seat objects now expose CanGraphical, CanTTY properties
1147 which is required to deal with very fast bootups where the
1148 display manager might be running before the graphics drivers
1149 completed initialization.
1151 * Seat objects now expose a State property.
1153 * We now include RPM macros for service enabling/disabling
1154 based on the preset logic. We recommend RPM based
1155 distributions to make use of these macros if possible. This
1156 makes it simpler to reuse RPM spec files across
1159 * We now make sure that the collected systemd unit name is
1160 always valid when services log to the journal via
1163 * There's a new man page kernel-command-line(7) detailing all
1164 command line options we understand.
1166 * The fstab generator may now be disabled at boot by passing
1167 fstab=0 on the kernel command line.
1169 * A new kernel command line option modules-load= is now understood
1170 to load a specific kernel module statically, early at boot.
1172 * Unit names specified on the systemctl command line are now
1173 automatically escaped as needed. Also, if file system or
1174 device paths are specified they are automatically turned
1175 into the appropriate mount or device unit names. Example:
1177 systemctl status /home
1178 systemctl status /dev/sda
1180 * The SysVConsole= configuration option has been removed from
1181 system.conf parsing.
1183 * The SysV search path is no longer exported on the D-Bus
1186 * The Names= option is been removed from unit file parsing.
1188 * There's a new man page bootup(7) detailing the boot process.
1190 * Every unit and every generator we ship with systemd now
1191 comes with full documentation. The self-explanatory boot is
1194 * A couple of services gained "systemd-" prefixes in their
1195 name if they wrap systemd code, rather than only external
1196 code. Among them fsck@.service which is now
1197 systemd-fsck@.service.
1199 * The HaveWatchdog property has been removed from the D-Bus
1202 * systemd.confirm_spawn= on the kernel command line should now
1205 * There's a new man page crypttab(5) which details all options
1206 we actually understand.
1208 * systemd-nspawn gained a new --capability= switch to pass
1209 additional capabilities to the container.
1211 * timedated will now read known NTP implementation unit names
1212 from /usr/lib/systemd/ntp-units.d/*.list,
1213 systemd-timedated-ntp.target has been removed.
1215 * journalctl gained a new switch "-b" that lists log data of
1216 the current boot only.
1218 * The notify socket is in the abstract namespace again, in
1219 order to support daemons which chroot() at start-up.
1221 * There is a new Storage= configuration option for journald
1222 which allows configuration of where log data should go. This
1223 also provides a way to disable journal logging entirely, so
1224 that data collected is only forwarded to the console, the
1225 kernel log buffer or another syslog implementation.
1227 * Many bugfixes and optimizations
1229 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1230 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1231 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1232 Shawn Landden, Tom Gundersen
1236 * "systemctl help <unit>" now shows the man page if one is
1239 * Several new man pages have been added.
1241 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1242 MaxLevelConsole= can now be specified in
1243 journald.conf. These options allow reducing the amount of
1244 data stored on disk or forwarded by the log level.
1246 * TimerSlackNSec= can now be specified in system.conf for
1247 PID1. This allows system-wide power savings.
1249 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1250 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1255 * logind is now capable of (optionally) handling power and
1256 sleep keys as well as the lid switch.
1258 * journalctl now understands the syntax "journalctl
1259 /usr/bin/avahi-daemon" to get all log output of a specific
1262 * CapabilityBoundingSet= in system.conf now also influences
1263 the capability bound set of usermode helpers of the kernel.
1265 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1266 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1267 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1268 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1272 * Note that we skipped 139 releases here in order to set the
1273 new version to something that is greater than both udev's
1274 and systemd's most recent version number.
1276 * udev: all udev sources are merged into the systemd source tree now.
1277 All future udev development will happen in the systemd tree. It
1278 is still fully supported to use the udev daemon and tools without
1279 systemd running, like in initramfs or other init systems. Building
1280 udev though, will require the *build* of the systemd tree, but
1281 udev can be properly *run* without systemd.
1283 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1284 should be used to create dead device nodes as workarounds for broken
1287 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1288 no longer supported. udev_monitor_new_from_netlink() needs to be
1289 used to subscribe to events.
1291 * udev: when udevd is started by systemd, processes which are left
1292 behind by forking them off of udev rules, are unconditionally cleaned
1293 up and killed now after the event handling has finished. Services or
1294 daemons must be started as systemd services. Services can be
1295 pulled-in by udev to get started, but they can no longer be directly
1296 forked by udev rules.
1298 * udev: the daemon binary is called systemd-udevd now and installed
1299 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1300 to adapt to that, create symlink, or rename the binary after building
1303 * libudev no longer provides these symbols:
1304 udev_monitor_from_socket()
1305 udev_queue_get_failed_list_entry()
1306 udev_get_{dev,sys,run}_path()
1307 The versions number was bumped and symbol versioning introduced.
1309 * systemd-loginctl and systemd-journalctl have been renamed
1310 to loginctl and journalctl to match systemctl.
1312 * The config files: /etc/systemd/systemd-logind.conf and
1313 /etc/systemd/systemd-journald.conf have been renamed to
1314 logind.conf and journald.conf. Package updates should rename
1315 the files to the new names on upgrade.
1317 * For almost all files the license is now LGPL2.1+, changed
1318 from the previous GPL2.0+. Exceptions are some minor stuff
1319 of udev (which will be changed to LGPL2.1 eventually, too),
1320 and the MIT licensed sd-daemon.[ch] library that is suitable
1321 to be used as drop-in files.
1323 * systemd and logind now handle system sleep states, in
1324 particular suspending and hibernating.
1326 * logind now implements a sleep/shutdown/idle inhibiting logic
1327 suitable for a variety of uses. Soonishly Lennart will blog
1328 about this in more detail.
1330 * var-run.mount and var-lock.mount are no longer provided
1331 (which prevously bind mounted these directories to their new
1332 places). Distributions which have not converted these
1333 directories to symlinks should consider stealing these files
1334 from git history and add them downstream.
1336 * We introduced the Documentation= field for units and added
1337 this to all our shipped units. This is useful to make it
1338 easier to explore the boot and the purpose of the various
1341 * All smaller setup units (such as
1342 systemd-vconsole-setup.service) now detect properly if they
1343 are run in a container and are skipped when
1344 appropriate. This guarantees an entirely noise-free boot in
1345 Linux container environments such as systemd-nspawn.
1347 * A framework for implementing offline system updates is now
1348 integrated, for details see:
1349 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1351 * A new service type Type=idle is available now which helps us
1352 avoiding ugly interleaving of getty output and boot status
1355 * There's now a system-wide CapabilityBoundingSet= option to
1356 globally reduce the set of capabilities for the
1357 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1358 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1359 even CAP_NET_ADMIN system-wide for secure systems.
1361 * There are now system-wide DefaultLimitXXX= options to
1362 globally change the defaults of the various resource limits
1363 for all units started by PID 1.
1365 * Harald Hoyer's systemd test suite has been integrated into
1366 systemd which allows easy testing of systemd builds in qemu
1367 and nspawn. (This is really awesome! Ask us for details!)
1369 * The fstab parser is now implemented as generator, not inside
1372 * systemctl will now warn you if .mount units generated from
1373 /etc/fstab are out of date due to changes in fstab that
1374 haven't been read by systemd yet.
1376 * systemd is now suitable for usage in initrds. Dracut has
1377 already been updated to make use of this. With this in place
1378 initrds get a slight bit faster but primarily are much
1379 easier to introspect and debug since "systemctl status" in
1380 the host system can be used to introspect initrd services,
1381 and the journal from the initrd is kept around too.
1383 * systemd-delta has been added, a tool to explore differences
1384 between user/admin configuration and vendor defaults.
1386 * PrivateTmp= now affects both /tmp and /var/tmp.
1388 * Boot time status messages are now much prettier and feature
1389 proper english language. Booting up systemd has never been
1392 * Read-ahead pack files now include the inode number of all
1393 files to pre-cache. When the inode changes the pre-caching
1394 is not attempted. This should be nicer to deal with updated
1395 packages which might result in changes of read-ahead
1398 * We now temporaritly lower the kernel's read_ahead_kb variable
1399 when collecting read-ahead data to ensure the kernel's
1400 built-in read-ahead does not add noise to our measurements
1401 of necessary blocks to pre-cache.
1403 * There's now RequiresMountsFor= to add automatic dependencies
1404 for all mounts necessary for a specific file system path.
1406 * MountAuto= and SwapAuto= have been removed from
1407 system.conf. Mounting file systems at boot has to take place
1410 * nspawn now learned a new switch --uuid= to set the machine
1411 ID on the command line.
1413 * nspawn now learned the -b switch to automatically search
1416 * vt102 is now the default TERM for serial TTYs, upgraded from
1419 * systemd-logind now works on VT-less systems.
1421 * The build tree has been reorganized. The individual
1422 components now have directories of their own.
1424 * A new condition type ConditionPathIsReadWrite= is now available.
1426 * nspawn learned the new -C switch to create cgroups for the
1427 container in other hierarchies.
1429 * We now have support for hardware watchdogs, configurable in
1432 * The scheduled shutdown logic now has a public API.
1434 * We now mount /tmp as tmpfs by default, but this can be
1435 masked and /etc/fstab can override it.
1437 * Since udisks doesn't make use of /media anymore we are not
1438 mounting a tmpfs on it anymore.
1440 * journalctl gained a new --local switch to only interleave
1441 locally generated journal files.
1443 * We can now load the IMA policy at boot automatically.
1445 * The GTK tools have been split off into a systemd-ui.
1447 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1448 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1449 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1450 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1451 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1452 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1453 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1454 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1455 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1460 * This is mostly a bugfix release
1462 * Support optional initialization of the machine ID from the
1463 KVM or container configured UUID.
1465 * Support immediate reboots with "systemctl reboot -ff"
1467 * Show /etc/os-release data in systemd-analyze output
1469 * Many bugfixes for the journal, including endianess fixes and
1470 ensuring that disk space enforcement works
1472 * sd-login.h is C++ comptaible again
1474 * Extend the /etc/os-release format on request of the Debian
1477 * We now refuse non-UTF8 strings used in various configuration
1478 and unit files. This is done to ensure we don't pass invalid
1479 data over D-Bus or expose it elsewhere.
1481 * Register Mimo USB Screens as suitable for automatic seat
1484 * Read SELinux client context from journal clients in a race
1487 * Reorder configuration file lookup order. /etc now always
1488 overrides /run in order to allow the administrator to always
1489 and unconditionally override vendor supplied or
1490 automatically generated data.
1492 * The various user visible bits of the journal now have man
1493 pages. We still lack man pages for the journal API calls
1496 * We now ship all man pages in HTML format again in the
1499 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1500 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1501 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1506 * This is mostly a bugfix release
1508 * systems lacking /etc/os-release are no longer supported.
1510 * Various functionality updates to libsystemd-login.so
1512 * Track class of PAM logins to distuingish greeters from
1515 Contributions from: Kay Sievers, Lennart Poettering, Michael
1520 * This is an important bugfix release for v41.
1522 * Building man pages is now optional which should be useful
1523 for those building systemd from git but unwilling to install
1526 * Watchdog support for supervising services is now usable. In
1527 a future release support for hardware watchdogs
1528 (i.e. /dev/watchdog) will be added building on this.
1530 * Service start rate limiting is now configurable and can be
1531 turned off per service. When a start rate limit is hit a
1532 reboot can automatically be triggered.
1534 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1536 Contributions from: Benjamin Franzke, Bill Nottingham,
1537 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1538 Schmidt, Michał Górny, Piotr Drąg
1542 * The systemd binary is installed /usr/lib/systemd/systemd now;
1543 An existing /sbin/init symlink needs to be adapted with the
1546 * The code that loads kernel modules has been ported to invoke
1547 libkmod directly, instead of modprobe. This means we do not
1548 support systems with module-init-tools anymore.
1550 * Watchdog support is now already useful, but still not
1553 * A new kernel command line option systemd.setenv= is
1554 understood to set system wide environment variables
1555 dynamically at boot.
1557 * We now limit the set of capabilities of systemd-journald.
1559 * We now set SIGPIPE to ignore by default, since it only is
1560 useful in shell pipelines, and has little use in general
1561 code. This can be disabled with IgnoreSIPIPE=no in unit
1564 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1565 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1570 * This is mostly a bugfix release
1572 * We now expose the reason why a service failed in the
1573 "Result" D-Bus property.
1575 * Rudimentary service watchdog support (will be completed over
1576 the next few releases.)
1578 * When systemd forks off in order execute some service we will
1579 now immediately changes its argv[0] to reflect which process
1580 it will execute. This is useful to minimize the time window
1581 with a generic argv[0], which makes bootcharts more useful
1583 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1584 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1585 Mike Kazantsev, Ray Strode
1589 * This is mostly a test release, but incorporates many
1592 * New systemd-cgtop tool to show control groups by their
1595 * Linking against libacl for ACLs is optional again. If
1596 disabled, support tracking device access for active logins
1597 goes becomes unavailable, and so does access to the user
1598 journals by the respective users.
1600 * If a group "adm" exists, journal files are automatically
1601 owned by them, thus allow members of this group full access
1602 to the system journal as well as all user journals.
1604 * The journal now stores the SELinux context of the logging
1605 client for all entries.
1607 * Add C++ inclusion guards to all public headers
1609 * New output mode "cat" in the journal to print only text
1610 messages, without any meta data like date or time.
1612 * Include tiny X server wrapper as a temporary stop-gap to
1613 teach XOrg udev display enumeration. This is used by display
1614 managers such as gdm, and will go away as soon as XOrg
1615 learned native udev hotplugging for display devices.
1617 * Add new systemd-cat tool for executing arbitrary programs
1618 with STDERR/STDOUT connected to the journal. Can also act as
1619 BSD logger replacement, and does so by default.
1621 * Optionally store all locally generated coredumps in the
1622 journal along with meta data.
1624 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
1625 writing short strings to files (for usage for /sys), and for
1626 creating symlinks, character and block device nodes.
1628 * New unit file option ControlGroupPersistent= to make cgroups
1629 persistent, following the mechanisms outlined in
1630 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
1632 * Support multiple local RTCs in a sane way
1634 * No longer monopolize IO when replaying readahead data on
1635 rotating disks, since we might starve non-file-system IO to
1636 death, since fanotify() will not see accesses done by blkid,
1639 * Don't show kernel threads in systemd-cgls anymore, unless
1640 requested with new -k switch.
1642 Contributions from: Dan Horák, Kay Sievers, Lennart
1643 Poettering, Michal Schmidt
1647 * This is mostly a test release, but incorporates many
1650 * The git repository moved to:
1651 git://anongit.freedesktop.org/systemd/systemd
1652 ssh://git.freedesktop.org/git/systemd/systemd
1654 * First release with the journal
1655 http://0pointer.de/blog/projects/the-journal.html
1657 * The journal replaces both systemd-kmsg-syslogd and
1658 systemd-stdout-bridge.
1660 * New sd_pid_get_unit() API call in libsystemd-logind
1662 * Many systemadm clean-ups
1664 * Introduce remote-fs-pre.target which is ordered before all
1665 remote mounts and may be used to start services before all
1668 * Added Mageia support
1670 * Add bash completion for systemd-loginctl
1672 * Actively monitor PID file creation for daemons which exit in
1673 the parent process before having finished writing the PID
1674 file in the daemon process. Daemons which do this need to be
1675 fixed (i.e. PID file creation must have finished before the
1676 parent exits), but we now react a bit more gracefully to them.
1678 * Add colourful boot output, mimicking the well-known output
1679 of existing distributions.
1681 * New option PassCredentials= for socket units, for
1682 compatibility with a recent kernel ABI breakage.
1684 * /etc/rc.local is now hooked in via a generator binary, and
1685 thus will no longer act as synchronization point during
1688 * systemctl list-unit-files now supports --root=.
1690 * systemd-tmpfiles now understands two new commands: z, Z for
1691 relabelling files according to the SELinux database. This is
1692 useful to apply SELinux labels to specific files in /sys,
1695 * Output of SysV services is now forwarded to both the console
1696 and the journal by default, not only just the console.
1698 * New man pages for all APIs from libsystemd-login.
1700 * The build tree got reorganized and a the build system is a
1701 lot more modular allowing embedded setups to specifically
1702 select the components of systemd they are interested in.
1704 * Support for Linux systems lacking the kernel VT subsystem is
1707 * configure's --with-rootdir= got renamed to
1708 --with-rootprefix= to follow the naming used by udev and
1711 * Unless specified otherwise we'll now install to /usr instead
1712 of /usr/local by default.
1714 * Processes with '@' in argv[0][0] are now excluded from the
1715 final shut-down killing spree, following the logic explained
1717 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1719 * All processes remaining in a service cgroup when we enter
1720 the START or START_PRE states are now killed with
1721 SIGKILL. That means it is no longer possible to spawn
1722 background processes from ExecStart= lines (which was never
1723 supported anyway, and bad style).
1725 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1726 reloading of units together.
1728 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1729 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1730 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1731 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1732 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek