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 still is 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 not explicitly flush the journal files to disk
36 the latest 5min after each write. This will also mark the
37 files as offline then until the next write. This should
38 increase reliability. The synchronization delay can be
39 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 Contributions from: Auke Kok, Colin Walters, Cristian
54 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
55 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
56 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
57 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
58 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
59 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
60 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
61 Zbigniew Jędrzejewski-Szmek
65 * Configuration of unit files may now be extended via drop-in
66 files without having to edit/override the unit files
67 themselves. More specifically, if the administrator wants to
68 change one value for a service file foobar.service he can
69 now do so by dropping in a configuration snippet into
70 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
71 will load all these snippets and apply them on top of the
72 main unit configuration file, possibly extending or
73 overriding its settings. Using these drop-in snippets is
74 generally nicer than the two earlier options for changing
75 unit files locally: copying the files from
76 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
77 them there; or creating a new file in /etc/systemd/system/
78 that incorporates the original one via ".include". Drop-in
79 snippets into these .d/ directories can be placed in any
80 directory systemd looks for units in, and the usual
81 overriding semantics between /usr/lib, /etc and /run apply
84 * Most unit file settings which take lists of items can now be
85 reset by assigning the empty string to them. For example,
86 normally, settings such as Environment=FOO=BAR append a new
87 environment variable assignment to the environment block,
88 each time they are used. By assigning Environment= the empty
89 string the environment block can be reset to empty. This is
90 particularly useful with the .d/*.conf drop-in snippets
91 mentioned above, since this adds the ability to reset list
92 settings from vendor unit files via these drop-ins.
94 * systemctl gained a new "list-dependencies" command for
95 listing the dependencies of a unit recursively.
97 * Inhibitors are now honored and listed by "systemctl
98 suspend", "systemctl poweroff" (and similar) too, not only
99 GNOME. These commands will also list active sessions by
102 * Resource limits (as exposed by the various control group
103 controllers) can now be controlled dynamically at runtime
104 for all units. More specifically, you can now use a command
105 like "systemctl set-cgroup-attr foobar.service cpu.shares
106 2000" to alter the CPU shares a specific service gets. These
107 settings are stored persistently on disk, and thus allow the
108 administrator to easily adjust the resource usage of
109 services with a few simple commands. This dynamic resource
110 management logic is also available to other programs via the
111 bus. Almost any kernel cgroup attribute and controller is
114 * systemd-vconsole-setup will now copy all font settings to
115 all allocated VTs, where it previously applied them only to
118 * libsystemd-login gained the new sd_session_get_tty() API
121 * This release drops support for a few legacy or
122 distribution-specific LSB facility names when parsing init
123 scripts: $x-display-manager, $mail-transfer-agent,
124 $mail-transport-agent, $mail-transfer-agent, $smtp,
125 $null. Also, the mail-transfer-agent.target unit backing
126 this has been removed. Distributions which want to retain
127 compatibility with this should carry the burden for
128 supporting this themselves and patch support for these back
129 in, if they really need to. Also, the facilities $syslog and
130 $local_fs are now ignored, since systemd does not support
131 early-boot LSB init scripts anymore, and these facilities
132 are implied anyway for normal services. syslog.target has
135 * There are new bus calls on PID1's Manager object for
136 cancelling jobs, and removing snapshot units. Previously,
137 both calls were only available on the Job and Snapshot
140 * systemd-journal-gatewayd gained SSL support.
142 * The various "environment" files, such as /etc/locale.conf
143 now support continuation lines with a backslash ("\") as
144 last character in the line, similar in style (but different)
145 to how this is supported in shells.
147 * For normal user processes the _SYSTEMD_USER_UNIT= field is
148 now implicitly appended to every log entry logged. systemctl
149 has been updated to filter by this field when operating on a
150 user systemd instance.
152 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
153 CAP_AUDIT_CONTROL capabilities to the capabilities set for
154 the container. This makes it easier to boot unmodified
155 Fedora systems in a container, which however still requires
156 audit=0 to be passed on the kernel command line. Auditing in
157 kernel and userspace is unfortunately still too broken in
158 context of containers, hence we recommend compiling it out
159 of the kernel or using audit=0. Hopefully this will be fixed
160 one day for good in the kernel.
162 * nspawn gained the new --bind= and --bind-ro= parameters to
163 bind mount specific directories from the host into the
166 * nspawn will now mount its own devpts file system instance
167 into the container, in order not to leak pty devices from
168 the host into the container.
170 * systemd will now read the firmware boot time performance
171 information from the EFI variables, if the used boot loader
172 supports this, and takes it into account for boot performance
173 analysis via "systemd-analyze". This is currently supported
174 only in conjunction with Gummiboot, but could be supported
175 by other boot loaders too. For details see:
177 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
179 * A new generator has been added that automatically mounts the
180 EFI System Partition (ESP) to /boot, if that directory
181 exists, is empty, and no other file system has been
182 configured to be mounted there.
184 * logind will now send out PrepareForSleep(false) out
185 unconditionally, after coming back from suspend. This may be
186 used by applications as asynchronous notification for
187 system resume events.
189 * "systemctl unlock-sessions" has been added, that allows
190 unlocking the screens of all user sessions at once, similar
191 how "systemctl lock-sessions" already locked all users
192 sessions. This is backed by a new D-Bus call UnlockSessions().
194 * "loginctl seat-status" will now show the master device of a
195 seat. (i.e. the device of a seat that needs to be around for
196 the seat to be considered available, usually the graphics
199 * tmpfiles gained a new "X" line type, that allows
200 configuration of files and directories (with wildcards) that
201 shall be excluded from automatic cleanup ("aging").
203 * udev default rules set the device node permissions now only
204 at "add" events, and do not change them any longer with a
205 later "change" event.
207 * The log messages for lid events and power/sleep keypresses
208 now carry a message ID.
210 * We now have a substantially larger unit test suite, but this
211 continues to be work in progress.
213 * udevadm hwdb gained a new --root= parameter to change the
214 root directory to operate relative to.
216 * logind will now issue a background sync() request to the kernel
217 early at shutdown, so that dirty buffers are flushed to disk early
218 instead of at the last moment, in order to optimize shutdown
221 * A new bootctl tool has been added that is an interface for
222 certain boot loader operations. This is currently a preview
223 and is likely to be extended into a small mechanism daemon
224 like timedated, localed, hostnamed, and can be used by
225 graphical UIs to enumerate available boot options, and
226 request boot into firmware operations.
228 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
229 the rest of the package. It also has been updated to work
230 correctly in initrds.
232 * Policykit previously has been runtime optional, and is now
233 also compile time optional via a configure switch.
235 * systemd-analyze has been reimplemented in C. Also "systemctl
236 dot" has moved into systemd-analyze.
238 * "systemctl status" with no further parameters will now print
239 the status of all active or failed units.
241 * Operations such as "systemctl start" can now be executed
242 with a new mode "--irreversible" which may be used to queue
243 operations that cannot accidentally be reversed by a later
244 job queuing. This is by default used to make shutdown
245 requests more robust.
247 * The Python API of systemd now gained a new module for
248 reading journal files.
250 * A new tool kernel-install has been added that can install
251 kernel images according to the Boot Loader Specification:
253 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
255 * Boot time console output has been improved to provide
256 animated boot time output for hanging jobs.
258 * A new tool systemd-activate has been added which can be used
259 to test socket activation with, directly from the command
260 line. This should make it much easier to test and debug
261 socket activation in daemons.
263 * journalctl gained a new "--reverse" (or -r) option to show
264 journal output in reverse order (i.e. newest line first).
266 * journalctl gained a new "--pager-end" (or -e) option to jump
267 to immediately jump to the end of the journal in the
268 pager. This is only supported in conjunction with "less".
270 * journalctl gained a new "--user-unit=" option, that works
271 similar to "--unit=" but filters for user units rather than
274 * A number of unit files to ease adoption of systemd in
275 initrds has been added. This moves some minimal logic from
276 the various initrd implementations into systemd proper.
278 * The journal files are now owned by a new group
279 "systemd-journal", which exists specifically to allow access
280 to the journal, and nothing else. Previously, we used the
281 "adm" group for that, which however possibly covers more
282 than just journal/log file access. This new group is now
283 already used by systemd-journal-gatewayd to ensure this
284 daemon gets access to the journal files and as little else
285 as possible. Note that "make install" will also set FS ACLs
286 up for /var/log/journal to give "adm" and "wheel" read
287 access to it, in addition to "systemd-journal" which owns
288 the journal files. We recommend that packaging scripts also
289 add read access to "adm" + "wheel" to /var/log/journal, and
290 all existing/future journal files. To normal users and
291 administrators little changes, however packagers need to
292 ensure to create the "systemd-journal" system group at
293 package installation time.
295 * The systemd-journal-gatewayd now runs as unprivileged user
296 systemd-journal-gateway:systemd-journal-gateway. Packaging
297 scripts need to create these system user/group at
300 * timedated now exposes a new boolean property CanNTP that
301 indicates whether a local NTP service is available or not.
303 * systemd-detect-virt will now also detect xen PVs
305 * The pstore file system is now mounted by default, if it is
308 * In addition to the SELinux and IMA policies we will now also
309 load SMACK policies at early boot.
311 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
312 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
313 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
314 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
315 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
316 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
317 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
318 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
319 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
320 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
321 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
322 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
323 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
324 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
328 * Timer units now support calendar time events in addition to
329 monotonic time events. That means you can now trigger a unit
330 based on a calendar time specification such as "Thu,Fri
331 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
332 or fifth day of any month of the year 2013, given that it is
333 a thursday or friday. This brings timer event support
334 considerably closer to cron's capabilities. For details on
335 the supported calendar time specification language see
338 * udev now supports a number of different naming policies for
339 network interfaces for predictable names, and a combination
340 of these policies is now the default. Please see this wiki
341 document for details:
343 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
345 * Auke Kok's bootchart implementation has been added to the
346 systemd tree. It's an optional component that can graph the
347 boot in quite some detail. It's one of the best bootchart
348 implementations around and minimal in its code and
351 * nss-myhostname has been integrated into the systemd source
352 tree. nss-myhostname guarantees that the local hostname
353 always stays resolvable via NSS. It has been a weak
354 requirement of systemd-hostnamed since a long time, and
355 since its code is actually trivial we decided to just
356 include it in systemd's source tree. It can be turned off
357 with a configure switch.
359 * The read-ahead logic is now capable of properly detecting
360 whether a btrfs file system is on SSD or rotating media, in
361 order to optimize the read-ahead scheme. Previously, it was
362 only capable of detecting this on traditional file systems
365 * In udev, additional device properties are now read from the
366 IAB in addition to the OUI database. Also, Bluetooth company
367 identities are attached to the devices as well.
369 * In service files %U may be used as specifier that is
370 replaced by the configured user name of the service.
372 * nspawn may now be invoked without a controlling TTY. This
373 makes it suitable for invocation as its own service. This
374 may be used to set up a simple containerized server system
375 using only core OS tools.
377 * systemd and nspawn can now accept socket file descriptors
378 when they are started for socket activation. This enables
379 implementation of socket activated nspawn
380 containers. i.e. think about autospawning an entire OS image
381 when the first SSH or HTTP connection is received. We expect
382 that similar functionality will also be added to libvirt-lxc
385 * journalctl will now suppress ANSI color codes when
388 * systemctl will no longer show control group information for
389 a unit if a the control group is empty anyway.
391 * logind can now automatically suspend/hibernate/shutdown the
394 * /etc/machine-info and hostnamed now also expose the chassis
395 type of the system. This can be used to determine whether
396 the local system is a laptop, desktop, handset or
397 tablet. This information may either be configured by the
398 user/vendor or is automatically determined from ACPI and DMI
399 information if possible.
401 * A number of PolicyKit actions are now bound together with
402 "imply" rules. This should simplify creating UIs because
403 many actions will now authenticate similar ones as well.
405 * Unit files learnt a new condition ConditionACPower= which
406 may be used to conditionalize a unit depending on whether an
407 AC power source is connected or not, of whether the system
408 is running on battery power.
410 * systemctl gained a new "is-failed" verb that may be used in
411 shell scripts and suchlike to check whether a specific unit
412 is in the "failed" state.
414 * The EnvironmentFile= setting in unit files now supports file
415 globbing, and can hence be used to easily read a number of
416 environment files at once.
418 * systemd will no longer detect and recognize specific
419 distributions. All distribution-specific #ifdeffery has been
420 removed, systemd is now fully generic and
421 distribution-agnostic. Effectively, not too much is lost as
422 a lot of the code is still accessible via explicit configure
423 switches. However, support for some distribution specific
424 legacy configuration file formats has been dropped. We
425 recommend distributions to simply adopt the configuration
426 files everybody else uses now and convert the old
427 configuration from packaging scripts. Most distributions
428 already did that. If that's not possible or desirable,
429 distributions are welcome to forward port the specific
430 pieces of code locally from the git history.
432 * When logging a message about a unit systemd will now always
433 log the unit name in the message meta data.
435 * localectl will now also discover system locale data that is
436 not stored in locale archives, but directly unpacked.
438 * logind will no longer unconditionally use framebuffer
439 devices as seat masters, i.e. as devices that are required
440 to be existing before a seat is considered preset. Instead,
441 it will now look for all devices that are tagged as
442 "seat-master" in udev. By default framebuffer devices will
443 be marked as such, but depending on local systems other
444 devices might be marked as well. This may be used to
445 integrate graphics cards using closed source drivers (such
446 as NVidia ones) more nicely into logind. Note however, that
447 we recommend using the open source NVidia drivers instead,
448 and no udev rules for the closed-source drivers will be
449 shipped from us upstream.
451 Contributions from: Adam Williamson, Alessandro Crismani, Auke
452 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
453 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
454 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
455 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
456 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
457 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
458 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
459 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
460 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
461 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
462 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
467 * udev gained support for loading additional device properties
468 from an indexed database that is keyed by vendor/product IDs
469 and similar device identifiers. For the beginning this
470 "hwdb" is populated with data from the well-known PCI and
471 USB database, but also includes PNP, ACPI and OID data. In
472 the longer run this indexed database shall grow into
473 becoming the one central database for non-essential
474 userspace device metadata. Previously, data from the PCI/USB
475 database was only attached to select devices, since the
476 lookup was a relatively expensive operation due to O(n) time
477 complexity (with n being the number of entries in the
478 database). Since this is now O(1), we decided to add in this
479 data for all devices where this is available, by
480 default. Note that the indexed database needs to be rebuilt
481 when new data files are installed. To achieve this you need
482 to update your packaging scripts to invoke "udevadm hwdb
483 --update" after installation of hwdb data files. For
484 RPM-based distributions we introduced the new
485 %udev_hwdb_update macro for this purpose.
487 * The Journal gained support for the "Message Catalog", an
488 indexed database to link up additional information with
489 journal entries. For further details please check:
491 http://www.freedesktop.org/wiki/Software/systemd/catalog
493 The indexed message catalog database also needs to be
494 rebuilt after installation of message catalog files. Use
495 "journalctl --update-catalog" for this. For RPM-based
496 distributions we introduced the %journal_catalog_update
497 macro for this purpose.
499 * The Python Journal bindings gained support for the standard
500 Python logging framework.
502 * The Journal API gained new functions for checking whether
503 the underlying file system of a journal file is capable of
504 properly reporting file change notifications, or whether
505 applications that want to reflect journal changes "live"
506 need to recheck journal files continously in appropriate
509 * It is now possible to set the "age" field for tmpfiles
510 entries to 0, indicating that files matching this entry
511 shall always be removed when the directories are cleaned up.
513 * coredumpctl gained a new "gdb" verb which invokes gdb
514 right-away on the selected coredump.
516 * There's now support for "hybrid sleep" on kernels that
517 support this, in addition to "suspend" and "hibernate". Use
518 "systemctl hybrid-sleep" to make use of this.
520 * logind's HandleSuspendKey= setting (and related settings)
521 now gained support for a new "lock" setting to simply
522 request the screen lock on all local sessions, instead of
523 actually executing a suspend or hibernation.
525 * systemd will now mount the EFI variables file system by
528 * Socket units now gained support for configuration of the
529 SMACK security label.
531 * timedatectl will now output the time of the last and next
532 daylight saving change.
534 * We dropped support for various legacy and distro-specific
535 concepts, such as insserv, early-boot SysV services
536 (i.e. those for non-standard runlevels such as 'b' or 'S')
537 or ArchLinux /etc/rc.conf support. We recommend the
538 distributions who still need support this to either continue
539 to maintain the necessary patches downstream, or find a
540 different solution. (Talk to us if you have questions!)
542 * Various systemd components will now bypass PolicyKit checks
543 for root and otherwise handle properly if PolicyKit is not
544 found to be around. This should fix most issues for
545 PolicyKit-less systems. Quite frankly this should have been
546 this way since day one. It is absolutely our intention to
547 make systemd work fine on PolicyKit-less systems, and we
548 consider it a bug if something doesn't work as it should if
549 PolicyKit is not around.
551 * For embedded systems it is now possible to build udev and
552 systemd without blkid and/or kmod support.
554 * "systemctl switch-root" is now capable of switching root
555 more than once. I.e. in addition to transitions from the
556 initrd to the host OS it is now possible to transition to
557 further OS images from the host. This is useful to implement
558 offline updating tools.
560 * Various other additions have been made to the RPM macros
561 shipped with systemd. Use %udev_rules_update() after
562 installing new udev rules files. %_udevhwdbdir,
563 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
564 %_sysctldir are now available which resolve to the right
565 directories for packages to place various data files in.
567 * journalctl gained the new --full switch (in addition to
568 --all, to disable ellipsation for long messages.
570 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
571 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
572 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
573 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
574 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
575 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
576 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
577 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
578 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
582 * journalctl gained new --since= and --until= switches to
583 filter by time. It also now supports nice filtering for
584 units via --unit=/-u.
586 * Type=oneshot services may use ExecReload= and do the
589 * The journal daemon now supports time-based rotation and
590 vacuuming, in addition to the usual disk-space based
593 * The journal will now index the available field values for
594 each field name. This enables clients to show pretty drop
595 downs of available match values when filtering. The bash
596 completion of journalctl has been updated
597 accordingly. journalctl gained a new switch -F to list all
598 values a certain field takes in the journal database.
600 * More service events are now written as structured messages
601 to the journal, and made recognizable via message IDs.
603 * The timedated, localed and hostnamed mini-services which
604 previously only provided support for changing time, locale
605 and hostname settings from graphical DEs such as GNOME now
606 also have a minimal (but very useful) text-based client
607 utility each. This is probably the nicest way to changing
608 these settings from the command line now, especially since
609 it lists available options and is fully integrated with bash
612 * There's now a new tool "systemd-coredumpctl" to list and
613 extract coredumps from the journal.
615 * We now install a README each in /var/log/ and
616 /etc/rc.d/init.d explaining where the system logs and init
617 scripts went. This hopefully should help folks who go to
618 that dirs and look into the otherwise now empty void and
621 * When user-services are invoked (by systemd --user) the
622 $MANAGERPID env var is set to the PID of systemd.
624 * SIGRTMIN+24 when sent to a --user instance will now result
625 in immediate termination of systemd.
627 * gatewayd received numerous feature additions such as a
628 "follow" mode, for live syncing and filtering.
630 * browse.html now allows filtering and showing detailed
631 information on specific entries. Keyboard navigation and
632 mouse screen support has been added.
634 * gatewayd/journalctl now supports HTML5/JSON
635 Server-Sent-Events as output.
637 * The SysV init script compatibility logic will now
638 heuristically determine whether a script supports the
639 "reload" verb, and only then make this available as
642 * "systemctl status --follow" has been removed, use "journalctl
645 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
646 have been removed since they are hardly useful to be
649 * And I'd like to take the opportunity to specifically mention
650 Zbigniew for his great contributions. Zbigniew, you rock!
652 Contributions from: Andrew Eikum, Christian Hesse, Colin
653 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
654 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
655 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
656 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
657 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
658 Jędrzejewski-Szmek, Сковорода Никита Андреевич
662 * If /etc/vconsole.conf is non-existent or empty we will no
663 longer load any console font or key map at boot by
664 default. Instead the kernel defaults will be left
665 intact. This is definitely the right thing to do, as no
666 configuration should mean no configuration, and hard-coding
667 font names that are different on all archs is probably a bad
668 idea. Also, the kernel default key map and font should be
669 good enough for most cases anyway, and mostly identical to
670 the userspace fonts/key maps we previously overloaded them
671 with. If distributions want to continue to default to a
672 non-kernel font or key map they should ship a default
673 /etc/vconsole.conf with the appropriate contents.
675 Contributions from: Colin Walters, Daniel J Walsh, Dave
676 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
677 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
681 * journalctl gained a new --cursor= switch to show entries
682 starting from the specified location in the journal.
684 * We now enforce a size limit on journal entry fields exported
685 with "-o json" in journalctl. Fields larger than 4K will be
686 assigned null. This can be turned off with --all.
688 * An (optional) journal gateway daemon is now available as
689 "systemd-journal-gatewayd.service". This service provides
690 access to the journal via HTTP and JSON. This functionality
691 will be used to implement live log synchronization in both
692 pull and push modes, but has various other users too, such
693 as easy log access for debugging of embedded devices. Right
694 now it is already useful to retrieve the journal via HTTP:
696 # systemctl start systemd-journal-gatewayd.service
697 # wget http://localhost:19531/entries
699 This will download the journal contents in a
700 /var/log/messages compatible format. The same as JSON:
702 # curl -H"Accept: application/json" http://localhost:19531/entries
704 This service is also accessible via a web browser where a
705 single static HTML5 app is served that uses the JSON logic
706 to enable the user to do some basic browsing of the
707 journal. This will be extended later on. Here's an example
708 screenshot of this app in its current state:
710 http://0pointer.de/public/journal-gatewayd
712 Contributions from: Kay Sievers, Lennart Poettering, Robert
713 Milasan, Tom Gundersen
717 * The bash completion logic is now available for journalctl
720 * We don't mount the "cpuset" controller anymore together with
721 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
722 started if no parameters are assigned to it. "cpuset" hence
723 broke code that assumed it it could create "cpu" groups and
726 * journalctl -f will now subscribe to terminal size changes,
727 and line break accordingly.
729 Contributions from: Dave Reisner, Kay Sievers, Lennart
730 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
734 * nspawn will now create a symlink /etc/localtime in the
735 container environment, copying the host's timezone
736 setting. Previously this has been done via a bind mount, but
737 since symlinks cannot be bind mounted this has now been
738 changed to create/update the appropriate symlink.
740 * journalctl -n's line number argument is now optional, and
741 will default to 10 if omitted.
743 * journald will now log the maximum size the journal files may
744 take up on disk. This is particularly useful if the default
745 built-in logic of determining this parameter from the file
746 system size is used. Use "systemctl status
747 systemd-journald.service" to see this information.
749 * The multi-seat X wrapper tool has been stripped down. As X
750 is now capable of enumerating graphics devices via udev in a
751 seat-aware way the wrapper is not strictly necessary
752 anymore. A stripped down temporary stop-gap is still shipped
753 until the upstream display managers have been updated to
754 fully support the new X logic. Expect this wrapper to be
755 removed entirely in one of the next releases.
757 * HandleSleepKey= in logind.conf has been split up into
758 HandleSuspendKey= and HandleHibernateKey=. The old setting
759 is not available anymore. X11 and the kernel are
760 distuingishing between these keys and we should too. This
761 also means the inhibition lock for these keys has been split
764 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
765 Poettering, Lukas Nykryn, Václav Pavlín
769 * Whenever a unit changes state we'll now log this to the
770 journal and show along the unit's own log output in
773 * ConditionPathIsMountPoint= can now properly detect bind
774 mount points too. (Previously, a bind mount of one file
775 system to another place in the same file system could not be
776 detected as mount, since they shared struct stat's st_dev
779 * We will now mount the cgroup controllers cpu, cpuacct,
780 cpuset and the controllers net_cls, net_prio together by
783 * nspawn containers will now have a virtualized boot
784 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
785 over with a randomized ID at container initialization). This
786 has the effect of making "journalctl -b" do the right thing
789 * The JSON output journal serialization has been updated not
790 to generate "endless" list objects anymore, but rather one
791 JSON object per line. This is more in line how most JSON
792 parsers expect JSON objects. The new output mode
793 "json-pretty" has been added to provide similar output, but
794 neatly aligned for readability by humans.
796 * We dropped all explicit sync() invocations in the shutdown
797 code. The kernel does this implicitly anyway in the kernel
798 reboot() syscall. halt(8)'s -n option is now a compatibility
801 * We now support virtualized reboot() in containers, as
802 supported by newer kernels. We will fall back to exit() if
803 CAP_SYS_REBOOT is not available to the container. Also,
804 nspawn makes use of this now and will actually reboot the
805 container if the containerized OS asks for that.
807 * journalctl will only show local log output by default
808 now. Use --merge (-m) to show remote log output, too.
810 * libsystemd-journal gained the new sd_journal_get_usage()
811 call to determine the current disk usage of all journal
812 files. This is exposed in the new "journalctl --disk-usage"
815 * journald gained a new configuration setting SplitMode= in
816 journald.conf which may be used to control how user journals
817 are split off. See journald.conf(5) for details.
819 * A new condition type ConditionFileNotEmpty= has been added.
821 * tmpfiles' "w" lines now support file globbing, to write
822 multiple files at once.
824 * We added Python bindings for the journal submission
825 APIs. More Python APIs for a number of selected APIs will
826 likely follow. Note that we intend to add native bindings
827 only for the Python language, as we consider it common
828 enough to deserve bindings shipped within systemd. There are
829 various projects outside of systemd that provide bindings
830 for languages such as PHP or Lua.
832 * Many conditions will now resolve specifiers such as %i. In
833 addition, PathChanged= and related directives of .path units
834 now support specifiers as well.
836 * There's now a new RPM macro definition for the system preset
839 * journald will now warn if it can't foward a message to the
840 syslog daemon because it's socket is full.
842 * timedated will no longer write or process /etc/timezone,
843 except on Debian. As we do not support late mounted /usr
844 anymore /etc/localtime always being a symlink is now safe,
845 and hence the information in /etc/timezone is not necessary
848 * logind will now always reserve one VT for a text getty (VT6
849 by default). Previously if more than 6 X sessions where
850 started they took up all the VTs with auto-spawned gettys,
851 so that no text gettys were available anymore.
853 * udev will now automatically inform the btrfs kernel logic
854 about btrfs RAID components showing up. This should make
855 simple hotplug based btrfs RAID assembly work.
857 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
858 (but not for its children which will stay at the kernel
859 default). This should allow setups with a lot more listening
862 * systemd will now always pass the configured timezone to the
863 kernel at boot. timedated will do the same when the timezone
866 * logind's inhibition logic has been updated. By default,
867 logind will now handle the lid switch, the power and sleep
868 keys all the time, even in graphical sessions. If DEs want
869 to handle these events on their own they should take the new
870 handle-power-key, handle-sleep-key and handle-lid-switch
871 inhibitors during their runtime. A simple way to achiveve
872 that is to invoke the DE wrapped in an invocation of:
874 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
876 * Access to unit operations is now checked via SELinux taking
877 the unit file label and client process label into account.
879 * systemd will now notify the administrator in the journal
880 when he over-mounts a non-empty directory.
882 * There are new specifiers that are resolved in unit files,
883 for the host name (%H), the machine ID (%m) and the boot ID
886 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
887 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
888 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
889 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
890 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
891 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
892 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
896 * Support for reading structured kernel messages from
897 /dev/kmsg has now been added and is enabled by default.
899 * Support for reading kernel messages from /proc/kmsg has now
900 been removed. If you want kernel messages in the journal
901 make sure to run a recent kernel (>= 3.5) that supports
902 reading structured messages from /dev/kmsg (see
903 above). /proc/kmsg is now exclusive property of classic
904 syslog daemons again.
906 * The libudev API gained the new
907 udev_device_new_from_device_id() call.
909 * The logic for file system namespace (ReadOnlyDirectory=,
910 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
911 require pivot_root() anymore. This means fewer temporary
912 directories are created below /tmp for this feature.
914 * nspawn containers will now see and receive all submounts
915 made on the host OS below the root file system of the
918 * Forward Secure Sealing is now supported for Journal files,
919 which provide cryptographical sealing of journal files so
920 that attackers cannot alter log history anymore without this
921 being detectable. Lennart will soon post a blog story about
922 this explaining it in more detail.
924 * There are two new service settings RestartPreventExitStatus=
925 and SuccessExitStatus= which allow configuration of exit
926 status (exit code or signal) which will be excepted from the
927 restart logic, resp. consider successful.
929 * journalctl gained the new --verify switch that can be used
930 to check the integrity of the structure of journal files and
931 (if Forward Secure Sealing is enabled) the contents of
934 * nspawn containers will now be run with /dev/stdin, /dev/fd/
935 and similar symlinks pre-created. This makes running shells
936 as container init process a lot more fun.
938 * The fstab support can now handle PARTUUID= and PARTLABEL=
941 * A new ConditionHost= condition has been added to match
942 against the hostname (with globs) and machine ID. This is
943 useful for clusters where a single OS image is used to
944 provision a large number of hosts which shall run slightly
945 different sets of services.
947 * Services which hit the restart limit will now be placed in a
950 Contributions from: Bertram Poettering, Dave Reisner, Huang
951 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
952 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
956 * When running in --user mode systemd will now become a
957 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
958 tree a lot more organized.
960 * A new PartOf= unit dependency type has been introduced that
961 may be used to group services in a natural way.
963 * "systemctl enable" may now be used to enable instances of
966 * journalctl now prints error log levels in red, and
967 warning/notice log levels in bright white. It also supports
968 filtering by log level now.
970 * cgtop gained a new -n switch (similar to top), to configure
971 the maximum number of iterations to run for. It also gained
972 -b, to run in batch mode (accepting no input).
974 * The suffix ".service" may now be ommited on most systemctl
975 command lines involving service unit names.
977 * There's a new bus call in logind to lock all sessions, as
978 well as a loginctl verb for it "lock-sessions".
980 * libsystemd-logind.so gained a new call sd_journal_perror()
981 that works similar to libc perror() but logs to the journal
982 and encodes structured information about the error number.
984 * /etc/crypttab entries now understand the new keyfile-size=
987 * shutdown(8) now can send a (configurable) wall message when
988 a shutdown is cancelled.
990 * The mount propagation mode for the root file system will now
991 default to "shared", which is useful to make containers work
992 nicely out-of-the-box so that they receive new mounts from
993 the host. This can be undone locally by running "mount
994 --make-rprivate /" if needed.
996 * The prefdm.service file has been removed. Distributions
997 should maintain this unit downstream if they intend to keep
998 it around. However, we recommend writing normal unit files
999 for display managers instead.
1001 * Since systemd is a crucial part of the OS we will now
1002 default to a number of compiler switches that improve
1003 security (hardening) such as read-only relocations, stack
1004 protection, and suchlike.
1006 * The TimeoutSec= setting for services is now split into
1007 TimeoutStartSec= and TimeoutStopSec= to allow configuration
1008 of individual time outs for the start and the stop phase of
1011 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
1012 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
1013 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
1014 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
1015 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
1016 Gundersen, Zbigniew Jędrzejewski-Szmek
1020 * The journal and id128 C APIs are now fully documented as man
1023 * Extra safety checks have been added when transitioning from
1024 the initial RAM disk to the main system to avoid accidental
1027 * /etc/crypttab entries now understand the new keyfile-offset=
1030 * systemctl -t can now be used to filter by unit load state.
1032 * The journal C API gained the new sd_journal_wait() call to
1033 make writing synchronous journal clients easier.
1035 * journalctl gained the new -D switch to show journals from a
1038 * journalctl now displays a special marker between log
1039 messages of two different boots.
1041 * The journal is now explicitly flushed to /var via a service
1042 systemd-journal-flush.service, rather than implicitly simply
1043 by seeing /var/log/journal to be writable.
1045 * journalctl (and the journal C APIs) can now match for much
1046 more complex expressions, with alternatives and
1049 * When transitioning from the initial RAM disk to the main
1050 system we will now kill all processes in a killing spree to
1051 ensure no processes stay around by accident.
1053 * Three new specifiers may be used in unit files: %u, %h, %s
1054 resolve to the user name, user home directory resp. user
1055 shell. This is useful for running systemd user instances.
1057 * We now automatically rotate journal files if their data
1058 object hash table gets a fill level > 75%. We also size the
1059 hash table based on the configured maximum file size. This
1060 together should lower hash collisions drastically and thus
1061 speed things up a bit.
1063 * journalctl gained the new "--header" switch to introspect
1064 header data of journal files.
1066 * A new setting SystemCallFilters= has been added to services
1067 which may be used to apply blacklists or whitelists to
1068 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
1070 * nspawn gained a new --link-journal= switch (and quicker: -j)
1071 to link the container journal with the host. This makes it
1072 very easy to centralize log viewing on the host for all
1073 guests while still keeping the journal files separated.
1075 * Many bugfixes and optimizations
1077 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
1078 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
1079 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
1084 * Several tools now understand kernel command line arguments,
1085 which are only read when run in an initial RAM disk. They
1086 usually follow closely their normal counterparts, but are
1089 * There's a new tool to analyze the readahead files that are
1090 automatically generated at boot. Use:
1092 /usr/lib/systemd/systemd-readahead analyze /.readahead
1094 * We now provide an early debug shell on tty9 if this enabled. Use:
1096 systemctl enable debug-shell.service
1098 * All plymouth related units have been moved into the Plymouth
1099 package. Please make sure to upgrade your Plymouth version
1102 * systemd-tmpfiles now supports getting passed the basename of
1103 a configuration file only, in which case it will look for it
1104 in all appropriate directories automatically.
1106 * udevadm info now takes a /dev or /sys path as argument, and
1107 does the right thing. Example:
1109 udevadm info /dev/sda
1110 udevadm info /sys/class/block/sda
1112 * systemctl now prints a warning if a unit is stopped but a
1113 unit that might trigger it continues to run. Example: a
1114 service is stopped but the socket that activates it is left
1117 * "systemctl status" will now mention if the log output was
1118 shortened due to rotation since a service has been started.
1120 * The journal API now exposes functions to determine the
1121 "cutoff" times due to rotation.
1123 * journald now understands SIGUSR1 and SIGUSR2 for triggering
1124 immediately flushing of runtime logs to /var if possible,
1125 resp. for triggering immediate rotation of the journal
1128 * It is now considered an error if a service is attempted to
1129 be stopped that is not loaded.
1131 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
1133 * systemd-analyze now supports Python 3
1135 * tmpfiles now supports cleaning up directories via aging
1136 where the first level dirs are always kept around but
1137 directories beneath it automatically aged. This is enabled
1138 by prefixing the age field with '~'.
1140 * Seat objects now expose CanGraphical, CanTTY properties
1141 which is required to deal with very fast bootups where the
1142 display manager might be running before the graphics drivers
1143 completed initialization.
1145 * Seat objects now expose a State property.
1147 * We now include RPM macros for service enabling/disabling
1148 based on the preset logic. We recommend RPM based
1149 distributions to make use of these macros if possible. This
1150 makes it simpler to reuse RPM spec files across
1153 * We now make sure that the collected systemd unit name is
1154 always valid when services log to the journal via
1157 * There's a new man page kernel-command-line(7) detailing all
1158 command line options we understand.
1160 * The fstab generator may now be disabled at boot by passing
1161 fstab=0 on the kernel command line.
1163 * A new kernel command line option modules-load= is now understood
1164 to load a specific kernel module statically, early at boot.
1166 * Unit names specified on the systemctl command line are now
1167 automatically escaped as needed. Also, if file system or
1168 device paths are specified they are automatically turned
1169 into the appropriate mount or device unit names. Example:
1171 systemctl status /home
1172 systemctl status /dev/sda
1174 * The SysVConsole= configuration option has been removed from
1175 system.conf parsing.
1177 * The SysV search path is no longer exported on the D-Bus
1180 * The Names= option is been removed from unit file parsing.
1182 * There's a new man page bootup(7) detailing the boot process.
1184 * Every unit and every generator we ship with systemd now
1185 comes with full documentation. The self-explanatory boot is
1188 * A couple of services gained "systemd-" prefixes in their
1189 name if they wrap systemd code, rather than only external
1190 code. Among them fsck@.service which is now
1191 systemd-fsck@.service.
1193 * The HaveWatchdog property has been removed from the D-Bus
1196 * systemd.confirm_spawn= on the kernel command line should now
1199 * There's a new man page crypttab(5) which details all options
1200 we actually understand.
1202 * systemd-nspawn gained a new --capability= switch to pass
1203 additional capabilities to the container.
1205 * timedated will now read known NTP implementation unit names
1206 from /usr/lib/systemd/ntp-units.d/*.list,
1207 systemd-timedated-ntp.target has been removed.
1209 * journalctl gained a new switch "-b" that lists log data of
1210 the current boot only.
1212 * The notify socket is in the abstract namespace again, in
1213 order to support daemons which chroot() at start-up.
1215 * There is a new Storage= configuration option for journald
1216 which allows configuration of where log data should go. This
1217 also provides a way to disable journal logging entirely, so
1218 that data collected is only forwarded to the console, the
1219 kernel log buffer or another syslog implementation.
1221 * Many bugfixes and optimizations
1223 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
1224 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
1225 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
1226 Shawn Landden, Tom Gundersen
1230 * "systemctl help <unit>" now shows the man page if one is
1233 * Several new man pages have been added.
1235 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
1236 MaxLevelConsole= can now be specified in
1237 journald.conf. These options allow reducing the amount of
1238 data stored on disk or forwarded by the log level.
1240 * TimerSlackNSec= can now be specified in system.conf for
1241 PID1. This allows system-wide power savings.
1243 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
1244 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
1249 * logind is now capable of (optionally) handling power and
1250 sleep keys as well as the lid switch.
1252 * journalctl now understands the syntax "journalctl
1253 /usr/bin/avahi-daemon" to get all log output of a specific
1256 * CapabilityBoundingSet= in system.conf now also influences
1257 the capability bound set of usermode helpers of the kernel.
1259 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
1260 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
1261 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
1262 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
1266 * Note that we skipped 139 releases here in order to set the
1267 new version to something that is greater than both udev's
1268 and systemd's most recent version number.
1270 * udev: all udev sources are merged into the systemd source tree now.
1271 All future udev development will happen in the systemd tree. It
1272 is still fully supported to use the udev daemon and tools without
1273 systemd running, like in initramfs or other init systems. Building
1274 udev though, will require the *build* of the systemd tree, but
1275 udev can be properly *run* without systemd.
1277 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
1278 should be used to create dead device nodes as workarounds for broken
1281 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
1282 no longer supported. udev_monitor_new_from_netlink() needs to be
1283 used to subscribe to events.
1285 * udev: when udevd is started by systemd, processes which are left
1286 behind by forking them off of udev rules, are unconditionally cleaned
1287 up and killed now after the event handling has finished. Services or
1288 daemons must be started as systemd services. Services can be
1289 pulled-in by udev to get started, but they can no longer be directly
1290 forked by udev rules.
1292 * udev: the daemon binary is called systemd-udevd now and installed
1293 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
1294 to adapt to that, create symlink, or rename the binary after building
1297 * libudev no longer provides these symbols:
1298 udev_monitor_from_socket()
1299 udev_queue_get_failed_list_entry()
1300 udev_get_{dev,sys,run}_path()
1301 The versions number was bumped and symbol versioning introduced.
1303 * systemd-loginctl and systemd-journalctl have been renamed
1304 to loginctl and journalctl to match systemctl.
1306 * The config files: /etc/systemd/systemd-logind.conf and
1307 /etc/systemd/systemd-journald.conf have been renamed to
1308 logind.conf and journald.conf. Package updates should rename
1309 the files to the new names on upgrade.
1311 * For almost all files the license is now LGPL2.1+, changed
1312 from the previous GPL2.0+. Exceptions are some minor stuff
1313 of udev (which will be changed to LGPL2.1 eventually, too),
1314 and the MIT licensed sd-daemon.[ch] library that is suitable
1315 to be used as drop-in files.
1317 * systemd and logind now handle system sleep states, in
1318 particular suspending and hibernating.
1320 * logind now implements a sleep/shutdown/idle inhibiting logic
1321 suitable for a variety of uses. Soonishly Lennart will blog
1322 about this in more detail.
1324 * var-run.mount and var-lock.mount are no longer provided
1325 (which prevously bind mounted these directories to their new
1326 places). Distributions which have not converted these
1327 directories to symlinks should consider stealing these files
1328 from git history and add them downstream.
1330 * We introduced the Documentation= field for units and added
1331 this to all our shipped units. This is useful to make it
1332 easier to explore the boot and the purpose of the various
1335 * All smaller setup units (such as
1336 systemd-vconsole-setup.service) now detect properly if they
1337 are run in a container and are skipped when
1338 appropriate. This guarantees an entirely noise-free boot in
1339 Linux container environments such as systemd-nspawn.
1341 * A framework for implementing offline system updates is now
1342 integrated, for details see:
1343 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
1345 * A new service type Type=idle is available now which helps us
1346 avoiding ugly interleaving of getty output and boot status
1349 * There's now a system-wide CapabilityBoundingSet= option to
1350 globally reduce the set of capabilities for the
1351 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
1352 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
1353 even CAP_NET_ADMIN system-wide for secure systems.
1355 * There are now system-wide DefaultLimitXXX= options to
1356 globally change the defaults of the various resource limits
1357 for all units started by PID 1.
1359 * Harald Hoyer's systemd test suite has been integrated into
1360 systemd which allows easy testing of systemd builds in qemu
1361 and nspawn. (This is really awesome! Ask us for details!)
1363 * The fstab parser is now implemented as generator, not inside
1366 * systemctl will now warn you if .mount units generated from
1367 /etc/fstab are out of date due to changes in fstab that
1368 haven't been read by systemd yet.
1370 * systemd is now suitable for usage in initrds. Dracut has
1371 already been updated to make use of this. With this in place
1372 initrds get a slight bit faster but primarily are much
1373 easier to introspect and debug since "systemctl status" in
1374 the host system can be used to introspect initrd services,
1375 and the journal from the initrd is kept around too.
1377 * systemd-delta has been added, a tool to explore differences
1378 between user/admin configuration and vendor defaults.
1380 * PrivateTmp= now affects both /tmp and /var/tmp.
1382 * Boot time status messages are now much prettier and feature
1383 proper english language. Booting up systemd has never been
1386 * Read-ahead pack files now include the inode number of all
1387 files to pre-cache. When the inode changes the pre-caching
1388 is not attempted. This should be nicer to deal with updated
1389 packages which might result in changes of read-ahead
1392 * We now temporaritly lower the kernel's read_ahead_kb variable
1393 when collecting read-ahead data to ensure the kernel's
1394 built-in read-ahead does not add noise to our measurements
1395 of necessary blocks to pre-cache.
1397 * There's now RequiresMountsFor= to add automatic dependencies
1398 for all mounts necessary for a specific file system path.
1400 * MountAuto= and SwapAuto= have been removed from
1401 system.conf. Mounting file systems at boot has to take place
1404 * nspawn now learned a new switch --uuid= to set the machine
1405 ID on the command line.
1407 * nspawn now learned the -b switch to automatically search
1410 * vt102 is now the default TERM for serial TTYs, upgraded from
1413 * systemd-logind now works on VT-less systems.
1415 * The build tree has been reorganized. The individual
1416 components now have directories of their own.
1418 * A new condition type ConditionPathIsReadWrite= is now available.
1420 * nspawn learned the new -C switch to create cgroups for the
1421 container in other hierarchies.
1423 * We now have support for hardware watchdogs, configurable in
1426 * The scheduled shutdown logic now has a public API.
1428 * We now mount /tmp as tmpfs by default, but this can be
1429 masked and /etc/fstab can override it.
1431 * Since udisks doesn't make use of /media anymore we are not
1432 mounting a tmpfs on it anymore.
1434 * journalctl gained a new --local switch to only interleave
1435 locally generated journal files.
1437 * We can now load the IMA policy at boot automatically.
1439 * The GTK tools have been split off into a systemd-ui.
1441 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
1442 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
1443 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
1444 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
1445 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
1446 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
1447 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
1448 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
1449 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
1454 * This is mostly a bugfix release
1456 * Support optional initialization of the machine ID from the
1457 KVM or container configured UUID.
1459 * Support immediate reboots with "systemctl reboot -ff"
1461 * Show /etc/os-release data in systemd-analyze output
1463 * Many bugfixes for the journal, including endianess fixes and
1464 ensuring that disk space enforcement works
1466 * sd-login.h is C++ comptaible again
1468 * Extend the /etc/os-release format on request of the Debian
1471 * We now refuse non-UTF8 strings used in various configuration
1472 and unit files. This is done to ensure we don't pass invalid
1473 data over D-Bus or expose it elsewhere.
1475 * Register Mimo USB Screens as suitable for automatic seat
1478 * Read SELinux client context from journal clients in a race
1481 * Reorder configuration file lookup order. /etc now always
1482 overrides /run in order to allow the administrator to always
1483 and unconditionally override vendor supplied or
1484 automatically generated data.
1486 * The various user visible bits of the journal now have man
1487 pages. We still lack man pages for the journal API calls
1490 * We now ship all man pages in HTML format again in the
1493 Contributions from: Dave Reisner, Dirk Eibach, Frederic
1494 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
1495 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
1500 * This is mostly a bugfix release
1502 * systems lacking /etc/os-release are no longer supported.
1504 * Various functionality updates to libsystemd-login.so
1506 * Track class of PAM logins to distuingish greeters from
1509 Contributions from: Kay Sievers, Lennart Poettering, Michael
1514 * This is an important bugfix release for v41.
1516 * Building man pages is now optional which should be useful
1517 for those building systemd from git but unwilling to install
1520 * Watchdog support for supervising services is now usable. In
1521 a future release support for hardware watchdogs
1522 (i.e. /dev/watchdog) will be added building on this.
1524 * Service start rate limiting is now configurable and can be
1525 turned off per service. When a start rate limit is hit a
1526 reboot can automatically be triggered.
1528 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
1530 Contributions from: Benjamin Franzke, Bill Nottingham,
1531 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
1532 Schmidt, Michał Górny, Piotr Drąg
1536 * The systemd binary is installed /usr/lib/systemd/systemd now;
1537 An existing /sbin/init symlink needs to be adapted with the
1540 * The code that loads kernel modules has been ported to invoke
1541 libkmod directly, instead of modprobe. This means we do not
1542 support systems with module-init-tools anymore.
1544 * Watchdog support is now already useful, but still not
1547 * A new kernel command line option systemd.setenv= is
1548 understood to set system wide environment variables
1549 dynamically at boot.
1551 * We now limit the set of capabilities of systemd-journald.
1553 * We now set SIGPIPE to ignore by default, since it only is
1554 useful in shell pipelines, and has little use in general
1555 code. This can be disabled with IgnoreSIPIPE=no in unit
1558 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
1559 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
1564 * This is mostly a bugfix release
1566 * We now expose the reason why a service failed in the
1567 "Result" D-Bus property.
1569 * Rudimentary service watchdog support (will be completed over
1570 the next few releases.)
1572 * When systemd forks off in order execute some service we will
1573 now immediately changes its argv[0] to reflect which process
1574 it will execute. This is useful to minimize the time window
1575 with a generic argv[0], which makes bootcharts more useful
1577 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
1578 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
1579 Mike Kazantsev, Ray Strode
1583 * This is mostly a test release, but incorporates many
1586 * New systemd-cgtop tool to show control groups by their
1589 * Linking against libacl for ACLs is optional again. If
1590 disabled, support tracking device access for active logins
1591 goes becomes unavailable, and so does access to the user
1592 journals by the respective users.
1594 * If a group "adm" exists, journal files are automatically
1595 owned by them, thus allow members of this group full access
1596 to the system journal as well as all user journals.
1598 * The journal now stores the SELinux context of the logging
1599 client for all entries.
1601 * Add C++ inclusion guards to all public headers
1603 * New output mode "cat" in the journal to print only text
1604 messages, without any meta data like date or time.
1606 * Include tiny X server wrapper as a temporary stop-gap to
1607 teach XOrg udev display enumeration. This is used by display
1608 managers such as gdm, and will go away as soon as XOrg
1609 learned native udev hotplugging for display devices.
1611 * Add new systemd-cat tool for executing arbitrary programs
1612 with STDERR/STDOUT connected to the journal. Can also act as
1613 BSD logger replacement, and does so by default.
1615 * Optionally store all locally generated coredumps in the
1616 journal along with meta data.
1618 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
1619 writing short strings to files (for usage for /sys), and for
1620 creating symlinks, character and block device nodes.
1622 * New unit file option ControlGroupPersistent= to make cgroups
1623 persistent, following the mechanisms outlined in
1624 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
1626 * Support multiple local RTCs in a sane way
1628 * No longer monopolize IO when replaying readahead data on
1629 rotating disks, since we might starve non-file-system IO to
1630 death, since fanotify() will not see accesses done by blkid,
1633 * Don't show kernel threads in systemd-cgls anymore, unless
1634 requested with new -k switch.
1636 Contributions from: Dan Horák, Kay Sievers, Lennart
1637 Poettering, Michal Schmidt
1641 * This is mostly a test release, but incorporates many
1644 * The git repository moved to:
1645 git://anongit.freedesktop.org/systemd/systemd
1646 ssh://git.freedesktop.org/git/systemd/systemd
1648 * First release with the journal
1649 http://0pointer.de/blog/projects/the-journal.html
1651 * The journal replaces both systemd-kmsg-syslogd and
1652 systemd-stdout-bridge.
1654 * New sd_pid_get_unit() API call in libsystemd-logind
1656 * Many systemadm clean-ups
1658 * Introduce remote-fs-pre.target which is ordered before all
1659 remote mounts and may be used to start services before all
1662 * Added Mageia support
1664 * Add bash completion for systemd-loginctl
1666 * Actively monitor PID file creation for daemons which exit in
1667 the parent process before having finished writing the PID
1668 file in the daemon process. Daemons which do this need to be
1669 fixed (i.e. PID file creation must have finished before the
1670 parent exits), but we now react a bit more gracefully to them.
1672 * Add colourful boot output, mimicking the well-known output
1673 of existing distributions.
1675 * New option PassCredentials= for socket units, for
1676 compatibility with a recent kernel ABI breakage.
1678 * /etc/rc.local is now hooked in via a generator binary, and
1679 thus will no longer act as synchronization point during
1682 * systemctl list-unit-files now supports --root=.
1684 * systemd-tmpfiles now understands two new commands: z, Z for
1685 relabelling files according to the SELinux database. This is
1686 useful to apply SELinux labels to specific files in /sys,
1689 * Output of SysV services is now forwarded to both the console
1690 and the journal by default, not only just the console.
1692 * New man pages for all APIs from libsystemd-login.
1694 * The build tree got reorganized and a the build system is a
1695 lot more modular allowing embedded setups to specifically
1696 select the components of systemd they are interested in.
1698 * Support for Linux systems lacking the kernel VT subsystem is
1701 * configure's --with-rootdir= got renamed to
1702 --with-rootprefix= to follow the naming used by udev and
1705 * Unless specified otherwise we'll now install to /usr instead
1706 of /usr/local by default.
1708 * Processes with '@' in argv[0][0] are now excluded from the
1709 final shut-down killing spree, following the logic explained
1711 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
1713 * All processes remaining in a service cgroup when we enter
1714 the START or START_PRE states are now killed with
1715 SIGKILL. That means it is no longer possible to spawn
1716 background processes from ExecStart= lines (which was never
1717 supported anyway, and bad style).
1719 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
1720 reloading of units together.
1722 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
1723 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
1724 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
1725 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
1726 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek